RNF126 as a Biomarker of a Poor Prognosis in Invasive Breast Cancer and CHEK1 Inhibitor Efficacy in Breast Cancer Cells Free

Ionizing radiation (IR) and most chemotherapies damage DNA as a major part of their mechanism of action. These remain standard therapies for all types of breast cancers, the most common cancer affecting women and the second most common cause of death due to cancers (1). The choice of radiotherapy or chemotherapy for breast cancer is currently made according to clinical factors. However, a subtype of breast cancer may be intrinsically resistant due to an upregulation of the DNA damage response (DDR), a major mechanism antagonizing DNA damage caused by radiotherapy/chemotherapy and involving cell-cycle checkpoints and DNA repair. This would result in a breast cancer patient subtype receiving unnecessary, aggressive treatments with minimal benefit. In addition, resistance to radiotherapy/chemotherapy may lead to tumor recurrence that can cause considerable morbidity, the dissemination of disease and an increased probability of mortality due to breast cancer (2, 3). Thus, there is an important need for identifying patients who are more likely to fail therapy and to improve treatment plans for those patients. By conducting large-scale profiling of cellular survival after exposure to radiation in a diverse collection of 533 genetically annotated human tumor cell lines, including breast cancer cell lines, a recent study demonstrated a broad variation in the response to radiotherapy, and perhaps also chemotherapy (4), as a result of genetic alterations. Thus, it is critical to identify patients whose breast cancer subtypes are intrinsically resistant to radiotherapy/chemotherapy and to explore new approaches to target their cancers.

RNF126 is a ring E3 ligase. Recent studies have suggested that RNF126 may have broad functions by targeting a variety of proteins for degradation; these may range from a role in endosomal sorting to the BAG6-dependent quality control of mislocalized proteins (5–9). RNF126 also promotes the proliferation of breast cancer by ubiquitinating CDKN1A and targeting it for degradation (10). By promoting nonhomologous end joining (NHEJ) and homologous recombination (HR; refs. 11, 12) RNF126 promotes the repair of DNA double-stranded breaks (DSB), the most dangerous type of DNA damage and that can be caused by endogenous and exogenous sources such as replication stress, radiotherapy and chemotherapeutic drugs. RNF126 promotes NHEJ via the ubiquitination of Ku80 (11). Interestingly, we recently reported that RNF126 facilitates HR by promoting the expression of BRCA1 in a manner independent of its E3 ligase activity but dependent on its interaction with E2F1 (12). A member of the family of E2F transcription factors, E2F1, is required for the expression of genes involved in a wide range of cellular processes, including cell-cycle progression, DNA replication, DNA repair, differentiation, and apoptosis. Consistent with its role in HR and NHEJ, RNF126 expression is associated with resistance to ionizing radiation (IR) and PARP inhibition (12) as both pathways are required to repair DSBs caused by IR and/or a PARP inhibitor. Thus, RNF126 appears to be associated with a diverse set of cellular processes in which its E3 ligase activity may or may not be involved. RNF126 has a close relative, BCA2, that shares 46% overall amino acid identity, and 75% identity in RING domains. Although BCA2 is highly expressed and is a prognostic biomarker in breast cancer (13–15), the pattern of RNF126 protein expression and its association with outcomes of breast cancer have not yet been evaluated.

The cell-cycle checkpoint kinase, ataxia telangiectasia mutated, and Rad3-related kinase (ATR) and its key downstream effector, CHEK1, can be activated by RPA (replication protein A)-coated elongated ssDNA. The ATR/CHEK1 pathway prevents the entry of cells with damaged or incompletely replicated DNA into mitosis when cells are challenged by DNA-damaging agents, such as IR or chemotherapeutic drugs, the major modalities used to treat cancers. This regulation is particularly evident in cells with a defective G₁ checkpoint, a common feature of cancer cells, owing to mutations in TP53. In addition, ATR and CHEK1 suppress replication stress by inhibiting excess origin firing, particularly in cells with activated oncogenes (16–18). Thus, ATR and CHEK1 inhibitors have been developed and are currently used either as single agents, or paired with radiotherapy or a variety of genotoxic chemotherapies in preclinical and clinical studies. Although CHEK1 inhibitors were initially thought to enhance the effects of radiotherapy and genotoxic drugs, particularly in TP53-deficient cells (19–21), recent preclinical studies suggest that CHEK1 inhibitors may function as signal agents since cancer cells are more reliant on ATR/CHEK1 for survival (17). Indeed, ATR/CHEK1 inhibition specifically target cancer cells expressing MYC, CYCLIN E, and H-RAS (18, 22–25). In addition, we recently also reported that CHEK1 inhibitors, as single agents, have antitumor activity in radioresistant breast cancer by enhancing replication stress (24). Radioresistant breast cancer cells carry high levels of C-MYC/CDC25A/C-SRC/H-RAS/E2F1 oncogenes and ATR/CHEK1/BRCA1/CtIP DDR proteins, indicating that upregulation of DDR proteins, including cell-cycle checkpoint proteins and oncogenes, may be characteristic of being targeted by CHEK1 inhibitors (24). As we have demonstrated that RNF126 binds to E2F1 (12), a transcription regulator that controls the expression of several hundred genes, including oncogenes and CHEK1 (26–29), we hypothesize that RNF126 promotes CHEK1 expression and that cells expressing high levels of RNF126 may be targeted by CHEK1 inhibitors. Thus, the aims of the current study are to investigate the clinical significance of RNF126 expression in breast cancer, and to determine the role of RNF126 in promoting CHEK1 expression, with particular attention to its influence on the efficacy of CHEK1 inhibitors.

Here, we demonstrate that RNF126 protein is highly expressed in invasive breast cancer and is associated with a poor prognosis. Critically, RNF126 is an independent factor for a poor prognosis. We also reveal that RNF126 controls CHEK1 expression via a direct interaction with E2F1. A strong correlation between RNF126 and CHEK1 protein expression exists in both breast cancer tissues and cell lines. Treatment with a CHEK1 inhibitor led to increased cell killing by enhancing replication stress in cells expressing a higher level of RNF126. Our studies suggest that RNF126 is a potential biomarker for the poor prognosis of invasive breast cancer and for the efficacy of CHEK1 inhibitors.

Two cohorts of breast cancer samples were included. The first cohort consisted of 110 paraffin-embedded tumor tissues from patients with invasive breast cancer, as well as 78 paired adjacent normal tissues as negative controls. The samples were taken from The First Affiliated Hospital of Sun Yat-sen University from January 1, 2004 to December 31, 2006. The patients in this study all had primary operable invasive breast cancer and were under the care of a single surgeon, Z. Ma. The most common histologic breast cancer type was invasive ductal carcinoma that comprised 88.18% (97/110) of cases, with the other 13 cases being invasive lobular, medullary, or mucinous carcinomas. Treatment included mastectomy or local excision, with or without adjuvant systemic chemotherapy and/or radiotherapy. Follow-up involved clinical reviews at 6-monthly intervals for the first 5 years, and then annually. All samples were confirmed histologically by two pathologists. Histologic diagnosis was determined according to American Joint Commission on Cancer Staging (AJCC) criteria. Conventional pathologic data were collected retrospectively and characteristics of primary breast cancers are summarized in Supplementary Table S1. A second cohort that consisted of 67 invasive breast cancer cases was prepared for TMA analysis of the expression of both RNF126 and CHEK1. The patients of the second cohort were treated at the same hospital between January 1, 2014 and April 1, 2015.

All patients were female and signed informed consent forms. The protocol was approved by the ethics committee of The First Affiliated Hospital of Sun Yat-sen University (application ID: [2016]060). Inclusive criteria were: (i) all patients had unilateral invasive breast cancer and underwent either a radical mastectomy or modified radical mastectomy. Adjacent normal breast tissues were selected from an area more than 5 cm from the edge of the tumor and were confirmed by two pathologists. (ii) Patients who received preoperative radiotherapy, chemotherapy, hormonal therapy, or any other anticancer therapy before resection were excluded. (iii) adjuvant treatments such as chemotherapy, radiotherapy, or endocrine therapy were chosen based on the patient's condition after surgery in accordance with the relevant The National Comprehensive Cancer Network guidelines (NCCN). (iv) All patients were followed up with medical appointments or by telephone. Cancer recurrence, metastasis, or death were end events. The follow-up deadline was April 18, 2016.

Full tissue sections of 110 paraffin embedded invasive breast cancer and 78 normal tissues were processed for IHC staining of RNF126. TMA blocks was constructed containing 67 invasive breast cancer. Serial 4-μm sections were cut from the TMA blocks for both RNF126 and CHEK1 staining. Antigen retrieval, blocking procedures, and a modified ImmunoMax method were used as described previously (30). In brief, slides were heated to 60°C and then deparaffinized in xylene. The slides were rehydrated in descending alcohol concentrations. Antigen retrieval was performed by incubating slides in a retrieval solution of citrate buffer. Hydrogen peroxide was added to block endogenous peroxidase activity to decrease unwanted background staining. Primary antibody (ab183102, 1:100; Abcam; 25887-1-AP, 1:150; Proteintech) was added at an optimum dilution. Negative controls were performed by the substitution of primary antibody with phosphate-buffered saline (PBS). To guarantee consistent IHC evaluation, slides from a reference tumor previously determined as positive were included in each staining procedure.

Evaluations of staining reactions were performed in accordance with the immunoreactive score (IRS) proposed by Remmele and Stegner: IRS = staining intensity (SI) × percentage of positive cells (PP). Staining intensity was marked as nongranulated (0); low grade (light yellow; 1); moderate (brownish yellow; 2); or strong (reddish brown; 3). The PP was scored as negative (<5%; 0); weak (5%–10%; 1); moderate (11%–50%; 2); strong (51%–80%; 3); or very strong (>81%; 4). Specimens scoring beyond 3 were considered positive overexpression (31). All slides were independently evaluated by two pathologists blind to patients and their corresponding clinical information.

MCF7, MDA-MB-231, SKBR3, MDA-MB-361, MCF10A, HCC202, ZR751, T47D, MDA-MB-468, HCC1187, HCC1569, HCC70, BT549, HCC1143, BT474, HCC38, and HCC1954 were cultured in DMEM (Invitrogen) supplemented with 10% FBS (Gibco/Thermo Fisher Scientific), in a humidified atmosphere containing 5% CO₂ at 37°C. The shRNA of RNF126 was purchased from Sigma-Aldrich. Full-length wild-type RNF126 and RING-domain mutated (C229A/C232A) RNF126 have been described previously (10). The CHEK1-promoter reporter was a gift from Dr. Pier Paolo Pandolfi (Beth Israel Deaconess Cancer Center, Boston, MA). All DNA-plasmid transfections were performed using Lipofectamine 2000 according to the manufacturer's recommendations (Invitrogen). Flag-RNF126 construct, the full-length RNF126 fragment, has been described previously (12). Two CHEK1 inhibitors were used in this study, including LY2603618 from ApexBio (A8638) and AZD7762 from Selleckchem (S1532).

For the MTT assay, cells were plated into 96-well plates and incubated overnight. Cells were then exposed to various doses of CHEK1 or ATR inhibitors for 72 hours. MTT (20 μL of 5 mg/mL) was added to each well and cells incubated for a further 3.5 hours in an incubator. MTT solvent was added after removing the medium and the cells in plates were agitated on an orbital shaker for 15 minutes. The absorbance was read at 590 nm with a reference filter of 620 nm. For clonogenic survival assays, cells plated into petri dishes (60 mm × 15 mm) were exposed to various doses of CHEK1 or ATR inhibitors for 24 hours, and then replaced with fresh medium. After 13–15 days of incubation at 37°C, the cells were stained using Giemsa. The number of colonies (>50 cells) was counted.

Quantitative reverse transcription-PCR (qRT-PCR) was conducted as described previously (12). Total RNA was isolated using an RNeasy Kit (Qiagen). Experiments were carried out in triplicate for each data point. Reactions were performed using SYBR Green mix and a MyiQ real-time PCR detection system (Bio-Rad). Relative mRNA levels were calculated using the comparative C_t method (ΔC_t).

• GAPDH forward/reverse primers: 5′-CTCTGCTCCTCCTGTTCGAC-3′/5′-TTAAAAGCAGCCCTGGTGAC-3′.

• CHEK1 forward/reverse primers: 5′-CCAGATGCTCAGAGATTCTTCCA-3′/ 5′-TGTTCAACAAACGCTCACGATTA-3′.

• E2F1 forward/reverse primers: 5′-GTGGACTCTTCGGAGAACTT-3′/5′-TGTTCTCCTCCTCAGAAGTG-3′.

• CYCLIN E forward/reverse primers: 5′-TTTCTTGAGCAACACCCTC-3′/5′-TGTCACATACGCAAACTGG-3′.

• RNF126 forward/reverse primers: 5′-TATCGAGGAGCTTCCGGAAGAGA-3′/5′-AAAGCAAACTGTCCGTAGCCCT-3′.

Chromatin immunoprecipitation (ChIP) was performed using a Simple ChIP Enzymatic Chromatin IP kit (#9002, Cell Signaling Technology). In brief, 5 × 10⁷ cells were fixed with 1% final concentration of formaldehyde for 10 minutes at room temperature. The formaldehyde was quenched by adding 125 mmol/L glycine for 5 minutes at room temperature. Cells were washed with cold PBS containing a protease inhibitor cocktail and were then lysed with cold Buffer A. Collected pellets were resuspended in cold Buffer B and then treated with 7 μL of micrococcal nuclease for 30 minutes at 37°C. Digested chromatin was sonicated and purified according to kit instructions. Chromatin (10 μg) was incubated with the following antibodies: 2 μg E2F1 (Santa Cruz Biotechnology, sc193), 2 μg H3 (provided by the kit), and 2 μg IgG (provided by the kit). The primers used to amplify the regions containing the putative consensus DNA-binding sites of RNF126 in the CHEK1 promoter by PCR were as follows: forward 5′-AGCACTCTGCTTCACCGACT-3′, reverse 5′-CTGGGCCCAAATATGAAGTG-3′.

Immunofluorescence assays were performed as described previously (24). Cells growing on slides were fixed directly in 3%–4% paraformaldehyde. For unextractable CDC45 staining, cells were extracted for 5 minutes on ice with 0.5% Triton X-100 in cytoskeletal (CSK) buffer (10 mmol/L PIPES, 300 mmol/L sucrose, 100 mmol/L NaCl, 3 mmol/L MgCl₂; pH = 6.8) supplemented with 1 mmol/L phenylmethylsulfonyl fluoride, 0.5 mmol/L sodium vanadate, and proteasome inhibitor for 10 minutes at 4°C. Then, extracted cells were fixed with 3%–4% paraformaldehyde. The cells were permeabilized for 10 minutes with PBS containing 0.5% Triton X-100 for 15 minutes at room temperature, followed by blocking with 1% BSA, and then incubated with primary antibodies. The bound antibodies were revealed with goat anti-mouse IgG Alexa Fluor 594 and chicken anti-rabbit IgG Alexa Fluor 488. Slides were viewed at 1,000× magnification with a NIKON 90i fluorescence microscope (photometric cooled mono CCD camera).

Cellular extracts were prepared by resuspending cells in RIPA lysis buffer and proteins were resolved by 5%, 12%, or 15% SDS-PAGE. For chromatin CDC45 isolation, chromatin-bound proteins were prepared according to a previous publication (32). In brief, 3 × 10⁶ cells were resuspended in 200 μL of buffer A [10 mmol/L HEPES (pH 7.9), 10 mmol/L KCl, 1.5 mmol/L MgCl₂, 0.34 mol/L sucrose, 10% glycerol, 1 mmol/L dithiothreitol, and protease inhibitor mixture (Roche Molecular Biochemicals)]. Triton X-100 was added to a final concentration of 0.1%, and the cells were incubated for 10 minutes on ice. Nuclei were collected in the pellet (P1) by low speed centrifugation (1,500 × g, 4 minutes, 4°C). The supernatant (S1) was further clarified by high speed centrifugation (13,000 × g, 10 minutes, 4°C) to remove cell debris and insoluble aggregates. Nuclei (P1) were washed once with buffer A and then lysed in 200 μL of buffer B (3 mmol/L EDTA, 0.2 mmol/L EGTA, 1 mmol/L DTT, and protease inhibitor mixture). After 10-minute incubation on ice, soluble nuclear proteins (S2) were separated from chromatin by centrifugation (2,000 × g, 4 minutes). Insoluble chromatin (P2) was washed once in buffer B and centrifuged again under the same conditions. The final chromatin pellet (P3) was resuspended in 30 μL Laemmli buffer and sonicated for 30 seconds in a sonicator using a microtip at 25% amplitude. The fractioned chromatin-bound protein was denatured by boiling the sample for 5–10 minutes, and analyzed by immunoprecipitation.

Dual-luciferase assays were conducted as described previously (33) Cell extracts were prepared according to the instructions of the manufacturer and assayed in a TD-20/20 luminometer (Turner Designs) using the Dual-Luciferase Reporter assay System (E1910, Promega). Briefly, cells were cotransfected with vector control, RNF126-WT, RNF126-Δf, or RNF126-C229A/C232A and CHEK1-reporter vector in the ratios of 10:1 for 6 hours, then replaced with fresh medium and continually cultured for additional 48 hours. The cells were washed with PBS, and then lysed with PLB reagent. Finally, the lysate were detected with a TD-20/20 luminometer according to the manufacturer's instructions.

Cell-cycle analyses were conducted as described previously (24). Cells were collected and fixed with cold 70% ethanol. Approximately 10⁶ cells/mL were incubated for 30 minutes with staining solution containing RNase A (10 μg/mL, Sigma), and propidium iodide (20 μg/mL, Sigma) for 30 minutes. The DNA content was measured by flow cytometry.

DNA fiber assays were performed as published with some modifications (24, 34). Cells were pulse-labeled with 50 μmol/L IdU (Sigma-Aldrich, I7125) for 40 minutes and then pulse-labeled with 200 μmol/L CldU (Sigma–Aldrich #C6891) for 40 minutes in the presence or absence of CHEK1 inhibitor. At the end of the CldU pulse, cell suspensions (2.5 μL) were mixed with 7.5 μL of lysis buffer [0.5% SDS, 200 mmol/L Tris-HCl (pH 7.4), 50 mmol/L EDTA]. Each mixture was dropped on the top of an uncoated regular glass slide. Slides were inclined at 25° to spread the suspension on the glass. Once dried, DNA spreads were fixed by incubation for 10 minutes in a 3:1 solution of methanol-acetic acid. The slides were dried and placed in precooled 70% ethanol at 4°C for at least 1 hour. DNA was denatured with 2.5 mol/L HCl for 30 minutes at 37°C. The slides were blocked in 1% BSA in PBS for 30 minutes at room temperature and then incubated with mouse anti-BrdUrd antibody (BD Biosciences, #347580) at a 1:200 dilution and rat anti-CldU antibody (Abcam, #ab6326) at a 1:400 dilution. The slides were incubated with secondary fluorescent antibodies [goat anti-mouse IgG [H+L] Alexa Fluor 594 secondary antibody (A-11032, 1:400; Thermo Fisher Scientific); or chicken anti-rabbit IgG [H+L] Alexa Fluor 488 secondary antibody [A-21441, 1:400]; Thermo Fisher Scientific]. Replication fibers were viewed at 1,000 × magnification on a NIKON 90i fluorescence microscope (photometric cooled mono CCD camera; Nikon). Signals were measured using ImageJ software (NCI/NIH), with some modifications made specifically to measure DNA fibers.

Primary antibodies used for Western blots were against: BRCA1 (Clone D-9, 1:200; Santa Cruz Biotechnology); RPA2 (Clone NA18, 1:100; Calbiochem/EMD Millipore); E2F1 (Clone KH95, 1:200; Santa Cruz Technology); β-Actin (Clone AC-74, 1:50,000; Sigma-Aldrich); CHEK1 (G-4, 1:200; Santa Cruz Biotechnology); phospho-CHEK1 antibody (#2344, CHEK1-pSer317, 1:500; Cell Signaling Technology); phospho-CHEK1 antibody (#133D3, CHEK1-p345, 1:500; Cell Signaling Technology); E2F1 (clone KH95 sc-251, 1:500; Cell Signaling Technology); CDC45 (G-12 sc55569, 1:200; Santa Cruz Biotechnology); γ-H2AX (ser139 JBC301, 1:500; Millipore clone); p-RPA2(S4/S8; rabbit polyclonal, BL647, 1: 1000; Bethyl Laboratories), CDC25A (clone DCS-120, 1:100; Thermo Fisher Scientific); ORC2 (sc13238, 1:200; Santa Cruz Biotechnology); CDK2 (610146, 1:200; BD Biosciences); and CYCLIN E (sc247, 1:200; Santa Cruz Biotechnology); PARP (65995,1:400; BD Biosciences); CASPASE 9 (M044232,1:1000; BD Biosciences); cleaved CASPASE 9 (D35427, 1:500; Calbiochem); CASPASE 8 (M043764,1:250; BD Biosciences); CASPASE 7 (SC56063,1:500; Santa Cruz Biotechnology); CASPASE 3 (#9665,1:500; Cell Signaling Technology); cleaved CASPASE 8 (#2008,1:500; Upstate Biotechnology/Thermo Fisher Scientific); cleaved CASPASE 6 (D35426,1:500; Calbiochem); and cleaved CASPASE 3 (76658,1:100; BD Biosciences). Secondary antibodies were goat anti-mouse IgG-horseradish peroxidase (HRP)–conjugated (#7076S, 1:1,000; Cell Signaling Technology), goat anti-rabbit IgG-HRP–conjugated (#7074S, 1:1000; Cell Signaling Technology), and donkey anti-goat IgG-HRP–conjugated (A2216, 1:1,000; Santa Cruz Biotechnology).

The primary antibodies used for immunofluorescence were against: γH2AX (clone JBW301, 1:500; Millipore); RPA2 (S4/S8; A300-245A, 1:500; Bethyl Laboratories); CDC45 (H-300 clone, SC20685, 1:50; Santa Cruz Biotechnology); phospho-HISTONE H3 (Ser10; #9706, 1:100; Cell Signaling Technology); goat anti-mouse IgG (H+L) Alexa Fluor 594 secondary antibody (A-11032, 1:400; Thermo Fisher Scientific); and chicken anti-rabbit IgG (H+L) Alexa Fluor 488 secondary antibody (A-21441, 1:400; Thermo Fisher Scientific).

MCF7 and MDA-MB-231, the two major cell lines used in this study, were authenticated via Short Tandem Repeat profiling by Genetica DNA Laboratories (a LabCorp brand) using a PowerPlex 16HS amplification kit (Promega Corporation) and GeneMapper ID v3.2.1 software (Applied Biosystems). The authentication of each cell line was confirmed by a 100% match to the reference STR profile of the respective cell lines from the ATCC.

Statistical analyses were undertaken using the statistical software package, R version 3.3.4 and stata12.0 (StataCorp). Comparisons between RNF126 staining and various existing prognostic factors were performed using a χ² test and logistic regression. Analyses of cumulative survival probability were performed using the Kaplan–Meier method and differences between groups were tested by log-rank test. Multivariate analysis was undertaken using the Cox proportional hazard regression model. The effects of CHEK1 inhibition on DNA repair recruitment/foci, DSB formation, and replication dynamics were examined using t test (two groups) or ANOVA (more than two groups). Tukey honest significant difference (HSD) test was further used to compare the difference between groups. Correlation analysis was examined using Spearman rank correlation.

To determine RNF126 protein expression in cases of invasive breast cancer, we collected 110 early-stage operable primary invasive breast cancer specimens and 78 adjacent normal tissues for study. All patients were female. The clinicopathologic features of patients with breast cancer enrolled in this study are shown in Supplementary Table S1. RNF126 expression was detected by IHC (Fig. 1A and B). Because of the lack of any study to define positivity according the expression level of RNF126, we determined RNF126 staining in tissues in accordance with an immunoreactive score (IRS) proposed by Remmele and Stegner (31). Of all samples, 55.45% (61 cases) of tumors were positive for RNF126 staining while 44.55% (49 cases) showed negative staining. In comparison, only 7.69% (6 cases) of adjacent tissue samples showed positive immunoreactivity to RNF126 and 92.31% (72 cases) displayed negative staining. Thus, the difference in RNF126 immunoreactivity between tumor samples and adjacent tissues was significant (χ² = 45.3894, P < 0.001; Fig. 1A). Representative RNF126 staining in both normal and tumor tissues is shown in Fig. 1B. RNF126 staining was found in both the nucleus and cytoplasm of cancer cells, a result consistent with a previous report (10) and that of our unpublished data showing that RNF126 is located in both the cytoplasm and nucleus of cultured cancer cells. In addition, RNF126 protein expression was further compared with several clinicopathologic variables in breast cancer, such as age, TNM stage, histologic grade, menstruation status, and molecular subtypes (Supplementary Table S2). With regard to cases with luminal A tumors, 68.75% (11/16), as well as 58.49% (31/53) of cases with luminal B tumors, which were both positive for ER/PR, displayed positive RNF126 staining, whereas only 42.86% (9/21) of cases with triple-negative breast cancer and 50.00% (10/20) of cases with HER2-enriched tumors were positive for RNF126 staining. Nevertheless, differences in RNF126 expression between triple-negative, HER2, and hormone receptor–positive tumors were not statistically significant (χ² = 2.9327 P = 0.402). In addition, logistic regression analysis was also established to measure the relationship between RNF126 expression and clinicopathologic parameters, including patient age, TNM stage, histologic grade, menstruation, and molecular subtypes. In this multivariable regression analysis, the ORs were 1.57, 1.07, 1.03, 0.64, and 0.67, respectively. The P values for all parameters were more than 0.05 (Fig. 1C), indicating that RNF126 expression had no obvious relationship with these well-known clinicopathologic factors.

Next, using recurrence, metastasis, or deaths as endpoints that reflect a low cumulative survival probability and poor prognosis, Kaplan–Meier plots for negative versus positive RNF126 expression showed that RNF126–positive staining was associated with a poor prognosis (log-rank test, P = 0.003; Fig. 1D). The median follow-up was 102 months (range 14–145 months). To determine whether RNF126 expression was associated with outcomes in the group of patients who received adjuvant therapies, 90 patients who received adjuvant chemotherapy after surgical resection based on RNF126 staining of their tumors were sorted into subgroups. Patients who showed RNF126-positive staining of breast cancer tumors displayed a lower cumulative survival probability compared with patients who had negative RNF126 staining of breast cancer tumors (log-rank test, P = 0.001; Fig. 1E), indicating that RNF126-positive staining was associated with a poor outcome in the group of patients who received adjuvant chemotherapy. Finally, we used a Cox proportional hazard model to determine the prognostic value of RNF126. RNF126 immunoreactivity, patient's age, TNM stage, histologic grade, menstruation, and molecular subtypes were chosen as risk variables since all are potential factors affecting a low cumulative survival probability of breast cancer. HRs are indicated in Fig. 1F. The HR values for RNF126 immunoreactivity and TNM stage were 7.3 (P = 0.009) and 3.8 (P = 0.002), respectively. This indicates that in multivariate analyses, RNF126 expression and TNM stage are two independent factors related to a poor outcome in patients with invasive breast cancer (Fig. 1F). Thus, high RNF126 expression is associated with a poor prognosis and is an independent predictor of a poor prognosis in breast cancer.

That RNF126 is associated with a poor prognosis highlights the clinical significance of this protein in breast cancer. However, a specific inhibitor of RNF126 is not currently available. Studying the role of RNF126 in the regulation of CHEK1 expression will provide new opportunities for therapeutic intervention in breast cancer. RNF126 knockdown by two shRNAs targeting different regions of RNF126 led to decreased CHEK1 protein levels in MCF7 (Fig. 2A) and MDA-MB-231 cells (Fig. 2B). Of note, downregulation of RNF126 was tolerated well by MCF7 and MDA-MB-231 cells, without a significant alteration in cell-cycle profiles being observed (ref. 12; Supplementary Fig. S1). Thus, the decreased expression of CHEK1 in cells depleted of RNF126 was not caused by cell-cycle changes in our experimental conditions. However, it has been suggested that RNF126 knockdown can cause cell arrest (10). This discrepancy may be caused by differences in the magnitude of RNF126 knockdown. RNF126 likely regulates CHEK1 at the transcriptional level because RNF126 depletion resulted in a decrease in CHEK1 mRNA in both cell lines (Fig. 2C). In accordance with these results, overexpression of Flag-RNF126 led to increased CHEK1 protein (Fig. 2D and E) and mRNA levels (Fig. 2F) in both MCF7 and MDA-MB-231 cells. Most importantly, the E3 ligase activity of RNF126 appears to be dispensable for the regulation of CHEK1 expression because the expression of a validated RNF126 E3 ligase mutant (RNF126 C229A/C232A; ref. 10) retained the ability to increase CHEK1 protein expression in both MCF7 and MDA-MB-231 cells (Fig. 2G and H). This result was consistent with our previous study where E3 ligase activity of RNF126 was found not to be required for BRCA1 expression (12).

We next investigated whether interaction between RNF126 and E2F1 was required for controlling CHEK1 expression by using a CHEK1 promoter–driven luciferase reporter (35), given that promotion of E2F1 mediated–transactivation by RNF126 depends on the direct interaction of these two proteins via a 185–195 (f) region in RNF126. Flag-RNF126 overexpression increased luciferase activity, indicating that RNF126 promotes transactivation of the CHEK1 promoter. In contrast, RNF126-Δf overexpression failed to induce a significant increase in luciferase activity but reduced luciferase activity compared with control cells in both MCF7 and MDA-MB-231 cell lines (Supplementary Fig. S2A), indicating that RNF126-Δf expression may interfere with the function of endogenous RNF126. These results are consistent with our previous report suggesting that a RNF126-Δf mutant lacking an association with E2F1 leads to a loss of function of RNF126 in promoting the E2F1-mediated transactivation of BRCA1; it also has a dominant-negative effect (12). This result was further supported by a chromatin immunoprecipitation (ChIP) assay showing that RNF126 overexpression enhanced the enrichment of E2F1 on the CHEK1 promoter; however, RNF126-Δf overexpression reduced the binding of E2F1 protein to the CHEK1 promoter (Supplementary Fig. S2B). Moreover, the decreased expression of CHEK1 at both mRNA and protein levels was observed in cells expressing RNF126-Δf compared with control cells, whereas increased CHEK1 mRNA (Supplementary Fig. S2C) and protein expression (Supplementary Fig. S2D) was found in cells expressing Flag-RNF126-WT. Again, in support of the idea that the E3 ligase activity of RNF126 is dispensable for the regulation of CHEK1 expression, the expression of RNF126 C229A/C232A led to increased luciferase activity of CHEK1 (Supplementary Fig. S2A), enrichment of RNF126 at a promoter of CHEK1 (Supplementary Fig. S2B), and elevated CHEK1 expression at both mRNA (Supplementary Fig. S2C) and protein (Supplementary Fig. S2D) levels, similar to that observed in cells expressing wild-type Flag-RNF126. In addition to CHEK1, RNF126 also promoted the expression of CYCLIN E, another downstream factor of E2F in both MCF7 and MDA-MB-231 cells (Supplementary Fig. S3A and S3B). Thus, we conclude that by interacting with E2F1, RNF126 promoted CHEK1 expression at the mRNA transcription level.

Next we were interested in determining any association between RNF126 and CHEK1 in breast cancer tissues. We assessed immunoreactive staining of these two proteins by analyzing a second cohort of breast cancer cases that consisted of samples from 67 patients with early-stage primary invasive breast cancer prepared as tissue microarrays (TMA; n = 67). Both RNF126 and CHEK1 staining were determined by IHC using TMA and quantified by IRS scores. Of note, CHEK1 immunoreactivity was predominantly located in the cytoplasm and was granular in appearance, although nuclear staining was also observed. CHEK1–positive staining was found in 94.59% (35/37) of RNF126-positive staining breast cancer samples. CHEK1 staining was negative in 80% (24/30) of RNF126-negative staining breast cancer samples. The expression of RNF126 in tissues was related to that of CHEK1 (χ² = 38.82, P < 0.001, Cramér V = 0.7612; Fig. 3A). Representative staining of these two proteins is shown in Fig. 3B. Thus, in invasive breast cancer, there was a strong and statistically significant correlation between RNF126 and CHEK1 protein expression.

In addition, we further analyzed the expression of RNF126 and CHEK1 proteins in a panel of 16 human mammary carcinoma–derived cell lines that comprised: luminal A, ER⁺ breast cancer (MCF7, ZR-75-1 and T47D), luminal B (MDA-MB-361, BT474), HER2⁺ breast cancer (HCC202, SK-BR-3, HCC1569) and triple-negative breast cancer (MDA-MB-231, HCC1143, HCC1954, HCC38, HCC1187, HCC70, BT549, MDA-MB-468) cells. We set MCF10A, a normal immortalized breast epithelial cell line, as a control. The expression of RNF126 and CHEK1 proteins was determined by Western blotting. Band intensities were quantified using ImageJ software, and normalized to β-actin (n = 3; Fig. 3C and D). RNF126 expression was increased in a large majority of breast cancer cell lines when compared with a MCF10A cell line used as a control. The highest level of RNF126 protein was found in highly tumorigenic MDA-MB-231 cells (Fig. 3C and D). Correspondingly, CHEK1 protein expression was also relatively high in these cells. The cell lines, BT474 and ZR751, showed lower or undetectable levels of RNF126 and CHEK1 protein expression compared with MDA-MB-231 cells (Fig. 3C and D). Therefore, a positive correlation between RNF126 and CHEK1 protein expression was observed in breast cancer cell lines (Fig. 3E; correlation coefficient of Spearman rank correlation, r = 0.682, P = 0.004), which is consistent with observations from breast cancer tissues (Fig. 3A). In addition, we also measured RNF126 and CHEK1 mRNA by qRT-PCR (n = 3; Fig. 3F). Levels of RNF126 protein essentially paralleled mRNA levels of CHEK1 (Fig. 3G; Spearman rank correlation, r = 0.532, P = 0.034). This result is consistent with Fig. 2 and Supplementary Fig. S2 showing that RNF126 promoted CHEK1 mRNA expression. However, interestingly, RNF126 protein and mRNA transcripts did not correlate in tested breast cancer cell lines (Fig. 3H; Spearman rank correlation, r = 0.300, P = 0.259), indicating that the high expression of RNF126 may not be a consequence of transcriptional regulation. Similarly, a corelationship between the expression of RNF126 protein and CYCLIN E mRNA existed (Spearman rank correlation, r = 0.624, P = 0.009; Supplementary Fig. S3D) that aligns with the result described in Supplementary Fig. S3A and S3B where RNF126 facilitated the expression of CYCLIN E at both mRNA and protein levels (Supplementary Fig. S3A and S3B). Thus, we concluded that RNF126 and CHEK1 protein expression positively correlated in both breast cancer tissue and cell lines.

We chose two pairs of breast cancer cell lines showing higher (MDA-MB-231, and MDA-MB-468), or lower/undetectable RNF126 expression (BT474, ZR751) for a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Fig. 4A). LY2603618, one of the first highly selective and potent CHEK1 inhibitors, was used in this study. We found that the two cell lines expressing higher levels of RNF126 were more sensitive to LY2603618 compared with the cells showing lower RNF126 expression, suggesting that RNF126 expression may determine sensitivity to CHEK1 inhibitors (Fig. 4A). To confirm these results, we determined the effect of RNF126 knockdown by RNF126 shRNA#1 (Fig. 4B and C) or #2 (Supplementary Fig. S4A and S4B) on the efficacy of LY2603618. LY2603618 exposure resulted in more killing of parental cells MCF7 and MDA-MB-231, compared with the corresponding cells with RNF126 knockdown by shRNA in MTT (Fig. 4B; Supplementary Fig. S4A and S4B) and/or colony-forming assays (Fig. 4C). Treatment with a second CHEK1 inhibitor, AZD7762, also decreased RNF126-expressing cell numbers compared with both MCF7 (Supplementary Fig. S4C) and MDA-MB-231 (Supplementary Fig. S4D) cells with knocked down RNF126, as determined by MTT assay. CHEK1 inhibition was monitored by measuring protein levels of CHEK1-p-S345 and its downstream factor CDC25A (Fig. 4D). Thus, we concluded that RNF126 depletion abrogated CHEK1-inhibited cell killing.

The reduction in cell viability was accompanied by an increase in CHEK1 inhibitor–induced apoptosis as determined by measuring optimal biomarkers of apoptosis, such as cleaved CASPASE 3, 6, 7, 8, 9, as well as PARP. Cleaved PARP and cleaved CASPASE 7 increased in LY2603618 treated MCF7 (Supplementary Fig. S5A) and MDA-MB-231cells (Supplementary Fig. S5B), whereas an obvious increase was not found in corresponding cells with RNF126 depletion under the same conditions. Interestingly, cleaved CASPASE 8 was observed in MDA-MB-231 cells with intact RNF126, but a reduced effect on cleaved CASPASE 8 was observed in RNF126–depleted cells (Supplementary Fig. S5B). However, cleaved CASPASE 8 was not seen in MCF7 cells (Supplementary Fig. S5A). The differences in the response of apoptosis proteins in MCF7 and MDA-MB-231 cells may be due to differences in basal levels of apoptotic proteins. For instance, CASPASE 3 was absent in MCF7 cells, whereas CASPASE 3 was present in MDA-MB-231 cells (36). Interestingly, according to immunofluorescence (IF) results, we found that as a single agent, CHEK1 inhibition by LY2603618 did not increase the rate of mitotic cells in MCF7 cells, with or without RNF126 knockdown. Instead, CHEK1 inhibition resulted in a decrease in the proportion of mitotic cells in MCF7 (Supplementary Fig. S5C) and MDA-MB-231 cell lines (Supplementary Fig. S5D), but not in cell lines showing RNF126 depletion (representative staining of mitotic cells, as determined by IF of p-HISTONE H3, is shown in Supplementary Fig. S5E). This may be explained by the fact that when RNF126 is intact, DNA damage induced by CHEK1 inhibition triggers ATR activity and G₂–M arrest, preventing cells from entering the next stage. However, in cells with depleted RNF126, less DNA damage is induced by CHEK1 inhibition. An insufficient amount of DNA damage may trigger a G₂–M checkpoint by CHEK1 inhibition. Thus, we conclude that CHEK1 inhibition was more effective in cells expressing higher levels of RNF126. Of note, a similar result was also seen with an ATR inhibitor. ATR inhibition by AZD6738 was more toxic in MCF7 (Supplementary Fig. S6A and S6B) and MDA-MB-231 cells (Supplementary Fig. S6C and S6D) with intact RNF126, compared with cells with RNF126 knocked down by RNF126 shRNA #1 (Supplementary Fig. S6A and S6C) or #2 (Supplementary Fig. S6B and S6D). These results support the notion that ATR has a similar function to that of CHEK1 in terms of suppressing oncogenic stress/checkpoints/HR.

We next determined the extent of replication stress following CHEK1 inhibition in cells, with or without RNF126 knockdown. We analyzed foci of phosphorylated RPA2 (p-RPA2), a marker for replication stress, in response to exogenous DNA-damaging agents by immunofluorescence staining. A more profound increase in the proportion of cells with p-RPA2 foci was observed in MCF7 cells compared with cells depleted of RNF126 by RNF126shRNA #1 (Fig. 5A). The greater increase in p-RPA2 foci in LY2603618-treated MCF7 cells were also confirmed by Western blot analysis (Fig. 5B). In addition, LY2603618 treatment led to a greater increase in γH2AX foci (Fig. 5C) and protein levels (Fig. 5D) in parental MCF-7 cells compared with MCF-7 cells with RNF126 knockdown. Of note, although CHEK1 inhibition caused an increase in CHEK1-p-S345 in cells depleted of RNF126, the extent was much less than that seen in cells with intact RNF126. This result supported our hypothesis that CHEK1 inhibition leads to a reduced amount of DNA damage in RNF126-depleted cells. A similar result was observed in parental MDA-MB-231 cells (Fig. 5E and F), and in MCF7 and MDA-MB-231 cells treated with the second CHEK1 inhibitor, AZD7762 (Supplementary Fig. S7A and S7B). These results suggest that CHEK1 inhibition suppresses the proliferation of breast cancer cells expressing higher levels of RNF126. A similar result was also seen using a second RNF126 shRNA #2 (Supplementary Fig. S7C and S7D). RNF126 knockdown by RNF126 shRNA #2 abrogated CHEK1 inhibition–induced replication stress in both MCF7 and MDA-MB-231 cells.

Deregulated origin firing contributes to oncogene-induced replication stress (37). We next determined how treatment with CHEK1 inhibitor affects the initiation of DNA replication by analyzing DNA fiber spreads, according to the protocol illustrated in Fig. 6A and our previous publication (24). The percentage of new origins increased when cells were treated with LY2603618 in both parental MCF7 cells and MCF7 cells with RNF126 knockdown (Fig. 6B). However, the magnitude of the increase was greater in parental cells compared with cells with RNF126 knockdown. A similar result was seen in MDA-MB-231 cells (Fig. 6C). CHEK1 is involved in controlling replication initiation via regulating CDC45 (38), a protein that is implicated in initiation rather than elongation processes. We next measured the amount of CDC45 in a nonextractable chromatin fraction. LY2603618 treatment caused a remarkable increase in the amount of nonextractable CDC45 protein in control cells compared with RNF126-depleted cells (Fig. 6D), although overall CDC45 levels were comparable (Fig. 6D). The effect of CHEK1 inhibition on chromatin loading of CDC45 was further confirmed by IF assay (Fig. 6E). As a result of an alteration in replication initiation, the elongation ratio was decreased when CHEK1 activity was inhibited, particularly in MCF7 cells and MDA-MB-231 with intact RNF126 (Fig. 6F and G). Representative DNA fiber staining is presented in Supplementary Fig. S8. Cumulatively, the results presented in Fig. 6 suggest that CHEK1 inhibition led to a greater increase in replication initiation and a decrease in replication speed, particularly in cells with higher RNF126 expression. This result was consistent with the results described in Fig. 5 where CHEK1 inhibition caused greater replication stress in cells with RNF126 expression compared with RNF126-depleted cells.

The biological functions of RNF126 have been explored recently (6, 8, 10–12). However, to date, a report determining RNF126 expression in human cancers is lacking. Our results suggest that RNF126 protein was highly expressed in invasive breast cancer (Fig. 1). Although the mechanism contributing to increased RNF126 expression is not clear, the data obtained from a panel of breast cancer cell lines suggest levels of RNF126 protein and mRNA transcripts are not correlated (Fig. 3). Thus, it is postulated that the increased RNF126 protein measured in breast cancer tissues may not necessarily be a consequence of an alteration in RNF126 mRNA transcripts. In addition, RNF126-positive staining appears to be slightly higher in the ER⁺ cohort compared with the other cohorts, such as the triple-negative subtype. However, this difference did not reach statistical significance. Thus, a further study with a larger number of invasive breast cancer cases is required. The RNF126 gene maps to chromosome 19p13.3, which is a commonly deleted region in ovarian cancer (39–41). Interestingly, a genome-wide study of breast cancer also detected a high frequency of loss of heterozygosity (LOH) in the 19p13 genomic region (42). It is not clear whether the LOH of 19p13 in breast cancer led to the decreased RNF126 expression observed. However, it is likely that RNF126 may be a context-dependent signaling molecule and that the expression of RNF126 in cancer may be contingent on the biological context. The high expression of RNF126 in breast cancer suggests that RNF126 may contribute to breast cancer development, although the molecular mechanisms behind this are, as yet, unclear.

In this study, we demonstrate that RNF126 expression is associated with a poor prognosis, such recurrence, metastasis, or deaths, in patients with invasive breast cancer (Fig. 1). Nevertheless, the relationship between RNF126 protein expression and each end point of a poor prognosis needs to be investigated further. Our most significant finding is that high RNF126 expression is an independent predictor for a poorer patient prognosis, which is independent from established prognostic markers such as patients' age, TNM stage, histologic grade, menstruation, and molecular subtypes (Fig. 1). Further analysis using adjuvant chemotherapy as a stratification criterion suggested that patients with RNF126-positive breast cancer tumors had a significantly lower cumulative survival probability compared with those with RNF126-negative tumors (Fig. 1). Although conclusions from our observations are limited due to the small number of patients who received adjuvant therapies (n = 90), the differences in survival probabilities are striking and suggest that RNF126 expression levels may influence the response to adjuvant therapies. As DSB repair proteins have been suggested to play an important role in the cellular response to chemotherapy as well as to radiotherapy, the role of RNF126 in the repair of DSBs by promoting HR and NHEJ may contribute to its poor prognosis. The association of RNF126 with a poor prognosis in breast cancer highlights the clinical significance of this protein.

In our study, we identify a relationship between RNF126 and CHEK1 by demonstrating that RNF126 promotes E2F1-mediated expression of CHEK1 transcripts (Fig. 2), which is consistent with our previous publication that outlined how RNF126 promoted the activity of the transcriptional factor, E2F1 (12). Breast cancer tumors expressing higher levels of RNF126 often show elevated CHEK1 protein expression in both breast cancer tissues and cell lines (Fig. 3). Most importantly, a correlation between RNF126 protein levels and CHEK1 transcripts in breast cancer cell lines was also observed, supporting our finding that RNF126 promotes CHEK1 expression at transcriptional levels (Fig. 2). Nevertheless, the positive relationship between RNF126 protein and CHEK1 transcripts needs to be verified in breast tumor tissues in future.

It is well established that ATR/CHEK1 suppress oncogene-induced replication stress. Cancer cells often harbor some degree of replication stress due to oncogene activities, which can be lethal to cells. Thus, they often upregulate ATR and CHEK1 activity to mediate survival because ATR/CHEK1 suppress replication stress to an intolerable level by the suppression of replication initiation and/or promoting HR (24, 43, 44). In support of this concept, increased ATR/CHEK1 expression was frequently observed in a variety of cancer cells, including lung cancer, ovarian cancer, head neck cancer, triple-negative breast cancer, neuroblastoma, T-cell acute lymphoblastic leukemia, acute myeloid leukemia, and hepatocellular carcinoma. Although the biological significance of the correlation between RNF126 and CHEK1 expression remains unknown, it may be related to the inhibition of replication stress by CHEK1 in cells expressing high levels of RNF126. Thus, increased CHEK1 protein expression in RNF126-positive breast cancer cells is likely related to the suppression of replication stress because RNF126 also promotes oncogene expression such as CYCLIN E (Supplementary Fig. S3), an oncogene that causes replication stress. It is most likely that RNF126-positive breast cancer upregulates oncogenes in addition to CHEK1, rendering cells dependent on ATR/CHEK1 for survival. Indeed, CHEK1 inhibition causes greater killing in cells expressing RNF126, whereas a lesser effect was found in cells with RNF126 depletion. Thus, in addition to MYC, CYCLIN E, and H-RAS that have been reported to affect the outcome of CHEK1 or ATR inhibitors, RNF126 is also a potential factor determining the efficacy of CHEK1 inhibitors (Fig. 6H). Using RNF126 expression as a biomarker for a CHEK1 inhibitor has a greater advantage than CHEK1 expression alone because the high expression of CHEK1 may not be functionally important. Indeed, p-CHEK1, instead of CHEK1 expression levels, is a biomarker for CHEK1 inhibitors (45). We also reported that radioresistant breast cancer cells that carry high levels of oncogene and DDR proteins, including ATR/CHEK1, are more sensitive to CHEK1 inhibition (24), suggesting that expression of both oncogene and cell-cycle checkpoint proteins are features that could be targeted by CHEK1 inhibitors. Thus, the role of RNF126 in promoting CHEK1 expression, and perhaps also oncogene expression, determine the sensitivity of RNF126-positive breast cancer to CHEK1 inhibitors.

The current model for oncogene-induced replication stress is related to deregulated replication initiation, because an excess of ongoing replication forks will consume the limited dNTP pool and cause fork stalling (37). This will generate extensive ssDNA regions that are protected by RPA coating. With a limited supply of RPA, uncoated ssDNA causes DSB. However, ATR/CHEK1 can be activated during replication stress, which, in turn, suppresses oncogene-induced replication by targeting CDC25A for degradation. Our studies provide evidence that further support the notion that CHEK1 suppresses replication stress by inhibiting replication initiation, particularly in cells expressing RNF126 (Fig. 6). As ATR/CHEK1 can also promote the repair of DSBs by facilitating HR, increased DSBs induced by CHEK1 inhibition may also be related to the impaired HR repair of collapsed replication forks. Thus, multiple mechanisms are involved in CHEK1 inhibition–induced replication stress in cells expressing RNF126.

We have reported that RNF126 promotes the expression of HR protein BRCA1 at the transcriptional level (12). The probability that BRCA1 affects the efficacy of the CHEK1 inhibitor on breast cancer cells expressing relatively high levels of RNF126 is very low. Transient RNF126 overexpression increases mRNA expression of the BRCA1 (12). However, coexpression of these two proteins may not be seen in tumor tissues or cancer cell lines, as it is well-known that the BRCA1 promoter is frequently methylated, leading to low expression (46, 47). Even if some high RNF126–expressing cell lines have high BRCA1 protein expression, its effect may be to reduce, rather than increase, the sensitivity to CHEK1 inhibitors, as HR-defective cells are more sensitive to ATR/CHEK1 inhibition (48).

Of note, despite the initial hypothesis that CHEK1 inhibitors can increase efficacy in combination with IR and chemotherapy drugs, particularly in cells with TP53 deficiency, our studies show that RNF126 promotes CHEK1 expression and affects sensitivity to CHEK1 inhibitors in cells, with or without wild-type TP53. Our results are consistent with previous publications showing that ATR/CHEK1 inhibition can target cancer cells as single agents independent of TP53 (24, 49). Thus, acting as single agents and in combination with other chemotherapy drugs/IR, the mechanisms by which CHEK1 inhibitors lead to cell death may be distinct.

In summary, we identify that RNF126 is highly expressed in invasive breast cancer and is an independent predictor of a poor outcome for this disease. High RNF126 expression may be used as a potential biomarker for CHEK1 inhibitors. Our study provides proof of concept in preclinical models for a new paradigm for treating breast cancer expressing high levels of RNF126 by CHEK1 inhibitors (Fig. 6H). Identifying breast cancers with high levels of RNF126 expression that can then be targeted by CHEK1 inhibitors will significantly improve the efficacy of such agents. It will be necessary to validate our findings in breast cancer using large randomized clinical trials. This may be done by assessing whether a simple IHC assay of RNF126 expression performed on routine paraffin-embedded tissue would be able to predict a patient's response to CHEK1 inhibitors. We also need to evaluate whether RNF126-positive breast tumors are more responsive to CHEK1 or ATR inhibitors.

No potential conflicts of interest were disclosed.

Conception and design: X. Yang, Y. Pan, Y. Zhang, S. Li, R.A. Keri, Z. Ma, J. Zhang

Development of methodology: X. Yang, Y. Pan, Z. Qiu, Z. Du, Y. Zhang, P. Fa, S. Ma, S. Li, J. Zhang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Yang, Z. Qiu, Y. Zhang, P. Fa, S. Gorityala, Y. Xu, Z. Ma

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Yang, Z. Qiu, Y. Zhang, P. Fa, S. Gorityala, H. Wang, Y. Xu, C. Yan, Z. Ma, J. Zhang

Writing, review, and/or revision of the manuscript: X. Yang, Y. Pan, H. Wang, Z. Ma, J. Zhang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Pan, S. Ma, C. Chen, Y. Xu, R.A. Keri

Study supervision: Z. Ma, J. Zhang

The authors apologize to colleagues whose work was not cited because of space limitations or ignorance. Our thanks for the service provided by BioMed Proofreading LLC.

The work described was supported by a grant (R01CA154625) from the National Cancer Institute and seed grants from the Case Comprehensive Cancer Center and VeloSano Bike to Cure Foundation (to J. Zhang); a National Natural Science Foundation of China grant (31571452 and 31271503) and Guangdong Provincial Natural Science Foundation of China grant (S2012010008368) and a startup fund from The First Affiliated Hospital of Sun Yat-sen University (to Z. Ma), and scholarships from the Chinese Scholarship Council (CSC). This research was also supported by the Radiation Resources Core Facility and Cytometry & Imaging Microscopy Core Facility of the Case Comprehensive Cancer Center (P30 CA43703).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.