Ezrin accelerates breast cancer liver metastasis through promoting furin-like convertase-mediated cleavage of Notch1

Abstract

Background

Ezrin, known as a crosslinker between the plasma membrane and actin cytoskeleton, is closely associated with breast cancer (BC) progression. Here, we explored a novel role of ezrin in breast cancer liver metastasis (BCLM).

Methods

The clinical relevance of ezrin was evaluated using in silico tools and confirmed in BC specimens. The effect of ezrin on proliferation, migration and invasion was examined in vitro and in vivo using murine primary liver-metastatic breast cancer cells (mLM). The molecular mechanism involved in ezrin-mediated activation of the Notch1 signaling pathway was elucidated using in vitro models.

Results

Data-mining demonstrated that ezrin mRNA and protein expression is up-regulated in breast cancer cohorts and has prognostic significance. Ezrin overexpression promotes cell proliferation, migration and invasion in vitro and in vivo. Hairy and enhancer of split-1 (Hes1) is one of the most significantly enriched candidates of differentially expressed genes in ezrin overexpression and control mLM cells. Ezrin can positively regulate Hes1 mRNA and protein expression, and their coexpression was associated with poor prognosis in BC patients. Ezrin promoted BC cell proliferation in a Hes1-dependent manner without directly interacting with Hes1. The functional link between ezrin and Hes1 is dependent on Notch1 activation through promotion of furin-like convertase cleavage.

Conclusion

Our results demonstrated that ezrin drives BCLM through activation of the Notch signaling pathway via furin-like convertase. These findings provide a better understanding of the mechanism of ezrin in breast cancer progression, with the goal of discovering a novel target for the treatment of BCLM in the future.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-022-00761-x.

Introduction

Breast cancer (BC) is the most commonly diagnosed cancer and the leading cause of cancer-related deaths among females worldwide. BC has at least five subtypes, including luminal A, luminal B, human epithelial growth receptor type 2 (HER-2), basal-like, and claudin-low [1]. With the advance of early diagnosis and early treatment of BC, its prognosis has improved significantly. The 5-year survival rate of primary breast cancer is more than 90% and the mortality rate has decreased, but the 5-year survival rate still remains at 25% once distant metastases have developed [2]. BC can metastasize to bone, lung, liver and brain via the circulatory system [3]. The 5-year overall survival rate of bone, lung, liver metastases is 22.8%, 16.8% and 8.5% respectively and brain metastases has an even shorter survival length with one-year survival rate of 20% [4–9]. Most of these patients have a poor prognosis even when treated with chemotherapy after the detection of metastasis [10]. One of the reasons why clinical treatment can not achieve the desired result is that breast cancer metastases to different sites involving unique molecular mechanisms. Therefore it is necessary to adopt specialized therapeutics according to the tissue distribution of metastases. At present, many studies have focused on breast cancer bone and lung metastasis, but there are few studies on liver metastasis. In recent years our research has concentrated on breast cancer liver metastasis (BCLM) to determine its intrinsic regulation mechanism.

Ezrin, a member of ezrin-radixin-moesin (ERM) proteins, is a crucial actin-binding protein (ABP) that is involved in organization of the cell cortex and plays a major role in a variety of cellular functions including cell morphology regulation, cell adhesion and motility, protein localization, vesicle traffic and receptor signal transduction [11]. Ezrin can link actin filaments to plasma membrane components that directly interact with the cytosolic region of membrane receptors (e.g. integrins, chemokine receptors) or it can indirectly bind to transmembrane proteins (e.g. CD44, CD43, ICAM-2) through scaffolding proteins [12]. Under physiological conditions, ezrin exists in a closed conformation due to intramolecular or intermolecular interactions between N- and C- terminal domains that mask the membrane and F-actin binding sites. Following activation, phosphorylation at T567 in ezrin transforms it into an active state with binding to the cell membrane and F-actin [13]. Increasing evidence has shown that ezrin not only acts as a cytoskeletal cross-linker, but also as a scaffolding protein in various cellular processes, for example, regulation of gene translation, mediating vesicle trafficking to the plasma membrane [14, 15]. These findings imply that ezrin plays an important role in various physiological and pathological processes and is a potential therapeutic target.

Ezrin is closely associated with BC metastasis and poor prognosis [16–18]. Chen et al. demonstrated a positive role for ezrin in regulation of the metastasis and angiogenesis of BC, and that ezrin interacted with AKT to activate the downstream signaling pathway leading to increased BC metastasis and angiogenesis [19]. Thus, ezrin may be a potential biomarker for predicting clinical prognosis and can be used as a therapeutic target in BC metastasis. However, many molecular mechanisms underlying the involvement of ezrin in the development and progression of BC remain unclear. In the present study, we explored the novel role of ezrin in BC and showed that ezrin can accelerate BCLM through activation of the Notch signaling pathway.

Materials and methods

Cell isolation and animal model establishment

Highly aggressive cells were isolated from lung metastatic lesions of the MMTV-PyMT mouse and injected into the mammary fat pad of six-week-old female BALB/c nude mice for orthotopic metastasis. Approximately six weeks later, liver metastases were discovered and identifed using HE staining (Supplementary Fig. 1A). A piece of freshly isolated liver metastatic lesion (about 200 mg) was put into DMEM/F12 supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine and 50 μg/mL ciprofloxacin, and minced into very small pieces with a tissue grinder under sterile conditions. After treatment with ACK lysis buffer (160 mM NH₄Cl, 130 mM EDTA, 12 mM NaHCO₃), the tissue pieces were collected and transferred to a 15 mL Falcon tube. After standing for 5–10 min and centrifugation at 500 g for 5 min, the supernatant was removed and the pellet resuspended in PBS. To obtain a single cell suspension, the resuspended cells were passed through a 100-μm cell strainer. After washing twice with PBS, the pellet was pipetted up and down using a 1 mL pipette for final single cell dissociation and then plated into 6 well plates. Culture medium was exchanged the next day and subsequently exchanged every three to four days until there was no fibroblast contamination. After verification, murine primary liver-metastatic breast cancer cells (murine LM, mLM) were successfully isolated and the breast cancer liver metastasis animal model was established using mLM cells (Supplementary Fig. 1B).

Cell cultures

mLM cells were cultured in DMEM/F12 medium with 10% FBS and 1% penicillin/streptomycin. MDA-MB-231 cells were obtained from Sciences Cell Resource Center (Shanghai, China) and cultured in L-15 medium with 10% FBS and 1% penicillin/streptomycin. 4T1 cells were purchased from Procell (Wuhan, China) and cultured in DMEM with 10% FBS and 1% penicillin/streptomycin. All cells were cultured at 37℃ in a humidified atmosphere with 5% CO₂.

Lentiviral packaging and stable cell lines

The lentiviral vector carrying the ezrin gene or the control vector were purchased from VectorBuilder and assembled in 293FT cells, and the viruses collected. After infecting mLM cells or MDA-MB-231 cells with lentivirus, the stable transfected cell lines were established by puromycin selection.

Transient siRNA transfection

Human ezrin siRNA and murine ezrin siRNA were obtained from Shanghai GenePharma (GenePharma, Shanghai, China). ON-TARGET plus mouse Hes1 siRNA smartpool and negative control siRNA pool were purchased from Horizon/Dharmacon (GE Healthcare Dharmacon Inc., USA). The sequences of all siRNAs used are presented in Supplementary Table 1. All siRNAs were dissolved to a final concentration of 20 μM and stored at -20℃. For RNA interference, siRNAs were transfected using Lipofectamine 3000 (Invitrogen, Carlsbad, USA) according to the manufacturer’s instructions.

Western blotting

Cells were lysed with RIPA buffer containing 1% protease inhibitor mixture (Millipore, USA) and 1 mM PMSF. Protein samples were separated on 10% or 12% denaturing polyacrylamide gels and transferred to PVDF membrane. The membrane was incubated with 5% fat-free milk for 1 h, and overnight with primary antibodies at 1:1000 dilution. The antibodies used are presented in Supplementary Table 2. After washing, the membrane was probed with secondary antibody at room temperature (RT) for 1 h. Protein bands on the membrane were detected by enzyme-like chemiluminescence (ECL) reagents (Millipore, USA).

Immunoprecipitation (Co-IP)

Total protein was obtained using WB/IP lysis buffer (Beyotime Biotechnology, Shanghai, China) supplemented with 1% protease inhibitor mixture and 1 mM PMSF. After pre-clearance, the supernatant was incubated with primary antibody or IgG overnight at 4℃. Protein A/G plus beads (Beyotime Biotechnology) were added and incubated for 4 h at 4℃. Precipitates were washed with lysis buffer four times, and added with 1 × loading buffer. Western blotting was performed to detect the presence of the indicated protein.

Immunofluorescence

Cells grown on a confocal dish were fixed with 4% paraformaldehyde for 30 min at RT. After permeabilizing with 0.1% TritonX-100 for 10 min at RT and blocked with 10% normal goat serum for 1 h, they were incubated with the primary antibodies at 1:100 dilution overnight at 4℃. After washing with PBS, the cells were incubated with the secondary antibodies and stained with DAPI. Images of cells were acquired using a confocal microscope and prepared with ImageJ software.

Proliferation assay using RTCA

An E-plate was used for measuring cell proliferation with the xCelligence® DP system as described previously [20]. Briefly, cells were trypsinized, counted with a haemacytometer, and resuspended in culture medium. Background was taken from the wells by adding 50 μL culture medium to the E-plate. 4 × 10³ or 5 × 10³ cells were plated with fresh medium and incubated for 30 min in the RTCA cradle at 37℃ and 5% CO₂ before signal recording. The signals were recorded every 15 min until the end of the experiment (up to 72 h).

Migration and invasion assay using RTCA

The CIM plate, a two-chambered system is used for measuring migration and invasion with the xCelligence® DP system as described previously [21]. 2 × 10⁴ cells were suspended in serum-free medium and added to the top chambers. The lower chambers contained culture media with 10% FBS. The cells were incubated for 30 min in the RTCA cradle at 37℃ and 5% CO₂ before signal recording. The signals were taken every 15 min for cell migration until the end of the experiment (up to 48 h).

For invasion assays, the membrane in the top chamber was pre-coated with 20 μL of 1:40 dilution of Matrigel™ (BD Biosciences) in serum-free medium and incubated for 4 h at 37℃. Subsequent steps were performed in the same way as for the cell migration assay. The cell index was recorded up to 72 h.

Colony-formation assay

Cells were plated in 6-well plates at a concentration of 1000 cells per well and the culture medium changed every three days. After incubating for 2 weeks, cells were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet and the total number of visible colonies per well counted.

Animal studies

MMTV-PyMT mice were purchased from the Model Animal Research Center of Nanjing University and female BALB/c nude mice from the Guangdong Medical Laboratory Animal Center. The mice were housed at 25℃ with 50–60% humidity under a 12-h light/dark cycle with free access to food and water.

For orthotopic injection, six-week-old female BALB/c nude mice were randomly divided into two groups (ezrin overexpression group, n = 10; vector control group, n = 5). 5 × 10⁶ ezrin overexpression or vector control mLM cells were injected into the mammary fat pad of mice. The mean tumor volume was measured after injection and the tumor volume was calculated using the formula of length × width × height × π/6 (mm³). After 30 days, the mice were sacrificed by cervical dislocation under deep anesthesia with isoflurane and the tumor tissues obtained and fixed with 4% paraformaldehyde for further study.

For tail vein metastasis, 5 × 10⁵ ezrin overexpression or vector control mLM cells were injected into the tail veins of six-week-old female BALB/c nude mice (n = 5 each group). After three weeks, the mice were sacrificed by cervical dislocation under deep anesthesia with isoflurane. The lung tissue and the liver tissues obtained and fixed with 4% paraformaldehyde for further study.

For orthotopic metastasis assays, 1 × 10⁶ mLM cells were injected into the mammary fat pad of six-week-old female BALB/c nude mice (n = 5 each group). The mice were sacrificed by cervical dislocation under deep anesthesia with isoflurane six weeks later. The primary site tumor and liver metastatic lesion were obtained and fixed with 4% paraformaldehyde for further study.

For intrasplenic injection-based metastasis assay, 1 × 10⁶ ezrin overexpression or vector control mLM cells were injected into the spleen of six-week-old female BALB/c nude mice with a 30-gauge needle (n = 5 each group). After two weeks, the mice were sacrificed by cervical dislocation under deep anesthesia with isoflurane and the liver tissue obtained and fixed with 4% paraformaldehyde for further study.

For drug treatment, on day 24 (orthotopic injection) or day 3 (intrasplenic injection-based metastasis and tail vein metastasis), NSC305787(0.240 mg/kg/day) and decRVKR-CMK (630 μM) were injected intraperitoneally, and PBS injection was used in controls.

Clinical samples

A total of 140 breast cancer tissue microarray samples (HBreD140Su07) were collected from Shanghai Outdo Biotech between Aug. 2004 and Dec. 2008. All tissues were fixed in 10% buffered formalin and embedded in paraffin blocks. The pathological parameters, including age, tumor size, clinical stage, differentiation, nodal metastasis and survival data, were carefully reviewed in all 140 breast cancer cases.

Hematoxylin–eosin staining

4-μm sections were oven-dried at 65℃ for 1 h and cooled to reach RT. The sections were dehydrated in gradient alcohol, washed, stained by hematoxylin for 3–5 min, followed by washing and differentiation in hydrochloric acid alcohol for 10 s. After being rinsed in water and dehydrated in 85% and 95% alcohol, sections were counterstained in eosin for 5 min, dehydrated in gradient alcohol, and mounted with neutral balsam. The histological appearance was observed under an optical microscope.

Immunohistochemistry

The sections were deparaffinized in xylene and rehydrated with graded ethanol (100% alcohol for 10 min, 85% alcohol for 5 min, 75% alcohol for 5 min, ddH₂O for 5 min). After antigen retrieval with EDTA buffer for 15 min, sections were blocked in 3% BSA for 30 min and incubated with primary antibody (1:1000 dilution for ezrin, 1:8500 dilution for Hes1) overnight at 4℃. After washing, slides were incubated with the secondary antibody for 1 h, washed and then stained with 3, 3’-diaminobenzidine (DAB) peroxidase staining kit (DAKO). The sections were counterstained with hematoxylin, the slides sealed with coverslips and monitored using an optical microscope.

The intensity and extent of immunostaining for all samples were evaluated by two pathologists under double-blind conditions. The staining intensity was graded as negative (0), weak (1), moderate (2) or strong (3), and the extent was graded according to the percentage of positive cell as 0 (no staining cells), 1 + (staining cells < 10%), 2 + (10% > staining cells < 50%), 3 + (50% < staining cells < 80%) and 4 + (staining cells > 80%). The score of immunostained tissues were obtained by multiplying the intensity and extent of staining scores.

Microarray analysis

RNA quantity and quality were detected using Nanodrop ND-1000 spectrophotometer, and RNA integrity measured by agarose gel electrophoresis. Total RNA from each sample was linearly amplified and labeled with Cy3-UTP. After purification, the concentration and specific activity of the labeled cRNAs were measured using NanoDrop ND-1000 spectrophotometer. After adding 11 μL 10 × blocking agent and 2.2 μL of 25 × fragmentation buffer, 1 μg of each labeled cRNA was fragmented, heated at 60 °C for 30 min, and finally 55 μL 2 × GE hybridization buffer was added to dilute the labeled cRNA. 100 μL of hybridization solution was dispensed into the gasket slide and assembled to the gene expression microarray slide (Agilent Mouse GE 4 × 44 K v2 Microarray, Design ID: 026,655). The slides were incubated for 17 h at 65 °C in an Agilent hybridization oven. After washing and fixation, the hybridized arrays were scanned using the Agilent DNA Microarray Scanner. Acquired array images were analyzed using Agilent Feature Extraction software (version 11.0.1.1). Quantile normalization and subsequent data processing were performed using the GeneSpring GX v12.1 software package (Agilent Technologies).

Bioinformatics analysis

The mRNA levels of ezrin in different cancers based on TCGA expression profile data were analyzed using Tumor Immune Estimation Resource (TIMER, http://cistrome.shinyapps.io/timer/). Data on ezrin mRNA expression in breast cancer and corresponding normal tissue from The Cancer Genome Atlas (TCGA) database and ezrin protein expression from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) was assessed using the UALCAN website (http://ualcan.path.uab.edu). Kaplan–Meier survival analysis of patients that were divided into two groups (low ezrin expression and high ezrin expression) was performed using a Kmplotter online tool. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were applied to identify the underlying functions of ezrin and some vital pathways. Three categories including biological process (BP), cellular component (CC) and molecular function (MF) were explored from GO analysis. The P value less than 0.05 denotes the significant GO term enrichment or the significant KEGG pathways. Gene set enrichment analysis (GSEA) was performed in control and ezrin overexpression groups. The GSEA procedures were performed based on the ClusterProfiler package in R software. The microarray dataset GSE62598 was used to screen the gene expression levels of claudin family members in parental 4T1 cell and liver aggressive 4T1 cell.

RT-PCR

Total RNA was extracted from cultured cells with RNAiso Plus. Reverse transcription was performed using PrimeScript™ RT reagent kit at 37 ℃ for 15 min with 1 μg total RNA, 4 μL 5 × PrimeScript RT mix and DEPC-H₂O up to 20 μL. The expression of target genes was analyzed using SYBR Master Mixture. The thermocycling conditions were as follows: pre-denaturation for 90 s at 95 ℃, 40 cycles of denaturation for 5 s at 95 ℃, annealing at 60 ℃ for 30 s and extension at 60 ℃ for 30 s. The sequences of all primers used are presented in Supplementary Table 3. The relative expression of the target genes was normalized to Rn18s using the 2^−ΔΔCT analysis method.

Furin-like enzyme activity assay

Fluorogenic substrate peptides (Boc-RVRR-AMC, ALX-260–040, Enzo) were used to measure furin-like proprotein convertase activity. The experiment was performed as described previously [22].

Statistical analysis

Statistical analysis was performed with SPSS software version 10.0. All the data are from more than three independent experiments and expressed as mean ± standard deviation (SD). Student’s t test was used for only two groups, and one-way ANOVA followed by Dunnett’s test for three or more groups. Differences were considered to be significant at P < 0.05. According to the expression level of ezrin and Hes1 in breast cancer tissues, the patients were divided into low and high expression groups. Survival analysis (Kaplan–Meier curve) and log-rank test were used to calculate the difference of total survival between low and high expression groups. The χ² test was used to analyze the association between categorical variables. Spearman correlation analysis was used to evaluate the correlation between ezrin and Hes1 scores.

Results

Ezrin promotes mLM cell proliferation, migration and invasion in vitro and in vivo

Using the TIMER database, the expression levels of ezrin in different cancers was compared between normal and tumor tissue samples. Ezrin was significantly upregulated in breast invasive carcinoma (BRCA), cholangiocarcinoma (CHOL), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), stomach adenocarcinoma (STAD) and uterine corpus endometrial carcinoma (UCEC), downregulated in colon adenocarcinoma (COAD), kidney chromophobe (KICH), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD) and thyroid carcinoma (THCA), and was not altered in bladder urothelial carcinoma (BLCA), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), liver hepatocellular carcinoma (LIHC), pheochromocytoma and paraganglioma (PCPG) and rectum adenocarcinoma (READ) (Supplementary Fig. 2). In breast cancer cohorts, ezrin mRNA and protein expression is up regulated and has prognostic significance, demonstrating ezrin is closely related to BC progression (Fig. 1A-B). However, the novel role and precise molecular mechanism of ezrin in BCLM remains unclear. We successfully isolated mLM cells from a liver metastatic lesion of orthotopic breast cancer mice. mLM cells could rapidly metastasize to liver. Obvious liver metastases could present after 3–4 weeks of orthotopic implantation, which is a versatile tool to study BCLM in vitro and in vivo (Supplementary Fig. 1). To further confirm the role of ezrin in BCLM and uncover the underlying mechanism, we used mLM cells in cell and animal models. mLM cells were stably transfected with ezrin and with vector-only as a negative control. Ezrin overexpression in mLM cells was confirmed by qPCR and Western blot (Fig. 1C). To identify changes in mLM cell progression associated with ezrin overexpression, cell proliferation, migration and invasion were investigated using RTCA assays. Ezrin overexpression markedly increased mLM cell proliferation (P = 0.04 at 60 h, P = 0.004 at 72 h), migration (P = 0.007 at 24 h, P = 0.01 at 36 h) and invasion (P = 0.05 at 60 h) compared with controls (Fig. 1D). It is interesting to note that migration of the ezrin overexpression cells starts to increase from around 15 h, suggesting that this is an early response to the increased level of ezrin in the cells. Ezrin overexpression also increased the number of colonies formed compared with controls (P = 0.001, Fig. 1E). In addition, ezrin overexpression cells showed increased protein expression of the mesenchymal markers N-cadherin and vimentin in contrast to control cells (Supplementary Fig. 3). These results indicated that ezrin increased the growth and metastatic abilities of mLM cells. To further determine the role of ezrin in vivo, ezrin overexpression and vector control cells were injected into the mammary fat pad of nude mice. The tumor growth kinetics demonstrated that ezrin overexpression promoted tumor growth, while the small molecule inhibitor NSC305787 that directly interacts with ezrin to inhibit its function could suppress tumor growth, indicating that ezrin could promote tumor growth. (Fig. 1F). Moreover, ezrin overexpression also promoted liver metastases of mLM cells after implantation (P = 0.0001, Fig. 1G). The expression of ezrin in the liver metastatic lesion of ezrin overexpression group was higher than that in control group (P = 0.009 vs control group, Fig. 1H). The result also showed that the positive rate of ki67, a marker of cell proliferative ability, in liver metastatic lesions of the ezrin overexpression group increased significantly compared with that in the control group, indicating that ezrin overexpression promoted the growth of metastatic tumors (P = 0.004 vs control group, Fig. 1I). These results further demonstrated that ezrin promotes BCLM progression in vitro and in vivo.

Fig. 1.

(A) Ezrin promotes mLM cell proliferation, migration and invasion in vitro and in vivo. Ezrin mRNA and protein expression is up-regulated in a breast cancer cohort. The violin plot (truncated) on the left is ezrin mRNA expression in breast cancer (n = 1097) and normal breast tissue (n = 114), P < 0.001; The violin plot (truncated) on the right is ezrin protein expression in breast cancer (n = 125) and normal tissue (n = 18), P < 0.001. (B) The curve plot is overall survival between ezrin high and low transcript groups in breast cancer. The red line indicates patients with high ezrin levels (n = 616) and the black line indicates patients with low ezrin levels (n = 473), log rank P = 0.005. (C) Ezrin expression was measured by qPCR (left) and western blot (right) in parental, vector-overexpression, and ezrin-overexpression mLM cells. ^***P < 0.01 ezrin overexpression mLM cells versus control mLM cells. Hsp70 from the same loading was used as a loading control. (D) Cell proliferation, migration and invasion were determined using RTCA. n = 3, ^*P < 0.05, ^**P < 0.01, ezrin overexpression mLM cells versus control mLM cells. (E) Cell colony formation ability was determined by colony formation assay and quantified. n = 3, ^**P < 0.01 ezrin overexpression mLM cells versus control mLM cells. Data represent the mean ± SD. (F) Ezrin overexpression in mLM cells enhanced the tumor growth and NSC305787 treatment suppressed the promotion of ezrin in tumor growth. The mice were sacrificed 30 days after orthotopic injection and the tumors were collected (left). Statistical results showed the tumor growth kinetics (right). n = 5, ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. Grey star symbol, ezrin overexpression group versus vector control group. Black star symbol, ezrin overexpression group versus ezrin overexpression group with NSC305787 treatment. (G) Ezrin overexpression promoted liver metastasis. The mice were sacrificed three weeks after tail vein injection and the liver metastatic lesions were collected. Liver metastasis area in the ezrin overexpression group was larger than in the control group (n = 5, P = 0.0001). (H) Representative images of ezrin staining. Scale bar, 50 μm. Positive rate of ezrin expression in ezrin overexpression group and control group (n = 5, P = 0.009). (I) Representative images of Ki67 staining. Scale bar, 50 μm. Positive rate of Ki67 in ezrin overexpression group and control group (n = 5, P = 0.005). Data represent the mean ± SD. ^**P < 0.01, ezrin overexpression group versus control group. RTCA, real time cell analysis

Identification of differentially expressed genes in ezrin overexpression mLM cells

To explore the unique molecular mechanism of ezrin in BCLM, differentially expressed genes (DEGs) between ezrin overexpression mLM cells and control cells were determined by microarray analysis. A total of 8964 oligo probes were upregulated, while 10,013 probes were downregulated (Fig. 2A). Approximately 58.15% of the transcriptome changed significantly in expression levels, as measured by > twofold change and P < 0.05. To explore the novel functions of ezrin and the related vital pathways, GO and KEGG enrichment analysis were performed to deduce possible enrichment functions and pathways. Biological processes (BP) including multicellular organismal process (GO:0,032,501), biological regulation (GO:0,065,007), anatomical structure development (GO:0,048,856), multicellular organism development (GO:0,007,275), system development (GO:0,048,731) for upregulated genes and celluar metabolic process (GO:0,044,237), metabolic process (GO:0,008,152), primary metabolic process (GO:0,044,238), organic substance metabolic process (GO:0,071,704), nitrogen compound metabolic process (GO:0,006,807) for downregulated genes were significantly associated with ezrin overexpression (Fig. 2B). Ezrin also prominently affected the cellular component (CC), such as cell periphery (GO:0,071,944), plasma membrane (GO:0,005,886), organelle (GO:0,043,226), cytoplasm (GO:0,005,737). In addition, molecular functions (MF) such as binding (GO:0,005,488), signaling receptor activity (GO:0,038,023), catalytic activity (GO:0,003,824), protein binding (GO:0,005,515) were highly regulated by ezrin. KEGG analysis showed that the most important pathways included neuroactive ligand-receptor interaction (mmu04080), signaling pathways regulating pluripotency of stem cells (mmu04550), Wnt signaling pathway (mmu04310), oxidative phosphorylation (mmu00190) (Fig. 2C). To further find the key genes that are associated with ezrin action, we selected the top 100 upregulated and downregulated genes according to the log ratio expression values (Fig. 2D and Supplementary Table 4) and then overlapped the common genes from these top 100 genes and KEGG analysis (Fig. 2E). Seven genes including Bcl2a1c, Colla1, Hes1, Col2a1, Shc4, Vangl1 and Aph1b were potentially core elements in both analyses (Table 1). Among these significantly enriched candidates, Hes1, a transcription factor that is involved in cancer development, appeared to be a good candidate for further investigation [23].

Fig. 2.

Identification of differentially expressed genes in ezrin overexpression-mLM cells. (A) Volcano plot of differentially expressed genes as measured by > twofold change and P < 0.05. (B) Top 5 significantly enriched BP, CC and MF for the upregulated and downregulated differentially expressed genes respectively. (C) The top 10 KEGG enrichment pathways of the upregulated and downregulated differentially expressed genes. (D) Heatmap of the top 100 upregulated and downregulated genes according to the log ratio expression values. (E) The common genes were overlapped from these top 100 genes and KEGG analysis. BP, biological process; CC, cellular component; MF, molecular function; KEGG, Kyoto Encyclopedia of Genes and Genomes

Table 1.

The significantly enriched candidates

Accession	Gene symbol	Log2 ratio	Main function
NM_001113515	Col2a1	300.497	Protein binding and platelet-derived growth factor binding
NM_007742	Colla1	258.966	Protein binding and platelet-derived growth factor binding
NM_008235	Hes1	224.577	DNA-binding transcription factor activity and sequence-specific DNA binding
NM_177545	Vangl1	4.225	Intestinal trefoil factor induced wound healing
NM_199022	Shc4	3.119	Protein kinase binding and receptor tyrosine kinase binding
NM_177583	Aph1b	-9.218	Peptidase activity
NM_007535	Bcl2a1c	-9.221	Protein heterodimerization and homodimerization activity

Ezrin regulates Hes1 expression and their coexpression is associated with poor prognosis in human BC

Aberrant Hes1 is present in breast cancer, but the relationship between ezrin and Hes1 remains unkown. To elucidate the relationship between ezrin and Hes1 and the related underlying molecular mechanisms, we selected MDA-MB-231 as a human BCLM cell model in addition to the mLM cells. MDA-MB-231 cells were stably transfected with ezrin and with vector-only as a negative control. Ezrin overexpression in MDA-MB-231 cells was confirmed by western blot (Supplementary Fig. 4). Then to confirm the relationship between ezrin and Hes1, Hes1 expression in ezrin overexpression mLM cells and MDA-MB-231 cells was detected by qPCR and western blot. The results showed that Hes1 mRNA and protein expression was upregulated in ezrin overexpression-mLM cells and -MDA-MB-231 cells (Fig. 3A-B). Hes1 expression was also investigated in ezrin knockdown cells. Knockdown of ezrin expression by siRNA pool was confirmed using western blot (Supplementary Fig. 5). Hes1 expression decreased after ezrin knockdown in MDA-MB-231 cells and 4T1 cells, further indicating ezrin can regulate Hes1 expression (Supplementary Fig. 6). To further analyze the correlation between ezrin and Hes1 and the prognosis of patients with breast cancer, immunohistochemistry was applied to assess the expression of ezrin and Hes1 in human BC tissues. The relationship between ezrin and Hes1 expression and clinicopathologic parameters was examined using Pearson χ² test. No correlation was found between ezrin or Hes1 expression levels and age, tumor size, lymph node metastasis, or TMN stage, but there was a significant correlation between ezrin and Hes1 expression (P < 0.0001, Table 2). The correlation between ezrin and Hes1 score was also analyzed by Spearman correlation analysis, showing a positive correlation between ezrin and Hes1 with r = 0.3152 and P = 0.0001 (Fig. 3C). Furthermore, we performed survival analysis of 140 patients using clinical follow-up results to evaluate the prognostic potential of ezrin and Hes1 in BC (Fig. 3D). The patient cumulative survival rates for a 3-year period with low and high ezrin expression were 89.3% and 76.8%, respectively. Likewise, the patient cumulative survival rates for a 5-year period with low and high expression of ezrin were 82.1% and 62.5%, respectively. Thus, patients with high ezrin expression had a significantly worse prognosis (P = 0.001). The 3- and 5- year cumulative survival rates of Hes1 high patients were 75.8% and 54.5%, respectively, which were lower than the survival rates of Hes1 low patients (both 91.9%). Apparently, high expression of Hes1 was linked to worse prognosis for BC patients (P < 0.0001). Then we explored the relationship between different combinations of ezrin and Hes1 expression and the prognosis of BC patients. Based on ezrin and Hes1 expression, the patients were divided into four groups: (1) ezrin(High)/Hes1(High); (2) ezrin(High)/Hes1(Low); (3) ezrin(Low)/Hes1(High); (4) ezrin(Low)/Hes1(Low). Among the four groups, the worst prognosis was observed in ezrin(High)/Hes1(High) patients, whereas the best prognosis was observed in the ezrin(Low)/Hes1(Low) patients. Moreover, the prognosis of ezrin(Low)/Hes1(High) patients was worse than that of the ezrin(High)/Hes1(Low) group (Fig. 3E-F). These results indicated that high coexpression of ezrin and Hes1 is closely correlated with poor prognosis, and ezrin and Hes1 functionally cross-talk in the prognosis of BC patients.

Fig. 3.

Ezrin regulates Hes1 expression and their coexpression associates with poor prognosis in human BC. (A, B) The mRNA and protein expression of ezrin and Hes1 were confirmed by qPCR and western blot. ^**P < 0.01, ^***P < 0.001, ezrin overexpression mLM cells versus control mLM cells, ezrin overexpression MDA-MB-231 cells versus control MDA-MB-231 cells. (C) The correlation between ezrin and Hes1 expression was analyzed using Spearman correlation analysis. n = 140, r = 0.3152, P = 0.0001. (D) The overall survival of patients with low Hes1 and high Hes1 or low ezrin and high ezrin. Total n = 140. Log rank P < 0.0001 for low Hes1 versus high Hes1, Log rank P = 0.001 for low ezrin versus high ezrin. (E) Representative images of ezrin(High)/Hes1(High) group and ezrin(Low)/Hes1(Low). (F) The overall survival of four groups, including ezrin(High)/Hes1(High), ezrin(High)/Hes1(Low), ezrin(Low)/Hes1(High) and ezrin(Low)/Hes1(Low). Total n = 140. Log rank P = 0.002 for ezrin(High)/Hes1(High) versus ezrin(High)/Hes1(Low), Log rank P < 0.0001 for ezrin(Low)/Hes1(High) versus ezrin(Low)/Hes1(Low)

Table 2.

Correlation between ezrin and Hes1 expression and clinicopathologic factors in BC

Parameter	Ezrin n(%)		P	Hes1 n(%)		P
	Low	High		Low	High
Overall	84(60.0)	56(40.0)		74(52.9)	66(47.1)
Age(years)			0.625			0.454
< 60	50(59.5)	31(55.3)		45(60.8)	36(54.5)
≥ 60	34(40.5)	25(44.7)		29(39.2)	30(45.5)
Tumor size			0.120			0.555
≤ 5	73(86.9)	43(76.8)		60(81.1)	56(84.8)
> 5	11(13.1)	13(23.2)		14(18.9)	10(15.2)
TNM stage			0.445			0.422
I-II	58(69.0)	42(75.0)		55(74.3)	45(68.2)
Ш-IV	26(31.0)	14(25.0)		19(25.7)	21(31.8)
Lymph node metastasis			0.890			0.403
No	43(51.2)	28(50.0)		40(54.1)	31(47.0)
Yes	41(48.8)	28(50.0)		34(45.9)	35(53.0)
Hes1			< 0.0001
Low	56	18
High	28	38

Collectively, these results suggested that Hes1 is closely related to ezrin–regulation of breast cancer progression, implying Hes1 associated signaling pathways may play a critical role in this process.

Ezrin requires Hes1 to promote BC cell proliferation but does not directly interact with Hes1

We have shown ezrin can regulate Hes1 expression, and then we investigated if there is a functional connection between ezrin and Hes1. To confirm the link between ezrin and Hes1 in cellular and pathological processes, Hes1 expression was knocked down by siRNA pool in ezrin overexpression mLM cells and the proliferation and migration of these siRNA-transfected cells examined. Knockdown of Hes1 expression in mLM cells by siRNA pool was confirmed by qPCR and western blot (Supplementary Fig. 7). When Hes1 expression was knocked down in the ezrin overexpression cells, cell proliferation dropped significantly (P = 0.01 at 60 h, P = 0.002 at 72 h), but cell migration did not change (Fig. 4A). Moreover, the expression levels of N-cadherin and vimentin remained unchanged in ezrin overexpression mLM cells with Hes1 knockdown, further confirming that the functional connection between ezrin and Hes1 is not regulation of cell migration (Supplementary Fig. 8). These results demonstrated that the regulatory effects of ezrin and Hes1 were linked in regulating cell proliferation. To determine if there is a direct physical interaction between ezrin and Hes1, Co-IP was used to confirm the interaction between ezrin and Hes1. The results showed that ezrin did not interact with Hes1 directly (Fig. 4B). In addition, the sub-cellular distribution of ezrin and Hes1 were determined using confocal analysis. The results showed that ezrin was mainly localized in the cell membrane and cytoplasm while Hes1 was distributed in the cytoplasm and nucleus (Fig. 4C). The results of the overlay images excluded their colocalization in the same position, further suggesting there was no direct interaction between ezrin and Hes1, implying other proteins play a role in mediating the functional connection of ezrin and Hes1 in regulation of BCLM cell proliferation.

Fig. 4.

Ezrin requires Hes1 to promote BC cell proliferation but does not directly interact with Hes1. (A) The proliferation and migration of ezrin overexpression mLM cells were determined by RTCA after transfection with the NC siRNA or Hes1 siRNA pools. n = 3, ^*P < 0.05, ^**P < 0.01, ezrin overexpression cells transfected with the Hes1 siRNA pool versus ezrin overexpression cells transfected with the NC siRNA pool. (B) Co-immunoprecipitation experiments were performed using MDA-MB-231 cell lysates or mLM cell lysates with anti-ezrin antibody, with anti-Hes1 antibody or with non-immune IgG as negative control. The immunocomplexes with the whole cell lysates were analyzed by western blot with anti-ezrin and anti-Hes1 antibodies. (C) Subcellular distribution of ezrin and Hes1 in MDA-MB-231 cells. RTCA, real time cell analysis; NC, negative control

Ezrin upregulation of Hes1 expression is dependent on Notch activation through promoting furin-like converase-mediated cleavage

The Notch signaling pathway is the canonical pathway for the regulation of Hes1, so we speculated that ezrin may act on an upstream component of the Notch signaling pathway [23]. To determine whether the involvment of ezrin in BCLM progression is dependent on the Notch-Hes1 axis, ezrin overexpression mLM cells and MDA-MB-231 cells were treated with the Notch signaling pathway inhibitor FLI-06 at different concentrations for 48 h. The results revealed that the expression of Hes1 was significantly decreased after inhibition of Notch pathway activation (Fig. 5A). To further confirm that ezrin regulates the function of BCLM through the Notch-Hes1 axis, cell proliferation and migration were investigated using RTCA assays. As shown in Fig. 5B-C, FLI-06 could inhibit cell proliferation and migration of ezrin overexpression mLM cells and MDA-MB-231 cells in a dose-dependent manner. In addition, FLI-06 could markedly reduce the number of colonies formed compared with ezrin overexpression cells (Fig. 5D). These results suggested that ezrin regulated breast cancer cell proliferation, migration and colony formation via the Notch-Hes1 axis. However, how does ezrin upregulate Hes1 expression via Notch? Cleavage is required for Notch activation. Therefore, we determined whether ezrin affected Notch cleavage. As shown in Fig. 5E-F, ezrin overexpression significantly increased the level of the Notch transmembrane fragment (NTM) and the Notch intracellular domain (NICD), while the ezrin functional inhibitor NSC305787 suppressed the production of NTM and NICD, confirming that ezrin affects the cleavage of Notch1. NSC30578 can directly bind to ezrin and inhibit its function through inhibiting the phosphorylation of ezrin T567, so the results further indicated that the promotion of ezrin in Notch1 cleavage is dependent on ezrin activation by its phosphorylation at T567. Moreover, immunofluorescence cell staining revealed that expression of cleaved Notch1 in the nucleus of ezrin overexpression cells and the number of endosomes containing cleaved Notch1 increased compared with that in control cells (Fig. 5G and Supplementary Fig. 9), suggesting ezrin could promote Notch1 cleavage. We speculated that ezrin may act on the protease at the S1 site of Notch. Furin-like convertase is the protease for the S1 site. To confirm the role of ezirn in the activity of furin-like convertase, furin-like enzyme activity was detected in MDA-MB-231 cells treated with or without NSC305787 or ezrin overexpression mLM cells. The results showed that the furin-like enzyme activity decreased after the inhibition of ezrin activation, while ezrin overexpression could promote the activity of furin-like convertase, indicating ezrin is involved in the regulation of furin-like convertase activation (Fig. 5H). These results demonstrated that the funtional link between ezrin and Hes1 is dependent on Notch activation through promoting furin-like convertase activation. In additon, in vivo experiments showed that inhibition of ezrin function decreased liver metastatic lesions (P = 0.04), while inhibiton of furin did not significantly reduce liver metastatic lesions, suggesting ezrin is an important regulatory gene in BCLM that is required for its activation (Fig. 5I).

Fig. 5.

Ezrin upregulating Hes1 expression is dependent on Notch activation through promoting furin-like convertase activity. (A) The expression of ezrin and Hes1 in ezrin overexpression mLM cells or MDA-MB-231 cells with or without the treatment of FLI-06 were determined by western blot. GAPDH from the same loading was used as a loading control. (B, C) Cell proliferation and migration were determined by RTCA with or without the treatment of FLI-06. n = 3, ^*P < 0.05, ^**P < 0.01, FLI-06 treated-ezrin overexpression mLM cells versus DMSO-treated ezrin overexpression mLM cells, FLI-06 treated-ezrin overexpression MDA-MB-231 cells versus DMSO-treated ezrin overexpression MDA-MB-231 cells. (D) Cell colony formation ability was determined by colony formation assay and quantified. n = 3, ^**P < 0.01, ^***P < 0.001. Data represent the mean ± SD. (E) The expression of Notch1 (NTM), cleaved Notch1 (NICD) and Hes1 in ezrin overexpression-mLM cells or -MDA-MB-231 cells and the corresponding control cells were detected using western blot. GAPDH from the same loading was used as a loading control. (F) The expression of Notch1 (NTM), cleaved Notch1 (NICD), Hes1, ezrin and p-ezrin in MDA-MB-231 cells and mLM cells treated with or without NSC305787 were detected using western blot. GAPDH from the same loading was used as a loading control. (G) Subcellular distribution of Notch1 (NTM) in ezrin overexpression MDA-MB-231 cells and control MDA-MB-231 cells. (H) The furin-like enzyme activity were measured in MDA-MB-231 with or without NSC305787 treatment or in ezrin overexpression mLM cells and control mLM cells. ^***P < 0.001, NSC305787 treated-MDA-MB-231 cells versus DMSO-treated MDA-MB-231 cells, ezrin overexpression mLM cells versus control mLM cells. (I) The effect of NSC305787 and decRVKR-CMK treatment on the promotion of ezrin in liver metastatic lesions. The mice were sacrificed two weeks after intrasplenic injection and the liver metastatic lesions were collected (left). Statistical results showed the number of liver surface lesions (right). n = 5, ^*P < 0.05. NTM, Notch transmembrane fragment; NICD, Notch intracellulardomain

Discussion

While the 5-year relative survival rate has improved due to improvements in treatment and early diagnosis of disease, the mortality rate of metastatic breast cancer is still very high [24]. Greater understanding of the molecular mechanisms of breast cancer metastasis will help guide development of better treatments. Increasing evidence has shown that ezrin plays key roles in cancer metastasis and the resistance of tumor cells to chemotherapy [25, 26]. We have previously demonstrated that ezrin is involved in regulating cell migration [27]. In the present study, we further explored the novel role of ezrin in BCLM and its related molecular mechanism, demonstrating that ezrin can promote BC malignancy through activating the Notch signaling pathway.

Ezrin, a crosslinker between plasma membrane and actin, connects the cell cortex to the plasma membrane, as well as acting as a signal transducer in many pathways. Increasing evidence has demonstrated ezrin is involved in tumor progression and is closely associated with breast cancer progression [16–18]. It is not only highly expressed in breast cancer tissue, but is also correlated with the degree of malignancy and prognosis of breast cancer patients. Our clinical samples also showed that patients with high expression of ezrin have worse survival, further confirming ezrin is an oncogene in breast cancer. Ezrin can promote the proliferation, migration, invasion, adhesion and angiogenesis of breast cancer cells via mechanisms associated with Akt, PKC and other signaling pathways [19, 28]. However, other novel functions of ezrin and the associated molecular mechanism in BC remain largely unclear.

Breast cancer can metastasize to several different sites. The success of metastatic growth in the secondary sites is dependent on the unique tissue microenvironment, termed metastatic niche [29, 30]. The metastatic niche of different secondary tissues contains specific influencing factors and signal pathways. Liver is one of the most common secondary sites of breast cancer metastasis, so elucidating the molecular mechanisms of BCLM is of paramount importance. Our data showed that MDA-MB-231 stem cells orthotopically implanted for about 5 weeks can metastasize to liver with stable repeatability. Other cell lines either cannot metastasize to liver, such as MCF7, or have poor reproducibility, such as SKBR3 (Data not shown). The immunoblot result showed that ezrin expression and its phosphorylation at T567 in MDA-MB-231 cells is higher than that in MCF7 cells (Supplemantary Fig. 10), implying that there is a positive relationship between ezrin and liver metastatic potential of breast cancer. Therefore, we used mLM cells and MDA-MB-231 cells as cell models to study the role of ezrin in BCLM and its related molecular mechanism both in vitro and in vivo. The findings demonstrated that ezrin promotes breast cancer cell liver metastasis. Then DEGs between ezrin overexpression-mLM cells and the corresponding control cells were determined by microarray analysis, determining the potential molecular mechanism by which ezrin promotes liver metastasis of BC. These results increase the understanding of liver-specific metastasis of BC and provide more evidence for ezrin as a new target of anti-breast cancer therapies. In addition, in tail vein injection experiments, we found that the liver metastasis was weakened after inhibition of ezrin function, but the lung metastasis was enhanced, implying that ezrin plays a different role in lung metastasis. However, the role of ezrin in lung metastasis and its specific molecular mechanism still need to be further elucidated (Supplementary Fig. 11).

Hes1, a transcription factor, is involved in regulating cell differentiation, cell cycle arrest, apoptosis and self-renewal [31]. It is regulated as a central target as it lies at the crossroads of multiple signaling pathways. Hes1 is primarily regulated by canonical and non-canonical signaling pathways. The Notch pathway is one of the major canonical signaling pathways. Li et al. showed that high expression of Hes1 is closely related to poor prognosis in BC [32], which is also confirmed in our clinical samples. Hes1 can promote BC cell proliferation and invasion via the Akt pathway and inducing EMT, indicating that it is a potential target for the treatment of BC. In the present study, Hes1, as one of the most significantly enriched candidates, attracted our attention. To the best of our knowledge, no previous publication mentions the relationship between ezrin and Hes1. In this study, we found that ezrin can positively regulate Hes1 expression. Though ezrin does not directly interact with Hes1, ezrin promotes BC cell proliferation in a Hes1-dependent manner. These results implied that other proteins may be intermediary in this process.

The Notch signaling pathway is the canonical pathway for the regulation of Hes1. Notch1 is the upstream molecule of Hes1, and can regulate the expression of Hes1, so we speculated that ezrin may act on the upstream components of the Notch signaling pathway. Using the Notch signaling inhibitor FLI-06, ezrin overexpression induced-cell proliferation, migration and colony formation could be suppressed, suggesting the functional link between ezrin and Hes1 is dependent on Notch activation. It was noted that inhibition of the Notch pathway affected the regulation of cell migration in ezrin overexpression cells, but knockdown of Hes1 expression had no effect, indicating that ezrin regulates breast cancer cell migration through the Notch pathway independent of Hes1. It may be mediated by other downstream genes of Notch such as Hey1, Myc or HER2, although this requires further investigation. These results also showed that ezrin can increase the level of NTM and NICD that are produced by the cleavage of full length Notch1, while their level was reduced after treatment with NSC305787. Theses results demonstrated that ezrin can promote the cleavage of Notch1. Several studies have shown that ERM proteins play a cleavage-promoting role [33, 34]. Hartmann et al. showed that ERM proteins promote CD44 cleavage by the metalloprotease ADAM10 [33]. Darmellah et al. demonstrated that activation of ERM proteins is required for the P2X7R-dependent proteolytic processing of Amyloid precursor protein [34]. If ezrin is involved in Notch activation, how does ezrin regulate Notch cleavage? During maturation, full length Notch are cleaved by a furin-like protease at the S1 site in the Golgi apparatus to produce two fragments. The two fragments, the extracellular domain (NECD) and the transmembrane domain (NTM) are held together in a heterodimer manner and redistribute at the plasma membrane. Under ligand-receptor interaction, the Notch receptor is subjected to sequential cleavage by an ADAM protease at the S2 site and by ϒ-secretase at the S3 site, leading to the release of the intracellular domain of Notch (NICD) and its translocation to the nucleus. In the nucleus, NICD interacts with CSL DNA-binding protein and Mastermind/Lag3 to form a complex and then activates target gene expression, such as Hes1, Hey, Myc [35–37]. According to the results in the present study and some pilot experiments (data not shown), we speculated that ezrin may promote the cleavage of Notch1 at the S1 site (Fig. 6). To test this hypothesis, we examined the role of ezrin in furin-like convertase activity. The results showed that ezrin participates in promoting furin activity, which demonstrated that the funtional link between ezrin and Hes1 is dependent on Notch activation through promoting furin-like convertase activation.

Fig. 6.

Scheme of the involvement of ezrin in breast cancer through activation of the Notch signaling pathway

In addition, cluster analysis of DEGs showed that ezrin regulates other important signaling pathways in addition to the Notch signaling pathway. The finding showed that many DEGs are related to cell junctions, including tight junctions (TJ) (Supplementary Fig. 12). Changes in the expression and/or distribution of TJ proteins can result in the metastasis of cancer [38]. Claudins are crucial structural and functional components of TJ [39]. Multiple claudins have a causal role in the regulation of EMT, the acquisition of cancer stem cell phenotype, and the therapy resistance of cancer cells [40]. Data mining of the microarray data from Siegel et al. showed that claudin family members have different expressions in the liver metastatic cell line derived from 4T1 cell and may play different roles (Supplementary Fig. 13A) [41, 42]. Among them, claudin-2 was found to promote BCLM through engagement of intergrin complexes [43]. Our study demonstrated overexpression of ezrin changed the expression of claudin family members, indicating a close relationship between ezrin and claudin family members (Supplementary Fig. 13B). However, how ezrin regulates the expression of claudin family members remains unclear. Previous studies found that Hes1 can regulate the claudin-1 promoter [44]. Further study is need to determine whether ezrin regulates the expression of claudins through Hes1.

In summary, our study demonstrated that ezrin drives BC malignancy by activating the Notch signaling pathway. These findings provide a better understanding of the mechanism of ezrin in the progression of BC, with the goal of discovering a novel target for the treatment of BCLM in the future.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Experimentation, M.J.C., Y.P., H.B.L., F.N., X.W.G., H.M.H., M.Z., Y.Y.D.; Cell culture, S.J.L.; IHC analysis, Q.S.L.; Conceiving and designing the study, Y.T.; Writing original draft, M.J.C.; Writing, review and editing, G.E.L. All authors have read and agreed to the final version of the manuscript.

Data availability

The data presented in this study are available in this article.

Declarations

Ethics approval

The present study was approved by the Ethics Committee of Guangzhou Women and Children’s Medical Center. The animal experiments were approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University.

Consent for publication

The authors declare that they agree to submit the article for publication.

Competing interests

The authors declare that they have no conflicts of interest.