Genome-Wide Rnai Screening Identifies RFC4 As a Factor That Mediates Radioresistance in Colorectal Cancer by Facilitating Nonhomologous End Joining Repair

Published OnlineFirst April 12, 2019; DOI: 10.1158/1078-0432.CCR-18-3735

Translational Cancer Mechanisms and Therapy Clinical Cancer Research Genome-wide RNAi Screening Identiﬁes RFC4 as a Factor That Mediates Radioresistance in Colorectal Cancer by Facilitating Nonhomologous End Joining Repair Xue-Cen Wang1, Xin Yue1, Rong-Xin Zhang2,Ting-Yu Liu1, Zhi-Zhong Pan2, Meng-Jie Yang3, Zhen-Hai Lu2, Zi-Yang Wang1, Jian-Hong Peng2, Li-Yuan Le4, Gao-Yuan Wang1, Qi-Hua Peng1, Yuan Meng1, Wenlin Huang1,5, and Ran-Yi Liu1

Abstract

Purpose: Neoadjuvant chemoradiotherapy (neoCRT) is a Results: RFC4, NCAPH, SYNE3, LDLRAD2, NHP2, and standard treatment for locally advanced rectal cancer FICD were identiﬁed as potential candidate radioresistance (LARC); however, resistance to chemoradiotherapy is one genes. RFC4 protected colorectal cancer cells from X-ray– of the main obstacles to improving treatment outcomes. The induced DNA damage and apoptosis in vitro and in vivo. goal of this study was to identify factors involved in Mechanistically, RFC4 promoted nonhomologous end join- the radioresistance of colorectal cancer and to clarify the ing (NHEJ)-mediated DNA repair by interacting with Ku70/ underlying mechanisms. Ku80 but did not affect homologous recombination–medi- Experimental Design: A genome-wide RNAi screen was ated repair. Higher RFC4 expression in cancer tissue was used to search for candidate radioresistance genes. After associated with weaker tumor regression and poorer prognosis RFC4 knockdown or overexpression, colorectal cancer cells in patients with LARC treated with neoCRT, which likely exposed to X-rays both in vitro and in a mouse model were resulted from the effect of RFC4 on radioresistance, not assayed for DNA damage, cytotoxicity, and apoptosis. More- chemoresistance. over, the regulatory effects and mechanisms of RFC4 in DNA Conclusions: RFC4 was identiﬁed as a radioresistance fac- repair were investigated in vitro. Finally, the relationships tor that promotes NHEJ-mediated DNA repair in colorectal between RFC4 expression and clinical parameters and out- cancer cells. In addition, the expression level of RFC4 predicted comes were investigated in 145 patients with LARC receiving radiotherapy responsiveness and the outcome of neoadjuvant neoCRT. radiotherapy in patients with LARC.

Introduction Although the incidence of colorectal cancer has declined in developed countries in recent years, it has been dramatically Colorectal cancer is the third most commonly diagnosed cancer rising in China (1–3). Neoadjuvant radiotherapy/chemora- and the second leading cause of cancer mortality worldwide, diotherapy (neoRT/CRT) is advocated by National Comprehen- accounting for more than 9% of all cancer-related deaths (1). sive Cancer Network (NCCN) guidelines as the standard treatment for locally advanced rectal cancer (LARC) because it has been found to significantly increase local control and cancer- 1State Key Laboratory of Oncology in South China, Collaborative Innovation specific survival in these patients (4–6). Most patients with LARC Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, treated with neoRT/CRT exhibit variable degrees of a tumor China. 2Department of Colorectal Surgery, Sun Yat-sen University Cancer response, but resistance still occurs in a proportion of patients. 3 Center, Guangzhou, China. School of Pharmaceutical Sciences (Shenzhen), To classify these patients, the neoadjuvant rectal (NAR) score, Sun Yat-sen University, Guangzhou, China. 4Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou, China. 5Guangdong Provincial Key Laboratory which incorporates weighted cT, ypT, and ypN, has been pro- of Tumor Targeted Drugs and Guangzhou Enterprise Key Laboratory of Gene posed (7). However, application of the NAR score may be limited Medicine, Guangzhou Doublle Bioproducts Co. Ltd., Guangzhou, China. by its inability to explain why some patients classified within Note: Supplementary data for this article are available at Clinical Cancer the same NAR category demonstrate different treatment Research Online (http://clincancerres.aacrjournals.org/). responses (7, 8). An increasing number of investigations have revealed that the treatment response of individual patients with X.-C. Wang, X. Yue, and R.-X. Zhang contributed equally to this article. colorectal cancer is associated with the genetic characteristics and Corresponding Authors: Ran-Yi Liu, Sun Yat-sen University Cancer Center, 651 gene expression pattern of cancer cells (9) and that the combined Dong-feng Road East, Guangzhou, Guangdong 510060, China. Phone: 8620- use of molecular biomarkers and clinical parameters would 8734-2303; Fax: 8620-8734-3178; E-mail: [email protected]; Wenlin Huang, [email protected]; and Xin Yue, [email protected] improve the prediction of neoadjuvant therapeutic responses and prognosis for patients with colorectal cancer (10, 11). Neverthe- Clin Cancer Res 2019;25:4567–79 less, few biomarkers have been clinically validated as predictive of doi: 10.1158/1078-0432.CCR-18-3735 the radiotherapy response. Thus, there is an urgent need to 2019 American Association for Cancer Research. elucidate the mechanisms underlying radiotherapy resistance and

www.aacrjournals.org 4567

Wang et al.

transient transfections were performed with Lipofectamine 2000 Translational Relevance or 3000 (Invitrogen) according to the manufacturer's guidelines. Neoadjuvant radiotherapy (neoRT) significantly increases the resection rate and local control of locally advanced rectal Generation of stable cell lines cancer (LARC) and the cancer-specific survival of affected Human RFC4 cDNA was cloned into the pCDH-EF1-MCS-T2A- patients. Although most patients with LARC treated with Puro plasmid (System Bioscience) tagged with 3 Flag at the neoRT exhibit variable degrees of a tumor response, a pro- N-terminus; shRNAs (Supplementary Table S2) were cloned into portion of patients still suffer from resistance. Therefore, it is the pLKO.1 vector (Sigma-Aldrich), according to the manufac- necessary to elucidate the underlying mechanisms of radio- turer's guidelines. These plasmids were cotransfected into 293T resistance and to discover reliable biomarkers for predicting cells with a Lentiviral Packaging Kit (FulenGen) to obtain recom- radiosensitivity in patients with LARC. This study identified binant lentiviruses. Colorectal cancer cells with stable RFC4 over- RFC4 as a radioresistance factor that promotes nonhomolo- expression or knockdown were obtained by lentiviral transduc- gous end joining (NHEJ)-mediated DNA repair in colorectal tion and puromycin selection, as described previously (14, 15). cancer. RFC4 upregulation in cancer cells was found to be associated with reduced tumor regression and poor prognosis High-throughput RNA interference screening in patients with LARC treated with neoRT. Therefore, a high Screening was performed using the LentiPlex human whole- RFC4 expression level may suggest radioresistance, and RFC4 genome shRNA library targeting 15,000 human genes (an average silencing/inhibition is a potential strategy for radiosensitiza- of 5 shRNAs per target gene; SHPH01; Sigma-Aldrich) based on tion in these patients; these findings may contribute to the the manufacturer's instructions. The relative copy number of each development of precise treatment strategies in clinical practice shRNA in the X-ray–treated and -untreated groups was compared, in the future. and MAGeCK-negative selection was performed to identify candidate radioresistance genes (16, 17). As the knockdown of radioresistance genes confers sensitivity to radiation, cells containing radioresistance gene–targeted shRNAs would be more susceptible to discover reliable biomarkers that effectively predict radiosen- to death upon X-ray exposure, and the copy number of these sitivity in patients with CRC. shRNAs would be reduced in the residual cell populations after Recently, high-throughput screening using shRNA and X-ray treatment. Thus, radioresistance gene candidates should CRISPR/Cas9 libraries has been applied for functional genomics meet the following criteria: (i) log2(fold change) < 2.0; (ii) FDR in human cells (12, 13). In this study, 6 candidate radioresistance- (computed from the empirical permutation P values using the related genes were chosen by high-throughput screening of a Benjamini–Hochberg procedure; ref. 17) 0.05 and FDR rank < genome-wide shRNA library targeting 15,000 human genes. 1,000; and (iii) at least three shRNAs with a copy number reduced Among these candidate genes, replication factor C subunit 4 by more than 50% in X-ray–treated cells compared with untreated (RFC4) was identified as the most prominent radioresistance cells (Fig. 1A). factor protecting colorectal cancer cells from radiation-induced DNA strand breaks (DSB). Mechanistically, RFC4 promotes non- Reverse transcription and quantitative PCR homologous end joining (NHEJ)-mediated DNA repair by inter- Total RNA was isolated from cells or tissues using TRIzol acting with the Ku70/Ku80 complex. Moreover, a higher level of reagent (Invitrogen) and reverse transcribed into cDNA using an RFC4 expression was found to be associated with weaker tumor M-MLV Reverse Transcriptase Kit (Promega). Gene expression regression and poorer prognosis in patients with LARC treated levels were assessed using qPCR with iTaq SYBR Green Mix with neoCRT. (Bio-Rad) and specific primers (Supplementary Table S3). Rela- DC tive mRNA levels are presented as 2 t (GAPDH was used as an internal control) after normalization to the control group (18). Materials and Methods Cells, plasmids, siRNAs, and transfections Cytotoxicity assays The HCT116, HT29, HCT15, and SW480 colorectal cancer cell Cytotoxicity or cell sensitivity to X-rays or drugs was analyzed lines, the NCM460 normal human colon mucosal epithelial cell using colony formation or Cell Counting Kit-8 (CCK-8) assays. line, and 293T cells were purchased from ATCC and cultured For the colony formation assay (19, 20), cells were seeded into according to ATCC guidelines. All cell lines were authenticated by 6-well plates at a density of 1,000 cells/well. Twenty-four hours short tandem repeat analysis at China Center for Type Culture later, the cells were treated with X-rays or drugs at the indicated Collection, and the absence of Mycoplasma contamination was doses and then cultured for 10 to 14 days, followed by staining verified using a PCR Detection Kit (Shanghai Biothrive Sci. & Tech. with 0.5% crystal violet. Colonies containing more than 50 cells Ltd.). Radioresistant HCT1166GyR cells were generated by treating were counted. For the CCK-8 assay (21), cells (3,000/well) were parental HCT116 cells with stepwise increasing doses of X-rays in cultured in 96-well plates for 24 hours and then treated with drugs our laboratory (illustrated in Supplementary Fig. S1). All cell lines for 72 hours. Cell viability was determined by OD450 nm using were frozen in liquid nitrogen and used for experiments at CCK-8 (Dojindo). The IC50 was calculated using GraphPad Prism passages 3 to 10 after thawing. 5.0 (GraphPad Software). To construct RFC4-expressing plasmids, human RFC4 cDNA (a gift from Prof. Jiahuai Han's laboratory at Xiamen University, Apoptosis assay China) was amplified and inserted into pcDNA4.0 for expression Apoptosis was evaluated with an Annexin V/propidium iodide with a Myc or Flag tag. Negative control (NC) and specific siRNAs (PI) double-staining assay, as described previously (22, 23). (Supplementary Table S1) were synthesized by GenePharma. All Briefly, cells were treated with X-rays, cultured for 48 hours,

4568 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

Figure 1. RFC4 is identified as a candidate radioresistance factor facilitating colorectal cancer cell survival in response to X-ray exposure in vitro. A, The schematic procedure of genome-wide RNAi screening that identified 6 candidate radioresistance genes. B–D, Colony formation assays were used to assess survival at 10 days after exposure to X-rays in HCT116 cells with stable gene knockdown (shNC, negative control shRNA; shGENE, GENE-specific shRNA; B), colorectal cancer cells with RFC4 knockdown or overexpression (top, WB; Vec, control lentiviral vector; RFC4, RFC4 overexpression lentiviral vector; sh1/2, RFC4 knockdown by shRFC4-1/2; C), and colorectal cancer cells with endogenous RFC4 silencing by stable transfection of sh5 (a shRNA targeting the RFC4 30-UTR) after restoring RFC4 expression by transient transfection of the RFC4 expression plasmid (left, representative colony formation pictures and WBs; right, bar charts indicating the mean number of colonies/well; Vec, pcDNA4.0; RFC4, RFC4 cDNA in pcDNA4.0; D). E and F, Assessment of the effects of RFC4 on apoptosis 48 hours after cells were treated with X-rays by Western blot (loading control: actin; E) and Annexin V/PI double-staining assays with flow cytometry (the percentage of apoptotic cells is presented as the mean SD of at least three independent experiments; F; , P < 0.05; , P < 0.01; , P < 0.001).

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4569

Wang et al.

harvested, and stained using an Annexin V/PI Apoptosis Kit resistant cells were utilized as models to analyze chromosomal (BestBio). The samples were then analyzed by flow cytometry. DSB repair. These cells were cotransfected with pCMV-NLS-I-SceI and the indicated shRNAs or siRNAs. After 48 hours of incubation, þ Coimmunoprecipitation and Western blot analyses the cells were harvested to assess the proportion of GFP cells Coimmunoprecipitation (Co-IP) and Western blot (WB) anal- using a CytoFLEX flow cytometer (Beckman). The NHEJ- or HR- yses were performed as described previously (14). Briefly, cells mediated DNA repair efficiency (NHEJ or HR activity) is presented þ were lysed using lysis buffer (Cell Signaling Technology) supple- as the percentage of GFP cells normalized to the respective mented with protease (Roche) and phosphatase (KeyGen Biotech) control. inhibitors, and the protein concentration was quantified using a BCA Protein Assay Kit (KeyGen Biotech). For co-IP, cell lysates Random plasmid integration assay were incubated with M2 anti-flag agarose (A2220; Sigma-Aldrich) NHEJ-mediated repair was assessed using a cell-based plasmid or anti-Ku70 (D10A7; Cell Signaling Technology)/anti-Ku80 integration assay, as previously described by Feng and Chen (26), (111; Invitrogen) antibodies plus protein G agarose beads (Santa with minor modifications. Briefly, cells were transiently cotrans- Cruz Biotechnology). After washing, the pull-down products were fected with an EcoRV/BamHI-linearized pcDNA3.1/Zeo(þ) plas- examined by mass spectrometry or Western blot analysis. mid (Invitrogen) and the pEGFP-C1 plasmid (Clontech). Twenty- For Western blot analysis, samples were separated by SDS- four hours later, the cells were collected, counted, and plated into PAGE and transferred onto a polyvinylidene difluoride mem- two dishes. One day after plating, the cells in one dish were used to brane (Sigma-Aldrich), which was blocked with 5% milk in Tris- determine the transfection efficiency by assessing EGFP expression buffered saline and incubated overnight at 4C with specific (green fluorescence), and the cells in the other dish were incu- antibodies against the following proteins: RFC4 (PA5-21538) bated in selective media containing 300 mg/mL zeocin (Synth- and Ku80 (111; Invitrogen); and GAPDH (D16H11), b-actin gene) for 14 days at 37C to allow colony formation. Colonies (8H10D10), Ku70 (D10A7), cleaved PARP-1 (D64E10), cleaved indicating random plasmid integration were counted after crystal caspase-3 (#9661), gH2AX (D17A3), p-ATM (Ser1981) (D25E5), violet staining. Colony number was normalized to the respective and proliferating cell nuclear antigen (PCNA; D3H8P; Cell Sig- transfection efficiency, and NHEJ activity is shown as a percentage naling Technology). Subsequently, the membrane was incubated relative to the control group. with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (#7074) or anti-mouse IgG (#7076P2; Cell Signaling Tech- Nude mouse xenograft model and radiotherapy nology). Signals were visualized using an ECL detection system Animal experiments were approved by the SYSUCC Institu- (Amersham Biosciences). tional Animal Care and Usage Committee following the Animal Welfare and Rights in China (Approval No.: L102012018006K). Immunofluorescence Female BALB/c nude mice (4–5 weeks old, 15–18 g; SLRC Lab- Cells in confocal dishes were fixed in 4% paraformaldehyde, oratory Animal Co.) were used to generate xenograft models via permeabilized using 0.5% Triton X-100, and blocked with 3% subcutaneous transplant of tumor sections (approximately BSA (Beyotime). The cells were then incubated with anti-gH2AX 5mm3) from colorectal cancer cell xenografts into the right flank. (1:50; D17A3; Cell Signaling Technology) or anti-RAD51 (1:50; Once the xenograft volume reached approximately 150 mm3, the 3C10; Thermo Fisher Scientific) antibodies at 4C overnight, mice were treated with X-rays at 2 Gy/day for 7 consecutive days. A followed by incubation with goat anti-rabbit Alexa Fluor 594 lead plate was used to cover the mouse body except the xenograft (#A11012) or goat anti-mouse Alexa Fluor 488 (#A11001; 1:200; region. Xenograft growth was monitored, and tumor volume was Invitrogen) for 1 hour at room temperature in the dark. The calculated using the following formula: tumor volume ¼ 0.52 samples were costained with DAPI (Beyotime) and examined by width2 length (14, 22). The mice were sacrificed 4 weeks after confocal laser scanning microscopy (20). radiotherapy, and the tumors were removed, weighed, and sub- jected to pathologic analysis. Comet assay DNA damage was assessed using the comet assay (24). Cells Patients were treated with X-rays (6 Gy), cultured for 6 hours, and then A total of 145 patients with LARC with newly diagnosed LARC collected by trypsinization for DNA damage analysis using a who received neoCRT from May 2005 to May 2016 at SYSUCC comet assay kit (KeyGen Biotech) according to the manufacturer's were included in this study. For inclusion, the patients were instructions. After PI staining, comet images were captured by required to have a single primary lesion, have completed standard fluorescence microscopy. The average comet tail moment (per- neoCRT and undergone radical surgical resection, and have sur- centage of DNA in tail tail length) was scored for three fields vived more than 1 month after surgery; complete information was using Comet Assay IV software (Perceptive Instruments). also an inclusion criterion (27). The study was approved by the SYSUCC Ethics Committee (Approval No.: GZR2018-164) and NHEJ and homologous recombination reporter assays conducted in accordance with the Declaration of Helsinki, Reporter constructs for NHEJ and homologous recombination and written informed consent was obtained from each patient. (HR; refs. 25, 26) and the I-SceI expression plasmid pCMV-NLS-I- Pre- and post-neoCRT tumor tissues were obtained by colonos- SceI were gifts from Dr. Lin Feng, Sun Yat-sen University Cancer copy or during surgery, respectively, and then paraffin-embedded Center (SYSUCC, Guangzhou, China). The structure and function for IHC. All patients received standard neoCRT with a total dose of the NHEJ and HR reporters are presented in Supplementary Fig. of 50 Gy (2 Gy/day, 5 days/week, 5 weeks) plus concurrent S2. For NHEJ and HR reporter assays (25), HCT116 and HT29 chemotherapy with fluorouracil (capecitabine 825 mg/m2 days cells were stably transfected with NHEJ or HR reporter constructs 1–5, repeated every week; or 5-FU 400 mg/m2 and leucovorin and selected using G418 (1 mg/mL; Beyotime) for 10 days. G418- 400 mg/m2 day 1 and 5-FU 1,200 mg/m2 continuous intravenous

4570 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

infusion (c.i.v.) 48 hours, repeated every 14 days) or fluorour- cancer cell lines with different genetic backgrounds, HCT116 and acilþoxaliplatin regimen [XELOX (oxaliplatin 100 mg/m2 day 1, HT29 cells, were infected with the lentiviral shRNA library and capecitabine 1,000 mg/m2 twice a day for days 1–14, repeated then exposed to X-rays after puromycin selection. shRNA every 21 days) or mFOLFOX6 (oxaliplatin 85 mg/m2, leucovorin sequences and copy numbers were examined in untreated cells 400 mg/m2, and 5-FU 400 mg/m2, day 1; 5-FU 1,200 mg/m2 c.i.v. and in those that survived 6 Gy of X-ray exposure; following 48 hours; repeated every 14 days)]. MAGeCK-negative selection, 757 and 565 candidate genes with a Tumor responsiveness to neoCRT was evaluated using the depleted shRNA copy number in surviving cells (Supplementary tumor regression grade (TRG), as proposed in the 8th Edition of Tables S5 and S6) were selected in HCT116 and HT29 cells, the American Joint Committee on Cancer (AJCC) Staging Manual respectively. The intersection of these two candidate gene sets System, which is based on residual neoplastic cells, cytologic identified 6 genes, RFC4, NCAPH, SYNE3, LDLRAD2, NHP2, and changes (including nuclear pyknosis or necrosis and/or eosino- FICD, as candidate radioresistance genes (Fig. 1A). The copy philia), and stromal changes (including dense or edematous number changes of shRNAs targeting these 6 genes in HCT116/ fibrosis with or without inflammatory infiltrate and giant cell HT29 cells are shown in Supplementary Fig. S4A, and these granulomatosis around ghost cells and keratin; ref. 28). shRNA sequences are listed in Supplementary Table S2. After surgery, 124 of the patients received and completed To confirm the screening results, we constructed HCT116 cell adjuvant chemotherapy with a capecitabine, XELOX, or mFOL- lines with stable gene knockdown (Supplementary Fig. S4B) FOX6 regimen. The detailed characteristics of the patients are via infection with lentiviruses encoding representative shRNAs presented in Supplementary Table S4. (Supplementary Table S2) targeting these 6 candidate genes; we then performed a clonogenic assay after exposure to X-rays. The IHC assay results showed that knockdown of RFC4, NCAPH,andFICD IHC assays were conducted to detect protein levels in tissues, as significantly impaired radioresistance in HCT116 cells; indeed, described previously (29). Briefly, 4-mm–thick formalin-fixed, cells with RFC4 knockdown were the most sensitive to X-rays paraffin-embedded tissue sections were deparaffinized and rehy- (Fig. 1B; Supplementary Fig. S4C). In addition, comparative drated, followed by antigen retrieval and endogenous peroxidase expression analysis of these 6 genes revealed significantly inactivation. After blocking, the slides were incubated overnight at increased mRNA levels of RFC4, NCAPH,andFICD in radio- 4C with anti-RFC4 (1:400; #PA5-21538; Thermo Fisher Scien- resistant HCT1166GyR cells compared with parental HCT116 tific), anti-cleaved caspase-3 (1:200, #9661; Cell Signaling Tech- cells, and the increase in RFC4 mRNA was the most notable nology), or anti-gH2AX (1:100; D17A3; Cell Signaling Technol- (Supplementary Fig. S4D). ogy) antibodies; incubated with a HRP-conjugated secondary Furthermore, radiation induced the upregulation of RFC4 antibody; visualized using an Envision Detection Kit (Dako); and expression in HCT116 cells in vitro in a dose-dependent manner, counterstained with hematoxylin. RFC4 levels were independent- and RFC4 protein levels were significantly higher in residual ly and semiquantitatively assessed by two pathologists using the tumor tissue after neoCRT than in biopsy specimens before immunoreactive score (IRS), which was calculated as the product neoCRT (Supplementary Fig. S5A–S5C). These findings indicate of the intensity score (0, absent; 1, weak; 2, moderate; and 3, that RFC4 may be an important factor in the radioresistance of strong; Supplementary Fig. S3) and the positive cell percentage colorectal cancer. score (0, 0%; 1, <10%; 2, 11%–50%; 3, 51% to 80%; and 4; 80%); thus, the IRS ranged from 0 to 12 (30). The cut-off value RFC4 facilitates colorectal cancer cell survival after X-ray for RFC4 level was deduced according to overall survival (OS) exposure in vitro using an ROC curve, and patients were categorized into two To further explore the effect of RFC4 on radioresistance in groups based on RFC4 level (high or low; ref. 19). colorectal cancer, we stably overexpressed RFC4 in HCT116 and HCT15 cells with lower RFC4 levels, knocked down RFC4 in HT29 Statistical analysis and SW480 cells with higher RFC4 levels (Supplementary Fig. All in vitro experiments were repeated at least three times, and S5D), and performed clonogenic assays after X-ray treatment. The the animal experiments were repeated twice. The data were results showed that RFC4 overexpression significantly enhanced analyzed using GraphPad Prism 5.0 or SPSS 20.0 software (SPSS colony formation in HCT116 and HCT15 cells after radiation, Inc.). Differences in the average among two or more groups were whereas RFC4 knockdown significantly weakened colony forma- compared using Student t test or ANOVA; associations between tion in HT29 and SW480 cells after radiation (Fig. 1C; Supple- RFC4 levels and clinicopathologic parameters were assessed using mentary Fig. S6). Moreover, restoring RFC4 expression in RFC4- the x2 test. Survival was analyzed using the Kaplan–Meier method silenced colorectal cancer cells via transient transfection of an and the log-rank test; associations of variables with survival were RFC4 expression plasmid rescued the impaired colony formation analyzed using a univariate or multivariate Cox proportional after radiation (Fig. 1D). hazards model. A P value <0.05 was considered to indicate a The Western blot results showed that PARP-1 and caspase-3 significant difference. cleavage induced by radiation was inhibited in RFC4-overexpressing HCT116 cells and enhanced in RFC4-silenced HT29 cells compared with the respective controls (Fig. 1E). Annexin V/PI Results double-staining assays revealed that RFC4 overexpression sup- Identification of RFC4 as a radioresistance factor in colorectal pressed radiation-induced apoptosis in HCT116 cells and that cancer RFC4 knockdown promoted radiation-induced apoptosis in To identify radioresistance factors in colorectal cancer, we HT29 cells (Fig. 1F; Supplementary Fig. S7). These findings developed a high-throughput RNA interference screen using the indicate that RFC4 protects colorectal cancer cells from radia- LentiPlex human whole-genome shRNA library. Two colorectal tion-induced apoptosis.

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4571

Wang et al.

Figure 2. RFC4 promotes colorectal cancer cell radioresistance in a nude mouse xenograft model. A and B, Assessment of radiotherapy efficacy in colorectal cancer xenografts in nude mice. Transplanted xenografts were established with RFC4-overexpressing (RFC4) and vector-transfected (Vec) HCT116 cells (A)orRFC4- knockdown (sh1/sh2) and negative control (shNC) HT29 cells in the right flank of BALB/c nude mice, which were treated with X-rays (2 Gy/day) for 7 consecutive days; tumor volume and weight were measured (left, removed xenografts; middle, tumor volume; right, average tumor weight; B). C, Representative images of IHC staining for RFC4 (top) and cleaved caspase-3 (middle) in the harvested xenografts and H&E staining (bottom) showing the distribution of tumor cells (red arrow) and the infiltration of inflammatory cells (green arrow; , P < 0.01; , P < 0.001; n.s., not significant).

RFC4 promotes the radioresistance of colorectal cancer cells in ment of subcutaneous colorectal cancer xenografts, the mice a nude mouse xenograft model were treated with X-rays (2 Gy/day for 7 consecutive days). As Next, we investigated the effect of RFC4 on the response to shown in Fig. 2A, the xenografts originating from HCT116 cells radiation in a nude mouse xenograft model. After establish- overexpressing RFC4 (HCT116-RFC4) were larger and heavier

4572 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

than those originating from vector-transfected HCT116 cells RFC4 is involved in NHEJ-mediated DNA repair (HCT116-vec) after radiotherapy (P < 0.001). Furthermore, the As NHEJ and HR are the two main DSB repair pathways, we first tumor inhibition ratio (radiotherapy effect) was significantly determined which pathway is influenced by RFC4 in DSB repair in attenuated in the HCT116-RFC4 group compared with the colorectal cancer cells using NHEJ and HR reporter systems HCT116-vec group (45.4% vs. 86.6%). Conversely, the xeno- (Supplementary Fig. S2; ref. 25). The results showed that RFC4 grafts derived from HT29 cells with RFC4 knockdown (HT29- knockdown significantly decreased NHEJ activity (Fig. 4A) but did sh1/sh2) were smaller and weighed less than those derived not affect HR activity (Fig. 4B), suggesting that RFC4 is likely from HT29-shNC cells after radiotherapy (P < 0.01), and the involved in NHEJ but not in HR during DSB repair in colorectal radiotherapy effect was markedly enhanced in the HT29-sh1/ cancer cells. sh2 groups compared with the HT29-shNC group (92.4% and Because the random integration of plasmids into genomic DNA 95.1% vs. 75.9%; Fig. 2B). IHC and hematoxylin and eosin is considered an NHEJ-mediated DNA repair mechanism (26, 31), (H&E) staining assays showed weaker cleaved caspase-3 stain- we performed a random plasmid integration assay to confirm ing and less inflammatory cell infiltration in RFC4-overexpres- whether RFC4 can regulate NHEJ repair. Consistent with the NHEJ sing HCT116 xenografts, and the opposite effects were found in reporter assay results, RFC4-overexpressing cells exhibited a sig- RFC4-silenced HT29 xenografts (Fig. 2C). These findings indi- nificantly enhanced NHEJ repair capacity, whereas RFC4-silenced cate that RFC4 confers radioresistance on colorectal cancer cells cells displayed reduced NHEJ activity (Fig. 4C; Supplementary Fig. by protecting them from radiation-induced apoptosis both in S9A). In addition, we detected RAD51 foci, a core sign of HR vitro and in vivo. repair (32), and found no difference in the number of RAD51 foci between RFC4-knockdown colorectal cancer cells and control RFC4 promotes radioresistance by facilitating DSB repair in cells (Supplementary Fig. S9B). These data suggest that RFC4 colorectal cancer cells facilitates NHEJ-mediated DSB repair in colorectal cancer cells. It is well known that radiation causes cellular DNA DSBs and We also investigated the effects of other components of the RFC that phospho-H2AX (g-H2AX) levels and foci are sensitive DSB complex on the NHEJ pathway. Similar to RFC4, knockdown of markers. Therefore, we investigated the influence of RFC4 on RFC5 significantly decreased NHEJ activity, but knocking down g-H2AX levels in X-ray–treated colorectal cancer cells by Western RFC1, RFC2,orRFC3 had no obvious effect on NHEJ pathway blot and IF assays. The results showed lower g-H2AX levels and activity (Supplementary Fig. S9C). fewer foci in HCT116-RFC4 cells than in HCT116-vec control cells at 12 hours posttreatment with 6 Gy X-rays; conversely, signifi- RFC4 regulates NHEJ repair by interacting with Ku70/80 cantly higher g-H2AX levels and more foci were observed in HT29- To investigate the molecular mechanism by which RFC4 reg- sh1/sh2 cells than in HT29-shNC cells at 12 hours after irradiation ulates NHEJ repair in colorectal cancer cells, we screened proteins (Fig. 3A and B). Moreover, IHC assays indicated significantly interacting with RFC4 via IP and mass spectrometry (Supplemen- lower g-H2AX levels in HCT116-RFC4 xenografts than in tary Tables S7 and S8) and found that both Ku70 and Ku80, which HCT116-vec xenografts at 4 weeks after radiotherapy but higher are core components of the NHEJ complex, bind to RFC4 in g-H2AX levels in HT29-sh1/sh2 xenografts than in control xeno- HCT116 and HT29 cells. Furthermore, co-IP–Western blot assays grafts (Fig. 3C). revealed that Ku70 and Ku80 were pulled down by RFC4 (FLAG) Next, we measured DSB levels by the comet assay, in which the in lysates of both radiation-treated and -untreated HCT116 cells mean tail moment after X-ray treatment was quantified using (Fig. 4D). Moreover, endogenous RFC4 was pulled down by both image analysis. We found that HCT116-RFC4 cells had signifi- Ku70 and Ku80 in lysates of HCT116 cells treated with or without cantly shorter comet tails and that HT29-sh1/sh2 cells displayed X-rays (Fig. 4E). To verify the impact of RFC4 binding to Ku70 and longer comet tails than did their control cells at 6 hours after Ku80, we examined the effects of RFC4 on radioresistance after radiation exposure (Fig. 3D). silencing Ku70 and Ku80. The results showed that Ku70 or Ku80 To clarify whether RFC4 affects radiation-induced DNA silencing almost completely abrogated the RFC4-induced radio- damage or DNA repair, we detected g-H2AX levels at different resistance of HCT116 cells (Fig. 4F), suggesting that the ability of times after radiation. The results demonstrated that g-H2AX RFC4 to promote radioresistance in colorectal cancer may depend levels immediately and dramatically increased after radiation, on its interaction with Ku70/80. peaking at 0.5 hour in HCT116 cells and at 2 hours in HT29 As it has been reported that the RFC complex interacts with cells, and then gradually decreased over time. No obvious PCNA and promotes its loading onto chromatin, which is difference in g-H2AX levels was found between HCT116- involved in DNA damage repair (33, 34), we investigated whether RFC4/HT29-sh1 cells and their relevant control cells before PCNA is involved in regulating RFC4 during DSB repair in g-H2AX levels peaked; however, g-H2AX levels decreased more colorectal cancer cells. Although RFC4 overexpression increased rapidly in colorectal cancer cells with higher RFC4 levels radioresistance similarly in PCNA-silenced HCT116 cells and (HCT116-RFC4 or HT29-shNC) than in those with lower RFC4 control cells, PCNA silencing significantly impaired radiation levels (HCT116-vec or HT29-sh1, respectively; Fig. 3E; Supple- resistance (Supplementary Fig. S9D). Taken together, our data mentary Fig. S8A). Further investigation revealed that ATM indicate that RFC4 interacts with Ku70 and Ku80 to promote the phosphorylation levels were similar in colorectal cancer cells NHEJ pathway, which is involved in DSB repair in colorectal with high or low RFC4 levels in the early stages after X-ray cancer cells. exposure (Supplementary Fig. S8B). These results suggest that RFC4 may not affect early activation of the DNA damage High RFC4 levels are associated with radioresistance and poor response but might promote consequent DNA repair. Taken prognosis in patients with LARC treated with neoRT together, the data indicate that RFC4 promotes DNA DSB repair Because RFC4 was identified as a colorectal cancer radioresis- in colorectal cancer cells. tance factor in our cellular and animal experiments, we sought to

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4573

Wang et al.

Figure 3. RFC4 promotes radioresistance by facilitating DNA DSB repair in colorectal cancer cells. A and B, The effects of RFC4 on g-H2AX levels and foci in colorectal cancer cells treated with X-rays by Western blot (A) and IF (B: left, representative IF images; right, foci number is presented as the mean SD of three independent experiments). C, IHC assay for g-H2AX levels in harvested xenografts (representative images). D, Comet assay for DNA damage in colorectal cancer cells at 6 hours after treatment with 6 Gy X-rays (left, representative pictures; right, bar charts indicating the average tail moment per cell). E, Western blot analysis of the time course of changes in g-H2AX levels in colorectal cancer cells treated with X-rays (loading control: GAPDH). Vec, lentiviral vector; RFC4, lentiviral vector expressing RFC4; shNC, negative control shRNA; sh1/2, shRNA shRFC4-1/2; , P < 0.01).

4574 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

Figure 4. RFC4 promotes NHEJ-mediated DSB repair by interacting with Ku70/Ku80. NHEJ (A) and HR reporter assays (B) showing the effects of RFC4 on DNA DSB repair in HT29 cells (top, NHEJ or HR repair activity presented as the relative GFPþ fraction normalized to the shNC group; bottom, Western blot analysis of RFC4 expression). C, Relative NHEJ activity in colorectal cancer cells was assessed using a random plasmid integration assay. D and E, Co-IP and Western blot analyses. HCT116-Vec () and HCT116-RFC4 (þ)(D) or HCT116 cells were treated with X-rays (0 or 6 Gy; E). After 12 hours, cell lysates were immunoprecipitated with Flag (D) or Ku70/Ku80 antibodies (E), followed by Western blot using the indicated antibodies. F, The radiosensitivity of HCT116 (Vec or RFC4) cells was evaluated by colony formation assays after silencing Ku70 or Ku80. Top, representative pictures; middle, colony number/well (mean SD of at least three independent experiments); bottom, Western blot analysis for RFC4, Ku70,andKu80 expression (loading control: actin; , P < 0.01; , P < 0.001; n.s., not significant). determine whether RFC4 has consistent effects on radiotherapy in Using ROC curve analysis, we obtained a cut-off value for RFC4 clinical patients with colorectal cancer. Thus, we investigated levels in colorectal cancer tissues, 6.8; this value optimally clas- RFC4 expression levels in pretreatment biopsy tumor specimens sified patients into two groups with high or low RFC4 levels. from 145 patients with LARC who received neoCRT followed by Further analysis showed that a high RFC4 level was significantly surgery with or without adjuvant chemotherapy and analyzed the associated with a high TRG (P < 0.001) and poor survival (P ¼ relationship between RFC4 levels and TRG. According to our 0.034; Supplementary Table S4). Kaplan–Meier survival analysis results, RFC4 levels were significantly lower in tumor tissues from indicated that patients with high RFC4 levels had a shorter patients with TRG 0 (pathologic complete response) than in those progression-free survival (PFS, P ¼ 0.035) and OS (P ¼ from patients with TRG 1–3 (Fig. 5A). 0.006; Fig. 5B and C). Moreover, multivariate Cox proportional

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4575

Wang et al.

Figure 5. RFC4 expression is associated with TRG and prognosis in patients with LARC treated with neoCRT. Relationships of RFC4 expression level in pretreatment biopsy cancer tissues with TRG and survival were investigated in 145 patients with LARC treated with neoCRT. A, RFC4 levels (IHC scores) in pretreatment biopsy tissues with different TRGs. PFS (B) and OS (C) according to high or low RFC4 expression levels (log-rank test). CI, conﬁdence interval (, P < 0.01; , P < 0.001).

hazard regression analysis revealed that the RFC4 level, together neoCRT. Altogether, RFC4 was found to potentially confer radio- with the pathologic grade and/or surgical procedure, could serve resistance in colorectal cancer; thus, a high RFC4 level might as an independent prognostic factor (Table 1). These findings predict poor tumor regression and prognosis in patients with indicate that high RFC4 levels in colorectal cancer tissues may LARC receiving preoperative neoRT. result in resistance to neoCRT. Regardless, we could not distinguish whether RFC4 confers resistance to radiotherapy, chemotherapy, or both based on these Discussion data. Therefore, we performed in vitro cell viability and colony RFC is a conserved eukaryotic complex consisting of five formation assays using colorectal cancer cells to investigate the distinct subunits, a large subunit, RFC1, and four small subunits, effects of RFC4 on resistance to the chemotherapy drugs 5-fluo- RFC2 to RFC5, which play essential roles in DNA replication and rouracil, oxaliplatin, doxorubicin, and etoposide. Overall, RFC4 postincision nucleotide excision repair (35, 36). RFC functions by promoted colorectal cancer cell resistance to doxorubicin and loading the PCNA sliding clamp complex onto double-stranded etoposide, which usually cause cellular DSBs, but not to 5-fluo- DNA at primer–template junctions (33, 34), and the ATP-binding rouracil and oxaliplatin, which were included in the chemother- domains of RFC2/3/4 are essential for DNA recognition and apy regimens for the abovementioned patients (Supplementary clamp loading (37). By interacting with RPA1, RFC4 is required Fig. S10). These findings suggest that high RFC4 expression for both DNA replication and DNA damage checkpoints in confers resistance to radiotherapy but not to chemotherapy in Saccharomyces cerevisiae (38).

Table 1. Univariate and multivariate Cox regression analyses of prognostic factors for patients with LARC PFS OS Univariate Multivariate Univariate Multivariate Variables HR (95% CI) P HR (95% CI) P HR (95% CI) P HR (95% CI) P Gender (male vs. female) 1.947 (1.035–3.662) 0.035a 1.341 (0.659–2.727) 0.417 Age (>57 years vs. 57 years) 0.993 (0.533–1.848) 0.982 1.195 (0.595–2.400) 0.616 Pathologic grade (III vs. II) 3.203 (1.698–6.042) <0.001b 3.101 (1.626–5.915) 0.001c 3.022 (1.491–6.125) 0.001c 2.805 (1.351–5.821) 0.006c cT (T4 vs. T2–T3) 1.138 (0.611–2.122) 0.683 1.252 (0.625–2.508) 0.526 cN (N1–N2 vs. N0) 0.675 (0.356–1.282) 0.227 0.818 (0.394–1.700) 0.590 CEA (ng/mL) (5 vs. <5) 1.599 (0.852–3.001) 0.141 1.718 (0.847–3.483) 0.129 CA199 (kU/L) (>35 vs. 35) 2.219 (1.092–4.508) 0.024a 1.822 (0.806–4.119) 0.144 Neoadjuvant chemotherapy 0.657 (0.248–1.737) 0.393 0.727 (0.213–2.478) 0.609 (FU vs. FUþOxa) Adjuvant chemotherapy 0.429 (0.209–0.881) 0.018a 0.435 (0.203–0.928) 0.031a 0.389 (0.178–0.852) 0.015a (yes vs. no) Surgical procedure 2.166 (1.155–4.060) 0.014a 2.392 (1.250–4.574) 0.008c 2.378 (1.173–4.819) 0.013a 2.474 (1.206–5.076) 0.013a (Dixon/Hartman vs. Miles) TRG (2–3 vs. 0–1) 2.141 (1.149–3.991) 0.014a 2.129 (1.058–4.282) 0.030a RFC4 level (high vs. low) 2.133 (1.038–4.381) 0.035a 2.266 (1.079–4.759) 0.031a 3.268 (1.337–7.989) 0.006c 3.421 (1.387–8.438) 0.008c Abbreviations: CI, conﬁdence interval; cN, clinical N stage; cT, clinical T stage; FU, capecitabine or 5-ﬂuorouracilþleucovorin; FUþOxa, XELOX, or mFOLFOX6. aP < 0.05. bP < 0.001. cP < 0.01.

4576 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

It is well known that radiation induces cellular DNA DSBs to kill colorectal cancer cells to these treatments with no obvious effect tumor cells and that intensive repair of DSBs results in radio- on other treatments. Considering that 5-fluorouracil and oxali- resistance in cancer cells (39, 40). HR and NHEJ are the major DSB platin, but not doxorubicin and etoposide, are commonly used in repair pathways in higher eukaryotes (41, 42). HR results in colorectal cancer chemotherapy, the impaired tumor regression precise repair of the DNA lesion but requires the presence of and poor prognosis of patients with LARC treated with neoCRT homologous sequences, whereas NHEJ results in error-prone are likely due to radiation resistance resulting from strong RFC4 repair because of its independence of DNA sequence homology. expression in cancer cells, not resistance to fluorouracil and/or In this study, RFC4 was identified as a radioresistance factor that oxaliplatin. facilitates NHEJ but not HR in DSB repair via interactions with RFC4 was found to be highly expressed in cancer tissues and Ku70/Ku80 (Fig. 4). The core NHEJ complex consists of Ku80 cells, including colorectal cancer (49), and RFC4 expression is (XRCC5), Ku70 (XRCC6), and DNA-dependent protein kinase, associated with tumor progression and prognosis in patients with and formation of the Ku70-Ku80 heterodimer and subsequent colorectal cancer (50). Nonetheless, according to the results of our binding to DNA DSB ends are the initiating events of NHEJ (43, investigation and a previous report (50), RFC4 levels were not 44). Therefore, we speculate that RFC4 may promote Ku80–Ku70 correlated with the T and N stages of patients with colorectal heterodimer recruitment and retention at DSB sites in colorectal cancer. Moreover, RFC4 did not affect the colony formation cancer cells. Knockdown of Ku70 or Ku80 almost completely ability of colorectal cancer cells in vitro (Supplementary Figs. S6 reversed RFC4-induced radiation resistance in colorectal cancer and S10B) or the growth of colorectal cancer xenografts in nude cells (Fig. 4F), suggesting that the interaction of RFC4 with Ku70/ mice (Fig. 2), even though RFC4 plays an important role in Ku80 is required for efficient NHEJ repair. radioresistance. Therefore, the influence of RFC4 on colorectal In general, RFC4 functions as a component of the RFC com- cancer tumorigenesis and progression is still unknown. In addi- plex (35). However, we found that knocking down RFC4 or RFC5, tion, an analysis of a cohort of patients with early-stage colorectal but not other RFC subunits, significantly reduced NHEJ activity cancer who received only surgery but not neoadjuvant or adjuvant (Supplementary Fig. S9C). Moreover, RFC4 was the only RFC therapy is needed to clarify the relationship between RFC4 component selected in our high-throughput RNA interference expression and colorectal cancer progression and prognosis. screen (Supplementary Tables S5 and S6). Thus, it is essential to Frequent amplification of the RFC4 gene has been reported further investigate whether RFC4 regulates NHEJ-mediated DNA in lung, breast, and cervical cancer, but RFC4 gene amplifica- repair by itself or as part of the complete RFC complex. tion has not been detected in colorectal cancer (49). We It has been reported that RFC is involved in nucleotide excision observed that radiation induced the upregulation of RFC4 repair by loading the PCNA sliding clamp complex onto expression in colorectal cancer tissues and cells and that RFC4 DNA (36). However, our study showed that PCNA knockdown levels were higher in radioresistant colorectal cancer cells than did not impair RFC4-enhanced radioresistance in colorectal can- in parental cells (Supplementary Figs. S4D and S5). We pre- cer cells (Supplementary Fig. S9D), suggesting that the effects of sume that the radiation-induced upregulation of RFC4 is likely RFC4 on the repair of radiation-induced DNA damage may be one of the causes of induced secondary radioresistance. How- independent of PCNA loading onto DNA. Our findings may ever, the exact causes and mechanisms of RFC4 upregulation in provide a novel regulatory mechanism for NHEJ-mediated DSB colorectal cancer tissues need to be further clarified in future repair. investigations. Resistance to CRT is the main cause of treatment failure or In summary, RFC4 was found to act as a resistance factor that relapse in colorectal cancer (45, 46). Our clinical investigation protects colorectal cancer cells from radiation-induced DSBs and revealed that higher RFC4 expression in cancer cells was signif- apoptosis in vitro as well as in nude mouse colorectal cancer icantly associated with weaker tumor regression after neoCRT xenografts. Mechanistically, RFC4 promoted NHEJ-mediated (Supplementary Table S4); moreover, RFC4 upregulation pre- DNA DSB repair by interacting with Ku70/Ku80. Moreover, RFC4 dicted a poor prognosis in patients with LARC who received upregulation in cancer cells was associated with reduced tumor neoCRT (Fig. 5). Indeed, patients with LARC with lower RFC4 regression and poor prognosis in patients with LARC treated with expression were more likely to experience better tumor regression neoRT. Therefore, RFC4 levels might serve as an effective predic- and survival outcomes. In addition to the long-course radiother- tive biomarker of radiation sensitivity and a target for radio- apy used in this study, short-course radiotherapy (5 Gy/day, sensitization in patients with LARC; these findings may contribute 5 days/week, 1 week) is administered as neoadjuvant therapy in to the development of precise treatment strategies in clinical some patients with stage T3 rectal cancer, based on NCCN guide- practice in the future. lines. According to data obtained from a nude mouse xenograft experiment, RFC4 promoted resistance to X-rays in colorectal Disclosure of Potential Conflicts of Interest cancer cells after only 7 consecutive days of treatment. Thus, we No potential conflicts of interest were disclosed. presume that RFC4 might also be involved in resistance to short- course radiotherapy, which should be examined in future studies. We also found that RFC4 overexpression greatly enhanced the Authors' Contributions resistance of colorectal cancer cells to doxorubicin and etoposide Conception and design: X. Yue, W. Huang, R.-Y. Liu but did not affect their sensitivity to 5-fluorouracil or oxaliplatin. Acquisition of data (provided animals, acquired and managed patients, Fluorouracil is an anticancer antimetabolite, and platinum causes provided facilities, etc.): X.-C. Wang, X. Yue, R.-X. Zhang, T.-Y. Liu, major single-strand DNA damage (47); doxorubicin and etopo- M.-J. Yang, Z.-H. Lu, J.-H. Peng, G.-Y. Wang, Q.-H. Peng, Y. Meng Analysis and interpretation of data (e.g., statistical analysis, biostatistics, side mainly cause DSBs in tumor cells (48). Therefore, we spec- computational analysis): X.-C. Wang, X. Yue, L.-Y. Le ulate that RFC4 might promote the repair of DSBs induced by Writing, review, and/or revision of the manuscript: X.-C. Wang, X. Yue, radiation or chemotherapy and then increase the resistance of R.-Y. Liu

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4577

Wang et al.

Administrative, technical, or material support (i.e., reporting or organizing by the National Natural Science Foundation of China (nos. 81871996 and data, constructing databases): R.-X. Zhang, Z.-Y. Wang 81803054) and Guangzhou Science and Technology Program key projects (no. Study supervision: Z.-Z. Pan, W. Huang, R.-Y. Liu 201704030037).

Acknowledgments The costs of publication of this article were defrayed in part by the We thank Prof. Jiahuai Han's laboratory (Xiamen University, Xiamen, China) payment of page charges. This article must therefore be hereby marked for providing RFC4 cDNA, Dr. Lin Feng (SYSUCC) for her technical guidance, advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate and the editorial ofﬁce of Cancer Communication and American Journal Experts this fact. for language editing. The authenticity of this article was validated by uploading the raw data to the Research Data Deposit public platform (www.researchdata. Received November 16, 2018; revised January 14, 2019; accepted April 9, org.cn) with approval number RDDB2019000520. This project was supported 2019; published ﬁrst April 12, 2019.

References 1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global 17. Li W, Xu H, Xiao T, Cong L, Love MI, Zhang F, et al. MAGeCK enables robust cancer statistics 2018: GLOBOCAN estimates of incidence and mortality identiﬁcation of essential genes from genome-scale CRISPR/Cas9 knock- worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68: out screens. Genome Biol 2014;15:554. 394–424. 18. Liu RY, Dong Z, Liu J, Zhou L, Huang W, Khoo SK, et al. Overexpression of 2. Gu X, Zheng R, Xia C, Zeng H, Zhang S, Zou X, et al. Interactions between asparagine synthetase and matrix metalloproteinase 19 confers cisplatin life expectancy and the incidence and mortality rates of cancer in China: a sensitivity in nasopharyngeal carcinoma cells. Mol Cancer Ther 2013;12: population-based cluster analysis. Cancer Commun 2018;38:44. 2157–66. 3. Zheng R, Zeng H, Zhang S, Chen W. Estimates of cancer incidence and 19. Yu H, Ye W, Wu J, Meng X, Liu RY, Ying X, et al. Overexpression of sirt7 mortality in China, 2013. Chin J Cancer 2017;36:66. exhibits oncogenic property and serves as a prognostic factor in colorectal 4. van Gijn W, Marijnen CA, Nagtegaal ID, Kranenbarg EM, Putter H, cancer. Clin Cancer Res 2014;20:3434–45. Wiggers T, et al. Preoperative radiotherapy combined with total mesor- 20. Du WY, Lu ZH, Ye W, Fu X, Zhou Y, Kuang CM, et al. The loss-of-function ectal excision for resectable rectal cancer: 12-year follow-up of the mutations and down-regulated expression of ASB3 gene promote the multicentre, randomised controlled TME trial. Lancet Oncol 2011;12: growth and metastasis of colorectal cancer cells. Chin J Cancer 2017;36:11. 575–82. 21. Yue X, Zuo Y, Ke H, Luo J, Lou L, Qin W, et al. Identiﬁcation of 4-arylidene 5. Ma B, Gao P, Wang H, Xu Q, Song Y, Huang X, et al. What has preoperative curcumin analogues as novel proteasome inhibitors for potential antican- radio(chemo)therapy brought to localized rectal cancer patients in terms of cer agents targeting 19S regulatory particle associated deubiquitinase. perioperative and long-term outcomes over the past decades? A systematic Biochem Pharmacol 2017;137:29–50. review and meta-analysis based on 41,121 patients. Int J Cancer 2017;141: 22. Kuang CM, Fu X, Hua YJ, Shuai WD, Ye ZH, Li Y, et al. BST2 confers cisplatin 1052–65. resistance via NF-kappaB signaling in nasopharyngeal cancer. Cell Death 6. Allegra CJ, Yothers G, O'Connell MJ, Beart RW, Wozniak TF, Pitot HC, et al. Dis 2017;8:e2874. Neoadjuvant 5-FU or capecitabine plus radiation with or without oxali- 23. Liu RY, Zhu YH, Zhou L, Zhao P, Li HL, Zhu LC, et al. Adenovirus-mediated platin in rectal cancer patients: a phase III randomized clinical trial. J Natl delivery of interferon-gamma gene inhibits the growth of nasopharyngeal Cancer Inst 2015;107:pii: djv248. carcinoma. J Transl Med 2012;10:256. 7. Fokas E, Fietkau R, Hartmann A, Hohenberger W, Grutzmann R, Ghadimi 24. Nandhakumar S, Parasuraman S, Shanmugam MM, Rao KR, Chand P, Bhat M, et al. Neoadjuvant rectal score as individual-level surrogate for disease- BV. Evaluation of DNA damage using single-cell gel electrophoresis (Com- free survival in rectal cancer in the CAO/ARO/AIO-04 randomized phase III et Assay). J Pharmacol Pharmacother 2011;2:107–11. trial. Ann Oncol 2018;29:1521–7. 25. Seluanov A, Mao Z, Gorbunova V. Analysis of DNA double-strand break 8. Smith JJ, Garcia-Aguilar J. Advances and challenges in treatment of locally (DSB) repair in mammalian cells. J Vis Exp 2010;43:e2002. advanced rectal cancer. J Clin Oncol 2015;33:1797–808. 26. Feng L, Chen J. The E3 ligase RNF8 regulates KU80 removal and NHEJ 9. Haan JC, Labots M, Rausch C, Koopman M, Tol J, Mekenkamp LJ, et al. repair. Nat Struct Mol Biol 2012;19:201–6. Genomic landscape of metastatic colorectal cancer. Nat Commun 2014;5: 27. Zhang RX, Zhou ZG, Lu SX, Lu ZH, Wan DS, Pan ZZ, et al. Pim-3 as a 5457. potential predictor of chemoradiotherapy resistance in locally advanced 10. D'Alterio C, Avallone A, Tatangelo F, Delrio P, Pecori B, Cella L, et al. A rectal cancer patients. Sci Rep 2017;7:16043. prognostic model comprising pT stage, N status, and the chemokine 28. Vecchio FM, Valentini V, Minsky BD, Padula GD, Venkatraman ES, Bal- receptors CXCR4 and CXCR7 powerfully predicts outcome in neoadjuvant ducci M, et al. The relationship of pathologic tumor regression grade (TRG) resistant rectal cancer patients. Int J Cancer 2014;135:379–90. and outcomes after preoperative therapy in rectal cancer. Int J Radiat Oncol 11. Ross JS, Torres-Mora J, Wagle N, Jennings TA, Jones DM. Biomarker-based Biol Phys 2005;62:752–60. prediction of response to therapy for colorectal cancer: current perspective. 29. Fu X, Hu J, Han HY, Hua YJ, Zhou L, Shuai WD, et al. High expression of Am J Clin Pathol 2010;134:478–90. XPA confers poor prognosis for nasopharyngeal carcinoma patients treated 12. Sims D, Mendes-Pereira AM, Frankum J, Burgess D, Cerone MA, Lombar- with platinum-based chemoradiotherapy. Oncotarget 2015;6:28478–90. delli C, et al. High-throughput RNA interference screening using pooled 30. Al-Azhri J, Zhang Y, Bshara W, Zirpoli G, McCann SE, Khoury T, et al. Tumor shRNA libraries and next generation sequencing. Genome Biol 2011;12: expression of vitamin D receptor and breast cancer histopathological R104. characteristics and prognosis. Clin Cancer Res 2017;23:97–103. 13. Zhou Y, Zhu S, Cai C, Yuan P, Li C, Huang Y, et al. High-throughput 31. Galanty Y, Belotserkovskaya R, Coates J, Polo S, Miller KM, Jackson SP. screening of a CRISPR/Cas9 library for functional genomics in human cells. Mammalian SUMO E3-ligases PIAS1 and PIAS4 promote responses to Nature 2014;509:487–91. DNA double-strand breaks. Nature 2009;462:935–9. 14. Li Y, Ye Z, Chen S, Pan Z, Zhou Q, Li YZ, et al. ARHGEF19 interacts with 32. Baumann P, West SC. Role of the human RAD51 protein in homologous BRAF to activate MAPK signaling during the tumorigenesis of non-small recombination and double-stranded-break repair. Trends Biochem Sci cell lung cancer. Int J Cancer 2018;142:1379–91. 1998;23:247–51. 15. Zhou QH, Deng CZ, Chen JP, Huang KB, Liu TY, Yao K, et al. Elevated serum 33. Yao NY, O'Donnell M. The RFC clamp loader: structure and function. LAMC2 is associated with lymph node metastasis and predicts poor Subcell Biochem 2012;62:259–79. prognosis in penile squamous cell carcinoma. Cancer Manag Res 2018; 34. Waga S, Stillman B. The DNA replication fork in eukaryotic cells. Annu Rev 10:2983–95. Biochem 1998;67:721–51. 16. Bassik MC, Lebbink RJ, Churchman LS, Ingolia NT, Patena W, LeProust EM, 35. Shiomi Y, Nishitani H. Control of genome integrity by RFC complexes; et al. Rapid creation and quantitative monitoring of high coverage shRNA conductors of PCNA loading onto and unloading from chromatin during libraries. Nat Methods 2009;6:443–5. DNA replication. Genes 2017;8:52.

4578 Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research

RFC4, a Radioresistance Factor, Promotes NHEJ Repair in Colorectal Cancer

36. Riedl T, Hanaoka F, Egly JM. The comings and goings of nucleotide excision 43. Burma S, Chen BP, Chen DJ. Role of non-homologous end joining (NHEJ) repair factors on damaged DNA. EMBO J 2003;22:5293–303. in maintaining genomic integrity. DNA Repair 2006;5:1042–8. 37. SchmidtSL,GomesXV,BurgersPM.ATP utilization by yeast replication 44. Lieber MR. The mechanism of human nonhomologous DNA end joining. factor C. III. The ATP-binding domains of Rfc2, Rfc3, and Rfc4 are J Biol Chem 2008;283:1–5. essential for DNA recognition and clamp loading. J Biol Chem 2001; 45. Fanale D, Castiglia M, Bazan V, Russo A. Involvement of non-coding RNAs 276:34784–91. in chemo- and radioresistance of colorectal cancer. Adv Exp Med Biol 2016; 38. Kim HS, Brill SJ. Rfc4 interacts with Rpa1 and is required for both DNA 937:207–28. replication and DNA damage checkpoints in Saccharomyces cerevisiae. 46. Marin JJ, Sanchez de Medina F, Castano B, Bujanda L, Romero MR, Mol Cell Biol 2001;21:3725–37. Martinez-Augustin O, et al. Chemoprevention, chemotherapy, and che- 39. Goldstein M, Kastan MB. The DNA damage response: implications for moresistance in colorectal cancer. Drug Metab Rev 2012;44:148–72. tumor responses to radiation and chemotherapy. Annu Rev Med 2015;66: 47. Wang D, Lippard SJ. Cellular processing of platinum anticancer drugs. 129–43. Nat Rev Drug Discov 2005;4:307–20. 40. Skvortsova I, Debbage P, Kumar V, Skvortsov S. Radiation resistance: cancer 48. Wijdeven RH, Pang B, van der Zanden SY, Qiao X, Blomen V, Hoogstraat M, stem cells (CSCs) and their enigmatic pro-survival signaling. Semin Cancer et al. Genome-wide identiﬁcation and characterization of novel factors Biol 2015;35:39–44. conferring resistance to topoisomerase II poisons in cancer. Cancer Res 41. Srivastava M, Raghavan SC. DNA double-strand break repair inhibitors as 2015;75:4176–87. cancer therapeutics. Chem Biol 2015;22:17–29. 49. Li Y, Gan S, Ren L, Yuan L, Liu J, Wang W, et al. Multifaceted regulation and 42. Takata M, Sasaki MS, Sonoda E, Morrison C, Hashimoto M, Utsumi H, et al. functions of replication factor C family in human cancers. Am J Cancer Res Homologous recombination and non-homologous end-joining pathways 2018;8:1343–55. of DNA double-strand break repair have overlapping roles in the main- 50. Xiang J, Fang L, Luo Y, Yang Z, Liao Y, Cui J, et al. Levels of human tenance of chromosomal integrity in vertebrate cells. EMBO J 1998;17: replication factor C4, a clamp loader, correlate with tumor progression and 5497–508. predict the prognosis for colorectal cancer. J Transl Med 2014;12:320.

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4579

Genome-wide RNAi Screening Identifies RFC4 as a Factor That Mediates Radioresistance in Colorectal Cancer by Facilitating Nonhomologous End Joining Repair

Xue-Cen Wang, Xin Yue, Rong-Xin Zhang, et al.

Clin Cancer Res 2019;25:4567-4579. Published OnlineFirst April 12, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-18-3735

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2019/04/12/1078-0432.CCR-18-3735.DC1

Cited articles This article cites 49 articles, 10 of which you can access for free at: http://clincancerres.aacrjournals.org/content/25/14/4567.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/25/14/4567.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/25/14/4567. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.