MYC Targeted Rad50 Drives Progression of High-Grade Serous Ovarian Cancer via NF-κB Activation

Background Rad50 is a component of MRN complex, which consists of Mre11-Rad50-Nbs1. The MRN complex participates in DNA double-strand break repair and DNA-damage checkpoint activation. We sought to investigate the clinical and functional signicance of Rad50 in high-grade serous ovarian cancer. Chromatin immunoprecipitation and luciferase assays were performed to evaluate the regulatory roles of MYC on Rad50 expression. Association between Rad50 expression and clinical outcome in HGSOCs was evaluated by Kaplan-Meier analysis. Invasion, clonogenic assay and xenograft mice model were conducted to determine to functional role of Rad50 in ovarian cancer. Protein immunoprecipitation and immunouorescence were used to explore the underlying mechanisms. Tissue microarrays (TMAs) Immunohistochemistry (IHC) staining was performed on formalin-xed paran-embedded HGSOC samples and xenograft tumors. Briey, tissue slides were deparanized and rehydrated in a graded series of ethanol. After that, antigenic retrieval was proceeded by microwave heating. Nonspecic antigens were blocked with the 1.5% normal goat serum. Then they were incubated at 4 °C in a moist chamber overnight with antibodies against human Rad50 (1: 300, Abcam, Cambridge, MA, USA) and Ki67 (1:150, DAKO, Tokyo, Japan). The next day, the slides were incubated with the corresponding secondary antibody. The staining was detected with DAB detection system basing on the Biotin-Streptavidin HRP Detection Systems (Zhongshan Biotechnology Company, China). The nal score of each sample was assessed by two independent pathologists based on intensity and extent of staining across the section. Staining intensity was graded according to the following standard: no staining (score 0); weak staining (score 1); moderate staining (score 2); strong staining (score 3). The proportion of every score was estimated for every sample as: 0 (0%), 1 (< 25%), 2 (25–50%), 3 (50–75%) and 4 (> 75%). The nal staining index (SI) was calculated as the staining intensity score multiplied by the proportion of stained tumor cells. The cases were then divided into low (SI ≤ 6) and high (SI > 9) groups. MYC, and was measured MYC knocked b-c luciferase activity d Luciferase activity in cells co-transfected MYC and pGL4 plasmids containing full length or truncated Rad50 promoters. e MYC binding peaks on Rad50 promoter region were shown in different cell lines based on ChIP-seq data in UCSC genome browser. f-g ChIP assays were in the of MYC on Rad50 promoter PCR. < 0.05, when in ovarian cancer cell lines. b Soft agar colony formation assay exhibited the effect of Rad50 on anchorage-independent growth and malignant transformation. c Invasion assays were performed in A2780, SKOV3 and HO8910 cells with Rad50 overexpression or knockdown. **P<0.01, ***P<0.001. d Morphological changes of A2780 cells with Rad50 overexpression or knockdown. E, EMT-related markers were detected by western blot in A2780, SKOV3 and HO8910 cells with Rad50 overexpression or knockdown.


Background
Ovarian cancer is the most lethal gynecologic malignancy and the overall 5-year survival rate is around 30% [1]. High-grade serous ovarian carcinoma (HGSOC) is the most common subtype and accounts for up to 70% of all ovarian cancer cases [2]. HGSOC is usually diagnosed at advanced stage and exhibits early peritoneal spread and lymph node metastasis, and resistance to conventional chemotherapy [3].
The development and molecular pathogenesis of HGSOC are largely unknown.
Approximately 50% of HGSOC cases were defective in homologous recombination (HR) DNA repair pathway. BRCA1/2 alterations, which can be derived from a combination of germline and somatic mutations, were seen in 22% of tumors [4]. Women with germline BRCA1/2 mutations have a 30-70% chance of developing high-grade serous ovarian carcinoma by age 70 [5]. Poly (ADP-ribose) polymerase (PARP) is an abundant nuclear enzyme involved in repairing single-stand breaks in DNA [6]. Impaired homologous recombination (HR) DNA repair in cancer cells would render synthetic lethality with PARP1 inhibition [7]. Olaparib was the rst PARP inhibitor approved in the European Union and the United States for the treatment of advanced BRCA-mutated ovarian cancer [8]. However, dysfunction of one DNA repair pathway may be compensated for by the function of another DNA damage response (DDR) pathway, which may be augmented and contribute to resistance to DNA-damaging chemotherapy and radiotherapy [9]. PARP inhibition in HR-de cient cells would result in increased activation of DNA-PK, increased nonhomologous end joining (NHEJ) activity [10]. High DNA-PKcs protein expression was associated with advanced stage and correlated with poor ovarian cancer speci c survival [11]. Some of DNA repair pathways are upregulated in cancer and contribute to the progression of malignancy [9].
The Mre11-Rad50-Nbs1 (MRN) complex plays important roles in the detection and signaling of DNA double-strand breaks (DSBs), as well as the repair pathways of HR and NHEJ [12]. MRN complex acts as a double-strand break sensor for ATM and recruits ATM to broken DNA ends [13]. The role of the MRN complex in the response to DSBs as well as its requirement for cellular survival makes it a potential target for cancer therapy [14]. MRE11 promotes tumorigenesis by facilitating resistance to oncogene-induced replication stress [15]. Overexpression of NBS1 contributes to transformation through the activation of phosphatidylinositol 3-kinase/Akt [16].
Rad50, an ATPase, binds to Mre11 and Nbs1 to form a repair complex that is transported to the nucleus in response to DNA double-strand breaks [17]. mRad50 was essential for cellular viability and disruption of mRAD50 causes ES cell lethality, abnormal embryonic development and sensitivity to ionizing radiation [18]. Frequent lack of RAD50 was found in endometrial [19] and low-grade epithelial ovarian cancer [20]. Whereas Rad50 was highly expressed in melanoma [21] and gastric cancer [22]. High level of Rad50 was associated with poor prognosis in gastric and colorectal cancer [22,23]. However, the expression pattern, functional role and clinical implication of Rad50 in HGSOC remain unclear.
In the present study, we found that oncogenic MYC directly transcriptionally activated Rad50 by binding its promoter. Upregulation of Rad50 is a frequent event in HGSOCs and enhanced Rad50 expression inversely correlated with patient survival. Furthermore, we elucidated that Rad50 overexpression promoted cell proliferation and invasiveness potentially via activation of the NFκB signaling pathway.
Our ndings indicate that MYC-regulated Rad50 may serve as a potential biomarker and therapeutic target for HGSOC.

Patients and tissue samples
This retrospective study included 151 cases of HGSOC and 33 cases of fallopian tube (FT) which were collected in Qilu hospital from November 2005 to December 2012. The HGSOC specimens were obtained from primary ovarian cancer patients and the FT tissues were from patients suffered from other benign pathologic changes or uterine disease. All participants in this study gave written informed consent as delineated by the protocol which was approved by the Ethics Committee of Shandong University.

Tissue microarray construction and immunohistochemistry
Tissue microarrays (TMAs) were constructed according to the method described previously [24]. Immunohistochemistry (IHC) staining was performed on formalin-xed para n-embedded HGSOC samples and xenograft tumors. Brie y, tissue slides were depara nized and rehydrated in a graded series of ethanol. After that, antigenic retrieval was proceeded by microwave heating. Nonspeci c antigens were blocked with the 1.5% normal goat serum. Then they were incubated at 4 °C in a moist chamber overnight with antibodies against human Rad50 (1: 300, Abcam, Cambridge, MA, USA) and Ki67 (1:150, DAKO, Tokyo, Japan). The next day, the slides were incubated with the corresponding secondary antibody. The staining was detected with DAB detection system basing on the Biotin-Streptavidin HRP Detection Systems (Zhongshan Biotechnology Company, China). The nal score of each sample was assessed by two independent pathologists based on intensity and extent of staining across the section. Staining intensity was graded according to the following standard: no staining (score 0); weak staining (score 1); moderate staining (score 2); strong staining (score 3). The proportion of every score was estimated for every sample as: 0 (0%), 1 (< 25%), 2 (25-50%), 3 (50-75%) and 4 (> 75%). The nal staining index (SI) was calculated as the staining intensity score multiplied by the proportion of stained tumor cells. The cases were then divided into low (SI ≤ 6) and high (SI > 9) groups.
Cell lines and cell culture HO8910 cells were obtained from China Type Culture Collection (CTCC, Shanghai, China); A2780, SKOV3 and HEK293T cells were originally purchased from American Type Culture Collection (ATCC, VA, USA). HO8910 and A2780 cells were routinely cultured in RPMI-1640 medium (Gibco, NY, USA). SKOV3 cells were cultured in McCoy's 5A medium (Gibco, NY, USA) and HEK293T cells were cultured in DMEM (Gibco, NY, USA). All media were supplemented with 10% FBS (Gibco, NY, USA), 100 U/ml penicillin and 100 µg/ml streptomycin. All the cells were cultured at 37 °C, 5% CO 2 in a humidi ed incubator (Thermo Fisher Scienti c, IL, USA).

Constructs and lentivirus infection
Rad50 overexpression and knockdown plasmids were obtained from Origene (Origene, MD, USA).
Lentivirus was produced in HEK293T cells. For stable infection, 1 × 10 5 cells were plated in 6-well plates with 2 ml medium without antibiotics. After overnight incubation, the medium was replaced by 1 ml Opti-MEM Reduced-Serum Medium (Gibco, NY, USA) containing 50 µl of concentrated lentiviral particles and 8 µg/ml of polybrene per well. Fresh medium containing 2 µg/ml of puromycin (Invitrogen, USA) was added to each well 24 hours later and then selected for two weeks.

RNA isolation and real-time PCR
Total RNA from cells was extracted with TRIZOL reagent (Invitrogen, USA) according to the manufacturer's protocol. RNA was reverse transcripted to cDNA using PrimeScript RT reagent Kit (Takara, JAPAN). Real-time PCR was performed with SYBR Green mix and detected by the Bio-Rad CFX96. GAPDH was served as an endogenous control. Primer information was shown in Table S2.
After incubation at designated times, 20 µl of MTT (5 mg/ml in PBS) was added to each well, and cells were incubated for another 4 h at 37 °C. The supernatants were carefully removed and 200 µl of dimethyl sulfoxide (DMSO, Sigma-Aldrich) was added to each well. The absorbance of each samples was measured at 490 nm with a Microplate Reader (Thermo Fisher Scienti c, IL, USA).

Clonogenic assay
Cells were seeded into 6-well plates (300-500 cells/well) and cultured for 2-3 weeks. The colonies were xed with methanol and stained with crystal violet. Colonies containing more than 50 cells were counted.
For soft agar colony formation, 3-8 × 10 3 cells were suspended in the upper layer (0.7%) of the two layers agar (0.7 and 1.0%) in 6 cm dish. After 3 weeks, the colonies were stained by MTT and the images of colonies were photographed.
Cell cycle analysis Cells (2-3 × 10 6 ) were cultured overnight in medium without FBS. Then the cells were cultured in complete medium and harvested at different times. Cell suspensions were incubated with 1 ml PBS containing 50 µg/ml PI with RNase A and permeabilization buffer for 30 min. The DNA content was analyzed by ow cytometer with a FACScalibur (BD Biosciences, USA). The results were analyzed with the ModFit software (Becton-Dickinson, USA).

Invasion and migration assay
Invasion and migration assays were performed in 24-well transwell chambers (BD Biosciences, USA) system with 8 µm pores coated with or without diluted matrigel (BD Biosciences, USA). Brie y, 1.5-2 × 10 5 cells were seeded into the upper chambers in certain medium containing no FBS, and lower chambers were lled with culture media containing 10-20% FBS as a chemo-attractant. After incubating at 37 °C for 12 to 48 h depending on the cell lines, cells that penetrated through the membrane were xed with methanol and stained with crystal violet.

Immuno uorescence
Cells were seeded on sterile glass coverslips and cultured overnight at 37℃. Cells were washed with PBS and then xed for 20 min with 4% paraformaldehyde. Following blocking with 10% goat serum in PBS for 2 h at room temperature, the cells were incubated with primary antibodies (anti-N-cadherin, anti-Ecadherin and anti-p65) overnight at 4℃. Cells were then incubated with Rhodamine/FITC-labeled secondary antibody (Beyotime Biotechnology, China) for 1 h at room temperature in the dark. Cells were washed with PBS three times before counterstaining with 1 µg/ml DAPI for 5 min in the dark. Finally, the slides were mounted with antifading reagent and examined with a uorescence microscope (Olympus, Japan).

Chromatin immunoprecipitation (ChIP)
ChIP assay was conducted using EZ-Magna ChIP A/G Chromatin Immunoprecipitation Kit (Millipore) following manufacturer's instructions. Brie y, cells were cross-linked using formaldehyde. Then the cells were lysed and DNA was sheared to 200-500 bp fragments by sonication. The cell lysate was incubated with the Myc antibody (CST 9402S) and magnetic beads overnight at 4 °C with rotation. DNA associated with MYC is represented as fold change of normal IgG, calculated byΔCq method. Primer sequences were shown in Supplementary Table S1.
Luciferase assay RAD50 wild type or truncated mutant promoter was cloned into pGL4.26 plasmid (Promega). Then RAD50 promotor report vector was co-transfected with pRL-TK and MYC plasmids into HEK293T cells. Luciferase activity was measured 24 hours after transfection using Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's protocol.

Xenograft tumor model
Female BALB/c nude mice (aged 4-6weeks) were injected subcutaneously in bilateral anks with cell suspensions (5 × 10 6 cells in 100 µl PBS). Tumor size was measured every three days after 1 week, and mice were killed with anesthesia after three weeks. Tumor volume was calculated according to the formula TV (cm 3 ) = a × b 2 × π/6, where 'a' is the longest diameter, and 'b' is the shortest diameter.
Hematoxylin and eosin (H/E) staining and immunohistochemistry were performed on sections from embedded samples. All animal experiments were performed with the approval of Shandong University Animal Care and Use Committee.

Statistical analysis
The software GraphPad Prism 5 was used for statistical analysis. The two-tailed Student's t-test and oneway ANOVA analysis were used to analyze the differences. The relationship between the expression of Rad50 and clinical pathological factors was analyzed by the chi-squared test. All error bars represent the standard error of three separate experiments. Differences with P < 0.05 were considered signi cant.

Results
MYC positively regulates Rad50 expression in high grade serous ovarian cancer MYC is the most ampli ed gene (30-60%) of human ovarian tumors in the TCGA cohort [4]. Coexpression analysis showed that the mRNA expression of MYC was positively correlated with Rad50 as well as NBS1 and MRE11 ( Fig. 1a and Fig. S1). To validate whether MYC regulates members of MRN complex, we knocked down MYC by siRNA in HEY and HO8910 cells. MYC depletion signi cantly downregulated expression of Rad50, NBS1 and MRE11 (Fig. 1b). In additional experiments, inhibition of MYC by JQ1 was observed to be markedly reduced the expression level of Rad50 and NBS1 in both cell lines (Fig. 1c). Rad50 was selected for further investigation since NBS1 was reported to be regulated by MYC previously [26]. Subsequently, we examined Rad50 expression by western blot and found forced MYC expression markedly enhanced protein level of Rad50 (Fig. 1d). Consistent with this, inhibition of MYC by siRNA or JQ1 reduced the protein level of Rad50 (Fig. 1d-e). These observations demonstrated that MYC positively regulates Rad50 expression in HGSOC.

MYC transcriptionally activates Rad50 through binding to its promoter
To determine whether MYC binds to the Rad50 promoter and transcriptionally regulates its expression. We analyzed Rad50 promoter region using JASPAR (http://jaspar.genereg.net/) and three putative MYCbinding sites were observed. Next, luciferase assay was performed and the result showed that forced expression of MYC increased luciferase activity of Rad50 promoter in SKOV3 and HO8910 cells (Fig. 2a). We further made truncated luciferase reporter vectors containing putative promoter region. Consistently, the luciferase activities of cells expressing truncated vectors were markedly lower than that of the fulllength vector (Fig. 2b-c). Ectopic expression of MYC in HEK293T cells led to signi cant induction of the luciferase activity of truncated luciferase reporters containing putative MYC-binding sites respectively (Fig. 2d). Cistrome analysis showed MYC binding peaks was enriched in the promoter regions of Rad50 in multiple cell lines (Fig. 2e). Subsequently, ChIP assays in HEY cells demonstrated that MYC directly bound to promoter regions of Rad50 ( Fig. 2f-g). Taken together, these data demonstrated that MYC transcriptionally activates Rad50 through binding to its promoter.
Rad50 is commonly elevated in HGSOCs and correlates with poor prognosis To determine the expression and clinical signi cance of Rad50 in HGSOCs, we rst examined Rad50 expression by western blot in HGSOCs and in fallopian tube (FT). As shown in Fig. 3a and Fig. S2a, Rad50 was overexpressed in a large proportion of the HGSOCs (n = 42) compared to FTs (n = 31). Consistently, immunohistochemical staining revealed a signi cant increase of Rad50 level in HGSOCs (n = 151) compared with FTs (n = 33) as illustrated in Fig. 3b. Furthermore, we analyzed clinicopathological and prognostic signi cance of Rad50 expression in HGSOCs (Table S1). Elevated Rad50 expression was correlated with lymph nodes metastasis (P = 0.0410) and the omentum metastasis (P = 0.0362). The association between Rad50 expression and overall survival was analyzed through Kaplan-Meier survival analysis of HGSOC patients. As shown in Fig. 3c-d, the high Rad50 level was signi cantly correlated with poor overall survival (P = 0.0002). Strikingly, Rad50 was upregulated in olaparib resistant ovarian cancer cells in GSE 117765 data (Fig. S2b). Taken together, these observations strongly indicate that Rad50 is commonly overexpressed and upregulation of Rad50 is associated with progression and poor prognosis in HGSOCs.

Ectopic expression of Rad50 promotes proliferation/invasion and induces epithelial-mesenchymal transition (EMT) of ovarian cancer cells
To explore the functional role of Rad50 in ovarian cancer, we established stable cell lines with Rad50 overexpression or knockdown (Fig. S2d). We then conducted clonogenic assay to evaluate the effect of Rad50 on the proliferation of ovarian cancer cells. Ectopic expression of Rad50 enhanced clonogenic capacity in HO8910 and A2780 cells while knockdown of Rad50 signi cantly reduced the colony-forming e ciency in A2780 and SKOV3 cells (Fig. 4a). We next performed soft agar colony formation assay and found Rad50 overexpression signi cantly increased whereas Rad50 knockdown reduced the colony number and the average colony size in soft agar (Fig. 4b). Moreover, Cell cycle analysis showed that Rad50 knockdown led to a decrease in the percentage of cells at S and G2-M phases compared with control group (Fig. S3a).
Because higher expression of Rad50 in HGSOC patients was associated with increased metastasis, we speculated that Rad50 may promote the invasiveness of ovarian cancer cells. We then measured migration and invasion abilities by transwell assay. Rad50 overexpression signi cantly enhanced the migration and invasion potential of A2780 and HO8910 cells while knockdown of Rad50 in A2780 and HO8910 cells led to reduced migration and invasion compared with control cells (Fig. 4c). Meanwhile, we found A2780 cells with Rad50 overexpression acquired more spindle-like morphology while Rad50 knockdown exhibited a more epithelial morphology compared to control cells (Fig. 4d). We further assessed the expression of EMT markers by Western blot and immuno uorescence. Ectopic expression of Rad50 in A2780 and HO8910 cells could obviously up-regulate mesenchymal phenotype markers (Ncadherin, Vimentin, Snail, Twist) and down-regulate epithelial phenotype marker (E-cadherin) while Rad50 knockdown upregulated epithelial marker and reduced the levels of mesenchymal markers in A2780 and SKOV3 cells (Fig. 4e and Fig. S4). Notably, knockdown of Rad50 in MYC-overexpressing cells attenuated the promotion of ovarian cancer cell invasion and colony-forming e ciency (Fig. S5a-d). These ndings suggest that Rad50 promotes migration /invasion and induces EMT in ovarian cancer cells.

Rad50 promotes tumor growth and metastasis of tumor xenografts
We next assessed the potential role of Rad50 on tumor growth and metastasis in vivo. Overexpression of Rad50 in A2780 cells signi cantly promoted ovarian tumor growth (Fig. 5a) whereas knockdown of Rad50 in SKOV3 cells remarkably reduced tumor burden in mice (Fig. 5b). Immunostaining of Ki-67 showed a higher proliferative index in tumors with Rad50 overexpression and a lower proliferative index in tumors with Rad50 knockdown than their corresponding control cells (Fig. 5c). We further evaluated metastasis-promoting potential of Rad50 in the lung metastasis model via tail vein injection and in the intraperitoneal metastatic model via intraperitoneal injection. As shown in Fig. 5d, knockdown of Rad50 in SKOV3 cells signi cantly reduced the number and size of lung metastasis nodules comparing with control. Meanwhile, overexpression of Rad50 in A2780 cells aggregated and formed much more dissemitated peritoneal tumors, while no peritoneal tumor was observed in the control group (Fig. S6).
These results strongly suggest that Rad50 possesses potent oncogenic properties in vivo.

Rad50 induces aggressiveness in HGSOC via activation of NF-κB signaling pathway
A study on dendritic cells revealed that Rad50-CARD9 interaction activates NF-κB pathway when transfected with dsDNA or infected a DNA virus [27]. We then postulated that Rad50 may activate NF-κB pathway in ovarian cancer. We rst examined the NF-κB pathway via immunoblot. Interestingly, overexpression of Rad50 increased the phosphorylation of p65 and decreased total IκBα protein levels while knockdown of Rad50 signi cantly decreased p-p65 protein level in ovarian cancer cell lines (Fig. 6a). We further examined the nuclear localization of p65 protein by Western blot of fractionated proteins and found levels of the activated nuclear form p-p65 were signi cantly higher in A2780 cells with Rad50 overexpression compared to control cells (Fig. 6b, left panel). While knockdown of Rad50 inhibited the translocation of p-p65 to the nucleus in both A2780 and SKOV3 cell lines (Fig. 6b, middle and right panels). Immuno uorescence data showed that ectopic expression of Rad50 induced, whereas downregulation of Rad50 inhibited nuclear localization of p-p65 (Fig. 6c). To determine whether Rad50 drives the HGSOC aggressive behavior through NF-κB activation, we blocked the NF-κB pathway by PDTC in Rad50-overexpressing cells. As expected, PDTC treatment decreased Rad50-induced invasiveness and EMT ( Fig. 6d and Fig. S7a). These results suggest that Rad50 activates NF-κB and the oncogenic function of Rad50 is dependent on the activation of NF-κB.
We proposed that Rad50 may also activate NF-κB pathway via the interaction with CARD9 in ovarian cancer cells. To this end, we performed immunoprecipitation (IP) -western blot to test whether Rad50 can interact with CARD9 in ovarian cancer cells. As indicated in Fig. 6e, CARD9 was detected in the immunoprecipitates with anti-Rad50 antibody, but not with the control mouse IgG in HEK293T and ovarian cancer cell lines. Consistently, Rad50 was detected in the immunoprecipitates with anti-CARD9 antibody (Fig. 6e). To verify whether Rad50-mediated activation of NF-κB pathway is CARD9 dependent, we knocked down CARD9 in Rad50-overexpressed ovarian cancer cell lines and found knockdown of CARD9 signi cantly reversed Rad50-induced migration activities (Fig. S7b). Consistent with this, activated nuclear form p-p65 induced by Rad50 was attenuated by siRNAs targeting CARD9 (Fig. S7b). Therefore, we identi ed CARD9 as an interacting protein of Rad50 in ovarian cancer cells and Rad50mediated activation of NF-κB pathway is CARD9 dependent. The diagram of the mechanism that MYC targeted Rad50 drives ovarian cancer progression via NF-κB activation was presented in Fig. 7.

Discussion
DDR is a double-edged sword in cancer prevention and cancer therapy. It acts as a positive guardian of genomic stability to prevent tumorigenesis. On the other hand, DDR also functions as a negative saboteur to resist chemo-and radiotherapy [28]. In our study, we reported that MYC proto-oncogene transcriptionally activated Rad50 expression in high-grade serous ovarian cancer. Rad50 was commonly upregulated in HGSOC and enhanced Rad50 expression inversely correlated with patient survival and metastasis. In agreement with our results, previous studies also reported that high expression of Rad50 were associated with poor prognosis in lung [29] and gastric [22] cancer.
In addition to conferring chemoresistance and radioresistance, high expression of DNA repair genes has been shown to exhibit oncogenic properties. NF-κB pathway in HGSOC. Future studies will be required to explore the molecular cooperation between Rad50 and CARD9 in promoting proliferation/invasion and induces EMT in ovarian cancer.

Conclusions
Our ndings showed that MYC-regulated Rad50 exhibits oncogenic property in ovarian cancer. We provided evidence that Rad50 ectopic expression promotes migration /invasion and induces EMT in ovarian cancer cells. Rad50 induces aggressiveness in HGSOC via activation NF-κB signaling pathway. Our data suggest that Rad50 could be a prognostic biomarker and potential therapeutic target for ovarian cancer.     MYC transcriptionally activates Rad50 through binding to its promoter. a luciferase activity was measured in SKOV3 and HO8910 cells cotransfected with MYC and pGL4 plasmids containing with or without Rad50 promoter. b-c luciferase activity was measured in HEK293T cells transfected with pGL4 plasmids containing full length or truncated Rad50 promoters. d Luciferase activity was measured in HEK293T cells co-transfected MYC and pGL4 plasmids containing full length or truncated Rad50 promoters. e MYC binding peaks on Rad50 promoter region were shown in different cell lines based on ChIP-seq data in UCSC genome browser. f-g ChIP assays were performed in HEK293T cells and the binding of MYC on Rad50 promoter was measured by PCR. *P < 0.05, **P<0.01, ***P<0.001, when compared with control group. promoters. e MYC binding peaks on Rad50 promoter region were shown in different cell lines based on ChIP-seq data in UCSC genome browser. f-g ChIP assays were performed in HEK293T cells and the binding of MYC on Rad50 promoter was measured by PCR. *P < 0.05, **P<0.01, ***P<0.001, when compared with control group.

Figure 2
Page 22/37 MYC transcriptionally activates Rad50 through binding to its promoter. a luciferase activity was measured in SKOV3 and HO8910 cells cotransfected with MYC and pGL4 plasmids containing with or without Rad50 promoter. b-c luciferase activity was measured in HEK293T cells transfected with pGL4 plasmids containing full length or truncated Rad50 promoters. d Luciferase activity was measured in HEK293T cells co-transfected MYC and pGL4 plasmids containing full length or truncated Rad50 promoters. e MYC binding peaks on Rad50 promoter region were shown in different cell lines based on ChIP-seq data in UCSC genome browser. f-g ChIP assays were performed in HEK293T cells and the binding of MYC on Rad50 promoter was measured by PCR. *P < 0.05, **P<0.01, ***P<0.001, when compared with control group.

Figure 4
Rad50 promotes proliferation and invasion of ovarian cancer cells.   Rad50-mediated activation of NF-κB pathway is CARD9 dependent. a Western blot analysis showing the expression of p65, p-p65 and IκBα in ovarian cancer cell lines with Rad50 overexpression or knockdown.
b Western blot analysis of indicated proteins in cytoplasmic and nuclear extracts from A2780 and SKOV3 cells with Rad50 overexpression or knockdown. c Immuno uorescence showing cytoplasmic or nuclear localization of the NF-κB complex in the indicated cell lines. d In the absence or presence of NF-κB inhibitor PDTC, migration/invasion ability and NF-κB downstream targets were measured with or without Rad50 overexpression. *P < 0.05, **P<0.01, ***P<0.001, when compared with the control group. e Interaction of Rad50 and CARD9 was determined by co-IP and western blot in HEK293T and ovarian cancer cell lines immunoprecipitated with either anti-Rad50 antibody or IgG control. f Interaction of Rad50 and CARD9 was determined by co-IP and western blotting in indicated cell lines immunoprecipitated with anti-CARD9 antibody or IgG control.

Figure 6
Page 35/37 Rad50-mediated activation of NF-κB pathway is CARD9 dependent. a Western blot analysis showing the expression of p65, p-p65 and IκBα in ovarian cancer cell lines with Rad50 overexpression or knockdown. b Western blot analysis of indicated proteins in cytoplasmic and nuclear extracts from A2780 and SKOV3 cells with Rad50 overexpression or knockdown. c Immuno uorescence showing cytoplasmic or nuclear localization of the NF-κB complex in the indicated cell lines. d In the absence or presence of NF-κB inhibitor PDTC, migration/invasion ability and NF-κB downstream targets were measured with or without Rad50 overexpression. *P < 0.05, **P<0.01, ***P<0.001, when compared with the control group. e Interaction of Rad50 and CARD9 was determined by co-IP and western blot in HEK293T and ovarian cancer cell lines immunoprecipitated with either anti-Rad50 antibody or IgG control. f Interaction of Rad50 and CARD9 was determined by co-IP and western blotting in indicated cell lines immunoprecipitated with anti-CARD9 antibody or IgG control.

Figure 7
The diagram of the mechanism that MYC targeted Rad50 drives ovarian cancer progression via NF-κB activation.