Notch3 Inhibits Epithelial-to-Mesenchymal Transition by Transactivating Glycogen Synthase Kinase-3-beta–Mediated Wnt/β-catenin Signaling in Breast Cancer


 As a critical transformational process in the attributes of epithelial cells, epithelial-to-mesenchymal transition (EMT) is involved in tumor invasion, metastasis, and resistance to treatment, which contributes to the ultimate death of some patients with breast cancer. Glycogen synthase kinase-3-beta (GSK3β) is thought to be an EMT suppressor that down-regulates the protein, snail, a zinc finger transcription inhibitor, and regulates E-cadherin expression and the Wnt signaling pathway. Our previous studies have shown that Notch3 also inhibits EMT in breast cancer. In mammary gland cells, GSK3β physically bound and phosphorylated the intracellular domain of two Notch paralogs: N1ICD was positively regulated but N2ICD was negatively regulated. However, the relationship between Notch3, GSK3β, and EMT in breast cancer is still unclear and crosstalk between Notch3 and GSK3β has not been widely investigated.In this study, we revealed that Notch3 was an essential antagonist of EMT in breast cancer cells by transcriptionally upregulating GSK3β. In breast cancer MCF-7 and MDA-MB-231 cell lines, the silencing of Notch3 reduced GSK3β expression, which is sufficient to induce EMT. Conversely, ectopic Notch3 expression re-activated GSK3β and E-cadherin. Mechanistically, Notch3 can bind to the GSK3β promoter directly and activate GSK3β transcription. In human breast cancer samples, Notch3 expression is positively associated with GSK3β (r=0.416, P = 0.001). Moreover, high expressions of Notch3 and GSK3β mRNA are correlated to better relapse-free survival in all breast cancer patients via analysis in “the Kaplan–Meier plotter” database.In summary, our preliminary results suggested that Notch3 might inhibit EMT by trans-activating GSK3β in breast cancer cells. The suppression of Notch3 expression may contribute to EMT by transcriptionally downregulating GSK3β in breast cancer.


Introduction
Breast cancer is the most common malignant tumor and main cause of death due to cancer worldwide in women. Moreover, local and/or distant metastasis is the leading cause of death in patients with this disease [1]. In recent decades, advances in early diagnosis, improved surgical techniques, and better adjuvant drug therapy have decreased the mortality rate of breast cancer [2]. Unfortunately, many patients exist who still have relapses or distant metastases, resulting in eventual treatment failure. The process of epithelial-to-mesenchymal transition (EMT) results in the transformation of epithelial cells into a mesenchymal phenotype, signi cantly reduces adhesion strength between cells, and drives the aggressiveness and cell migration of epithelial tumors [3]. In recent years, EMT has been associated with tumor genesis and progression, and is considered to be an important biological process leading to epithelial malignancy during embryonic development, with enhanced migration and invasion capacity [4].
The mechanisms involved in the occurrence and progression of EMT have been extensively explored, but other potential mechanisms, especially critical genes that drive EMT, still remain unclear [5]. As a highly conserved signaling pathway, Notch plays an essential role in numerous biological processes in metazoan development, including stem cell self-renewal, cell differentiation, proliferation, migration, adherence, survival, apoptosis, and cell-fate decisions [6,7]. Recently, increasing evidence has shown that Notch plays a fundamental role in development and tumorigenesis of mammary gland, including induction of the EMT process [8,9]. Notch3, a member of the Notch family of transmembrane receptors, is considered an oncogene or tumor suppressor in distinct cancers [10][11][12][13]. Our previous studies showed that estrogen receptor (ER)α inhibits EMT by suppressing Bmi1 [14], and Notch3 transcriptionally upregulates ERα in breast cancer. Furthermore, Notch3 can inhibit EMT in breast cancer epithelial cells by transactivating Kibra [15]. The aforementioned results indicate that Notch3 inhibits progression of EMT in breast cancer.
Glycogen synthase kinase-3-beta (GSK3β) is a serine/threonine kinase that can phosphorylate vast numbers of substrates, such as structural proteins, transcription factors, and signaling proteins [16]. Growing evidence demonstrates that GSK3β acts as an EMT suppressor via its downregulation of snail, a zinc-nger transcriptional repressor that regulates the epithelial marker, E-cadherin [17]. Both Notch and Wnt pathways are known to play key roles in development and disease [6]. Direct or indirect interactions have been shown between these two signaling pathways [18,19]. However, the relationship between Notch3 and GSK3β in EMT of breast cancer is still unclear.
The purpose of this study was to investigate whether Notch3 acts as a transcriptional activator of GSK3β in breast cancer. Notch3 was found to inhibit EMT by directly binding to the GSK3β promoter and transactivating GSK3β in breast cancer. In addition, Notch3 and GSK3β expression was positively associated in human breast cancer samples. Finally, over-expression of both Notch3 and GSK3β mRNA was found to be associated with better recurrence-free survival (RFS) in all patients with breast cancer.

Materials And Methods
Cell lines, plasmids, and reagents Human breast cancer cell lines T-47D, MCF-7, MDA-MB-231, and BT-549, were purchased from the American Type Culture Collection (Manassas, VI, USA). The pCMV and pCMV-Notch3 intracellular domain (N3ICD) plasmids were given by Professor Michael M. Wang from the Medical School of Michigan University. The psi-U6.1/eGFP/short hairpin (sh)RNA-GSK3β plasmid was purchased from GeneCopoeia (Guangzhou, China), and pCMV3-GSK3β-GFPSpark was purchased from Sino Biological Inc (Beijing, China). A wild-type reporter assay vector containing a GSK3β promoter region from -353 to -268 (p GL3-GSK3β-enhancer) was generated and a mutant reporter assay vector was constructed by replacing a TCC (-307 to -305) sequence in the wild-type reporter assay vector with a DNA-binding protein (CSL) antisense binding sequence, TTCCCA. All antibodies used in this study are summarized in Table S1.

Western blot
Western blotting was performed as previously described [20]. The antibodies used are outlined in Table  S1.

RNA extraction and RT-PCR analysis
Total RNA was isolated from cultured cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. A PrimeScript RT Reagent Kit (Takara, Dalian, China) was used to synthesize rst-strand cDNA. Gene expression was detected by quantitative real-time PCR analysis using a SYBR Select Master Mix (Applied Biosystems, Foster City, CA, USA). Sequences of all primers used in the study are summarized in Table S3.

Immuno uorescence
Immuno uorescence staining of MCF-7 and T-47D cells was performed as previously described [21]. Staining was visualized with a Zeiss microscope (Zeiss, Oberkochen, Germany). Notch3 and GSK3β antibodies are listed in Supplementary Table S1. Wound healing assay Wound healing assays were performed to evaluate the migration of MDA-MB-231 and MCF-7 cells under different transfection conditions. MDA-MB-231 cells were transfected with pCMV, pCMV-N3ICD or pCMV-N3ICD plus psi-U6.1/eGFP/shRNA-GSK3β, respectively, and inoculated in a 6-well cell culture plate at a density of 1×10 5 cells/well. Subsequently, cells were incubated in serum-free DMEM for 12 h. A pipette tip was used to establish a line in a monolayer of adherent cells. Photos were taken at 0, 24 h (MDA-  or 48 h (MCF-7). Similarly, MCF-7 cells were transfected with siRNA-negative control (NC), siRNA-Notch3, and siRNA-Notch3 plus pCMV3-GSK3β-GFPSpark. Experiments were performed in triplicate with three follow-up elds recorded for each well.
Migration and invasion assays BD Falcon Cell Culture Inserts (24-well plates, 8 µm; BD Biocoat, BD Biosciences, Franklin Lakes, NJ, USA) were used for cell migration assays as described previously [22,23]. In an invasion assay, cells were inoculated in the upper compartment of Matrigel-coated inserts (24-well plates, 8 µm; BD Biocoat, BD Biosciences). All experiments were conducted in three separate assays.
Sequences of all primers used in this study are shown in Table S3.

Construction of reporter assay vector and CSL-binding site mutation
A reporter assay vector, pGL3-GSK3β-enhancer, containing a CSL-binding site region of the GSK3β promoter was constructed to further explore the relationship between Notch3 and GSK3β by dualluciferase reporter assay. A pGL3-GSK3β-enhancer-promoter mutant construct was generated by deleting three bases, "TCC", of the CSL-binding site in the wild-type pGL3-GSK3β-enhancer-promoter vector using a site-directed mutagenesis kit (Cat. SDM-15, Saibaisheng Gene, China) according to the manufacturer's instructions.

Immunohistochemistry
Human breast cancer specimens were obtained from 68 patients who underwent breast cancer surgery at the Cancer Hospital of Shantou University Medical College in 2014. Immunohistochemical methods were undertaken as previously described [26].
The expression of Notch3 and GSK3β was detected, with anti-mouse or rabbit IgG antibody (Abcam) used as negative controls. The immunohistochemical analysis criteria of Notch3 and GSK3β were previously described [27]. Both Notch3 and GSK3β stains localized in the cell nucleus and cytoplasm. The mean percentages of positive cells were scored as 0 (≤5%), 1 (5-24%), 2 (25-49%), 3 (50-74%), or 4 (>75%). The staining intensity was scored as 1 (weak), 2 (moderate), or 3 (strong). Final histological (h) scores were obtained for each case by multiplying the percentage and intensity score. Protein expression levels were further analyzed by classifying h values as low (based on an h value <5) or high (based on an h value ≥5). The antibodies used are listed in Supplementary Table S1.

Database analysis
The patients' overall survival, relapse-free survival, and distal metastasis survival rate was based on Notch3 and GSK3β expression levels, and automatically drawn by the database website's own software for Kaplan-Meier plots from the Kaplan-Meier Plotter (http://kmplot.com). The entire analysis process was performed in strict accordance of the site's instructions. The Breast Cancer Gene-Expression Miner v4.7 database (http://bcgenex.centregauducheau.fr/BC-GEM/GEM-requete.php) was used to explore the relationship, at the DNA level, between Notch 3 and GSK3β.

Statistical analysis
Statistical analysis involved the use of SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Differences between variables were assessed by χ 2 analysis or two-tailed Student's t-test. Data are presented as the mean ± standard error of the mean (SEM) unless otherwise indicated. Two-sided P < 0.05 was considered statistically signi cant. Each experiment was done at least in triplicate.

Elevated expression of Notch 3 and GSK3β correlated with a luminal subtype in breast cancer cell lines
We rst performed an extensive analysis of the Breast Cancer Gene-Expression Miner v4.7 database in order to explore the relationship between Notch 3 and GSK3β. We found that Notch3 was highly correlated with GSK3β at the DNA level (r = 0.15, P < 0.0001; Fig. 1A, B). To further investigate the association of Notch3 and GSK3β in breast cancer, we evaluated their protein expression levels in different subtypes of breast cancer cell lines. Notch3 protein was mainly present in the luminal subtype cell lines (T-47D and MCF-7) and almost absent in the triple-negative breast cancer (TNBC) cell lines (MDA-MB-231 and BT-549). Interestingly, GSK3β expression paralleled Notch3 expression, with high levels observed in MCF-7 cells and nearly absent levels in MDA-MB-231 cells (Fig. 1C).

Ectopic Notch3 induces GSK3β expression and inhibits EMT in human breast cancer cells
Increasing evidence shows that GSK3β is an EMT suppressor and has been found to regulate Notch1 and Notch2 expression. We investigated whether cross-talk existed between Notch3 and GSK3β in the progression of EMT in breast cancer. We found that the protein level of GSK3β increased in ectopically overexpressed-Notch3 MDA-MB-231 cells but decreased in Notch3-knockdown MCF-7 cells (Fig. 2B, E). Similarly, quantitative real-time PCR showed that the GSK3β mRNA level was consistent with the protein level detected by western blot (Fig. 2A, D).
To explore the role of Notch3 in breast cancer and its effect on EMT progression in breast cancer cells, we silenced endogenous Notch3 in MCF-7 cells using siRNA. Suppression of activated Notch3 (N3ICD) with siRNA signi cantly downregulated E-cadherin and upregulated vimentin, shown in the gure (Fig. 2C).
Notch3 activates GSK3β by directly binding to CSL-binding sites in the GSK3β promoter Based on our results that GSK3β expression is regulated by Notch3 transcriptionally, we speculated on whether Notch3 bound the GSK3β promoter. To search for a speci c binding site in the GSK3β promoter we found a sense binding site at position -1131 to -1126 and three antisense binding sites at -1165 to -1160, -308 to -303, and -167 to -162 (Fig. 3A). To investigate whether Notch3 binds to the endogenous GSK3β promoter in chromosomal DNA, we performed a ChIP assay. We found that Notch3 interacts with the region between positions -368 to -268 containing a CSL antisense binding site (Fig. 3B).
To investigate the activity of the GSK3β promoter directly regulated by Notch3, we constructed a reporter assay vector pGL3-GSK3β-enhancer containing the region from -354 to -268 of the GSK3β promoter. A dual-luciferase reporter assay was carried out to determine GSK3β promoter activity in Notch3overexpressing MDA-MB-231 cells, and Notch3-knockdown MCF-7 cells by co-transfection with a reporter assay vector and Renilla luciferase reporter genes. After Notch3 silencing in MCF-7 cells, the luciferase activity of the GSK3β reporter decreased in a dose-dependent manner (Fig. 3C). Consistently, the luciferase activity of the GSK3β reporter increased following N3ICD overexpression in MDA-MB-231 cells (Fig. 3D). Additionally, GSK3β promoter activity showed no difference in a mutant CSL binding site reporter assay vector ( Fig. 3C and D). These results indicated that Notch3 upregulated GSK3β by directly binding to the GSK3β promoter.

Ectopic Notch3 expression inhibits migration and invasion in vitro, which is attenuated by GSK3β silencing
To investigate the effects of Notch3 overexpression or knockdown of Notch3 on migration and invasion by breast cancer cells, a wound healing assay, transwell in vitro migration, and invasion assays were performed. After cells were cultured for 48 h, the width of the wound in MCF-7 siRNA-Notch3 cells had reduced to only 25% compared to that at 0 h, while the width of the wound in siRNA-NC cells had reduced to about 54.5% compared to that at 0 h. This suggested that silencing of Notch3 disinhibited MCF-7 cells that had migrated into the wound area. Conversely, the effect was reversed by ectopic GSK3β expression as the width of the wound was restored to 43.5%, indicating that the increased migration caused by the silencing of Notch3 was mediated by GSK3β ( Fig. 4A and B). In contrast, we observed that after culturing for 24 h, overexpression of Notch3 inhibited MDA-MB-231 cell migration into the wound region compared with that of the control group (44% vs. 14.5%), while downregulation of GSK3β expression reversed this (21.5%; Fig. 4E and F). We initially evaluated the effects of recombined Notch3 and GSK3β expression on breast cancer cell invasion using migration and invasion assays. Notch3 knockdown increased cell migration by 3.87-fold and 5.35-fold, respectively, which was partially eliminated by GSK3β overexpression (Fig. 4C and D). In contrast, Notch3 overexpression reduced migration by nearly 80.9% and 78.43%, respectively, which was partially abrogated by GSK3β silencing (Fig. 4G and H). These data demonstrated that Notch3 induces GSK3β expression and inhibits wound healing and migration/invasion, an effect that can be attenuated by GSK3β knockdown in vitro.

Positive correlation of Notch3 and GSK3β expression in patients with breast cancer
We analyzed the protein levels of Notch3 and GSK3β in the pathological parameters of 68 human breast cancer samples. Forty-three of the 68 specimens (64.8%) were found to be positive for Notch3 expression. Notch3 negative and positive specimens are shown in Fig. 5A and B, respectively. Positive GSK3β expression was found in 40 of the 68 samples (58.8%). Glycogen synthase kinase-3-beta negative and positive samples are shown in Fig. 5C and D, respectively. Immunohistochemistry of Notch3 and GSK3β in breast cancer specimens revealed that Notch3 and GSK3β expression are positively associated (r=0.416, P = 0.001; Table). In general, these data demonstrated that Notch3 correlated in a positive manner with elevated GSK3β expression.
High expression of both Notch3 and GSK3β mRNA predicts an improved prognosis in breast cancer patients We rst examined the prognostic effect of the expression of Notch3 and GSK3β in www.kmplot.com.
High expression of Notch3 (P =4.1e-07) and GSK3β (P=3.8e-05) mRNAs correlated with an improved RFS for all patients with breast cancer (Fig. 6A, B). Restricting the analysis to intrinsic subtypes of breast cancer, we found that high expression of Notch3 correlated with a greater RFS for luminal A, luminal B, and human epidermal growth factor receptor 2 (Her2) subtypes but not the basal-like subtype compared to those with low expression of Notch3 (Fig. 6D, G, J,M). In addition, patients with luminal A, luminal B, and Her2 subtypes showed a superior RFS when GSK3β was overexpressed but not the basal-like subtype (Fig. 6E, H, N, K). Notably, high mRNA expression of both Notch3 and GSK3β predicted greater relapsefree survival in overall (P = 1.8e-05) and luminal A breast cancer patients (P = 0.0082; Fig. 6C, F),but not luminal B, basal-like or Her 2 subtypes (Fig. 6I,L,O). In conclusion, Notch3 and GSK3β mRNA overexpression suggests a good prognosis for patients with breast cancer.

Discussion
The Notch receptor is a highly conserved type I transmembrane glycoprotein that is involved in cell differentiation, proliferation, and survival, which play important roles in various tumors [28]. Notch1-4 has a highly similar structure and corresponding functional role in mammals, but has a distinct function in normal breast development and breast cancer [29]. It has been identi ed that inhibition of Notch1 expression can reverse the EMT process of breast cancer cells, thereby inhibiting cell migration and invasion [30]. Furthermore, a clinical study found that Notch1 correlated to a poor prognosis in patients with breast cancer [31]. Notch2 overexpression predicts better overall survival in women with breast cancer, which is completely different from Notch1 [32]. A recent study showed the carcinogenic effect of Notch4. Zhou et al. revealed that Notch4 overexpression can promote the EMT of breast cancer cells and maintain the self-renewal properties of mesenchymal breast cancer stem cells [33], which is of concern and is associated with endocrine drug resistance in breast cancer.
Although the basic structure of Notch3 bears a resemblance to other Notch receptor members, many differences exist in the intracellular domain as well as a special shorter transcription activation domain. Several studies have indicated that Notch3 may play a considerable role in the formation of mammary epithelial cells during breast development [34]. Daniel et al. [35] revealed that Notch3 could label luminal progenitor cells and inhibit the proliferation of tumor cells. Cui et al. [13] found that Notch3 regulated cell senescence by regulating P21, thus inhibiting the occurrence and development of tumors. Similarly, Notch3 is associated with the inhibition of cell proliferation and apoptosis in HER2-negative breast cancer [36]. Our previous studies revealed that ERα inhibits EMT by suppressing Bmi1 [14], and Notch3 upregulates ERα [21]. Meanwhile, another study from our group showed that Notch3 can inhibit EMT in breast cancer epithelial cells by transactivating Kibra [15]. Therefore, Notch3 may have a pivotal role in tumor suppression, especially in breast cancer EMT.
The loss of E-cadherin expression is regarded as an essential event in EMT. It leads to disruption of epithelial cell polarity induced EMT [37] thus maintaining the mesenchymal phenotype and enhancing the invasion and metastasis of cancer cells [38]. A variety of molecular mechanisms can repress E-cadherin expression via transcriptional inhibition and promotor methylation. Doble et al. demonstrated that GSK3β acts as an EMT suppressor of the zinc-nger transcriptional repressor, snail, which regulates epithelial marker E-cadherin [17]. These results indicated that both Notch3 and GSK3β inhibit EMT in breast cancer and GSK3β can physically bind and phosphorylate the intracellular domain of two Notch paralogs [18,19]. However, the interaction of Notch3 and GSK3β in EMT, a potential key step in breast cancer metastasis that contributes to breast cancer-related deaths, remains unclear.
Our results revealed that GSK3β was highly expressed in the luminal-type breast cancer cell lines, MCF-7 and T-47D, but was expressed at low levels in the TNBC cell lines, MDA-MB-231 and BT-549. This is consistent with our previous ndings that the expression pattern of Notch3 is similar to GSK3β in these cell lines. In addition, our previous studies found that the CSL-binding element, GGGAA, participated in regulating the ER and GATA3 [21,39]. In this study, we demonstrated that Notch3 bound to the GSK3β promoter, which contains CSL-binding elements. Here, we found that the luciferase reporter activity of GSK3β decreased in the form of a concentration gradient after Notch3 knockdown in MCF-7 cells. In contrast, GSK3β-mediated luciferase activity increased after N3ICD over-expression in MDA-MB-231 cells. Our study on its molecular mechanism revealed that inhibition of Notch3 downregulated GSK3β mRNA and protein levels in MCF-7 cells signi cantly. Conversely, Notch3ICD overexpression upregulated that of GSK3β in MDA-MB-231 cells as illustrated in Figure 7.
In summary, our research ndings indicated that Notch3 represses the processes of EMT in breast cancer, which is consistent with our previous reports. Remarkably, this nding demonstrates for the rst time that Notch3 may inhibit the EMT process in breast cancer cells through transcriptionally up-regulating GSK3β.
From data obtained from clinical tissue, Notch3 expression was signi cantly consistent with GSK3β expression in breast cancer tissue samples. In a prognostic analysis, high expression of mRNA of both Notch3 and GSK3β was related to better RFS in all patients with breast cancer studied, which implies that Notch3 and GSK3β are bene cial biomarkers in this disease. Our results highlight a novel mechanism for exploring how Notch3 regulates EMT as well as the crosstalk between Notch and Wnt signaling pathways. This may have important implications for identifying new biomarkers for the prognosis of and as therapeutic targets in breast cancer.

Ethics approval
The work was conducted in accordance with the Declaration of Helsinki. The study was approved by the Medical Ethics Committee of the Cancer Hospital of Shantou University Medical College and Xiang'an Hospital of Xiamen University.

Competing interests
None of the authors have any nancial or personal relationships with other people or organizations that inappropriately in uenced this research.

Informed consent
This article does not contain any studies with human participants or animals. The entire study protocol was approved by the Medical Ethical Committee of the Cancer Hospital of Shantou University Medical College and Xiang'an Hospital of Xiamen University. The need for obtaining informed consent was waived by this committee.

Reference
Page 12/22 References are not available with this version.   Notch3 transactivates GSK3β by directly binding to the GSK3β promoter A: A schema of the four CSL-binding element-containing primers (glycogen synthase kinase-3-beta [GSK3β] 1-1, GSK3β 2, and GSK3β regions containing respective single CSL-binding elements; region GSK3β 1-2 containing two CSL-binding elements) used for chromatin immunoprecipitation (ChIP) assays. B: Products of ChIP were ampli ed by PCR and analyzed by 2% agarose gel electrophoresis. A speci c band was seen in the GSK3β 2 region; C: MCF-7 cells co-transfected with gradient concentrations of siRNA-Notch3 ("+" 10 pmol, "++" 20 pmol, "+++" 40 pmol, "-" 0 pmol) or the same concentrations of subtypes but not the basal-like subtype (K). A superior RFS was observed for breast cancer patients expressing both high Notch3 and GSK3β levels in all patients, luminal A subtype (C, F) but not luminal B, basal-like, or Her 2 subtypes (I,L, and O).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.