SND1 Acts Downstream of Tgfb1 and Upstream of Smurf1 to Promote Breast Cancer Metastasis

Published OnlineFirst January 16, 2015; DOI: 10.1158/0008-5472.CAN-14-2387 Cancer Molecular and Cellular Pathobiology Research

SND1 Acts Downstream of TGFb1 and Upstream of Smurf1 to Promote Breast Cancer Metastasis Lin Yu1,2,3,4, Xin Liu1,2,3,4, Kang Cui1,2,3,4, Yanbo Di1,2,3,4, Lingbiao Xin1,2,3,4, Xiaoming Sun1,2,3,4, Wei Zhang1,2,3,4, Xi Yang5, Minxin Wei6, Zhi Yao2,3,4, and Jie Yang1,2,3,4

Abstract

SND1 is an AEG-1/MTDH/LYRIC-binding protein that is cific recognition domains (RD motifs), which were recognized upregulated in numerous human cancers, where it has been and bound by the Smad complex that enhanced the transcrip- assigned multiple functional roles. In this study, we report its tional activation of SND1. We found that SND1 promoted association with the TGFb1 signaling pathway, which promotes expression of the E3 ubiquitin ligase Smurf1, leading to RhoA epithelial–mesenchymal transition (EMT) in breast cancer. ubiquitination and degradation. RhoA degradation in breast SND1 was upregulated in breast cancer tissues, in particular cancer cells disrupted F-actin cytoskeletal organization, reduced in primary invasive ductal carcinomas. Transcriptional activa- cell adhesion, increased cell migration and invasion, and pro- tion of the SND1 gene was controlled by the TGFb1/Smad moted metastasis. Overall, our results define a novel role for pathway, specifically by activation of the Smad2/Smad3 com- SND1 in regulating breast tumorigenesis and metastasis. Cancer plex. The SND1 promoter region contained several Smad-spe- Res; 75(7); 1–12. 2015 AACR.

Introduction Smad5, and Smad8 are specifically activated by the bone morphogenetic protein (BMP) receptors, whereas Smad2 and Smad3 Breast cancer is the most commonly diagnosed cancer in are specifically activated by TGFb-related receptors (9). In the women and the leading cause of female cancer death worldwide activated pathway, the phosphorylated R-Smads enter the nucleus (1). Distant metastasis and invasion of advanced cancer are to form complex with co-Smad Smad4 and then combine with responsible for more than 90% of cancer-related deaths (2). specific Smad recognition domains (RD) in the promoter region However, the molecular mechanisms underlying the breast cancer of target genes (10). tumorigenesis and metastasis are largely unknown. Staphylococcal nuclease domain-containing 1 (SND1), also The TGFb signaling pathway is a key player in the epithelial– known as Tudor staphylococcal nuclease (Tudor-SN) or p100 mesenchymal transition (EMT) in breast cancer. The misregula- protein, ubiquitously exists in human and many other species. tion of TGFb signaling may promote tumor invasion and migra- Proteomic analysis of clinical breast cancer samples shows a close tion, leading to detrimental consequences for prognosis in correlation of elevated SND1 expression with cancer metastasis patients with breast cancer (3–5). The TGFb signaling pathway (11). It was also reported that SND1 could interact with astrocyte contains different R-Smads (Smad1, Smad2, Smad3, and Smad5) elevated gene-1 (AEG-1) in breast cancer and strongly promote and co-Smad (Smad4) transcription factors (6–8). Smad1, lung metastasis (12). In our earlier work using chromatin immunoprecipitation guided ligation and selection (ChIP-GLAS) 1Department of Biochemistry and Molecular Biology, School of Basic assays, we have demonstrated that SND1 is potentially involved Medical Sciences, Tianjin Medical University, Tianjin, China. 2Depart- in the TGFb signaling pathway in breast cancer cells (13). How- ment of Immunology, School of Basic Medical Sciences,Tianjin Medical ever, it remains unclear how SND1 is upregulated in breast cancer University, Tianjin, China. 3Laboratory of Molecular Immunology, Research Center of Basic Medical Science, Tianjin Medical University, and how SND1 participates in breast cancer metastasis. Tianjin, China. 4Tianjin Key Laboratory of Cellular and Molecular Immu- In the present study, we demonstrate for the first time that nology and Key Laboratory of the Educational Ministry of China, 5 SND1 is a novel target of the Smad complex. The gene tran- Tianjin Medical University,Tianjin, China. Department of Immunology, SND1 b University of Manitoba, Winnipeg, Canada. 6Department of Cardio- scriptional activation of could be regulated by the TGF / vascular Surgery, Tianjin Medical University General Hospital, Tianjin, Smad pathway. In addition, SND1 mediates the gene transcrip- China. tional activation of Smurf1, which is an E3 ubiquitin ligase, Note: Supplementary data for this article are available at Cancer Research promoting RhoA ubiquitination, causing F-actin rearrange- Online (http://cancerres.aacrjournals.org/). ment, and leading to increased invasiveness and metastasis of L. Yu and X. Liu contributed equally to this article. breast cancer. Corresponding Authors: Jie Yang, Tianjin Medical University, Heping District, Qixiangtai Road, No. 22, Tianjin 300070, China. Phone: 86-22-23542520; Fax: 86-22-23542581; E-mail: [email protected]; and Zhi Yao, Phone: 86-22- Materials and Methods 83332817; Fax: 86-22-83332817; E-mail: [email protected] Cell lines, siRNA, and cell transfection doi: 10.1158/0008-5472.CAN-14-2387 The cell lines MCF-7, T47D, SK-BR-3, MDA-MB-231, and 2015 American Association for Cancer Research. BT549 were obtained from the ATCC. MDA-MB-231 was

www.aacrjournals.org OF1

Yu et al.

maintained in L-15 medium with 10% FBS, BT549 was cul- After 24 hours, the cells migrated through the filter were fixed tured in RPMI1640 with 10% FBS, and MCF-7, T47D, and SK- stained with 10% crystal violet. For each chamber, the mean of BR-3 cells were cultured in DMEM with 10% FBS. All the migrated cell number from five randomly fields was calculated siRNAs used in the present study were designed according to and displayed in the plot. previous report. SiRNAs of Smad2, Smad3, Smad4, and MCF-7 and MDA-MB-231 cells were seeded at 90% confluent in Smurf1 (14, 15) were purchased from Santa Cruz Biotechnol- 6-well plates. Three vertical wounds were scratched per well. The ogy. Smad5 (16) siRNA was purchased from Abnova. Smad1 relative migration was assessed at 100 magnification using (17) and scramble control (18) siRNAs were purchased from inverted microscope at the designated time (0 and 24 hours), Cell Signaling Technology. The cells were transfected with then calculated by following formula: (0 h wound width 24 h Lipofectamine 2000 transfection reagent (Invitrogen) accord- wound width)/0 h wound width. ing to the manufacturer's protocol. RNA extraction and real-time RT-PCR Plasmids, lentivirus, and stable cell construction Total RNA was isolated from cells using TRIzol (Invitrogen). The pGL3 series of plasmids were purchased from Promega. Primers were synthesized by Sangon Biotechnology Co. Ltd. The pSG5-SND1 plasmid containing full-length SND1 and pSil- (Supplementary Table S1). The RevertAid First Strand cDNA SND1-sh plasmid containing shRNA targeting SND1 were con- Synthesis Kit (Fermentas) was used to synthesize cDNA. The PCR structed as described previously (19). reactions were performed with FastStart Universal SYBR Green Lentiviruses containing either the shRNA targeting SND1 Master Mix (Roche) on a StepOne Real-Time PCR System (SND1-sh, antisense sequence: 50-GGTACCATCCTTCATC- (Applied Biosystems). The PCR conditions were 95 C for 2 min- CAAAT-30) or the negative control shRNA (scramble control, utes and 40 cycles of 95 C for 30 seconds and 60 C for 1 minute. sequence: 50-TTCTCCGAACGTGTCACGT-30) were constructed GAPDH or b-actin (Supplementary Fig. S1) was used as reference and packed by GenePharma. The titers of the virus solutions were genes, and the mRNA data were presented as the ratio of target 1.0 to 1.2 109/mL. gene against reference gene. The breast cancer cell lines MCF-7 and MDA-MB-231 were infected with 30 mL of lentivirus solution per 6-cm dishes. Forty- Western blotting and antibodies eight hours after infection, the cells were treated with 4 mg/mL Total cell lysates (40 mg) were subjected to 12% SDS-PAGE gel puromycin for 14 days to select stable cell lines. The stable electrophoresis and were blotted with series of antibodies: rabbit cell lines were named as MCF-7-SND1-sh, MCF-7-scramble, monoclonal antibodies against RhoA, Cdc42, phosphorylated MDA-MB-231-SND1-sh, and MDA-MB-231-scramble. Smad3, and phosphorylated Smad5 (1:1,000; CST); mouse monoclonal antibodies against SND1, Rac1, Smad1, Smad2, IHC Smad3, Smad4, Smad5, and Smurf1 (1:1,000; Abcam); and a The malignant and normal breast tissues with histopathology mouse monoclonal antibody against b-actin (1:6,000 dilution; reports were requested from the Cancer Hospital of Tianjin Sigma). Medical University. Informed consent was obtained from each patient. Use of the tissue samples in this study was approved by Chromatin immunoprecipitation the Institutional Review Board. The fresh tissues were fixed in 4% The chromatin immunoprecipitation (ChIP) assay was per- formaldehyde and embedded in paraffin. After deparaffinization formed using the EZ-Magna-ChIP-Chromatin Immunoprecipita- and rehydration, five-micrometer sections were incubated with a tion Kit (Millipore) according to the manufacturer's protocol. primary antibody directed against SND1 (1:75 dilution; Abcam) Immunoprecipitation was performed with the anti-Smad3 anti- at room temperature for 2 hours. A biotinylated secondary antibody (8 mg; Abcam) or anti-IgG antibody (8 mg; Sigma). The body was added for 30 minutes at 37 C, followed by immuno- purified immunoprecipitated DNA samples were used in qPCR histochemical staining using a DAB kit (Zhongshan). reactions.

Nude mice and tumor implantation Luciferase assay All mouse procedures were approved by the committee on the MCF-7 cells were cotransfected with both pGL3-SND1-pro use and care of animals of Tianjin Medical University. Six-week- reporter construct and Renilla construct. After 24 hours, the old female BALB/c nude mice were used for the animal studies. MCF-7 cells were starved overnight and stimulated with 15 ng/mL The mice were fed with a standard rodent diet in an aseptic TGFb1 (CST) or vehicle control for 6 hours. The luciferase reporter laminar ﬂow room with 60% to 70% humidity at 25 C. Hundred assay was performed using the Dual-Luciferase Reporter Assay Kit microliters of cell solutions (containing 4 106 logarithmic according to the manufacturer's protocol (Promega). growth phase MDA-MB-231 scramble or SND1-sh cells) were subcutaneously injected into second or third mammary fat pad of Immunoprecipitation the mice. The mice were then observed daily for diet consump- MCF-7 cells were cotransfected with pSG5-SND1 and Smurf1 tion, stools, and mental state. The tumor sizes were measured siRNA or scramble siRNA. After 36 hours, the total cell lysates every 5 days. In 25 days, the formed tumors were surgically were incubated with a rabbit monoclonal anti-RhoA antibody removed, and the weight of tumor was measured. (8 mg/mL; CST) or a rabbit polyclonal IgG antibody as a control (8 mg/mL; Santa Cruz Biotechnology). The incubation was per- Transwell and wound-healing assay formed overnight at 4C. Then, the proteins were precipitated MCF-7 and MDA-MB-231 cells were seeded in transwell inserts by protein A/G Sepharose (Sigma). The bound proteins were with an 8-mm pore size (BD Biosciences) at 5 105 cells per well. blotted with a mouse monoclonal anti-RhoA antibody (1:1,000;

OF2 Cancer Res; 75(7) April 1, 2015 Cancer Research

TGFb1-Regulated SND1 Promotes Metastasis through Smurf1

Abcam) and a mouse monoclonal anti-ubiquitin antibody cally removed. The average diameter (Fig. 1G) and weight (1:1,000 dilution; CST). (Fig. 1H) of the implanted tumors with knockdown of SND1 were significantly less than those of the tumors from the MDA- Immunofluorescence MB-231 scramble cells. These data further verify the involvement Cells were seeded on poly-D-lysine–coated coverslips and of SND1 in breast cancer tumorigenesis. grown until 70% confluent. Then, the coverslips were fixed and blocked with 3% BSA/PBS. F-actin was detected using Alexa 488– SND1 promotes the metastasis phenomenon of breast cancer labeled phalloidin (1:600; Invitrogen), and RhoA was detected cells using a rabbit monoclonal antibody (1:200; CST), followed by The effect of gain-of-function or loss-of-function of SND1 was the Alexa 594–conjugated secondary antibody (1:600; Invitro- performed in MCF-7 and MDA-MB-231 cells to investigate the gen). Images were captured by confocal fluorescent microscope involvement of SND1 on metastatic ability. qRT-PCR and Western (FV1200, Olympus). blot analysis revealed that both mRNA (Fig. 2A) and protein level (Fig. 2B) of SND1 were efficiently overexpressed or knocked down Statistical analysis by transfecting of pSG5-SND1 plasmid or pSil-SND1-sh plasmid, Numerical data are presented as the mean SD from at least respectively. three independent experiments and compared by the Student The wound-healing assay was performed to detect the migra- t test or one-way ANOVA using SPSS 16.0 software. A P value < tion activity of different breast cancer cells (Fig. 2C). The results 0.05 was considered as statistical significance. showed that the migration ability was significantly increased with ectopically overexpression of SND1 (SND1) in both MCF-7 (left panel) and MDA-MB-231 (right panel), comparing with Results parental control or scramble control cells. In contrast, knock- SND1 expression is upregulated in breast cancer tissues and down of endogenous SND1 protein (SND1-sh) inhibited the positively correlates with the metastatic ability of breast migration ability. Similarly, the transwell invasion assay (Fig. cancer cells 2D) demonstrated that the invasion ability was increased in To investigate the relevance of SND1 expression to breast cancer SND1 overexpression cells (SND1), but decreased in SND1 development and metastasis, we used immunohistochemical knockdown cells (SND1-sh). Furthermore, we performed SND1 approach to detect SND1 in clinical samples of normal breast rescue experiment in MCF-7-SND1-sh and MDA-MB-231-SND1- epithelium, ductal carcinoma (DC) in situ, and primary invasive sh stable cells to confirm the effect of SND1 in breast cancer ductal carcinoma (IDC; Fig. 1A). For the normal breast epithelium metastasis. The wound-healing assays and the transwell invasion tissues, SND1 expression was negative in all 12 of the studied cases assays demonstrated that comparing with the basal control level, (100%). Among the 28 cases with DC in situ tissues, SND1 the migration and invasion ability was increased in cells with expression was highly positive in 6 cases (21.43%), weakly overexpression of SND1, but decreased in untreated SND1-sh positive in 19 cases (67.86%), and negative in three cases knockdown cells. Although the migration and invasion ability of (10.71%). The expression of SND1 in the IDC samples was SND1-sh knockdown cells with rescue expression of SND1 was significantly higher than the DC in situ samples (, P < 0.01 vs. increased compared with untreated SND1-sh cells and rehabil- DC in situ; Fig. 1B and Supplementary Table S2). SND1 expression itated to the same level as the basal control groups (Supple- was positive in all 23 studied cases: 18 were highly positive mentary Fig. S1B and S1C), migration and invasion abilities were (78.26%) and 5 were weakly positive (21.74%). The survival the key components of the metastatic cascades, and all these data time of the IDC cases was analyzed by the Kaplan–Meier assay, further indicate the effect of SND1 on the metastatic phenotype and the result was presented in Fig. 1C. Median survival time of of breast cancer cells. high SND1 staining samples (n ¼ 18, 57 months) was significantly shorter than low SND1 staining samples (n ¼ 5, >107 SND1 expression is upregulated by TGFb1 stimulation months, P ¼ 0.00306). According to the pathologic data of IDC To investigate the underline mechanisms of upregulation of tissues, there was no significant correlation with SND1 expression SND1 in breast cancer cells, we employed bioinformatics analysis and estrogen receptor (ER), progesterone receptor (PR), or HER2 and revealed that the promoter region of the SND1 gene contains level, thus the involvement of SND1 in the tumorigenesis of breast two Smad recognition domains (RD1 and RD2) and one sp1 cancer is likely to be independent with ER, PR, and HER2 (Sup- recognition domain. Smads are essential transcription factors in plementary Table S3). the TGFb signaling pathway, which are overactive in advanced We then examined SND1 expression in different established breast cancer (21), thus the hyperexpression of SND1 in breast breast cancer cell lines (20). SND1 was highly expressed at both cancer cells may be regulated by the activated TGFb pathway. We mRNA (Fig. 1D) and protein level (Fig. 1E) in the metastatic breast first detected the expression of SND1 in MCF-7 cells at different cancer cell line (MDA-MB-231 and BT549), comparing with the time points with TGFb1 treatment (15 ng/mL). qRT-PCR nonmetastatic (MCF-7, T47D, and SK-BR-3). These data further results (Fig. 3A) showed that with TGFb1 treatment, the mRNA demonstrate the correlation of SND1 expression with the metas- level of SND1 was gradually increased, and was 2.93-fold tasis of breast cancer cells. higher than the basal level (non TGFb1 treatment) after TGFb1 To investigate the involvement of SND1 in tumorigenesis treatment for 12 hours. Western blot analysis indicated that in vivo, the effect of the loss-of-function of SND1 on tumorigenesis the SND1 protein level was also increased in a time-dependent of MDA-MB-231-SND1-sh cells was assessed in nude mice by manner following TGFb1 treatment (Fig. 3B); for 24 hours orthotopic implantation. The results showed that knockdown treatment, the SND1 protein level was 2.5-fold higher than the of SND1 slowed down the growth of the tumors in situ than nontreated one, and was approximately 6.0-fold higher with scramble control (Fig. 1F). After 25 days, the tumors were surgi- treatment for 48 hours. We then performed luciferase reporter

www.aacrjournals.org Cancer Res; 75(7) April 1, 2015 OF3

Yu et al.

Normal 100 DC in situ IDC

50 Normal ×10 Normal ×40

Figure 1. The expression pattern of SND1 in

Percentage of SND1-positive cells of SND1-positive Percentage 0 different breast tissues and breast Normal DC in situ IDC cancer cells. A, IHC staining of SND1 in biopsy specimens of normal breast C SND1 low tissues, DC in situ tissues, and IDC SND1 high tissues. B, the percentage of SND1- 1.0 positive cell showed different SND1 DC in situ ×10 DC in situ ×40 expression in normal, DC in situ, and 0.8 IDC tissues (, P < 0.01 vs. DC in situ). See also Supplementary Table S2. 0.6 C, Kaplan–Meier survival curves based on SND1 expression for IDC 0.4 samples. Median survival times were Survival rate Survival >107 months (n ¼ 5) for low SND1 0.2 staining and 57 months (n ¼ 18) for high SND1 staining (P ¼ 0.00306). IDC 10 IDC 40 0.0 D, qRT-PCR detection of SND1 mRNA × × 050100 ﬁ Time (month) in ve different breast cancer cells. D E E, Western blot detection of SND1 protein in different breast cancer 6 cells. F, MDA-MB-231 cells were stably transfected with either the human SND1 shRNA (SND1-sh) or the 4 MCF-7 T-47D SK-BE-3 MDA-MB-231BT549 scramble control shRNA (scramble) SND1 lentivirus. Two stable cell lines were 2 transplanted into nude mice for Blot with anti-SND1 in vivo tumor formation assay. The mean of tumor diameters for every 0 β-Actin Relative SND1 mRNA level Relative 5 days is presented. G, the formed T-47D β tumors were surgically removed after MCF-7 SK-BE-3 BT549 Blot with anti– -actin MDA-MB-231 25 days, and the tumor diameters were measured. H, quantitative FGHanalysis of tumor weight (, P < 0.05 vs. scramble, n ¼ 8). 3 MDA-MB-231 MDA-MB-231 MDA-MB-231 Scramble 40 2 Scramble 30 SND1-sh 1 SND1-sh 20 10 0 Tumor diameter (cm) diameter Tumor 0 Mean of tumor weight 0 5 10 15 20 25 (Day) Scramble SND1-sh

assay to examine whether TGFb1 regulates SND1 expression at Activated Smad2 and Smad3 are required for TGFb1-mediated the gene transcription level. MCF-7 cells were transfected with SND1 expression in breast cancer cells the luc-reporter plasmids containing full length of the SND1 To identify the specific Smad responsible for regulating gene promoter region, and then treated with or without TGFb1 transcriptional activation of SND1, we used a series of siRNAs (15 ng/mL) for different time points. The results showed that to knock down different endogenous Smads, respectively, in (Fig. 3C) the gene transcriptional activation of SND1 was breast cancer cells to block the TGFb signaling pathway. The increased in a time-dependent manner with TGFb1 treatment. knockdown efficiency of siRNAs was detected by Western blot These data demonstrate that TGFb1 is likely to promote SND1 (Fig. 4A). The results showed that (Fig. 4B) the mRNA level expression via gene transcriptional activation in a time-sequence of SND1 was remarkably enhanced by TGFb1 treatment in the manner, the mRNA level of SND1 primarily increases, and the MCF-7 cells with knockdown of endogenous Smad1 or Smad5 protein level increases consequently. or the scramble control, but no significant enhancement in the

OF4 Cancer Res; 75(7) April 1, 2015 Cancer Research

TGFb1-Regulated SND1 Promotes Metastasis through Smurf1

AB 30 MCF-7 MDA-MB-231 20

con SND1-shpSG5-SND1 con SND1-shpSG5-SND1 10 SND1

mRNA level Blot with anti-SND1 Blot with anti-SND1 Relative SND1 Relative Figure 2. 1 β-Actin SND1 expression positively correlates Blot with anti–β-actin Blot with anti–β-actin 0 with malignant phenotype of breast con SND1-sh pSG5-SND1 MCF-7 MDA-MB-231 cancer cell. A, the MCF-7 and MBA- MB-231 breast cancer cells were C transfected with SND1 shRNA expression vector pSil-SND1-sh Scramble SND1-sh con SND1 Scramble SND1-sh con SND1 (SND1-sh), SND1 overexpression vector pSG5-SND1 (SND1), and control 0 h 0 h vector pSG5 (con), respectively. qRT- PCR analyzes the SND1 mRNA levels in 24 h 24 h different treated MCF-7 and MDA-MB- 231 cells (, P < 0.01 vs. con, n ¼ 3). 3 Scramble 3 B, Western blot analysis of the SND1 Scramble SND1-sh SND1-sh protein levels in different transfected con cells. C, wound-healing assay of the con Relative

SND1 Relative migration migration activity in different treated migration SND1 cells. The vertical axis of coordinate 0 0 h 24 h 0 0 h 24 h represents the relative migration MCF-7 MDA-MB-231 activity of cells (, P < 0.01; , P < 0.05 D vs. scramble, n ¼ 3). D, Transwell assay of the invasion activity in different Scramble SND1-sh con SND1 Scramble SND1-sh con SND1 treated cells. The vertical axis of coordinate represents the mean of invasive cell numbers (, P < 0.01; , P < 0.05 vs. scramble, n ¼ 5). 180 180 120 120 60 60 Mean of Mean of cell number cell 0 number cell 0 con SND1 con SND1 SND1-sh Scramble Scramble SND1-sh MCF-7 MDA-MB-231 cells with knockdown of endogenous Smad2, Smad3, or Smad4. tion of SND1 in breast cancer cells is mediated by Smad2/Smad3/ Correspondingly, Western blot analysis (Fig. 4C) also showed Smad4. that knockdown of Smad2, Smad3, or Smad4 (lanes 8, 10, and When TGFb signaling is activated, Smad2 and Smad3 can be 12), unlike Smad1 or Smad5 (lane 4 or 6), blocked TGFb1- phosphorylated and form a heterodimer, which then bind to induced upregulation of SND1 protein levels in the MCF-7 cells. Smad4 to form the Smad complex, being translocated into the Our current observations suggest that TGFb1-induced upregula- nucleus and regulating the target gene transcription (22), while AB C TGFβ1 treatment time (h) 160,000 3 02448 120,000 2 SND1 Blot with anti-SND1 80,000 1 β-Actin 40,000 Blot with anti–β-actin

SND1 relative mRNA level SND1 relative 0 Relative ﬂuorescence unit ﬂuorescence Relative 0 0812 0122448 TGFβ1 treatment time (h) TGFβ1 treatment time (h)

Figure 3. TGFb1 enhances SND1 expression. A, the MCF-7 cells were treated with 15 ng/mL TGFb1 for different time points (0, 8, and 12 hours). The mRNA levels of SND1 were detected by qRT-PCR. B, the protein levels of SND1 in TGFb1-treated MCF-7 cells were detected by Western blot. C, MCF-7 cells were transfected with pGL3-SND1-pro and Renilla reporters. Twenty-four hours after transfection, cells were stimulated with TGFb1 for different time points, and the SND1 gene transcription was determined by ﬁreﬂy luciferase expression and normalized against Renilla expression.

www.aacrjournals.org Cancer Res; 75(7) April 1, 2015 OF5

Yu et al.

A Scramble Different Smads siRNA Scramble Different Smads siRNA Knockdown Knockdown

ScrambleSmad1-siSmad5-siSmad2-siSmad3-siSmad4-si ScrambleSmad1-siSmad5-siSmad2-siSmad3-siSmad4-si Smad1 Smad3 Blot with anti-Smad1 Blot with anti-Smad3 Smad5 Smad4 Blot with anti-Smad5 Blot with anti-Smad4 Smad2 β-Actin Blot with anti-Smad2 Blot with anti–β-actin B C Figure 4. b 5 Scramble Different Smads siRNA The TGF 1/Smad2/3 pathway Knockdown enhances SND1 expression. A, MCF-7 4 Control cells were transfected with siRNAs TGFβ1 treatment ScrambleSmad1 Smad5 Smad2 Smad3 Smad4 targeting different Smads. Western 3 TGFβ1 – + – + – + – + – + – + blot analysis was performed to detect ﬁ 2 Lane 1234 56 789101112 the knockdown ef ciency. B, MCF-7 cells were transfected with different 1 SND1 Smad siRNAs. Twenty-four hours after transfection, cells were either SND1 relative mRNA level SND1 relative Blot with anti-SND1 0 stimulated with TGFb1orleft β-Actin untreated, cells were lysed after 24 Scramble β hours, and SND1 mRNA level was Smad1 siRNASmad5 Smad2siRNA Smad3siRNA siRNASmad4 siRNA Blot with anti– -actin determined by qRT-PCR (, P < 0.05 D vs. control, n ¼ 3). C, Western blot Inhibitor: Vehicle SB431542 LDN193189 Inhibitor: Vehicle SB431542 LDN193189 analysis was also performed to detect SND1 protein levels in differently TGFβ1 treatment BMP4 treatment + + + – + – + – + – – – treated MCF-7 cells. D, MCF-7 cells Lane 123456 Lane 789101112 were pretreated with 100 nmol/L SB431542 or LDN193189 for 30 pSmad3 pSmad5 minutes before TGFb1 or BMP4 Blot with anti-pSmad3 Blot with anti-pSmad5 stimulation. The inhibitor activities were detected by Western blot. E, Smad5 Smad3 MCF-7 cells were pretreated with Blot with anti-Smad3 Blot with anti-Smad5 SB431542 (SB) or LDN193189 (LN). Then the cells were stimulated with β β -Actin -Actin TGFb1. qRT-PCR was performed to detect SND1 mRNA level of different Blot with anti–β-actin Blot with anti–β-actin inhibitor-treated cells ( , P < 0.05 vs. E F TGþSB, n ¼ 3). F, Western blot analysis was also performed to detect 3 Inhibitor pretreatment: Vehicle SB431542 LDN193189 the SND1 protein level of different treated cells. G, the luciferase reporter TGFβ1 treatment – + – + – + 2 assays were performed to detect the Lane 123456 SND1 gene transcription in different SND1 treated cells (!, P < 0.01 vs. TGþSB, 1 Blot with anti-SND1 n ¼ 3).

β-Actin

SND1 relative mRNA level SND1 relative 0 Blot with anti–β-actin con+vehiTG+vehicon+SBTG+SBcon+LNTG+LN G con con+SB RD2 RD1 sp1 con+LN SND1 promoter-full lengh TG TG+SB TG+LN

0 50,000 100,000 150,000 200,000 Relative ﬂuorescence unit

Smad1/Smad5/Smad8 maintain unphosphorylated. MCF-7 cells (top, lane 4) was efﬁciently inhibited by SB, and Smad5 phos- were treated with Smad2/Smad3 phosphorylation inhibitor phorylation was inhibited by LN (bottom, lane 12). The same SB431542 (SB), or Smad1/Smad5/Smad8 phosphorylation inhib- amount of Smad3 or Smad5 protein was observed in different itor LDN193189 (LN), to further investigate whether Smad2 and samples (middle). The qRT-PCR results (Fig. 4E) showed that com- Smad3 are required for TGFb1-induced SND1 expression. Western paring with control cells (conþvehi), the mRNA level of SND1 in blot results showed that (Fig. 4D) the phosphorylation of Smad3 the MCF-7 cells treated with SB was not enhanced (conþSB), even

OF6 Cancer Res; 75(7) April 1, 2015 Cancer Research

TGFb1-Regulated SND1 Promotes Metastasis through Smurf1

after TGFb1 treatment (TGþSB). However, TGFb1 significantly (full-length promoter region). Although TGFb1 could still increase increased the SND1 mRNA level in the cells incubated with LN SND1 transcriptional activation with deletion of sp1 motif in the (TGþLN), which was at the same level as parental cells (TGþvehi). promoter region (Dsp1), it was not as efficient as the control Western blot showed similar results (Fig. 4F). Comparing with the group. These observations strongly support the idea that RD, control cells without (lane 1) or with (lane 2) TGFb1treatment,SB especially RD2, in the promoter region of the SND1 gene is significantly inhibited the enhancement of SND1 protein level essential for TGFb1-mediated gene transcription of SND1. induced by TGFb1 (lane 3 and lane 4), whereas LN did not block To further verify that the Smad complex could interact with TGFb1-mediated SND1 expression (lane 5 and lane 6). Further- the specific promoter region of SND1, we then employed ChIP more, luciferase reporter assays indicated (Fig. 4G) that TGFb1 experiments in MCF-7 cells with control IgG or antibodies against treatment significantly enhanced the transcriptional activation of Smad3, with or without TGFb1 treatment, and then real-time PCR SND1 with/without treatment of LN, but the enhancement of was performed with different primers, as indicated in Fig. 5B, to SND1 was completely blocked by SB. Collectively, all these data detect the corresponding regions associated with Smad3 protein. indicate that TGFb1-mediated upregulation of SND1 is dependent The results showed that, with TGFb1 treatment, Smad3 protein on the Smad2/Smad3 activation. interacted with the motif within the two major cis-elements of SND1 promoter, the 2100- to 2000-bp region and the 1100- SND1 gene is a novel target of the Smad complex to 1000-bp region, which correspond to RD2 and RD1, respec- fi To determine the speci c sequence in the SND1 promoter tively (9.31-fold for RD2; 4.93-fold for RD1). Consistently, inhi- region that could be recognized by the Smad complex, we gen- biting the phosphorylation state of Smad2/Smad3 with SB signif- erated different luc-reporter plasmids containing different icantly reduced the association of Smad3 to RD1 and RD2, whereas SND1 sequence of promoter regions as illustrated in Fig. 5A. the treatment of LN did not show obvious effect. These results MCF-7 cells were transfected with different luc-reporter plasmids, indicate that TGFb1 induces Smad2/Smad3 phosphorylation, respectively, and luciferase reporter assays were carried out. The which could recognize and bind with RD, especially RD2 motif results showed that (Fig. 5A) disruption of the upstream Smad in the promoter region of SND1, and then mediate gene tran- D recognition domain (RD2), either individually ( Smad RD2) or scriptional activation of SND1. in combination with RD1 (DSmad RD1þ2), significantly blocked the enhancement of gene transcriptional activation of SND1 SND1 promotes Smurf1 expression and RhoA ubiquitination induced by TGFb1. However, the deletion of RD1 motif (DSmad As SND1 is regulated by the TGFb1/Smad2/Smad3 pathway, to RD1) had no significant effect, comparing with the control cells gain mechanistic insights into the role of SND1 in the breast

A RD2 RD1 sp1 SND1 promoter-full lenght RD2 RD1 SND1 promoter Δsp1 RD2 sp1 SND1 promoter ΔSmad RD1

Figure 5. RD1 sp1 SND1 promoter ΔSmad RD2 Phosphorylated Smad3 binds to RD2 Control sp1 β in the promoter region of SND1 and SND1 promoter ΔSmad RD1+2 TGF 1 treatment promotes SND1 expression. A, the MCF-7 cells were transfected with 0 50,000 100,000 150,000 reporter plasmids containing SND1 Relative ﬂuorescence unit promoter fragments with different B mutations (deletion of sp1 response RD2 RD1 sp1 site, Dsp1; deletion of Smad- SND1 recognized site1, DSmad RD1; deletion of Smad-recognized site2, DSmad –2,400 –1,600 –800 0 RD2; deletion of both Smad- recognized site1 and site2, DSmad region 1 2 3 4 RD1þ2). Twenty-four hours after 10 transfection cells were stimulated Input con Ig with TGFb1, the SND1 gene con Smad3 transcription was determined by 8 TG Ig luciferase expression. ( , P < 0.01; TG Smad3 , P < 0.05 vs. control, n ¼ 3). B, ChIP TG Smad3+SB demonstrates different associations 6 TG Smad3+LN of Smad3 with SND1 promoters in MCF-7 cells treated with or without SB431542 or LDN193189 (!, P < 0.01 4 vs. TG Smad3þSB, n ¼ 3). Relative enrichment Relative 2

0 region 1 2 3 4

www.aacrjournals.org Cancer Res; 75(7) April 1, 2015 OF7

Yu et al.

cancer metastasis, we then analyzed the gene expression cor- r ¼ 0.5243, P < 0.00001, Fig. 6A, Supplementary Table S4), relation between SND1 and TGFb signaling pathway molecules indicating a strong positive correlation between the expression based on 151 subjects with useable gene expression data from of SND1 and Smurf1. Furthermore, the network analysis was The Cancer Genome Atlas (TCGA) breast IDC dataset BRCA performed by R software, and the result presented the involve- (http://cancergenome.nih.gov). Statistical analysis of SND1 ment of SND1 in breast cancer metastasis through promoting and Smurf1 expression found a Pearson correlation (n ¼ 151, Smurf1 expression (Fig. 6B). We then performed qRT-PCR to

AB Pearson r = 0.5243

In same component Interact with 2 State change Targeted by transcription Other Merged 0

Smurf1 mRNA expression –2 Figure 6. SND1 promotes Smurf1-induced –2 0 2 4 RhoA ubiquitination. A, TCGA cancer SND1 mRNA expression database was used to analyze 151 cases of the breast IDC dataset (BRCA). Smurf1 expression was CDSND1-sh rescue positively correlated with SND1 con SND1 SND1-sh n ¼ 5 Lane 123 4 expression in BRCA samples ( 151, Pearson r ¼ 0.5243, P < 0.00001). SND1 See also Supplementary Table S4. 4 Blot with anti-SND1 B, the network view plot shows the relationship between SND1 (thick Smurf1 3 gray circle labeled) and TGFb Blot with anti-Smurf1 signaling pathway members (thick 2 black circle labeled). C, the mRNA RhoA levels of Smurf1 were detected by Blot with anti-RhoA qRT-PCR in different MCF-7 cells Relative Smurf1 mRNA Relative 1 after SND1 rescue transfection β-Actin ( , P < 0.05; , P < 0.01 vs. con, n ¼ 3). 0 D, Western blot analysis of Blot with anti–β-actin con SND1 the protein level of SND1, Smurf1, and SND1-sh SND1-sh rescue RhoA in different MCF-7 cells. E, MCF-7 cells were cotransfected with pSG5 (control) or pSG5-SND1 EFIP IgG IP RhoA (SND1) plasmids and scramble (lanes 1, 3) or Smurf1-siRNA (lanes 2, Control SND1 Control SND1 Control SND1 4), respectively. Western blot was done to detect the protein level of Knockdown SND1, Smurf1, and RhoA. F, the Smurf1 Smurf1 Smurf1 Smurf1 Smurf1 Smurf1 Scramble Scramble Scramble Scramble total cell lysates of different Scramble -siRNAScramble -siRNA -siRNA -siRNA -siRNA -siRNA Lane 1234 Lane 1234 567 8 treated MCF-7 cells (top) were immunoprecipitated with anti IgG IB SND1 IB RhoA (lanes 1–4) or anti-RhoA (lanes 5–8) antibody, then subjected to anti- Blot with anti-SND1 RhoA (top) or antiubiquitination Blot with anti-RhoA (bottom) immunoblotting. Smurf1

Blot with anti-Smurf1

RhoA Ubiquitin Blot with anti-RhoA n(Ub)

β-Actin

Blot with anti–β-actin Position of RhoA band

Blot with anti-ubiquitin

OF8 Cancer Res; 75(7) April 1, 2015 Cancer Research

TGFb1-Regulated SND1 Promotes Metastasis through Smurf1

detect the correlation between SND1 and Smurf1 in the stable other two members of Rho family GTPase. However, there was MCF-7 cells with overexpression of SND1 (SND1), knockdown no detectable change of the protein level of Cdc42 or Rac1 with of endogenous SND1 (SND1-sh), or gain-of-function of SND1 overexpression of SND1 (lane 2). in SND1-sh cells (SND1-sh cells transfected with pSG5-SND1). RhoA plays essential roles in maintaining the assembly of The result showed that (Fig. 6C) the expression of Smurf1 is focal adhesions and contractile actin-myosin stress fibers (26). positively correlated with SND1 levels. The lack of RhoA can lead to F-actin cytoskeleton disorganiza- It has been reported that Smurf1 promotes the ubiquitination tion. We then performed immunofluorescence assays to exam- and degradation of the small GTP protein RhoA, thereby regulates ine the cytoskeletal structure in MCF7 cells with different cell polarity and protrusion (23), which are important events in expression levels of SND1 protein. As shown in Fig. 7D, in the the TGFb-induced EMT (24, 25). We therefore detected the effect control cells, the F-actin was distributed as a net (green, Fig. 7D, of SND1 on regulating the protein level of Smurf1 and RhoA in 1a) representing of normal cytoskeletal structure, and RhoA MCF-7 cells. As shown in Fig. 6D, compared with the control resided in both cytoplasm and nucleus (red, Fig. 7D, 2a). MCF-7 cells (lane 1), the ectopic overexpression of SND1 protein Overexpression of SND1 caused the alteration of the F-actin (SND1) enhanced the Smurf1 protein level but reduced the RhoA cytoskeleton with extended cellular protrusions (Fig. 7D, 1b), protein level (lane 2), whereas knockdown of endogenous meanwhile the protein level of RhoA was decreased. (red, Fig. SND1 protein (SND1-sh) decreased the Smurf1 protein level but 7D, 2b). These data further demonstrate that overexpression of increased the RhoA protein level (lane 3). In the rescue experiment SND1 protein can cause the morphologic change in MCF-7 cells with the gain-of-function of SND1 protein in SND1-sh cells, the and affect the normal function of F-actin through regulating the protein level of Smurf1 was increased but RhoA was decreased activity of the Smurf1–RhoA pathway. Thus, the functional (lane 4), which further indicates that SND1 could positively correlation of SND1 and Smurf1 expression illustrates that regulate the expression of Smurf1, as a result, reducing the protein SND1 can promote the metastatic ability of breast cancer cells level of RhoA. via Smurf1-mediated RhoA ubiquitination and degradation We then determined the RhoA ubiquitination level in the (Fig. 7E). manipulated breast cancer cells with knockdown of endogenous Smurf1, together with or without overexpression of SND1. Cor- related with the expression of SND1 and Smurf1 (Fig. 6E), the Discussion protein level of RhoA was decreased with overexpression of SND1 SND1 upregulation has been reported in various malignant (lane 3), but increased with knockdown of Smurf1 (lane 2 and tumors, such as hepatocellular carcinoma (27), prostate cancer lane 4), comparing with the control cells (lane 1). The same lysates (28), and colon cancer (29). The molecular mechanisms for the were immunoprecipitated with anti-RhoA antibody, or anti-IgG involvement of SND1 in tumorigenesis of different types of as negative control, and then blotted with anti-RhoA or anti- cancers appear to be diverse. Through regulating alternative ubiquitin antibody. The results showed that (Fig. 6F) the same splicing, SND1 promotes prostate cancer cell growth and survival amount of RhoA was immunoprecipitated from different cell (30). It is reported that overexpression of SND1 increases RNA- lysates (top, lanes 5–8). RhoA was slightly ubiquitinated in the induced silencing complex activity and contributes to cutaneous control MCF-7 cells (bottom, lane 5), but was strongly ubiqui- malignant melanoma metastases (31), and SND1 can also pro- tinated in the cells with overexpression of SND1 protein (lane 7). mote tumor angiogenesis via miR-221 in hepatocellular carcino- In contrast, there was no detectable ubiquitinated RhoA band in ma (32). In human colon cancer, SND1 is upregulated and the cells with knockdown of endogenous Smurf1 (lanes 6, 8), contributes to cancer development at the early stages via activa- even with overexpression of SND1 (lane 8). These data further tion of the Wnt signaling pathway (33). Meanwhile, SND1 could demonstrate that SND1-induced RhoA degradation is mediated also interact with AEG-1 in colon cancer (29). In the present study, by Smurf1. we demonstrate that SND1 plays important roles in breast cancer metastasis. Around 50% of DC in situ will progress to IDC or metastasis SND1 promotes the metastatic ability of breast cancer cells breast cancers if untreated (34). However, the molecular mechan- via Smurf1-induced RhoA degradation isms underlying the progression of DC in situ to invasive DC are Proteasome inhibitor MG132 is commonly used to inhibit largely unknown. ER and HER2 are the two markers currently used protein degradation caused by ubiquitination. The stable in routine assessment of ER-positive DC in situ, but there is no MCF-7 cells with overexpression of SND1 (SND1 cells) were marker for ER-negative DC in situ. Our present study indicates that transfected with Smurf1 siRNA to knockdown endogenous SND1 could be a potential novel biomarker to predict invasive Smurf1 protein or pretreated with MG132 to inhibit the deg- occurrence for DC in situ independent to ER and HER2 expression. radation of RhoA protein. The migration (Fig. 7A) and invasion Several TGFb/Smad-regulated genes control the metastasis (Fig. 7B) abilities were increased significantly with overexpres- processes (35, 36). Supportively, we identified SND1 as a novel sion of SND1 (SND1þscram-si), but remained at the similar downstream target gene of the TGFb/Smad pathway: (i) bioin- level as control cells together with knockdown of endo- formatics analysis reveals that the SND1 promoter region contains genous Smurf1 (SND1þSmurf1-si) or pretreated with MG132 two Smad RDs; (ii) knockdown of endogenous Smad2, Smad3, or (SND1þMG132). Western blot analysis showed that (Fig. 7C) Smad4, or inhibition of the phosphorylation of Smad2 and overexpression of SND1 (lane 2 and lane 3) increased the Smad3 could block TGFb-induced enhancement of SND1 at both expression of Smurf1 protein (lane 2 and lane 3), and remark- mRNA and protein level; (iii) the luciferase reporter assay and ably reduced the protein level of RhoA (lane 2), but not in the ChIP experiments further illustrate that Smad3 could specifically cells pretreated with MG132 (lane 3). Meanwhile, we also recognize the RD in the SND1 promoter region. Because TGFb detected the protein level of Cdc42 and Rac1, which are the signaling is constitutively activated in breast cancer cells, the

www.aacrjournals.org Cancer Res; 75(7) April 1, 2015 OF9

Yu et al.

AB SND1+ SND1+ SND1+ con scram-si Smurf1-si MG132 SND1+ SND1+ SND1+ 0 h con scram-si Smurf1-si MG132

24 h 4 200 2 100 Relative Mean of migration

0 number cell 0 con con SND1+ SND1+ SND1+ SND1+ SND1+ SND1+ MG132 MG132 scram-si Smurf1-si scram-si Smurf1-si con SND1+ C SND1 MG132 D Lane 123

SND1 Figure 7. SND1 promotes the metastatic ability of Blot with anti-SND1 breast cancer cells via Smurf1-induced Smurf1 RhoA degradation. A, wound-healing assay shows the invasion ability of Blot with anti-Smurf1 different treated MCF-7 cells (, P < 0.01 n ¼ RhoA vs. con, 3). B, Transwell assay shows the migration ability of different treated Blot with anti-RhoA MCF-7 cells (, P < 0.01 vs. con, n ¼ 5). C, Rac1 the Smurf1, RhoA, Rac1, and Cdc42 steady-state levels of different cells were Blot with anti-Rac1 determined by Western blot. D, Cdc42 immunoﬂuorescence assay showed the expression pattern and subcellular Blot with anti-Cdc42 location of F-actin (1a and 1b) and β-Actin RhoA (2a and 2b) in normal or SND1 Blot with anti–β-actin overexpression MCF-7 cells. E, schematic diagram shows the E mechanism of metastasis stimulative β TGF 1 function of SND1. F-actin disorganization metastasis TGFβ1

TGFBR RhoA

Smad2/3 Rh oA SND1 Smurf1 Smad2/3

SND1 mRNA Smurf1 mRNA Smad2/3

observed upregulation of SND1 expression in breast cancer tissues Smurf1 expression, leading to increased RhoA ubiquitination is likely to be the result of TGFb-induced activation of the Smad and proteasomal degradation, causes morphologic alterations pathway. of the actin cytoskeleton, and enhances cell motility. There The most novel and interesting ﬁnding in the study is the is evidence that strongly supports our novel ﬁndings: (i) promoting effect of SND1 on Smurf1 expression and conse- knockdown of endogenous Smurf1 protein blocked the quent RhoA degradation. Smurf1 is an E3 ubiquitin ligase, SND1-induced metastatic ability of breast cancer cells; (ii) the whichisinvolvedintheTGFb-induced EMT (25, 37), could invasive phenotype correlated with SND1 hyperexpression was bind to nucleotide-free RhoA, and promote RhoA ubiquitina- inhibited by blocking the proteasomal degradation of RhoA tion (38). The present study indicates that SND1 enhances with proteasome inhibitor. This novel function of SND1 is

OF10 Cancer Res; 75(7) April 1, 2015 Cancer Research

TGFb1-Regulated SND1 Promotes Metastasis through Smurf1

similar to the proto-oncogene c-Myc, which is upregulated in Analysis and interpretation of data (e.g., statistical analysis, biostatistics, many cancers causing cytoskeleton depolymerization and inhi- computational analysis): L. Yu, X. Liu, L. Xin bition of the RhoA signaling pathway (39). Writing, review, and/or revision of the manuscript: L. Yu, X. Yang, J. Yang Administrative, technical, or material support (i.e., reporting or organizing In the present study, we provide supportive evidences to data, constructing databases): X. Liu, Y. Di, W. Zhang indicate that the TGFb/Smad pathway could activate SND1 gene Study supervision: M. Wei transcription, in which the Smad complex recognizes and binds to RD motif in the SND1 promoter region, and consequently enhances the transcriptional activation of SND1. Furthermore, Grant Support SND1 promotes Smurf1 expression, leading to RhoA ubiquitina- This work was supported by grants (81202102 to L. Yu; 90919032, tion and degradation. Most importantly, the degradation of RhoA 31370749 and 31125012 to J. Yang) from the of National Natural Science Foundation of China, grant IRT13085 to J. Yang from Project for Innovative protein in the breast cancer cells could disrupt the distribution of Research Team of Ministry of Education, grant 13JCQNJC12100 to L. Yu and F-actin cytoskeleton, increase cell migration and invasion abili- 13JCYBJC21400 to X. Liu from Project of Applicative Basic Research ties, thus promoting breast cancer metastasis (Fig. 7E). Our study and Advanced Technology of Tianjin Municipal Science and Technology demonstrates a novel role of SND1 in regulating breast cancer Commission, grant 20121202120018 to L. Yu from Specialized Research tumorigenesis and metastasis. Fund for the Doctoral Program of Higher Education, and grant 20110102 to L. Yu from Tianjin Municipal High School Science & Technology Foundation ﬂ Project. Disclosure of Potential Con icts of Interest The costs of publication of this article were defrayed in part by the ﬂ No potential con icts of interest were disclosed. payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Authors' Contributions this fact. Conception and design: L. Yu, X. Yang, Z. Yao, J. Yang Development of methodology: L. Yu, J. Yang Acquisition of data (provided animals, acquired and managed patients, Received August 22, 2014; revised November 11, 2014; accepted November provided facilities, etc.): L. Yu, X. Liu, K. Cui, Y. Di, J. Yang 29, 2014; published OnlineFirst January 16, 2015.

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer 14. Zhao WJ, Wang K. NGX6 expression improves the sensitivity of tamoxifen- statistics. CA Cancer J Clin 2011;61:69–90. resistant MCF-7 cells through modulation of the Smad signaling pathway. 2. Laohavinij S, Ruikchuchit K, Maneechavakajorn J. Survival and prognostic Int J Oncol 2013;42:2060–8. factors of stage I-III breast cancer. J Med Assoc Thai 2013;96 Suppl 3: 15. Pan H, Ding E, Hu M, Lagoo AS, Datto MB, Lagoo-Deenadayalan SA. S23–34. SMAD4 is required for development of maximal endotoxin tolerance. 3. Kang Y, He W, Tulley S, Gupta GP, Serganova I, Chen CR, et al. Breast cancer J Immunol 2010;184:5502–9. bone metastasis mediated by the Smad tumor suppressor pathway. Proc 16. Naito Y, Yamada T, Matsumiya T, Ui-Tei K, Saigo K, Morishita S. dsCheck: Natl Acad Sci U S A 2005;102:13909–14. highly sensitive off-target search software for double-stranded RNA-medi- 4. Nam JS, Suchar AM, Kang MJ, Stuelten CH, Tang B, Michalowska AM, et al. ated RNA interference. Nucleic Acids Res 2005;33:W589–91. Bone sialoprotein mediates the tumor cell-targeted prometastatic activity 17. Chau JF, Jia D, Wang Z, Liu Z, Hu Y, Zhang X, et al. A crucial role for bone of transforming growth factor beta in a mouse model of breast cancer. morphogenetic protein-Smad1 signalling in the DNA damage response. Cancer Res 2006;66:6327–35. Nat Commun 2012;3:836. 5. Roberts AB, Wakefield LM. The two faces of transforming growth factor beta 18. Siemens N, Patenge N, Otto J, Fiedler T, Kreikemeyer B. Streptococcus in carcinogenesis. Proc Natl Acad Sci U S A 2003;100:8621–3. pyogenes M49 plasminogen/plasmin binding facilitates keratinocyte inva- 6. Gupta J, Robbins J, Jilling T, Seth P. TGFbeta-dependent induction of sion via integrin-integrin-linked kinase (ILK) pathways and protects from interleukin-11 and interleukin-8 involves SMAD and p38 MAPK pathways macrophage killing. J Biol Chem 2011;286:21612–22. in breast tumor models with varied bone metastases potential. Cancer Biol 19. Yang J, Aittomaki S, Pesu M, Carter K, Saarinen J, Kalkkinen N, et al. Ther 2011;11:311–6. Identification of p100 as a coactivator for STAT6 that bridges STAT6 with 7. Wang SE, Wu FY, Shin I, Qu S, Arteaga CL. Transforming growth factor RNA polymerase II. EMBO J 2002;21:4950–8. {beta} (TGF-{beta})-Smad target gene protein tyrosine phosphatase 20. Holliday DL, Speirs V. Choosing the right cell line for breast cancer receptor type kappa is required for TGF-{beta} function. Mol Cell Biol research. Breast Cancer Res 2011;13:215. 2005;25:4703–15. 21. Giampieri S, Manning C, Hooper S, Jones L, Hill CS, Sahai E. Localized and 8. Theohari I, Giannopoulou I, Magkou C, Nomikos A, Melissaris S, Nako- reversible TGFbeta signalling switches breast cancer cells from cohesive to poulou L. Differential effect of the expression of TGF-beta pathway inhi- single cell motility. Nat Cell Biol 2009;11:1287–96. bitors, Smad-7 and Ski, on invasive breast carcinomas: relation to biologic 22. Datto MB, Frederick JP, Pan L, Borton AJ, Zhuang Y, Wang XF. Targeted behavior. APMIS 2012;120:92–100. disruption of Smad3 reveals an essential role in transforming growth factor 9. Attisano L, Wrana JL. Signal transduction by the TGF-beta superfamily. beta-mediated signal transduction. Mol Cell Biol 1999;19:2495–504. Science 2002;296:1646–7. 23. Sahai E, Garcia-Medina R, Pouyssegur J, Vial E. Smurf1 regulates tumor cell 10. He Y, Zhang H, Yung A, Villeda SA, Jaeger PA, Olayiwola O, et al. ALK5- plasticity and motility through degradation of RhoA leading to localized dependent TGF-beta signaling is a major determinant of late-stage adult inhibition of contractility. J Cell Biol 2007;176:35–42. neurogenesis. Nat Neurosci 2014;17:943–52. 24. Wang HR, Zhang Y, Ozdamar B, Ogunjimi AA, Alexandrova E, Thomsen 11. Ho J, Kong JW, Choong LY, Loh MC, Toy W, Chong PK, et al. Novel breast GH, et al. Regulation of cell polarity and protrusion formation by targeting cancer metastasis-associated proteins. J Proteome Res 2009;8:583–94. RhoA for degradation. Science 2003;302:1775–9. 12. Blanco MA, Aleckovic M, Hua Y, Li T, Wei Y, Xu Z, et al. Identification of 25. Townsend TA, Wrana JL, Davis GE, Barnett JV. Transforming growth factor- staphylococcal nuclease domain-containing 1 (SND1) as a Metadherin- beta-stimulated endocardial cell transformation is dependent on Par6c interacting protein with metastasis-promoting functions. J Biol Chem regulation of RhoA. J Biol Chem 2008;283:13834–41. 2011;286:19982–92. 26. Aguilar-Rojas A, Huerta-Reyes M, Maya-Nunez G, Arechavaleta-Velasco F, 13. Liu X, Dong L, Zhang X, Wang B, Wang X, Li H, et al. Identification of p100 Conn PM, Ulloa-Aguirre A, et al. Gonadotropin-releasing hormone recep- target promoters by chromatin immunoprecipitation-guided ligation and tor activates GTPase RhoA and inhibits cell invasion in the breast cancer cell selection (ChIP-GLAS). Cell Mol Immunol 2011;8:88–91. line MDA-MB-231. BMC Cancer 2012;12:550.

www.aacrjournals.org Cancer Res; 75(7) April 1, 2015 OF11

Yu et al.

27. Yoo BK, Santhekadur PK, Gredler R, Chen D, Emdad L, Bhutia S, et al. human colon cancers and implicated in early stage colon carcinogenesis. Increased RNA-induced silencing complex (RISC) activity contributes to Cancer Res 2007;67:9568–76. hepatocellular carcinoma. Hepatology 2011;53:1538–48. 34. Jones JL, Shaw JA, Pringle JH, Walker RA. Primary breast myoepithelial cells 28. Kuruma H, Kamata Y, Takahashi H, Igarashi K, Kimura T, Miki K, et al. exert an invasion-suppressor effect on breast cancer cells via paracrine Staphylococcal nuclease domain-containing protein 1 as a potential tissue down-regulation of MMP expression in ﬁbroblasts and tumour cells. marker for prostate cancer. Am J Pathol 2009;174:2044–50. J Pathol 2003;201:562–72. 29. Wang N, Du X, Zang L, Song N, Yang T, Dong R, et al. Prognostic impact 35. Nakata D, Hamada J, Ba Y, Matsushita K, Shibata T, Hosokawa M, et al. of Metadherin-SND1 interaction in colon cancer. Mol Biol Rep 2012;39: Enhancement of tumorigenic, metastatic and in vitro invasive capacity of 10497–504. rat mammary tumor cells by transforming growth factor-beta. Cancer Lett 30. Cappellari M, Bielli P, Paronetto MP, Ciccosanti F, Fimia GM, Saarikettu 2002;175:95–106. J, et al. The transcriptional co-activator SND1 is a novel regulator 36. Xie W, Mertens JC, Reiss DJ, Rimm DL, Camp RL, Haffty BG, et al. of alternative splicing in prostatecancercells.Oncogene2014;33: Alterations of Smad signaling in human breast carcinoma are associated 3794–802. with poor outcome: a tissue microarray study. Cancer Res 2002;62: 31. Sand M, Skrygan M, Georgas D, Sand D, Gambichler T, Altmeyer P, et al. 497–505. The miRNA machinery in primary cutaneous malignant melanoma, cuta- 37. Ozdamar B, Bose R, Barrios-Rodiles M, Wang HR, Zhang Y, Wrana JL. neous malignant melanoma metastases and benign melanocytic nevi. Cell Regulation of the polarity protein Par6 by TGFbeta receptors controls Tissue Res 2012;350:119–26. epithelial cell plasticity. Science 2005;307:1603–9. 32. Santhekadur PK, Das SK, Gredler R, Chen D, Srivastava J, Robertson C, et al. 38. Wang HR, Ogunjimi AA, Zhang Y, Ozdamar B, Bose R, Wrana JL. Degra- Multifunction protein staphylococcal nuclease domain containing 1 dation of RhoA by Smurf1 ubiquitin ligase. Methods Enzymol 2006; (SND1) promotes tumor angiogenesis in human hepatocellular carcinoma 406:437–47. through novel pathway that involves nuclear factor kappaB and miR-221. 39. Shi JW, Liu W, Zhang TT, Wang SC, Lin XL, Li J, et al. The enforced J Biol Chem 2012;287:13952–8. expression of c-Myc in pig ﬁbroblasts triggers mesenchymal-epithelial 33. Tsuchiya N, Ochiai M, Nakashima K, Ubagai T, Sugimura T, Nakagama H. transition (MET) via F-actin reorganization and RhoA/Rock pathway SND1, a component of RNA-induced silencing complex, is up-regulated in inactivation. Cell Cycle 2013;12:1119–27.

OF12 Cancer Res; 75(7) April 1, 2015 Cancer Research

SND1 Acts Downstream of TGFβ1 and Upstream of Smurf1 to Promote Breast Cancer Metastasis

Lin Yu, Xin Liu, Kang Cui, et al.

Cancer Res Published OnlineFirst January 16, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-14-2387

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2015/01/17/0008-5472.CAN-14-2387.DC1

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2015/03/20/0008-5472.CAN-14-2387. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.