RAE1 mediated ZEB1 expression promotes epithelial–mesenchymal transition in breast cancer

Breast cancer metastasis accounts for most of the deaths from breast cancer. Since epithelial-mesenchymal transition (EMT) plays an important role in promoting metastasis of cancer, many mechanisms regarding EMT have been studied. We previously showed that Ribonucleic acid export 1 (RAE1) is dysregulated in breast cancer and its overexpression leads to aggressive breast cancer phenotypes by inducing EMT. Here, we evaluated the functional capacity of RAE1 in breast cancer metastasis by using a three-dimensional (3D) culture system and xenograft models. Furthermore, to investigate the mechanisms of RAE1-driven EMT, in vitro studies were carried out. The induction of EMT with RAE1-overexpression was confirmed under the 3D culture system and in vivo system. Importantly, RAE1 mediates upregulation of an EMT marker ZEB1, by binding to the promoter region of ZEB1. Knockdown of ZEB1 in RAE1-overexpressing cells suppressed invasive and migratory behaviors, accompanied by an increase in epithelial and a decrease in mesenchymal markers. Taken together, these data demonstrate that RAE1 contributes to breast cancer metastasis by regulating a key EMT-inducing factor ZEB1 expression, suggesting its potential as a therapeutic target.


RAE1 overexpression enhances cell spreading in 3D culture systems and metastasis in mouse xenograft models.
To investigate the precise effects of RAE1 overexpression in breast cancer, we carried out 3D cell culture analysis with stable MCF7 cell lines overexpressing RAE1 (MCF7:RAE1 #1, 2, and 3) and empty vector (MCF7:emp vec #1, and 2). The Matrigel-embedded 3D culture system is more appropriate for structural and functional studies than the 2D culture system 23 . The results of phalloidin and DAPI staining at day 10 showed that MCF7 cells stably overexpressing RAE1 spread outwards along the extracellular matrix, whereas the control MCF7 cell lines maintained a spherical morphology without extending along the bottom line of the 3D culture vessel (Fig. 1A). In addition, confocal images representing a cross-section of the colony revealed that RAE1overexpressing MCF7 cells were dispersed towards the outside, while control MCF7 cells gathered near the center (Fig. 1B). Serial confocal transverse section images of each stable cell line are provided in Fig. S1.
To further explore the functional role of RAE1 in breast cancer progression in vivo, we evaluated the effects of RAE1 on the metastasis in a breast cancer xenograft model. MDA-MB-231 cells were used in this study because of their high metastatic potential. Tumors were monitored over a period of 11 weeks, and then the distance between the injection site and final position was measured as shown in Fig. 2A. Short-term xenograft (6 hrs-1 week) did not show significant migration, while long-term xenograft (7-11 weeks) showed significant migratory abilities by RAE1-overexpressing cells. The investigation of tumor cell spreading from the primary tumor cells to other sites at 11 weeks showed that xenograft mice injected with control cells formed primary tumors at the injection site (2 out of 4 mice) but did not form metastatic tumors at any other organs throughout the body except the liver (Fig. 2B,C). On the other hand, xenograft mice with RAE1-overexpressing cells formed primary tumors at the injection site (4 out of 4 mice) and metastatic tumors at the fat pad opposite to the injection site (3 out of 4 mice) (Fig. 2B,C). Although the metastatic tumors were not found during this period of time in distal organs such as kidney and lung in both groups, all mice injected with RAE1-overexpressing cells showed a stronger signal for liver metastases (Fig. 2C). Together, these results support that RAE1 accelerates tumor metastasis in vivo.
Upregulation of RAE1 enhances the expression of ZEB1 by binding to the promoter region. To investigate the molecular mechanisms underlying the role of RAE1 in mediating cancer metastasis, we performed gain of function studies using in vitro models. Among various breast cancer cell lines, we found that RAE1 is expressed highly in BT474, but it is expressed relatively low in MDA-MB-453, T47D, and MDA-MB-231 (Fig. S2A,B). We confirmed the subcellular localization of endogenous and exogenous RAE1 in several different cell lines (Fig. S2C,D) and concluded that forced expression of RAE1 does not lead to mislocalization of abnormal protein product. Recent studies on the NPC components and their association with gene expression regulation suggest that high concentration of RAE1 at the peripheral portion of the nucleus may play a role as a transcription regulator [24][25][26] . As RAE1 has been shown to induce EMT signals and promote invasion and migration abilities, we determined the expression levels of several EMT-associated transcription factors (Fig. S3) and found that ZEB1 mRNA levels were significantly upregulated by RAE1 overexpression (Fig. 3A). Furthermore, in order to confirm the positive correlation between RAE1 and ZEB1 in an in vivo system, IHC was performed with anti-ZEB1 antibody in tumor tissues retrieved from the xenograft experiment. In the MDA-MB-231 xenograft tumor tissues, ZEB1 was expressed mainly in the nucleus. The number of ZEB1-positive cells decreased from 129.5 ± 4.42 to 44.6 ± 11.45 in RAE1-knockdowned tumors, but increased from 126.3 ± 2.80 to 199.6 ± 9.03 in RAE1-overexpressing tumors. This may be an indirect evidence for the altered expression of ZEB1 through RAE1 regulation (Fig. 3B,C).
To determine whether RAE1 may regulate for ZEB1 expression, the ability of RAE1 to bind the ZEB1 promoter was determined by ChIP assay. PCR amplicon sites were designed near the putative promoter region of ZEB1 (Fig. 3D). ChIP-qPCR data showed that overexpression of RAE1 led to increased binding of RAE1 in the −880 to −157 bp and −164 to +64 bp amplicon sites (pZEB1 #1 and #3) (Fig. 3E). To further delineate the effects of RAE1 on ZEB1 transcriptional activity, we performed dual luciferase assay by cloning ZEB1 promoter region (from −881 up to + 64 bp downstream of the TSS) and negative control (from −1500 up to −888 bp downstream of the TSS) into the luciferase vector. The ZEB1 promoter activity was increased by overexpression of RAE1 compared to negative control in HEK293T cells (Fig. 3F). Collectively, these results suggest RAE1 positively regulates ZEB1 expression during cancer progression.

ZEB1 is a mediator for RAE1-induced EMT, invasion and migration in breast cancer.
In a previous study, we have shown that overexpression of RAE1 in epithelial-like MCF7 and T47D induces EMT-like morphological changes and EMT marker expression 18 . In an opposite way, knockdown of RAE1 in mesenchymal-like MDA-MB-231 cells reduced cancer cell invasion and migration 18 . Here, to examine whether RAE1-induced metastatic capability was mediated by ZEB1, siRNA was used to silence ZEB1 gene expression. We have previously shown that overexpression of RAE1 in MCF7 cells also altered EMT-related marker levels and showed more distinct spindle-shape morphology (  Fig. 4B, rows 3 and 4). In addition, silencing ZEB1 repressed RAE1-driven invasion and migration abilities (Fig. 4C,D). The effect of ZEB1 knockdown on the morphological and molecular changes was similarly observed in RAE1-overexpressing T47D and MDA-MB-231cells (Figs S5 and S6). Together, these data suggest that ZEB1 functions as a key component of RAE1-mediated EMT in breast cancer.

Discussion
In this study, we evaluated the metastatic properties of RAE1-overexpressing breast cancer cells both in 3D culture systems and in vivo xenograft models, and further demonstrated the molecular mechanisms underlying RAE1-mediated tumor progression. Our data revealing RAE1-mediated ZEB1-transcription regulation suggest that ZEB1, an EMT inducing transcription activator, contributes to the development of RAE1-driven EMT.
RAE1 is a component of the NPC. NPCs are large multi-protein structures that are present in the double membrane of the nuclear envelope mediating trafficking between the nucleoplasm and cytoplasm [27][28][29] . Alterations in nucleoporins, which compose each NPC, are frequently associated with particular defects in development and disease, and the resulting phenotypes are typically thought to be consequences of disturbed activity of nuclear transport 24,29 . Particularly, primary human specimens derived from different forms of cancer revealed www.nature.com/scientificreports www.nature.com/scientificreports/ dysregulation of mRNA export. Various components of mRNA export and its related factors contribute to the preferential export of transcripts encoding proteins involved in proliferation, survival, metastases, and invasion 30 .
Alternatively, NPCs can regulate the transcription levels of particular genes in a transport-independent manner 24,26 . Several studies have suggested that interactions between genes and the NPC modulate both the definition of hetero-and euchromatin boundaries and transcription 25 . Thus, NPCs are essential for controlling not only transport between the nucleoplasm and cytoplasm, but also organization of the genome and specific gene transcription 31 . The potential function of RAE1 as a transcription activator has not been reported before. However, some previous studies demonstrated the role of nuclear pore proteins (Nups), which interact with RAE1, as potential regulator of gene transcription. For example, Nup210 positions its target genes at the nuclear periphery, an environment where gene transcription can be controlled 32,33 . Here, we provide evidence that RAE1 regulates gene transcription by binding to the promoter region of a particular gene, ZEB1. Although we have realized the analysis of RAE1 expression in the Cancer Cell Line Encyclopedia (CCLE) databank did not find any correlation between RAE1 and ZEB1 mRNA expression nor with the breast cancer subtypes (Fig. S7), it may seem quite plausible to say that the regulation of ZEB1 expression by RAE1 is cell context-specific and/or may depend on the threshold of RAE1 expression. Taking this into account, our data showing the possibility of modulating ZEB1 expression by RAE1 suggest that the association of these two molecules may be important under certain conditions.
Apart from its function as a transcription factor, it is possible that RAE1 interacts with cytoplasmic components such as the cytoskeleton and contributes to genome organization and gene expression. RAE1, as a nuclear exporter, is expected to only be present in nuclear pores. However, according to public available databases (e.g. The Human Protein Atlas), RAE1 is present in the nucleoli fibrillar center and in the nucleus (Fig. S8). In addition, our ICC results show that RAE1 can be detected in the cytoplasm. (Fig. S2). The cytoskeleton is a well-organized dynamic cellular architecture that is typically involved in signal transduction in the cytoplasm 34 . In this regard, www.nature.com/scientificreports www.nature.com/scientificreports/ NPCs can act as docking sites for chromatin, ultimately contributing to the organization of the global topology of chromosomes in close association with other elements of the nuclear envelope 25 .
Based on the facts that the dysregulation of many NPC components is found in the different types of cancer and that various nuclear factors interacting with them, such as chromosomal maintenance 1 (CRM1, also known as exportin1), karyopherin family (KPNA2 and KPNB1), and chromosome segregation gene, contribute to the progression of cancer 29,35-37 , more attention is being focused on developing therapeutic drugs using transport factors 35,38 . In particular, CRM1/XPO1 is proposed as a promising drug candidate 39,40 , and is likely to bind to RAE1, either directly or indirectly (Fig. S9). Therefore, understanding the RAE1 function and mechanisms of action will be valuable enough to consider the value of therapeutic use. In conclusion, our study demonstrated that RAE1 promotes progression of breast cancer cells by activating ZEB1 at the transcription level and revealed the positive correlation between RAE1 and ZEB1 in breast cancer metastasis. Three-dimensional (3D) in vitro cell culture. To analyze cellular growth in 3D culture, an eight-well chamber slide was pre-coated with 50 μL of Matrigel TM (BD, San Jose, CA, USA) and incubated at 37 °C for 30 min to allow for gel formation. While the Matrigel was solidifying, 5 × 10 3 cells were diluted in cell culture medium (final concentration: 2.5 × 10 4 cells/mL) and mixed with 4% Matrigel-containing medium in a 1:1 ratio. The cell mixture was placed on top of the solidified Matrigel. Medium containing 2% Matrigel was changed every 3-4 days. Confocal images were acquired on the 10th day of culture after staining with anti-phalloidin (Invitrogen, Carlsbad, CA, USA) and fluorochrome 4′,6-diamidino-2-phenylindole (DAPI). Images were captured with a 40 × C-Apochromat water immersion lens on a Zeiss LSM 700 Confocal, using Zen 2011 software (Carl Zeiss, Oberkochen, Germany). For 3D images, z-stack scans were collected by incremental stepping through the 3D sample using a focal drive. The step size was automatically calculated with Zen 2011 software.  Mice were sacrificed in a CO 2 chamber at 11 weeks after injection, and the liver, pancreas, spleen, kidney, lung, and heart tissues were collected to analyze metastasis. All animal procedures were approved by the Institutional Animal Care Committee at Yonsei University, and were carried out in accordance with the guidelines and regulations set by the ethics committee. Animals were then euthanized and selected tissues were processed for histology. Immunohistochemistry (IHC) analysis. IHC analyses were performed as previously described with minor modification 41 . Briefly, slides were deparaffinized and rehydrated through series of graded ethanol. Antigen retrieval was performed using a pressure cooker. Endogenous peroxidase activity was blocked by incubation with 3% H 2 O 2 for 30 min then incubated with protein blocking solution (Dako, Glostrup, Denmark) for 1 hour at room temperature. Primary antibody was incubated in a humid chamber at 4 °C overnight, and then slides were incubated with secondary rabbit IgG (Dako) for 15 min at room temperature, and developed with Dako Envision + System-HRP DAB (Dako). Anti-ZEB1 (Abcam, Cambridge, UK; dilution ratio 1:2000) was purchased. After counterstaining with Meyer's Hematoxylin (Sigma), slides were mounted with mounting solution (Electron Microscopy Sciences, Hatfield, PA, USA). Quantitation of ZEB1-positive cells was done and then statistical analyses were performed with JMP software. Student's t-tests were used to analyze differences between means. Data are represented as mean ± SEM. Immunocytochemistry (ICC) analysis. ICC analyses were performed, with Abcam's ICC protocol. An anti-RAE1 antibody (Abcam) was used to detect RAE1 protein. The nucleus was counterstained with DAPI. Images were captured with a 40 × C-Apochromat water immersion lens on a Zeiss LSM 700 Confocal, using Zen 2011 software.
Real-time PCR. Real-time PCR analysis was performed as previously described 42 . For quantitative PCR analysis, the StepOnePlus TM Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and Power SYBR Green PCR Master Mix (Applied Biosystems) kits were used. All samples were run in triplicate, and RAE1 and ZEB1 expression levels were normalized relative to that of β-Actin, which was used as an internal loading control. Primers for PCR are listed in Table 1.
Western blotting and antibodies. Western blot analyses were performed as previously described 18  Matrigel invasion and migration assays. The Matrigel TM (BD) invasion and migration assays were performed previously described 18 . Chromatin immunoprecipitation (ChIP) assay. ChIP analysis was performed as previously described 42 with minor modifications. Chromatin was prepared from stable RAE1-overexpressing and control breast cancer cell lines. Briefly, 1 × 10 6 cells were cross-linked with 1% formaldehyde for 15 min, followed by the addition of glycine at 125 mM. Chromatin was sheared by sonication to fragments averaging between 0.5 and 1 kb in buffer containing 1% SDS, 1% Triton X-100, 0.1% sodium deoxycholate, 10 mM EDTA, 50 mM Tris-HCl (pH 8.0), and protease inhibitor cocktail (Roche Applied Science, Basel, Switzerland). Chromatin was pre-cleared with protein A/G beads containing 50% slurry (Santa Cruz Biotechnology, Dallas, TX, USA) and salmon sperm DNA, followed by immunoprecipitation with anti-RAE1 antibody (Abcam) coupled to protein A/G beads under each experimental condition. Nonimmune mouse IgG (Santa Cruz Biotechnology) was used as a control. ChIP-PCR data are shown as the percentage of input after normalization with IgG. Primers for ChIP-PCR are listed in Table 2.
Dual luciferase assay. Dual luciferase assay was performed as described previously 43 . Genomic DNA fragment of the ZEB1 promoter region was cloned into the pGL3-Basic vector (Promega, Madison, WI, USA) using