Signal peptide-CUB-EGF domain-containing protein 2 (SCUBE2) belongs to a secreted and membrane-associated multi-domain SCUBE protein family. We previously demonstrated that SCUBE2 is a novel breast-tumor suppressor and could be a useful prognostic marker. However, the role of SCUBE2 in breast-cancer cell migration and invasion and how it is regulated during the epithelial–mesenchymal transition (EMT) remain undefined. In this study, we showed that ectopic SCUBE2 overexpression could enhance the formation of E-cadherin-containing adherens junctions by β-catenin–SOX-mediated induction of forkhead box A1 (a positive regulator of E-cadherin) and upregulation of E-cadherin, which in turn led to epithelial transition and inhibited migration and invasion of aggressive MDA-MB-231 breast-carcinoma cells. SCUBE2 expression was repressed together with that of E-cadherin in TGF-β-induced EMT; direct expression of SCUBE2 alone was sufficient to inhibit the TGF-β-induced EMT. Furthermore, quantitative DNA methylation, methylation-specific PCR, and chromatin immunoprecipitation analyses revealed that SCUBE2 expression was inactivated by DNA hypermethylation at the CpG islands by recruiting and binding DNA methyltransferase 1 during TGF-β-induced EMT. Together, our results suggest that SCUBE2 plays a key role in suppressing breast-carcinoma-cell mobility and invasiveness by increasing the formation of the epithelial E-cadherin-containing adherens junctions to promote epithelial differentiation and drive the reversal of EMT.

Signal peptide-CUB (complement protein C1r/C1s, Uegf, and Bmp1)-EGF (epidermal growth factor) domain-containing protein 2 (SCUBE2) belongs to a novel secreted and membrane-associated SCUBE protein family (Tsai et al., 2009; Yang et al., 2002). SCUBE genes are evolutionarily conserved from zebrafish to humans (Grimmond et al., 2000; Grimmond et al., 2001; Hollway et al., 2006; Kawakami et al., 2005; Tsai et al., 2009; Woods and Talbot, 2005; Wu et al., 2004; Yang et al., 2002). Three distinct members have been isolated and designated SCUBE1–3 by the order of their discovery (Grimmond et al., 2000; Grimmond et al., 2001; Hollway et al., 2006; Kawakami et al., 2005; Tsai et al., 2009; Woods and Talbot, 2005; Wu et al., 2004; Yang et al., 2002). These SCUBE proteins contain ∼1000 amino acids and are organized in a modular fashion with five distinct domain motifs: an NH2-terminal signal peptide sequence, followed by nine copies of EGF-like repeats, a spacer region, three cysteine-rich motifs, and one CUB domain at the COOH terminus (Hollway et al., 2006; Tsai et al., 2009). When overexpressed, SCUBE2 is a secreted glycoprotein that can form oligomers and is stably anchored on the cell surface (Tsai et al., 2009; Yang et al., 2002).

Large-scale microarray gene expression analyses have revealed a signature set of genes that can predict breast-cancer prognosis (Paik et al., 2004; Paik et al., 2006; Perou et al., 2000; van't Veer et al., 2002; van de Vijver et al., 2002; Wang et al., 2005). However, few genes overlapped in these assays, and only a few of the breast-associated genes have been validated at the protein level. Interestingly, a cross-platform comparison of breast-cancer gene sets from these profiling studies revealed SCUBE2 as the only common gene (Fan et al., 2006), so SCUBE2 must have important roles in breast-cancer progression. Consistent with this finding, quantitative RT-PCR analysis of the expression of SCUBE2, along with that of 15 other breast cancer-related genes, has been clinically used in the Oncotype DX Breast Cancer Assay to calculate the ‘recurrence score’ for predicting the likelihood of breast-cancer recurrence in early-stage, node-negative, estrogen-receptor (ER)-positive breast cancers (Kim and Paik, 2010; Sparano and Paik, 2008). Despite the clinical utility of SCUBE2 mRNA expression in guiding adjuvant treatment for breast tumor patients, little is known about the biological functions of SCUBE2 in breast cancer progression.

Our recent clinical association studies with anti-SCUBE2 immunohistochemistry showed that patients with SCUBE2-positive tumors had better disease-free survival than those with negative tumors (Cheng et al., 2009). Multivariate analysis confirmed SCUBE2 protein expression as an independent prognostic factor of disease-free survival (Cheng et al., 2009). Further domain and functional studies demonstrated that SCUBE2 could be a novel breast-tumor suppressor through its N-terminal EGF-like repeats or C-terminal CUB domain (Lin et al., 2011). Our molecular and biochemical analyses showed that the C-terminal CUB domain could directly bind and antagonize bone morphogenetic protein activity in an autocrine fashion, whereas the N-terminal EGF-like repeats could mediate cell–cell hemophilic adhesion in a calcium-dependent manner, associate with E-cadherin, and modulate β-catenin–TCF (T-cell factor) signaling in non-invasive, epithelial-like MCF-7 breast-carcinoma cells (Lin et al., 2011). However, whether SCUBE2 and E-cadherin directly interact and the specific domains that participate in their interactions remain unknown. In addition, whether the SOX transcription factors (structurally related to TCF) that can utilize β-catenin as a cofactor or antagonize β-catenin–TCF function (Kormish et al., 2010) are involved in SCUBE2-mediated signaling is unclear. In light of the crucial role of E-cadherin in the formation of adherens junctions and the maintenance of well-differentiated, minimally invasive epithelial traits, E-cadherin has been considered a potent invasion and/or tumor suppressor of breast cancer (Berx and Van Roy, 2001; Cowin et al., 2005; Strumane et al., 2004). Likewise, loss of E-cadherin in the epithelial–mesenchymal transition (EMT) during epithelial tumor progression leads to mesenchymal morphological features or a phenotype with increased cell migration and invasion (Perl et al., 1998; Vleminckx et al., 1991), as well as metastasis (Mareel et al., 1991; Osada et al., 1996).

However, whether SCUBE2 plays a role in controlling the migratory and invasive behaviors in breast-carcinoma cells remains unknown. In addition, the regulatory mechanism controlling the expression of SCUBE2 during EMT is poorly understood. In this study, we demonstrated that SCUBE2 forms a complex with E-cadherin and increases the formation of E-cadherin-containing adherens junctions by inducing the expression of forkhead box A1 (FOXA1; a positive regulator of E-cadherin) and subsequent transactivation of E-cadherin, which drives breast-carcinoma cells to undergo epithelial transition by reversing EMT. Furthermore, we showed that SCUBE2 was epigenetically inactivated by recruiting DNA methyltransferase 1 onto its CpG islands during EMT. Together, these results indicate that downregulation of SCUBE2 is part of the EMT program that plays important roles in modulating breast-cancer cell migration and invasion. Further exploration of SCUBE2 as a diagnostic or therapeutic target for human breast tumors is warranted.

SCUBE2 forms a complex and is co-expressed with E-cadherin in breast-carcinoma cells during EMT

We recently showed that SCUBE2, through its NH2-terminal EGF-like repeats, acts as a homophilic cell–cell adhesive module that can associate with E-cadherin and modulate β-catenin-TCF signaling in MCF-7 cells (Lin et al., 2011). E-cadherin is a single-transmembrane glycoprotein with a large extracellular domain comprising five cadherin-motif subdomains and a short cytoplasmic domain that associates with three catenins (α, β and p120) linking E-cadherin to the actin cytoskeleton (van Roy and Berx, 2008). We performed further analysis of the interaction between E-cadherin and SCUBE2 by using a series of E-cadherin cytoplasmic truncation forms (FL, D1, D2 and D3) and various SCUBE2 deletion mutant constructs (FL, ty97 and D4) including a mutant (D4) that cannot bind E-cadherin as a control (Fig. 1A–C). Our data showed that the membrane-proximal cytoplasmic region (residues 732–820) is crucial in its association with the N-terminal EGF-like repeats of SCUBE2 (Fig. 1C). Using confocal immunofluorescence, SCUBE2 was found to be consistently targeted to the plasma membrane at cell–cell contact sites where it colocalized with E-cadherin or β-catenin, two core proteins of the epithelial adherens junctions (Fig. 1D–G). Further immunoprecipitation analyses identified that it formed a stable complex with E-cadherin or β-catenin in breast cancer cells (Fig. 1H).

Fig. 1.

SCUBE2 colocalizes and forms a stable complex with E-cadherin and β-catenin at adherens junctions. (A) Domain organization of E-cadherin constructs (Myc epitope-tagged) including the full-length (FL) and the C-terminal deletion mutants (D1–3). SP, signal peptide; PRO, propeptide; EC, extracellular cadherin repeat; TM, transmembrane region; CD, cytoplasmic domain. Amino acid numbers start at the first methionine codon. (B) Domain organization of the SCUBE2 expression constructs (FL, ty97 and D4) used in this study. A FLAG epitope was added immediately after the signal peptide sequence at the N-terminus for easy detection. SP, signal peptide; E, EGF-like repeats; Cys-rich, cysteine-rich motifs; CUB, CUB domain. (C) The juxtamembrane region in the cytoplasmic domain of E-cadherin is crucial for its association with SCUBE2. Lysates of HEK-293T cells transiently transfected with the expression plasmid encoding the FLAG-tagged SCUBE2-FL, ty97 or D4 alone or together with the Myc-tagged E-cadherin FL, D1, D2 or D3 constructs were immunoprecipitated (IP) with anti-FLAG antibody, and associated protein levels were determined by western blot analysis (WB) with anti-Myc antibody (top panel) or by a reciprocal co-IP experiment (IP with anti-Myc antibody followed by anti-FLAG immunoblotting; second panel). The expression of FLAG–SCUBE2-FL, ty97 or D4 or all other Myc-tagged E-cadherin proteins was verified by anti-FLAG or anti-Myc western blot analysis, respectively (third panel and bottom panel). (D–G) SCUBE2 colocalizes with E-cadherin and β-catenin at cell–cell contact sites (arrows). Three-color confocal immunofluorescence microscopy of colocalization of SCUBE2 with E-cadherin or β-catenin (D, white). SCUBE2 (FLAG-tagged) localization was detected with rabbit anti-FLAG monoclonal antibody and Alexa-Fluor-488-conjugated goat anti-rabbit IgG (E, green). E-cadherin or β-catenin was detected with mouse anti-E-cadherin and Alexa-Fluor-594-conjugated goat anti-mouse IgG (F, red) or anti-β-catenin antibody conjugated with Alexa-Fluor-647 (G, blue). Scale bar: 10 µm. (H) SCUBE2 forms a complex with E-cadherin and β-catenin. Lysates of MDA-MB-231 control or SCUBE2-expressing cells were immunoprecipitated with anti-FLAG antibody to pull down FLAG-tagged SCUBE2 protein, and associated E-cadherin or β-catenin protein level was verified by western blot analysis with anti-E-cadherin or anti-β-catenin antibody (upper panel). Protein levels of SCUBE2, E-cadherin, β-catenin and β-actin (a loading control) were confirmed by western blot analysis (lower panel).

Fig. 1.

SCUBE2 colocalizes and forms a stable complex with E-cadherin and β-catenin at adherens junctions. (A) Domain organization of E-cadherin constructs (Myc epitope-tagged) including the full-length (FL) and the C-terminal deletion mutants (D1–3). SP, signal peptide; PRO, propeptide; EC, extracellular cadherin repeat; TM, transmembrane region; CD, cytoplasmic domain. Amino acid numbers start at the first methionine codon. (B) Domain organization of the SCUBE2 expression constructs (FL, ty97 and D4) used in this study. A FLAG epitope was added immediately after the signal peptide sequence at the N-terminus for easy detection. SP, signal peptide; E, EGF-like repeats; Cys-rich, cysteine-rich motifs; CUB, CUB domain. (C) The juxtamembrane region in the cytoplasmic domain of E-cadherin is crucial for its association with SCUBE2. Lysates of HEK-293T cells transiently transfected with the expression plasmid encoding the FLAG-tagged SCUBE2-FL, ty97 or D4 alone or together with the Myc-tagged E-cadherin FL, D1, D2 or D3 constructs were immunoprecipitated (IP) with anti-FLAG antibody, and associated protein levels were determined by western blot analysis (WB) with anti-Myc antibody (top panel) or by a reciprocal co-IP experiment (IP with anti-Myc antibody followed by anti-FLAG immunoblotting; second panel). The expression of FLAG–SCUBE2-FL, ty97 or D4 or all other Myc-tagged E-cadherin proteins was verified by anti-FLAG or anti-Myc western blot analysis, respectively (third panel and bottom panel). (D–G) SCUBE2 colocalizes with E-cadherin and β-catenin at cell–cell contact sites (arrows). Three-color confocal immunofluorescence microscopy of colocalization of SCUBE2 with E-cadherin or β-catenin (D, white). SCUBE2 (FLAG-tagged) localization was detected with rabbit anti-FLAG monoclonal antibody and Alexa-Fluor-488-conjugated goat anti-rabbit IgG (E, green). E-cadherin or β-catenin was detected with mouse anti-E-cadherin and Alexa-Fluor-594-conjugated goat anti-mouse IgG (F, red) or anti-β-catenin antibody conjugated with Alexa-Fluor-647 (G, blue). Scale bar: 10 µm. (H) SCUBE2 forms a complex with E-cadherin and β-catenin. Lysates of MDA-MB-231 control or SCUBE2-expressing cells were immunoprecipitated with anti-FLAG antibody to pull down FLAG-tagged SCUBE2 protein, and associated E-cadherin or β-catenin protein level was verified by western blot analysis with anti-E-cadherin or anti-β-catenin antibody (upper panel). Protein levels of SCUBE2, E-cadherin, β-catenin and β-actin (a loading control) were confirmed by western blot analysis (lower panel).

Because this juxtamembrane region is also the binding site for p120-catenin (van Roy and Berx, 2008), we examined whether overexpression of p120-catenin could affect the interaction of SCUBE2 with E-cadherin. As shown by co-immunoprecipitation assay (supplementary material Fig. S1), p120-catenin overexpression did not compete with, but further increased the association of SCUBE2 with E-cadherin. However, pull-down assays with a GST fusion protein containing the juxtamembrane region of the E-cadherin cytoplasmic domain (resides 732–820) and recombinant fragment of SCUBE2 (supplementary material Fig. S2) did not reveal a direct interaction, which suggests that as-yet-unknown protein(s) may connect the cytoplasmic juxtamembrane region of E-cadherin with the surface-tethered SCUBE2 into a complex. Furthermore, SCUBE2 expression is highly correlated with the expression of E-cadherin in an array of breast-cancer cell lines: high in MCF-7 and T-47D breast-cancer cells but absent in MDA-MB-231 and ZR-75-30 cells (supplementary material Fig. S3). Of note, the former two cell lines were non-invasive and had an epithelium-like morphology, whereas the latter were highly invasive and MDA-MB-231 cells had a spindle-shaped, fibroblastic mesenchymal phenotype (supplementary material Fig. S3) (Zajchowski et al., 2001).

To test whether SCUBE2 is co-regulated with E-cadherin and involved in EMT, we used a well-documented model of transforming growth factor β (TGF-β)-induced EMT (Fuxe et al., 2010) to assess expression of SCUBE2 and molecular markers associated with EMT by western blot analysis. TGF-β1 long-term treatment (14 or 21 days) had no apparent effect on apoptosis as evaluated by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay or caspase-3 activation (supplementary material Fig. S4A–C) but did induce MCF-7 cells to undergo EMT, as evidenced by the acquisition of fibroblastic mesenchymal morphological features (supplementary material Fig. S4D), loss of the epithelial marker E-cadherin, and gain of the mesenchymal marker vimentin (supplementary material Fig. S4E,F). Importantly, SCUBE2 protein level was suppressed together with decreased E-cadherin protein level during TGF-β-induced EMT in a dose- and time-dependent manner (supplementary material Fig. S4E,F), which suggests that downregulation of SCUBE2 is part of the EMT program.

Ectopic SCUBE2 expression drives an epithelial phenotype and inhibits cell migration and invasion

To test whether direct expression of SCUBE2 alone could inhibit the TGF-β-induced EMT, we used a stable MCF-7 cell line with the expression of exogenous FLAG-tagged SCUBE2 (FLAG–SCUBE2) under the control of an inducible Tet-Off promoter as described previously (Cheng et al., 2009). In addition, the MCF-7 Tet-Off vector clones containing stable integration of the empty expression vector were used as a control. These MCF-7 Tet-Off cells were first cultured in the absence of doxycycline for 5 days to induce the expression of ectopic FLAG–SCUBE2 protein, followed by treatment of TGF-β1 to induce EMT. Interestingly, in contrast to control cells, ectopic SCUBE2 overexpression prevented TGF-β-induced cell scattering (Fig. 2A, bottom panel), impeded TGF-β-promoted loss of the epithelial marker E-cadherin at the lateral cell surface (Fig. 2B,C), and inhibited accumulation of the mesenchymal marker vimentin (Fig. 2C). The mRNA expression levels of SCUBE2, E-cadherin and vimentin were generally in agreement with their protein expression in these cells (Fig. 2D).

Fig. 2.

Ectopic SCUBE2 overexpression inhibited transforming growth factor β (TGF-β)-induced EMT. (A) TGF-β1-induced EMT resulted in morphological changes. MCF-7 Tet-Off vector and FLAG–SCUBE2-expressing cells were treated with TGF-β1 (10 ng/ml) for 7, 14 and 21 days. Phase-contrast images show the morphological alterations associated with EMT. Scale bar: 50 µm. (B) Lateral cell-surface distribution of E-cadherin in MCF-7 vector- and SCUBE2-expressing cells before and after TGF-β1-induced EMT. Immunofluorescent staining of E-cadherin mainly localized to the lateral cell surface in confluent MCF-7 vector- and SCUBE2-expressing cells in the basal status. In contrast to control cells showing marked loss of the cell-surface expression of E-cadherin, ectopic SCUBE2 expression prevented TGF-β1-promoted decrease of E-cadherin after TGF-β1 treatment. Scale bar: 10 µm. (C,D) Western blot analysis of protein levels (C) or RT-PCR analysis (D) for mRNA levels of SCUBE2, E-cadherin, vimentin, or β-actin. Exogenous FLAG–SCUBE2 was detected by anti-FLAG antibody, whereas the total amount of SCUBE2 (a combination of endogenous and exogenous SCUBE2) was detected with anti-SCUBE2 antibody. RT-PCR analysis to determine the mRNA level of SCUBE2 (total or ectopic expression level), E-cadherin, vimentin, or GAPDH, which served as an internal control with first-strand cDNA from TGF-β1-treated MCF-7 vector- or SCUBE2-expressing cells at the indicated times.

Fig. 2.

Ectopic SCUBE2 overexpression inhibited transforming growth factor β (TGF-β)-induced EMT. (A) TGF-β1-induced EMT resulted in morphological changes. MCF-7 Tet-Off vector and FLAG–SCUBE2-expressing cells were treated with TGF-β1 (10 ng/ml) for 7, 14 and 21 days. Phase-contrast images show the morphological alterations associated with EMT. Scale bar: 50 µm. (B) Lateral cell-surface distribution of E-cadherin in MCF-7 vector- and SCUBE2-expressing cells before and after TGF-β1-induced EMT. Immunofluorescent staining of E-cadherin mainly localized to the lateral cell surface in confluent MCF-7 vector- and SCUBE2-expressing cells in the basal status. In contrast to control cells showing marked loss of the cell-surface expression of E-cadherin, ectopic SCUBE2 expression prevented TGF-β1-promoted decrease of E-cadherin after TGF-β1 treatment. Scale bar: 10 µm. (C,D) Western blot analysis of protein levels (C) or RT-PCR analysis (D) for mRNA levels of SCUBE2, E-cadherin, vimentin, or β-actin. Exogenous FLAG–SCUBE2 was detected by anti-FLAG antibody, whereas the total amount of SCUBE2 (a combination of endogenous and exogenous SCUBE2) was detected with anti-SCUBE2 antibody. RT-PCR analysis to determine the mRNA level of SCUBE2 (total or ectopic expression level), E-cadherin, vimentin, or GAPDH, which served as an internal control with first-strand cDNA from TGF-β1-treated MCF-7 vector- or SCUBE2-expressing cells at the indicated times.

In addition, we investigated whether SCUBE2 could drive mesenchymal-like breast-cancer cells to undergo an epithelial transformation when introduced into SCUBE2-negative MDA-MB-231 cells. MDA-MB-231 cells were engineered with the full-length SCUBE2 expression vector or empty vector (control) by a leniviral transduction system. Control cells were spindly, elongated, and dispersed, whereas overexpression of SCUBE2 transformed the fibroblastic, spindle-shaped MDA-MB-231 cells into epithelial clusters with a cobblestone growth pattern in monolayer culture much like MCF-7 cells (Fig. 3A), and formed multicellular aggregates in suspension (supplementary material Fig. S5). Consistent with the morphological phenotypes (Fig. 3A), MDA-MB-231 cells expressing ectopic SCUBE2 protein showed increased expression of E-cadherin (an epithelial marker) but decreased expression of the mesenchymal markers vimentin and N-cadherin at both the mRNA and protein levels, similar to epithelial-like MCF-7 control cells (Fig. 3B,C). Interestingly, overexpression of a SCUBE2 mutant (D4) unable to bind E-cadherin produced no effect on epithelial transformation and upregulation of FOXA1 (a positive transcriptional activator of E-cadherin, see below), suggesting that SCUBE2–E-cadherin interaction is important for mediating SCUBE2-drived epithelial transition (supplementary material Fig. S6).

Fig. 3.

Reversal of epithelial–mesenchymal transition (EMT) in MDA-MB-231 cells ectopically expressing SCUBE2. (A) Phase-contrast images showing the morphological features of MCF-7, MDA-MD-231 and MDA-MB-231 cells overexpressing SCUBE2. Scale bar: 50 µm. (B) Western blot and RT-PCR analyses. Cell lysates from MCF-7 and MDA-MB-231 control or SCUBE2-expressing cells were immunoblotted with antibodies for anti-SCUBE2, E-cadherin, vimentin, N-cadherin or β-actin. MCF-7 cells were an epithelial cell control. The first-strand cDNAs derived from MCF-7, MDA-MB-231 control or SCUBE2-expressing cells were used to perform RT-PCR analysis to determine the mRNA level of SCUBE2, E-cadherin, vimentin, N-cadherin, or GAPDH, which served as an internal control. (C) Immunocytochemical staining of cells with anti-E-cadherin (green) or anti-vimentin (red) antibody. Scale bars: 10 µm. Nuclei were visualized by DAPI staining (blue). (D,E) SCUBE2 overexpression suppressed MDA-MB-231 cell migration and invasion. (D) SCUBE2 inhibition of MDA-MB-231 cell migration was determined by the gap-closure migration assay. Spontaneous motility of MDA-MB-231 control or SCUBE2-expressing cells examined over 24 hours in confluent monolayers separated with cell culture insert. After removal of the insert the gap was photographed at 0, 12 and 24 hours (upper panel), and the mean proportion of the gap closure was calculated (lower panel). Scale bar: 500 µm. (E) MDA-MB-231 control or SCUBE2-expressing cells were seeded on Transwell plates coated with 30 µg Matrigel. After incubation for 18 hours, the invaded cells were fixed, stained and photographed (upper panel). Quantification of invasion experiment involved counting the number of invaded cells (n = 3, lower panel). All assays were performed in triplicate. Quantitative data are means ± s.d., *P<0.05.

Fig. 3.

Reversal of epithelial–mesenchymal transition (EMT) in MDA-MB-231 cells ectopically expressing SCUBE2. (A) Phase-contrast images showing the morphological features of MCF-7, MDA-MD-231 and MDA-MB-231 cells overexpressing SCUBE2. Scale bar: 50 µm. (B) Western blot and RT-PCR analyses. Cell lysates from MCF-7 and MDA-MB-231 control or SCUBE2-expressing cells were immunoblotted with antibodies for anti-SCUBE2, E-cadherin, vimentin, N-cadherin or β-actin. MCF-7 cells were an epithelial cell control. The first-strand cDNAs derived from MCF-7, MDA-MB-231 control or SCUBE2-expressing cells were used to perform RT-PCR analysis to determine the mRNA level of SCUBE2, E-cadherin, vimentin, N-cadherin, or GAPDH, which served as an internal control. (C) Immunocytochemical staining of cells with anti-E-cadherin (green) or anti-vimentin (red) antibody. Scale bars: 10 µm. Nuclei were visualized by DAPI staining (blue). (D,E) SCUBE2 overexpression suppressed MDA-MB-231 cell migration and invasion. (D) SCUBE2 inhibition of MDA-MB-231 cell migration was determined by the gap-closure migration assay. Spontaneous motility of MDA-MB-231 control or SCUBE2-expressing cells examined over 24 hours in confluent monolayers separated with cell culture insert. After removal of the insert the gap was photographed at 0, 12 and 24 hours (upper panel), and the mean proportion of the gap closure was calculated (lower panel). Scale bar: 500 µm. (E) MDA-MB-231 control or SCUBE2-expressing cells were seeded on Transwell plates coated with 30 µg Matrigel. After incubation for 18 hours, the invaded cells were fixed, stained and photographed (upper panel). Quantification of invasion experiment involved counting the number of invaded cells (n = 3, lower panel). All assays were performed in triplicate. Quantitative data are means ± s.d., *P<0.05.

To determine the functional changes in cell behavior that occurred following SCUBE2 overexpression, we performed gap-closure migration (wound-healing) and Matrigel-coated Transwell invasion assays to characterize the control and SCUBE2-expressing MDA-MB-231 cells. MDA-MB-231 control cells were highly motile and invasive, but SCUBE2 overexpression significantly decreased both motility and invasiveness (Fig. 3D,E). Importantly, cell proliferation did not differ between the control and SCUBE2-expressing cells (data not shown), which excluded general cytostatic and cytotoxic effects caused by SCUBE2 expression in MDA-MB-231 cells. Thus, expression of SCUBE2 induces breast-cancer cells to undergo epithelial transition or a reversal of the EMT and suppresses the motility and invasive ability of invasive, mesenchyme-like MDA-MB-231 cells.

Knockdown of endogenous SCUBE2 disrupted the adherens junctions and concomitantly increased the migratory and invasive ability of MCF-7 breast-cancer cells

To further confirm the role of SCUBE2 in increasing the epithelial E-cadherin-containing adherens junctions and modulating cancer cell motility and invasiveness, we used two independent SCUBE2-targeting short hairpin RNA (shRNA) lentiviruses (SCUBE2-shRNA #1 and #2) to inhibit endogenous SCUBE2 expression in MCF-7 cells, which exhibit non-invasive and epithelium-like features (Zajchowski et al., 2001). As a negative control, cells were infected with an shRNA lentivirus targeting bacterial β-galactosidase (control shRNA). The efficiency of shRNA-mediated SCUBE2 knockdown (∼90%) was confirmed by RT-PCR and western blot analyses (Fig. 4A). Control-shRNA-treated cells retained an epithelial-cluster appearance that was similar to the parental MCF-7 cells (Fig. 4B; Fig. 3A), but the two independent SCUBE2-shRNA knockdown cell clones were dispersed and scattered with an increase in cell spreading and reduced cell–cell contacts (Fig. 4B). In agreement with the morphological changes, knockdown of SCUBE2 induced EMT-like events by downregulating the epithelial E-cadherin and upregulating mesenchymal markers vimentin and N-cadherin at both the mRNA and protein levels (Fig. 4A). Consistently, the intensity of E-cadherin immunofluorescence was markedly reduced at cell junction sites, which indicates the disruption and loss of the E-cadherin-containing adherens junctions (Fig. 4C). As a consequence, SCUBE2 knockdown, in contrast to its overexpression, promoted the motility and invasiveness of MCF-7 breast-cancer cells (Fig. 4D,E). Again, the cell proliferation rate was comparable between the control and SCUBE2-silenced cells (data not shown). Together, these data indicate that SCUBE2 expression is essential to maintain the epithelial phenotype, with its long-term loss sufficient to induce subsequent mesenchymal transition.

Fig. 4.

SCUBE2 knockdown leads to the loss of the E-cadherin-containing adherens junctions, induces EMT-like events, and increases the migration and/or invasion of MCF-7 cells. (A) Effect of endogenous SCUBE2 knockdown on the expression of EMT markers. Endogenous SCUBE2 expression was inhibited by two independent SCUBE2-targeting short hairpin RNA (shRNA) lentiviruses (SCUBE2-shRNA #1 or #2) in MCF-7 cells. A luciferase shRNA lentivirus was used as a negative control (control-shRNA). Efficiency of SCUBE2 knockdown and its effect on the expression of EMT markers (E-cadherin, vimentin and N-cadherin) were assessed by RT-PCR or western blot analyses using GAPDH or β-actin expression as an internal or loading control. Omission of template cDNA served as a negative PCR reaction control (-). (B) Effect of SCUBE2 knockdown on cellular morphology. Cells cultured under subconfluent conditions were photographed using phase-contrast microscopy. SCUBE2-knockdown cells were dispersed and exhibited migratory features including protruding filopodia at the cell edge, whereas control cells retained an epithelial-like appearance that was similar to parental MCF-7 cells (Fig. 3A). Scale bar: 50 µm. (C) Effect of SCUBE2 knockdown on the formation of the E-cadherin-containing adherens junctions at the cell–cell contact sites. Confocal immunofluorescence microscopy with merged images of E-cadherin (green) and β-catenin (red) staining marking the adherens junctions seen in control-shRNA cells but not in two SCUBE2 knockdown cells (SCUBE2 shRNA #1 and #2). Scale bar: 10 µm. (D,E) Knockdown of endogenous SCUBE2 expression increased the migratory and invasive capabilities of MCF-7 breast-cancer cells. Migration and invasion assays were performed as described in Fig. 3, except the invaded MCF-7 cells were stained with Crystal Violet and the OD570 was measured to quantify the amount of DMSO-solubilized dye. All assays were performed in triplicate. Quantitative data are means ± s.d.; *P<0.05, **P<0.01.

Fig. 4.

SCUBE2 knockdown leads to the loss of the E-cadherin-containing adherens junctions, induces EMT-like events, and increases the migration and/or invasion of MCF-7 cells. (A) Effect of endogenous SCUBE2 knockdown on the expression of EMT markers. Endogenous SCUBE2 expression was inhibited by two independent SCUBE2-targeting short hairpin RNA (shRNA) lentiviruses (SCUBE2-shRNA #1 or #2) in MCF-7 cells. A luciferase shRNA lentivirus was used as a negative control (control-shRNA). Efficiency of SCUBE2 knockdown and its effect on the expression of EMT markers (E-cadherin, vimentin and N-cadherin) were assessed by RT-PCR or western blot analyses using GAPDH or β-actin expression as an internal or loading control. Omission of template cDNA served as a negative PCR reaction control (-). (B) Effect of SCUBE2 knockdown on cellular morphology. Cells cultured under subconfluent conditions were photographed using phase-contrast microscopy. SCUBE2-knockdown cells were dispersed and exhibited migratory features including protruding filopodia at the cell edge, whereas control cells retained an epithelial-like appearance that was similar to parental MCF-7 cells (Fig. 3A). Scale bar: 50 µm. (C) Effect of SCUBE2 knockdown on the formation of the E-cadherin-containing adherens junctions at the cell–cell contact sites. Confocal immunofluorescence microscopy with merged images of E-cadherin (green) and β-catenin (red) staining marking the adherens junctions seen in control-shRNA cells but not in two SCUBE2 knockdown cells (SCUBE2 shRNA #1 and #2). Scale bar: 10 µm. (D,E) Knockdown of endogenous SCUBE2 expression increased the migratory and invasive capabilities of MCF-7 breast-cancer cells. Migration and invasion assays were performed as described in Fig. 3, except the invaded MCF-7 cells were stained with Crystal Violet and the OD570 was measured to quantify the amount of DMSO-solubilized dye. All assays were performed in triplicate. Quantitative data are means ± s.d.; *P<0.05, **P<0.01.

SCUBE2 expression upregulates an E-cadherin transcriptional activator forkhead box A1 (FOXA1) to maintain the epithelial phenotype in breast-cancer cells

Because SCUBE2 was colocalized and associated with E-cadherin-positive adherens junctions (Fig. 1), and because several structural proteins associated with adherens junctions can mediate cellular signaling (Cowin et al., 2005; van Roy and Berx, 2008), SCUBE2 expression might induce a transcriptional activator to reinstate the expression of E-cadherin in aggressive mesenchyme-like breast-carcinoma cells. Because E-cadherin expression during EMT or reversal of EMT could be transcriptionally repressed by members of TWIST, SNAIL and ZEB protein families (van Roy and Berx, 2008) or positively upregulated by FOXA1 (Liu et al., 2005), we sought to profile the expression of these E-cadherin transcriptional regulators. Western blot and RT-PCR analyses showed that SCUBE2 overexpression or knockdown did not substantially alter these E-cadherin repressors at both the mRNA and protein levels, except for TWIST, which was upregulated but unable to bind the E-cadherin promoter with SCUBE2 overexpression in MDA-MB-231 (Fig. 5A–E). Because SCUBE2 overexpression induced E-cadherin expression and drove an epithelial phenotype in MDA-MB-231 cells (Fig. 3), such TWIST upregulation conflicts with its classical role as an EMT inducer, and its precise function in this context required further studies. Of note, a chromatin immunoprecipitation (ChIP) assay showed that SCUBE2 overexpression decreased the binding of SNAIL and ZEB1 to the E-cadherin promoter in MDA-MB-231 cells (Fig. 5E). However, further studies are required to elucidate the underlying mechanism and the biological relevance of these findings related to the SCUBE2-mediated reversal of EMT.

Fig. 5.

Effect of SCUBE2 on the expression of EMT inducers, on E-cadherin promoter binding of classical E-cadherin repressors and on competition with E-cadherin activator FOXA1 in breast-cancer cells. (A–D) RT-PCR or western blot analysis of the mRNA or protein expression of the classical EMT inducers TWIST, SNAIL, SLUG, ZEB1 and ZEB2 in SCUBE2-overexpressing MDA-MB-231 cells (A,B) or SCUBE2-silencing MCF-7 cells (C,D). RT-PCR analysis using first-strand cDNAs derived from MDA-MB-231 (A) or MCF-7 (C) breast-carcinoma cell lines. GAPDH expression was amplified as an internal control and lack of cDNA was a negative control (−). Western blot analysis of protein expression of the EMT inducers in the indicated MDA-MB-231 (B) or MCF-7 (D) cells with the specific antibody for each EMT-inducing E-cadherin repressor. Nuclear protein lamin A/C expression was used as a loading control. The expression of the EMT inducers remained basically unaltered at both the mRNA and protein levels, except TWIST was induced by SCUBE2 overexpression in MDA-MB-231 cells. Because SCUBE2 overexpression induces E-cadherin expression and drives an epithelial phenotype in MDA-MB-231, such upregulation of TWIST is contrary to its accepted function as an EMT inducer and its exact function in this context requires further studies. (E) Effect of SCUBE2 overexpression on the binding of the classical EMT inducers to the E-cadherin promoter. ChIP assay with antibodies specific to each E-cadherin repressor (TWIST, SNAIL, SLUG, ZEB1 or ZEB2) to determine the binding of these repressors to the E-cadherin promoter in MDA-MB-231 control and SCUBE2-expressing cells analyzed by PCR with specific primers for E-cadherin amplifying a 92-bp fragment (see Fig. 6B). n.s., non-specific PCR product. (F) FOXA1 overcomes suppression of the E-cadherin promoter mediated by SNAIL and ZEB1. In MDA-MB-231 SCUBE2 cells, the expression plasmid encoding FOXA1, SNAIL or ZEB1 was transfected alone or in combination as indicated. Two days after transfection, promoter activity was determined by dual-luciferase assays. Luciferase activity was corrected for Renilla luciferase activity (pRL-TK) to control for transfection efficiency. Data are means ± s.d. of three experiments; *P<0.01.

Fig. 5.

Effect of SCUBE2 on the expression of EMT inducers, on E-cadherin promoter binding of classical E-cadherin repressors and on competition with E-cadherin activator FOXA1 in breast-cancer cells. (A–D) RT-PCR or western blot analysis of the mRNA or protein expression of the classical EMT inducers TWIST, SNAIL, SLUG, ZEB1 and ZEB2 in SCUBE2-overexpressing MDA-MB-231 cells (A,B) or SCUBE2-silencing MCF-7 cells (C,D). RT-PCR analysis using first-strand cDNAs derived from MDA-MB-231 (A) or MCF-7 (C) breast-carcinoma cell lines. GAPDH expression was amplified as an internal control and lack of cDNA was a negative control (−). Western blot analysis of protein expression of the EMT inducers in the indicated MDA-MB-231 (B) or MCF-7 (D) cells with the specific antibody for each EMT-inducing E-cadherin repressor. Nuclear protein lamin A/C expression was used as a loading control. The expression of the EMT inducers remained basically unaltered at both the mRNA and protein levels, except TWIST was induced by SCUBE2 overexpression in MDA-MB-231 cells. Because SCUBE2 overexpression induces E-cadherin expression and drives an epithelial phenotype in MDA-MB-231, such upregulation of TWIST is contrary to its accepted function as an EMT inducer and its exact function in this context requires further studies. (E) Effect of SCUBE2 overexpression on the binding of the classical EMT inducers to the E-cadherin promoter. ChIP assay with antibodies specific to each E-cadherin repressor (TWIST, SNAIL, SLUG, ZEB1 or ZEB2) to determine the binding of these repressors to the E-cadherin promoter in MDA-MB-231 control and SCUBE2-expressing cells analyzed by PCR with specific primers for E-cadherin amplifying a 92-bp fragment (see Fig. 6B). n.s., non-specific PCR product. (F) FOXA1 overcomes suppression of the E-cadherin promoter mediated by SNAIL and ZEB1. In MDA-MB-231 SCUBE2 cells, the expression plasmid encoding FOXA1, SNAIL or ZEB1 was transfected alone or in combination as indicated. Two days after transfection, promoter activity was determined by dual-luciferase assays. Luciferase activity was corrected for Renilla luciferase activity (pRL-TK) to control for transfection efficiency. Data are means ± s.d. of three experiments; *P<0.01.

In contrast, the nuclear protein level of FOXA1, a positive regulator of E-cadherin, was upregulated and associated with SCUBE2-induced E-cadherin mRNA and protein expression (Fig. 6A,E). Furthermore, FOXA1 transactivates the E-cadherin promoter containing FOXA1 binding sites on a luciferase reporter assay (Fig. 6B,C). We further examined the direct effect of FOXA1 on the E-cadherin promoter in MDA-MB-231 control and MDA-MB-231 SCUBE2-expressing cells by transient transfection with plasmids encoding full-length FOXA1 or two different FOXA1-targeting shRNAs (#1 and #2), respectively. With FOXA1 overexpression, activity of the E-cadherin promoter containing FOXA1-binding sites [E-cad (WT)-Luc] but not mutated FOXA1-binding sites [E-cad (Mut)-Luc] was greatly increased in MDA-MB-231 control cells (Fig. 6C, middle panel). In contrast, knockdown of FOXA1 significantly suppressed the SCUBE2-mediated increase in the E-cadherin promoter activity (Fig. 6C, bottom panel). In addition, a ChIP assay confirmed that endogenous FOXA1 protein indeed interacted with the DNA region that harbors a consensus FOXA-binding site [5′-TGTTTG(T/C)-3′] within the E-cadherin promoter (Fig. 6D). Furthermore, we examined the effects of FOXA1 overexpression in MDA-MB-231 control and FOXA1 silencing in MDA-MB-231 SCUBE2-expressing cells on E-cadherin mRNA and protein expression as described above. FOXA1 overexpression could transactivate and upregulate E-cadherin in MDA-MB-231 control cells, but knockdown of FOXA1 to ∼10% of control levels in MDA-MB-231 SCUBE2-expressing cells concomitantly diminished E-cadherin expression at both the mRNA and protein levels by ∼90% as compared with control cells (Fig. 6E). Given that FOXA1 can activate the E-cadherin promoter in SNAIL- or ZEB1-expressing MDA-MB-231 cells, we investigated whether FOXA1 is a dominant activating factor capable of overcoming E-cadherin-suppressing signals. Using MDA-MB-231 SCUBE2 cells, we showed that ectopic overexpression of SNAIL or ZEB1 alone was able to repress E-cadherin promoter reporter activity compared with the control cells (Fig. 5F). Interestingly, co-expression of FOXA1 can overcome the suppression of the E-cadherin promoter mediated by SNAIL or ZEB1 (Fig. 5F), indicating that FOXA1 has a dominant regulatory role on the E-cadherin promoter in breast cancer cells. Together, these data suggest that FOXA1 is important in mediating the SCUBE2-induced upregulation of E-cadherin and promotes an epithelial character in invasive MDA-MB-231 breast-cancer cells.

Fig. 6.

FOXA1-mediated SCUBE2-induced upregulation of E-cadherin. (A) Induction of FOXA1 is associated with SCUBE2 expression. Western blot analysis of nuclear FOX1 levels in MCF-7, MDA-MB-231 control and SCUBE2-expressing cells. Nuclear lamin A/C level was used as a loading control. (B) Scheme of the E-cadherin promoter reporter constructs containing the native (WT) or mutated (Mut) FOXA1-binding sites. The location of the promoter (−995∼+1) of the E-cadherin gene is indicated by a lateral line. The FOXA1-binding sites are marked by diamonds, whereas the mutant sites are shown as dashed lines. The primers and the amplicon region used in the ChIP assay are labeled (upper panel). (C) FOXA1 transactivates the E-cadherin promoter-driven luciferase activity. Quantification of activity of E-cadherin WT or Mut promoter constructs transiently transfected into MDA-MB-231 control or SCUBE2-expressing cells. Luciferase activity was corrected for Renilla luciferase activity (pRL-TK) to control for transfection efficiency. Data are means ± s.d. of three experiments; *P<0.05 compared with the vector control. Note that transactivating activity of the E-cadherin promoter was higher when FOXA1 was overexpressed, whereas the FOXA1-medaited transactivation was completely abolished when the FOXA1-binding sites were mutated. (D) FOXA1 directly binds to the E-cadherin promoter. ChIP assay of in vivo binding of FOXA1 protein to the promoter of E-cadherin in MDA-MB-231 control and SCUBE2-expressing cells analyzed by PCR with specific primers for E-cadherin amplifying a 92-bp fragment (see Fig. 6B). Lack of cell lysates (−) or control IgG were used as negative controls. (E) Effects of FOXA1 overexpression or knockdown on the expression of E-cadherin. FOXA1 was overexpressed in MDA-MB-231 control cells (left panels), but its expression was suppressed by two independent FOXA1-targeting short hairpin RNAs (shRNAs; FOXA1-shRNA #1 or #2) in MDA-MB-231 SCUBE2 cells (right panels). An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). RT-PCR is shown in the upper panel, western blot (WB) analysis for FOXA1 overexpression or knockdown in nuclear fractions is shown in the middle panel, and its effect on the expression of E-cadherin in cell lysates is shown in the lower panel. Anti-lamin A/C blotting served as a nuclear marker, whereas anti-GAPDH blotting was a cytosolic marker and loading control.

Fig. 6.

FOXA1-mediated SCUBE2-induced upregulation of E-cadherin. (A) Induction of FOXA1 is associated with SCUBE2 expression. Western blot analysis of nuclear FOX1 levels in MCF-7, MDA-MB-231 control and SCUBE2-expressing cells. Nuclear lamin A/C level was used as a loading control. (B) Scheme of the E-cadherin promoter reporter constructs containing the native (WT) or mutated (Mut) FOXA1-binding sites. The location of the promoter (−995∼+1) of the E-cadherin gene is indicated by a lateral line. The FOXA1-binding sites are marked by diamonds, whereas the mutant sites are shown as dashed lines. The primers and the amplicon region used in the ChIP assay are labeled (upper panel). (C) FOXA1 transactivates the E-cadherin promoter-driven luciferase activity. Quantification of activity of E-cadherin WT or Mut promoter constructs transiently transfected into MDA-MB-231 control or SCUBE2-expressing cells. Luciferase activity was corrected for Renilla luciferase activity (pRL-TK) to control for transfection efficiency. Data are means ± s.d. of three experiments; *P<0.05 compared with the vector control. Note that transactivating activity of the E-cadherin promoter was higher when FOXA1 was overexpressed, whereas the FOXA1-medaited transactivation was completely abolished when the FOXA1-binding sites were mutated. (D) FOXA1 directly binds to the E-cadherin promoter. ChIP assay of in vivo binding of FOXA1 protein to the promoter of E-cadherin in MDA-MB-231 control and SCUBE2-expressing cells analyzed by PCR with specific primers for E-cadherin amplifying a 92-bp fragment (see Fig. 6B). Lack of cell lysates (−) or control IgG were used as negative controls. (E) Effects of FOXA1 overexpression or knockdown on the expression of E-cadherin. FOXA1 was overexpressed in MDA-MB-231 control cells (left panels), but its expression was suppressed by two independent FOXA1-targeting short hairpin RNAs (shRNAs; FOXA1-shRNA #1 or #2) in MDA-MB-231 SCUBE2 cells (right panels). An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). RT-PCR is shown in the upper panel, western blot (WB) analysis for FOXA1 overexpression or knockdown in nuclear fractions is shown in the middle panel, and its effect on the expression of E-cadherin in cell lysates is shown in the lower panel. Anti-lamin A/C blotting served as a nuclear marker, whereas anti-GAPDH blotting was a cytosolic marker and loading control.

β-catenin is involved in SCUBE2-mediated induction of FOXA1

Because SCUBE2 overexpression resulted in increased expression of β-catenin protein (Fig. 1H) and because β-catenin directly interacts with Sox17 to transcribe FoxA1 in Xenopus embryos (Sinner et al., 2004), we then investigated how SCUBE2 overexpression regulates the expression of β-catenin and whether β-catenin participates in upregulation of FOXA1 and E-cadherin in breast cancer cells. RT-PCR and western blot analyses with anti-β-catenin antibody or anti-β-catenin phosphorylation-specific antibodies showed that SCUBE2 expression is positively associated with β-catenin expression at the mRNA transcriptional level (Fig. 7A) but not by post-translational regulation of its stability and localization such as phosphorylation at Ser33, Ser37, Thr41 or Tyr654 (Fig. 7B). In addition, we found that both SOX4 and SOX7, which can interact with β-catenin and modulate β-catenin-mediated transcription (Sinner et al., 2007; Takash et al., 2001), are closely associated with β-catenin and SCUBE2 expression in breast cancer cells (Fig. 7C). We then examined whether β-catenin is involved in SCUBE2-induced FOXA1 upregulation by shRNA silencing experiments. Knockdown of β-catenin concomitantly suppressed FOXA1 and E-cadherin expression at both mRNA and protein levels in MDA-MB-231 SCUBE2-expressing cells (Fig. 7D,E), which suggests that β-catenin, possibly in cooperating with SOX protein, is an essential upstream transcriptional regulator of its target gene FOXA1 when SCUBE2 is overexpression.

Fig. 7.

β-catenin is involved in SCUBE2-induced FOXA1 upregulation in breast-cancer cells. (A,B) SCUBE2 expression is positively associated with β-catenin expression at the mRNA transcription level but not after post-translational phosphorylation regulation. RT-PCR or western blot (WB) analysis with the indicated antibodies or phosphorylation-specific antibodies, of the expression of β-catenin (A) or phosphorylation status of β-catenin at Ser33, Ser37, Thr41 or Tyr654 (B) with subcellular (cytosolic and nuclear) fractions, total cell lysates or first-strand cDNA from the indicated cell lines. (C) Expression of SOX proteins is associated with SCUBE2 and β-catenin expression. RT-PCR analysis of the mRNA levels of SOX members (SOX4, SOX7 and SOX17) in the indicated breast cancer cell lines. (D,E) Effect of β-catenin knockdown on FOXA1 and E-cadherin expression. β-catenin expression was suppressed by two different β-catenin-targeting shRNAs (#1 and #2) in MDA-MB-231 SCUBE2 cells. An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). Effect of β-catenin knockdown on nuclear expression of β-catenin and FOXA1 or total E-cadherin was verified by RT-PCR (D) or western blot (E) analyses. Anti-lamin A/C blotting served as a nuclear marker and loading control.

Fig. 7.

β-catenin is involved in SCUBE2-induced FOXA1 upregulation in breast-cancer cells. (A,B) SCUBE2 expression is positively associated with β-catenin expression at the mRNA transcription level but not after post-translational phosphorylation regulation. RT-PCR or western blot (WB) analysis with the indicated antibodies or phosphorylation-specific antibodies, of the expression of β-catenin (A) or phosphorylation status of β-catenin at Ser33, Ser37, Thr41 or Tyr654 (B) with subcellular (cytosolic and nuclear) fractions, total cell lysates or first-strand cDNA from the indicated cell lines. (C) Expression of SOX proteins is associated with SCUBE2 and β-catenin expression. RT-PCR analysis of the mRNA levels of SOX members (SOX4, SOX7 and SOX17) in the indicated breast cancer cell lines. (D,E) Effect of β-catenin knockdown on FOXA1 and E-cadherin expression. β-catenin expression was suppressed by two different β-catenin-targeting shRNAs (#1 and #2) in MDA-MB-231 SCUBE2 cells. An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). Effect of β-catenin knockdown on nuclear expression of β-catenin and FOXA1 or total E-cadherin was verified by RT-PCR (D) or western blot (E) analyses. Anti-lamin A/C blotting served as a nuclear marker and loading control.

Methylation of CpG sites near the first exon of SCUBE2 gene is crucial for repression of its expression

Because SCUBE2 is co-regulated with E-cadherin and loss of E-cadherin expression is regulated by transcriptional repressors such as SNAIL, SLUG and ZEB (van Roy and Berx, 2008) or by epigenetic regulation, including DNA methylation and histone deacetylation, in several cancer cell lines (Koizume et al., 2002), we investigated the role of classical repressors of E-cadherin on SCUBE2 expression and whether these epigenetic modifications are involved in suppression of SCUBE2 mRNA and protein expression in breast-cancer cells. To determine whether these E-cadherin repressors directly regulated SCUBE2 at the transcriptional level, we first performed a promoter assay in SCUBE2-expressing MCF-7 cells by transfecting the expression plasmids encoding these repressors individually. SCUBE2 promoter reporter activity remained relatively unaltered, whereas E-cadherin promoter activity was significantly suppressed by these EMT inducers (supplementary material Fig. S7A). Consistently, overexpression of these E-cadherin repressors has no effect on SCUBE2 expression at both mRNA and protein levels (supplementary material Fig. S7B–D). Our results suggest that these classical repressors of E-cadherin may not be involved in the downregulation of SCUBE2 in breast-carcinoma cells. We then treated SCUBE2-negative MDA-MB-231 cells with 5-aza-2′-deoxycytidine (AZA), a DNA demethylation agent; trichostatin A (TSA), a histone deacetylase inhibitor; or both (supplementary material Fig. S8). Treatment with AZA or AZA plus TSA had greater effects than TSA alone on upregulation of SCUBE2 and E-cadherin at both mRNA and protein levels (supplementary material Fig. S8A,C). Thus, DNA methylation but not histone acetylation may play a major role in regulating SCUBE2 gene expression in breast-carcinoma cells.

Genomic analysis revealed CpG islands at the 5′ flanking promoter and downstream of the first exon region of SCUBE2 (supplementary material Fig. S9A). To delineate which clusters of CpG sites are involved in regulating SCUBE2 expression, we generated three independent luciferase reporter gene constructs containing: (1) an upstream ∼1.5-kb promoter (Prom-Luc), (2) a genomic fragment containing exon 1 and a 5′ part of intron 1 (Ex1+In1-Luc) or (3) a 3′ part of intron 1 and exon 2 (In1+Ex2-Luc) (supplementary material Fig. S9A). We prepared methylated or unmethylated plasmid DNAs of these three constructs and transiently transfected them into MCF-7 cells to determine whether the transcriptional activity of these proximal SCUBE2 promoter regions was sensitive to DNA methylation (supplementary material Fig. S9B). Luciferase activity assay revealed that in vitro methylation of the Ex1+In1-Luc or In1+Ex2-Luc but not the Prom-Luc plasmid significantly inhibited transcriptional activity in MCF-7 cells (supplementary material Fig. S9C). Thus, SCUBE2 expression is indeed repressed by DNA methylation in breast-cancer cells, which may be specifically attributed to the CpG sites downstream of exon 1.

Methylation status and recruitment of DNA methyltransferase 1 to the SCUBE2 CpG sites during TGF-β1-induced EMT

Because the Ex1+In1-Luc reporter activity was much reduced by DNA methylation and contained more CpG sites, we focused on this regulatory region to quantify its DNA methylation levels. Mass spectrometry of base-specific cleaved amplification products (Sequenom EpiTYPER assay) of the four breast-cancer cell lines revealed the CpG island downstream of exon 1 in a single amplicon (position +159 to +592 nt; Fig. 8A). The level of DNA methylation was inversely associated with SCUBE2 mRNA expression (Fig. 8B): MDA-MB-231 and ZR-75-30 cells with no or low SCUBE2 expression showed a high degree of methylation (>70% of detectable CpG sites) compared with SCUBE2-positive breast-cancer cells MCF-7 or T47-D. Therefore, CpG methylation status downstream of exon 1 has an important role in SCUBE2 expression in breast-carcinoma cells. To further validate the CpG methylation status, we designed methylation-specific primers to target CpG sites downstream of exon 1 of SCUBE2 on the basis of quantitative DNA methylation results. We detected an unmethylated allele only in SCUBE2-positive MCF-7 and T-47D cells, but methylated bands were markedly amplified in SCUBE2-negative MDA-MB-231 and ZR-75-30 cells (Fig. 8C). Thus, the downstream region of exon 1 has an important role in methylation-related gene silencing of SCUBE2 in breast-carcinoma cells.

Fig. 8.

CpG methylation status and recruitment of DNMT1 to CpG sites of the SCUBE2 gene during TGF-β-activated EMT. (A) Sequence of CpG sites (+159 to +592) of human SCUBE2. The location of exons 1 and 2 of SCUBE2 is shown as a boxed number. Vertical lines represent a single CpG site. Methylation status of CpG sites 1–43, which span positions +159 to +592, was quantified by MassARRAY analysis. The number of CpG sites (+159 to +592) are indicated. The primer sequences are boxed. (B) Quantitative methylation status of CpG sites located in SCUBE2 in four breast-cancer cell lines. Different colors show relative methylation changes in 10% increments (red = 0%, yellow = 100% methylated). (C) Methylation-specific PCR (MSP) analysis of the SCUBE2 regulatory region in human breast-cancer cell lines. The PCR products labeled ‘M’ (methylated) were generated by methylation-specific primers and those labeled ‘UM’ (unmethylated) were generated by primers specific for unmethylated genomic DNA. (D) MSP analysis of the SCUBE2 regulatory region. (E) Western blot analysis of the protein level of DNMT1. Lamin A/C was a loading control. (F) Chromatin immunoprecipitation of in vivo binding of DNMT1 protein to the exon 1 region of SCUBE2. Primers amplified 246-bp fragments for SCUBE2 (+346–+592). (G) Effect of DNMT1 overexpression and knockdown on DNA methylation and expression of SCUBE2. The expression plasmid encoding DNMT1 was transiently transfected into MCF-7 cells. As well, DNMT1 expression was inhibited by transfection of plasmid encoding DNMT1-targeting shRNA (#1 and #2) into MDA-MB-231 cells. An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). Three days after transfection, RT-PCR, methylation-specific PCR (MSP) and western blot (WB) analyses were performed to determine the effects of DNMT1 gain- and loss-of-function on DNA methylation and SCUBE2 expression. The MSP products labeled M (methylated) were generated by methylation-specific primers and those labeled UM (unmethylated) were generated by primers specific for unmethylated genomic DNA.

Fig. 8.

CpG methylation status and recruitment of DNMT1 to CpG sites of the SCUBE2 gene during TGF-β-activated EMT. (A) Sequence of CpG sites (+159 to +592) of human SCUBE2. The location of exons 1 and 2 of SCUBE2 is shown as a boxed number. Vertical lines represent a single CpG site. Methylation status of CpG sites 1–43, which span positions +159 to +592, was quantified by MassARRAY analysis. The number of CpG sites (+159 to +592) are indicated. The primer sequences are boxed. (B) Quantitative methylation status of CpG sites located in SCUBE2 in four breast-cancer cell lines. Different colors show relative methylation changes in 10% increments (red = 0%, yellow = 100% methylated). (C) Methylation-specific PCR (MSP) analysis of the SCUBE2 regulatory region in human breast-cancer cell lines. The PCR products labeled ‘M’ (methylated) were generated by methylation-specific primers and those labeled ‘UM’ (unmethylated) were generated by primers specific for unmethylated genomic DNA. (D) MSP analysis of the SCUBE2 regulatory region. (E) Western blot analysis of the protein level of DNMT1. Lamin A/C was a loading control. (F) Chromatin immunoprecipitation of in vivo binding of DNMT1 protein to the exon 1 region of SCUBE2. Primers amplified 246-bp fragments for SCUBE2 (+346–+592). (G) Effect of DNMT1 overexpression and knockdown on DNA methylation and expression of SCUBE2. The expression plasmid encoding DNMT1 was transiently transfected into MCF-7 cells. As well, DNMT1 expression was inhibited by transfection of plasmid encoding DNMT1-targeting shRNA (#1 and #2) into MDA-MB-231 cells. An shRNA targeting bacterial β-galactosidase was used as a negative control (Control-shRNA). Three days after transfection, RT-PCR, methylation-specific PCR (MSP) and western blot (WB) analyses were performed to determine the effects of DNMT1 gain- and loss-of-function on DNA methylation and SCUBE2 expression. The MSP products labeled M (methylated) were generated by methylation-specific primers and those labeled UM (unmethylated) were generated by primers specific for unmethylated genomic DNA.

We further investigated whether epigenetic alterations were involved in repressing SCUBE2 expression in these cells with TGF-β-induced EMT. Epigenetic modifications in the regulatory region of the SCUBE2 gene were assessed by methylation-specific PCR (MSP). DNA methylation of the SCUBE2 locus started at day 7 and peaked at day 21 with TGF-β1 treatment in MCF-7 cells (Fig. 8D). Therefore, downregulation of SCUBE2, along with decreased E-cadherin level, is a dynamic event followed by a sequential and progressive increase in DNA hypermethylation of the regulatory CpG sites in the SCUBE2 gene during TGF-β-induced EMT.

Because DNA methyltransferase 1 (DNMT1), the major DNMT in adult cells, catalyzes DNA methylation and contributes to the CpG-island hypermethylation of tumor-suppressor genes, including E-cadherin (Suzuki et al., 2004; Wang et al., 2008), we determined whether DNMT1 is upregulated and binds to the CpG sites near exon 1 of SCUBE2. Western blot analysis and a ChIP assay revealed higher DNMT1 protein expression with predominant binding to the CpG island near exon 1 in MDA-MB-231 cells that did not express SCUBE2 than in SCUBE2-expressing MCF-7 cells (supplementary material Fig. S10). In line with these findings, treatment with TGF-β1 for 21 days to induce EMT increased the expression of DNMT1 (Fig. 8E) and binding of DNMT1 to the regulatory CpG sites at the SCUBE2 locus in MCF-7 cells (Fig. 8F). Furthermore, we performed gain- and loss-of-function experiments to confirm the effect of DNMT1 on DNA methylation and expression of SCUBE2. DNMT1 overexpression increased DNA methylation and suppressed SCUBE2 expression in MCF-7 cells, whereas DNMT1 knockdown decreased DNA methylation and upregulated SCUBE2 expression in MDA-MB-231 cells (Fig. 8G). Therefore, DNMT1 is upregulated and actively recruited to methylate the SCUBE2 CpG sites, which subsequently inactivates the expression of tumor-suppressor SCUBE2 during TGF-β-promoted EMT.

SCUBE2 was initially identified (van't Veer et al., 2002) and thereafter repeatedly reported as a breast-tumor-associated gene by several large-scale microarray gene expression profiling analyses (Paik et al., 2004; Paik et al., 2006; van de Vijver et al., 2002). Interestingly, SCUBE2 is the only gene commonly described in reports of these studies (Fan et al., 2006), so SCUBE2 must play a crucial role in breast-cancer biology. In support of this notion, the expression of SCUBE2 and 15 other breast-cancer-related genes have been clinically used to calculate a predictive score in guiding adjuvant treatment for breast-cancer patients (Kim and Paik, 2010; Sparano and Paik, 2008). In addition, our previous clinical-association study suggested that SCUBE2 is a prognostic marker for a favorable clinical outcome (Cheng et al., 2009) and is a novel breast-tumor suppressor (Lin et al., 2011). However, the functions of SCUBE2 in breast-cancer cell motility and invasion and the gene regulatory mechanism for this tumor suppressor during breast-cancer progression are largely unknown.

In this study, we first demonstrated that SCUBE2 was co-expressed with E-cadherin in a number of breast-cancer cell lines (supplementary material Fig. S3). Overexpressed SCUBE2 could interact with and increase the formation of the E-cadherin–β-catenin complex in invasive MDA-MB-231 breast-cancer cells (Fig. 1D–H). Consistent with these findings, introduction of SCUBE2 to these cells elevated the expression of epithelial E-cadherin and downregulated the mesenchymal markers vimentin and N-cadherin (Fig. 3), thus promoting an epithelial transition by reversing EMT. Consequently, SCUBE2 overexpression induced the formation of adherens junction-like strands on the membrane at cell–cell contacts (Fig. 1D–G) and transformed cells from a fibroblastic phenotype to an epithelia-like one that formed multicellular aggregates (Fig. 3A and supplementary material Fig. S5). These SCUBE2-expressing cells also showed markedly reduced cell motility and invasion activity (Fig. 3D,E). Because cell dissociation, motility and invasion are key early events in the metastatic process, SCUBE2 may be a breast-cancer invasion and metastasis suppressor by inhibiting breast-cancer progression at the initial step of metastasis. Overall, these findings are in agreement with the prognostic value of SCUBE2 in breast tumors, in that patients with SCUBE2-positive tumors had better disease-free survival than those with low or negative SCUBE2-protein-expressing tumors (Cheng et al., 2009). However, further studies are needed to elucidate the signaling pathway underlying the anti-migratory and anti-invasive functions of SCUBE2 in breast-carcinoma cells.

It is of interest to note that the E-cadherin-binding element (i.e. the EGF-like repeats) in SCUBE2 is extracellular, whereas SCUBE2 interacts with the cytoplasmic domain of E-cadherin (Fig. 1). In addition, our GST fusion protein pull-down assays did not reveal a direct interaction between SCUBE2 and E-cadherin, suggesting that as-yet-unknown protein(s) may connect the cytoplasmic juxtamembrane region of E-cadherin with the surface-tethered SCUBE2 into adherens junction complexes to elicit β-catenin–SOX signaling and FOXA1 upregulation. Similar to our previous study showing that SCUBE2 can modulate β-catenin signaling (Lin et al., 2011), here we found that SCUBE2 expression is positively associated with β-catenin expression at the mRNA transcriptional level but not by post-translational regulation of its stability and localization (Fig. 7). Unlike its conventional role with a reduced β-catenin and TCF transcriptional activity in epithelial cells, the SCUBE2-induced expression of SOX proteins may regulate β-catenin–TCF activity in the context of mesenchymal cells as seen in this study. In support of this notion, it has been reported that β-catenin and SOX17 can interact to transcribe endodermal gene expression of FoxA1 in Xenopus embryos (Sinner et al., 2004) and both SOX4 and SOX7 that are closely associated with β-catenin and SCUBE2 expression in breast cancer cells (Fig. 7) can interact with β-catenin and modulate β-catenin-mediated transcription (Sinner et al., 2007; Takash et al., 2001).

Previous molecular and cell biology studies revealed that FOXA1, a forkhead family transcription factor, can directly bind and transactivate the promoter of E-cadherin (Liu et al., 2005). When overexpressed, it induces the formation of E-cadherin-containing adherens junctions and transforms invasive MDA-MB-231 cells from being spindle-shaped to more epithelial like, with a concomitant decrease of motility (Liu et al., 2005). Furthermore, FOXA1 has been shown to be a ‘pioneer factor’ of chromatin remodeling that can open compacted chromatin through its C-terminal domain to facilitate further transcription (Carroll et al., 2005; Cirillo et al., 2002). Therefore, highly expressed FOXA1 protein probably can activate a silent E-cadherin promoter by its dual functions not only as a potent transactivator (Fig. 6) but also a chromatin remodeler to decrease the binding of the transcriptional repressors SNAIL and ZEB1 to the E-cadherin promoter (Fig. 5E,F). However, further investigations are needed to verify this hypothesis. In contrast, silencing of FOXA1 is sufficient to promote EMT (Song et al., 2010). Consistent with these findings, clinical studies showed that FOXA1 expression is a positive prognostic factor among patients with ER-positive tumors (Badve et al., 2007; Hisamatsu et al., 2012). In this study, we identified that SCUBE2 functions upstream, on the plasma membrane at adherens junctions to trigger induction of FOXA1 possibly by β-catenin–SOX transcriptional activity and subsequent E-cadherin upregulation (Fig. 7). Because this SCUBE2–β-catenin/SOX–FOXA1–E-cadherin pathway does not depend on the regulation of well-established E-cadherin repressors such as TWIST, SNAIL, SLUG, ZEB1 or ZEB2 (Fig. 5), it adds a novel mechanism to the regulatory circuits defining the differentiation state of breast-cancer cells.

Besides genetic mutations, epigenetic modifications such as DNA methylation of the E-cadherin gene promoter play a role in malignant tumor transformation (Grady et al., 2000; Graff et al., 1995; Yoshiura et al., 1995). Likewise, the loss or downregulation of SCUBE2 expression is regulated by DNA hypermethylation at the CpG islands downstream of exon 1, as evidenced by bisulphite sequencing and MSP (Fig. 8). Consistently, treating SCUBE2-negative MDA-MB-231 breast-cancer cells with AZA (a DNA methyltransferase inhibitor) reversed the hypermethylation status and restored both SCUBE2 and E-cadherin mRNA and protein expression (supplementary material Fig. S8). Most importantly, during TGFβ-induced EMT, SCUBE2 was downregulated by DNA hypermethylation with upregulation and recruitment of DNMT1 to the regulatory CpG sites at the SCUBE2 locus in MCF-7 breast-cancer cells (Fig. 8). DNMT1 may play a role in regulating SCUBE2, together with E-cadherin, for both EMT and the reversal of EMT, during breast-cancer progression.

Of note, unlike E-cadherin, which is silenced by 5′ promoter CpG-island methylation (Grady et al., 2000; Graff et al., 1995; Yoshiura et al., 1995), with SCUBE2 expression, the intrinsic CpG sites immediately downstream of exon 1 appear to play an essential role in DNA-methylation-mediated inactivation by active recruitment of DNMT1 in breast-carcinoma cells. Although CpG islands are generally found at the 5′ promoter region of genes, methylation of CpG islands far downstream of the transcription start site (Medvedeva et al., 2010) or within both exons and introns (Kim et al., 2008; Lu et al., 2001) can regulate gene expression. However, the precise mechanisms of the binding of DNMT1 to these SCUBE2 intrinsic CpG sites remain to be further explored. One possible mechanism, described for E-cadherin repression during EMT (Dong et al., 2012), is whether G9a (a histone lysine methyltransferase) could interact with DNMT1 directly for recruitment to the SCUBE2 CpG sites by associating with SNAIL, thus leading to SCUBE2 DNA hypermethylation and EMT induction.

In summary, our data unveiled previously unknown anti-migratory and anti-invasive functions of a newly discovered breast-tumor-associated gene, SCUBE2. Mechanistically, SCUBE2 acts as a tumor migratory and invasive suppressor by increasing the formation of E-cadherin–β-catenin adhesive complexes on the plasma membrane and drives epithelial differentiation through the reversal of EMT. Furthermore, TGF-β-mediated upregulation and recruitment of DNMT1 to regulatory CpG islands are crucial molecular events epigenetically inactivating this novel breast-tumor suppressor during EMT. This study highlights the important roles of SCUBE2 in breast-cancer progression and suggests further exploration of SCUBE2 as a diagnostic or therapeutic target for breast tumors.

Antibodies

Anti-SCUBE2, anti-DNMT1 and anti-N-cadherin polyclonal antibodies were from GeneTex (Irvine, CA, USA). E-cadherin and β-actin antibodies were from BD Biosciences (San Jose, CA, USA) and NOVUS Biologicals (Littleton, CO, USA), respectively. Vimentin, SNAIL, SLUG and lamin A/C antibodies were from Cell Signaling Technology (Danvers, MA, USA). FOXA1 and TWIST antibodies were from Abcam (Cambridge, MA, USA). ZEB1 and ZEB2 antibodies were from Santa Cruz Biotechnology (Dallas, Texas, USA).

Expression plasmids and the luciferase reporter vectors

The expression plasmids encoding SCUBE2 or E-cadherin were constructed as described previously (Lin et al., 2011). An expression plasmid encoding FOXA1 (pcDNA3-FOXA1) and the E-cadherin promoter luciferase reporter plasmid were kind gifts from Ji Hshiung Chen (Tzu Chi University, Taiwan) as described (Liu et al., 2005).

Cell culture

MCF-7, T-47D, MDA-MB-231 and ZR-75-30 human breast-cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA). MCF-7 and MDA-MB-231 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 100 µg/ml streptomycin. T-47D and ZR-75-30 breast cancer cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 µg/ml streptomycin.

TGF-β1 treatment

MCF-7 cells (5×105 cells) were seeded into 6-cm dishes for 18 hours. Cells were washed with phosphate-buffered saline (PBS) and incubated in serum-free medium for 24 hours. Cells were treated with 10 ng/ml TGF-β1 (in 1% fetal bovine serum medium) for 7, 14 and 21 days, with medium replaced with fresh TGF-β1 every 3 days.

Immunoprecipitation and western blot analysis

Cancer cells were washed once with PBS and lysed for 15 minutes on ice in lysis buffer. Lysates were clarified by centrifugation at 10,000 g for 15 minutes at 4°C. Samples underwent immunoprecipitation and western blot analysis as described (Tsai et al., 2009).

RT-PCR

Total RNA was prepared from cultured cells by the TRIzol method (Life Technologies, Grand Island, NY). First-strand cDNA synthesis with SuperScript II reverse transcriptase (Life Technologies) used 5 µg RNA. One-tenth of the first-strand cDNA reaction was used for each PCR as a template. Primers are listed in supplementary material Table S1. The PCR products were run on a 1% agarose gel.

Confocal immunofluorescence microscopy

MDA-MB-231 control, SCUBE2-expressing or MCF-7 cells were fixed in 4% formaldehyde, blocked with 2% fetal bovine serum for 1 hour, and incubated with mouse anti-E-cadherin antibody and rabbit anti-vimentin antibody for 1 hour. Slides were washed three times with PBS and stained with Alexa-Fluor-488-labeled anti-mouse IgG antibody and Alexa-Fluor-594-labeled anti-rabbit IgG antibody for 1 hour, then washed three times in PBS and mounted with VECTASHIELD mounting medium and DAPI (Vector Laboratories, Burlingame, CA, USA). Fluorescence images were captured at room temperature under a confocal microscope (model LSM 510; Carl Zeiss, Thornwood, NY, USA).

Cell aggregation assay

MDA-MB-231 control and SCUBE2-expressing cells were detached by treatment with trypsin (0.01%)-EDTA (2.5 mM) and suspended in DMEM containing 5 mM CaCl2 or Ca2+ free DMEM, at 1×106 cells/ml in polystyrene tubes. Then tubes were incubated on a rotating platform (10 rpm) at 37°C for 9 hours. The cell aggregation was viewed microscopically and photographed. For quantification, cell clusters of more than four cells were considered as aggregates. The data are means ± s.d. of three independent experiments performed in duplicate.

Cell migration assay

For the breast cancer cell migration assay, a culture insert (IBIDI GmbH, Martinsried, Germany) was placed into one well of a 24-well plate. An equal number of MDA-MB-231 control, SCUBE2-expressing, MCF-7 control-shRNA, or SCUBE2-shRNA cells (100 µl; 5×104 cells/ml) were added into the two reservoirs of the same insert and incubated at 37°C, 5% CO2. After 16 hours, the insert was gently removed, thus creating a gap of ∼500 µm. The well was filled with complete growth medium and migration was observed by live-cell imaging with a microscope (model IX71, Olympus, Center Valley, PA, USA).

Cell invasion assay

For invasion assays, cells were suspended in serum-free medium and plated in duplicate in the top well of Matrigel invasion chambers (8-µm pore size; BD, Franklin Lakes, NJ, USA). Complete medium was placed in the lower chamber and cells were allowed to invade for 18 hours at 37°C, 5% CO2. Cells in the upper chamber were removed using a cotton swab, and cells on the lower chamber were fixed in methanol and stained with Crystal Violet. The number of invasive cells was counted for three random fields per experiment from three independent experiments.

Luciferase reporter assay

MCF-7, MDA-MB-231 control, and SCUBE2 cells were plated in 24-well plates (1×105 cells/well) and incubated overnight at 37°C. The next day, cells were transfected with 0.5 µg SCUBE2 promoter, exon 1, exon 2 (MCF-7 cells) or E-cadherin promoter (MDA-MB-231 cells) luciferase constructs and internal control (0.05 µg pRL-TK Renilla luciferase plasmid) with FuGENE HD (Promega, Madison, WI, USA) according to the manufacturer's instructions. Cells were cultured for an additional 2 days, harvested and prepared for reporter assay with the Dual-Luciferase reporter assay system (Promega).

Pull-down assay

Recombinant HA-tagged p120-catenin protein (HA–p120-catenin), FLAG-tagged SCUBE2-FL, -ty97 or -D4 (FLAG–SCUBE2-FL, -ty97, -D4) protein was produced by overexpression from HEK-293T cells. The GST–E-cadherin protein was mixed with FLAG–SCUBE2 protein bound to anti-FLAG M2-antibody agarose beads or control beads in 0.5 ml binding buffer [40 mM HEPES (pH 7.5), 100 mM KCl, 0.1% Nonidet P-40 and 20 mM 2-mercaptoethanol]. After incubation for 4 hours at 4°C, the beads were washed extensively, and interacting protein was visualized by immunoblotting using anti-GST antibody.

Drug treatment

5-aza-2′-deoxycytidine (AZA) and trichostatin A (TSA) were obtained from CalBiochem (San Diego, CA, USA) and Sigma (St Louis, MO, USA), respectively; dissolved in DMSO and prepared as a 1000-fold concentrated stock solution; and added to the culture medium at 1∶1000 dilution according to the experimental procedure. MDA-MB-231 cells were treated with 10 µM AZA for 96 hours, with medium replaced with fresh AZA every 12 hours. TSA (300 nM) was added to cells 24 hours before the end of experiments.

Preparation of methylated and unmethylated SCUBE2 reporter-gene plasmid DNAs

The reporter plasmid DNAs were methylated in vitro by use of bacterial CpG methylase M.SssI (New England BioLabs, Ipswich, MA, USA) and 1 µM S-adenosylmethionine according to the manufacturer's instructions. Complete methylation of plasmid DNAs was confirmed by digestion with methylation-sensitive endonucleases (HpaII).

Quantitative DNA methylation analysis

Analysis of the regulatory regions of SCUBE2 involved the use of the EpiTYPER methylation assay (Sequenom, San Diego, CA, USA). We designed two bisulfite reactions that covered 43 CpG sites for SCUBE2. The DNA of breast-cancer cells was extracted, and 1 µg DNA was treated with bisulfite, in vitro transcribed, cleaved by RNase A, and subjected to matrix assisted laser desorption ionization time of flight (MALDI-TOF) mass spectroscopy analysis to determine methylation patterns. Primers used to amplify the regulatory regions for SCUBE2 are in supplementary material Table S1.

DNA extraction, bisulfite treatment and methylation-specific PCR

Genomic DNA was extracted from breast-cancer cells by use of the QIAamp DNA mini kit (Qiagen GmbH, Hilden, Germany). We used the EZ DNA Methylation Kit (ZYMO Research, Irvine, CA, USA). Briefly, extracted DNA samples were denatured and subjected to bisulfite treatment using CT conversion reagent for 16 hours at 50°C. Modified DNA was purified by use of the Zymo-Spin IC column according to the manufacturer's instructions. Modified DNA was then used for MSP to detect the methylation status of the SCUBE2 exon 1 to exon 2 region. The primer sequences designed to amplify specifically methylated (M) or unmethylated (UM) forms of SCUBE2 exon 1 are in supplementary material Table S1. The amplification products of M and UM forms were 217 and 220 bp, respectively. For both PCR reactions, 100 ng modified DNA was amplified in a 25-µl reaction at 94°C for 5 minutes; then 35 cycles at 94°C for 15 seconds, 60°C for 30 seconds and 72°C for 30 seconds; then 72°C for 5 minutes. The PCR products were run on a 1.5% agarose gel.

Chromatin immunoprecipitation

We used the EZ-Magna ChIP G Kit (Millipore, Billerica, MA, USA). Briefly, MCF-7 or MDA-MB-231 cells (1×107 cells) were cross-linked using 1% formaldehyde, lysed in 500 µl lysis buffer and sonicated to ∼500-bp fragments. ChIP involved antibodies against DNMT1, FOXA1 or IgG. The input control DNA or immunoprecipitated DNA was amplified in a 50-µl reaction volume consisting of 2 µl eluted DNA template with the primers (supplementary material Table S1). Taq polymerase was used for PCR with 35 cycles of 94°C for 30 seconds, 63.5°C for 30 seconds, 72°C for 50 seconds, then 5 minutes at 72°C. A 10-µl aliquot from each PCR reaction was separated on 1.5% agarose gel.

Statistical analyses

SPSS v10.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analyses. P<0.05 was considered statistically significant.

We thank the National Center for Genome Medicine at Academia Sinica, Taiwan, for DNA methylation analysis.

Author contributions

Y.C.L, Y.C.L., L.H.L. designed and performed experiments, and analyzed and interpreted data. C.J.C and R.B.Y. designed research, analyzed and interpreted data, and wrote the manuscript.

Funding

This work was supported by the Taiwan National Science Council [grant numbers NSC 101-2325-B-001-030, 102-2325-B-001-028, 102-2320-B-001-015-MY3 to R.B.Y., and NSC 100-2320-B-038-029 to C.J.C.]. The National Center for Genome Medicine at Academia Sinica, Taiwan is supported by grants from the National Core Facility Program for Biotechnology, National Science Council, Taiwan [grant numbers NSC 101-2319-B-001-001 and 102-2319-B-001-001].

Badve
S.
,
Turbin
D.
,
Thorat
M. A.
,
Morimiya
A.
,
Nielsen
T. O.
,
Perou
C. M.
,
Dunn
S.
,
Huntsman
D. G.
,
Nakshatri
H.
(
2007
).
FOXA1 expression in breast cancer – correlation with luminal subtype A and survival.
Clin. Cancer Res.
13
,
4415
4421
.
Berx
G.
,
Van Roy
F.
(
2001
).
The E-cadherin/catenin complex: an important gatekeeper in breast cancer tumorigenesis and malignant progression.
Breast Cancer Res.
3
,
289
293
.
Carroll
J. S.
,
Liu
X. S.
,
Brodsky
A. S.
,
Li
W.
,
Meyer
C. A.
,
Szary
A. J.
,
Eeckhoute
J.
,
Shao
W.
,
Hestermann
E. V.
,
Geistlinger
T. R.
 et al. (
2005
).
Chromosome-wide mapping of estrogen receptor binding reveals long-range regulation requiring the forkhead protein FoxA1.
Cell
122
,
33
43
.
Cheng
C. J.
,
Lin
Y. C.
,
Tsai
M. T.
,
Chen
C. S.
,
Hsieh
M. C.
,
Chen
C. L.
,
Yang
R. B.
(
2009
).
SCUBE2 suppresses breast tumor cell proliferation and confers a favorable prognosis in invasive breast cancer.
Cancer Res.
69
,
3634
3641
.
Cirillo
L. A.
,
Lin
F. R.
,
Cuesta
I.
,
Friedman
D.
,
Jarnik
M.
,
Zaret
K. S.
(
2002
).
Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4.
Mol. Cell
9
,
279
289
.
Cowin
P.
,
Rowlands
T. M.
,
Hatsell
S. J.
(
2005
).
Cadherins and catenins in breast cancer.
Curr. Opin. Cell Biol.
17
,
499
508
.
Dong
C.
,
Wu
Y.
,
Yao
J.
,
Wang
Y.
,
Yu
Y.
,
Rychahou
P. G.
,
Evers
B. M.
,
Zhou
B. P.
(
2012
).
G9a interacts with Snail and is critical for Snail-mediated E-cadherin repression in human breast cancer.
J. Clin. Invest.
122
,
1469
1486
.
Fan
C.
,
Oh
D. S.
,
Wessels
L.
,
Weigelt
B.
,
Nuyten
D. S.
,
Nobel
A. B.
,
van't Veer
L. J.
,
Perou
C. M.
(
2006
).
Concordance among gene-expression-based predictors for breast cancer.
N. Engl. J. Med.
355
,
560
569
.
Fuxe
J.
,
Vincent
T.
,
Garcia de Herreros
A.
(
2010
).
Transcriptional crosstalk between TGF-β and stem cell pathways in tumor cell invasion: role of EMT promoting Smad complexes.
Cell Cycle
9
,
2363
2374
.
Grady
W. M.
,
Willis
J.
,
Guilford
P. J.
,
Dunbier
A. K.
,
Toro
T. T.
,
Lynch
H.
,
Wiesner
G.
,
Ferguson
K.
,
Eng
C.
,
Park
J. G.
 et al. (
2000
).
Methylation of the CDH1 promoter as the second genetic hit in hereditary diffuse gastric cancer.
Nat. Genet.
26
,
16
17
.
Graff
J. R.
,
Herman
J. G.
,
Lapidus
R. G.
,
Chopra
H.
,
Xu
R.
,
Jarrard
D. F.
,
Isaacs
W. B.
,
Pitha
P. M.
,
Davidson
N. E.
,
Baylin
S. B.
(
1995
).
E-cadherin expression is silenced by DNA hypermethylation in human breast and prostate carcinomas.
Cancer Res.
55
,
5195
5199
.
Grimmond
S.
,
Larder
R.
,
Van Hateren
N.
,
Siggers
P.
,
Hulsebos
T. J.
,
Arkell
R.
,
Greenfield
A.
(
2000
).
Cloning, mapping, and expression analysis of a gene encoding a novel mammalian EGF-related protein (SCUBE1).
Genomics
70
,
74
81
.
Grimmond
S.
,
Larder
R.
,
Van Hateren
N.
,
Siggers
P.
,
Morse
S.
,
Hacker
T.
,
Arkell
R.
,
Greenfield
A.
(
2001
).
Expression of a novel mammalian epidermal growth factor-related gene during mouse neural development.
Mech. Dev.
102
,
209
211
.
Hisamatsu
Y.
,
Tokunaga
E.
,
Yamashita
N.
,
Akiyoshi
S.
,
Okada
S.
,
Nakashima
Y.
,
Aishima
S.
,
Morita
M.
,
Kakeji
Y.
,
Maehara
Y.
(
2012
).
Impact of FOXA1 expression on the prognosis of patients with hormone receptor-positive breast cancer.
Ann. Surg. Oncol.
19
,
1145
1152
.
Hollway
G. E.
,
Maule
J.
,
Gautier
P.
,
Evans
T. M.
,
Keenan
D. G.
,
Lohs
C.
,
Fischer
D.
,
Wicking
C.
,
Currie
P. D.
(
2006
).
Scube2 mediates Hedgehog signalling in the zebrafish embryo.
Dev. Biol.
294
,
104
118
.
Kawakami
A.
,
Nojima
Y.
,
Toyoda
A.
,
Takahoko
M.
,
Satoh
M.
,
Tanaka
H.
,
Wada
H.
,
Masai
I.
,
Terasaki
H.
,
Sakaki
Y.
 et al. (
2005
).
The zebrafish-secreted matrix protein you/scube2 is implicated in long-range regulation of hedgehog signaling.
Curr. Biol.
15
,
480
488
.
Kim
C.
,
Paik
S.
(
2010
).
Gene-expression-based prognostic assays for breast cancer.
Nat. Rev. Clin Oncol
7
,
340
347
.
Kim
S. K.
,
Jang
H. R.
,
Kim
J. H.
,
Kim
M.
,
Noh
S. M.
,
Song
K. S.
,
Kang
G. H.
,
Kim
H. J.
,
Kim
S. Y.
,
Yoo
H. S.
 et al. (
2008
).
CpG methylation in exon 1 of transcription factor 4 increases with age in normal gastric mucosa and is associated with gene silencing in intestinal-type gastric cancers.
Carcinogenesis
29
,
1623
1631
.
Koizume
S.
,
Tachibana
K.
,
Sekiya
T.
,
Hirohashi
S.
,
Shiraishi
M.
(
2002
).
Heterogeneity in the modification and involvement of chromatin components of the CpG island of the silenced human CDH1 gene in cancer cells.
Nucleic Acids Res.
30
,
4770
4780
.
Kormish
J. D.
,
Sinner
D.
,
Zorn
A. M.
(
2010
).
Interactions between SOX factors and Wnt/beta-catenin signaling in development and disease.
Dev. Dyn.
239
,
56
68
.
Lin
Y. C.
,
Chen
C. C.
,
Cheng
C. J.
,
Yang
R. B.
(
2011
).
Domain and functional analysis of a novel breast tumor suppressor protein, SCUBE2.
J. Biol. Chem.
286
,
27039
27047
.
Liu
Y. N.
,
Lee
W. W.
,
Wang
C. Y.
,
Chao
T. H.
,
Chen
Y.
,
Chen
J. H.
(
2005
).
Regulatory mechanisms controlling human E-cadherin gene expression.
Oncogene
24
,
8277
8290
.
Lu
A.
,
Gupta
A.
,
Li
C.
,
Ahlborn
T. E.
,
Ma
Y.
,
Shi
E. Y.
,
Liu
J.
(
2001
).
Molecular mechanisms for aberrant expression of the human breast cancer specific gene 1 in breast cancer cells: control of transcription by DNA methylation and intronic sequences.
Oncogene
20
,
5173
5185
.
Mareel
M. M.
,
Behrens
J.
,
Birchmeier
W.
,
De Bruyne
G. K.
,
Vleminckx
K.
,
Hoogewijs
A.
,
Fiers
W. C.
,
Van Roy
F. M.
(
1991
).
Down-regulation of E-cadherin expression in Madin Darby canine kidney (MDCK) cells inside tumors of nude mice.
Int. J. Cancer
47
,
922
928
.
Medvedeva
Y. A.
,
Fridman
M. V.
,
Oparina
N. J.
,
Malko
D. B.
,
Ermakova
E. O.
,
Kulakovskiy
I. V.
,
Heinzel
A.
,
Makeev
V. J.
(
2010
).
Intergenic, gene terminal, and intragenic CpG islands in the human genome.
BMC Genomics
11
,
48
.
Osada
T.
,
Sakamoto
M.
,
Ino
Y.
,
Iwamatsu
A.
,
Matsuno
Y.
,
Muto
T.
,
Hirohashi
S.
(
1996
).
E-cadherin is involved in the intrahepatic metastasis of hepatocellular carcinoma.
Hepatology
24
,
1460
1467
.
Paik
S.
,
Shak
S.
,
Tang
G.
,
Kim
C.
,
Baker
J.
,
Cronin
M.
,
Baehner
F. L.
,
Walker
M. G.
,
Watson
D.
,
Park
T.
 et al. (
2004
).
A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer.
N. Engl. J. Med.
351
,
2817
2826
.
Paik
S.
,
Tang
G.
,
Shak
S.
,
Kim
C.
,
Baker
J.
,
Kim
W.
,
Cronin
M.
,
Baehner
F. L.
,
Watson
D.
,
Bryant
J.
 et al. (
2006
).
Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer.
J. Clin. Oncol.
24
,
3726
3734
.
Perl
A. K.
,
Wilgenbus
P.
,
Dahl
U.
,
Semb
H.
,
Christofori
G.
(
1998
).
A causal role for E-cadherin in the transition from adenoma to carcinoma.
Nature
392
,
190
193
.
Perou
C. M.
,
Sørlie
T.
,
Eisen
M. B.
,
van de Rijn
M.
,
Jeffrey
S. S.
,
Rees
C. A.
,
Pollack
J. R.
,
Ross
D. T.
,
Johnsen
H.
,
Akslen
L. A.
 et al. (
2000
).
Molecular portraits of human breast tumours.
Nature
406
,
747
752
.
Sinner
D.
,
Rankin
S.
,
Lee
M.
,
Zorn
A. M.
(
2004
).
Sox17 and beta-catenin cooperate to regulate the transcription of endodermal genes.
Development
131
,
3069
3080
.
Sinner
D.
,
Kordich
J. J.
,
Spence
J. R.
,
Opoka
R.
,
Rankin
S.
,
Lin
S. C.
,
Jonatan
D.
,
Zorn
A. M.
,
Wells
J. M.
(
2007
).
Sox17 and Sox4 differentially regulate beta-catenin/T-cell factor activity and proliferation of colon carcinoma cells.
Mol. Cell. Biol.
27
,
7802
7815
.
Song
Y.
,
Washington
M. K.
,
Crawford
H. C.
(
2010
).
Loss of FOXA1/2 is essential for the epithelial-to-mesenchymal transition in pancreatic cancer.
Cancer Res.
70
,
2115
2125
.
Sparano
J. A.
,
Paik
S.
(
2008
).
Development of the 21-gene assay and its application in clinical practice and clinical trials.
J. Clin. Oncol.
26
,
721
728
.
Strumane
K.
,
Berx
G.
,
Van Roy
F.
(
2004
).
Cadherins in cancer.
Handb Exp Pharmacol
165
,
69
103
.
Suzuki
M.
,
Sunaga
N.
,
Shames
D. S.
,
Toyooka
S.
,
Gazdar
A. F.
,
Minna
J. D.
(
2004
).
RNA interference-mediated knockdown of DNA methyltransferase 1 leads to promoter demethylation and gene re-expression in human lung and breast cancer cells.
Cancer Res.
64
,
3137
3143
.
Takash
W.
,
Cañizares
J.
,
Bonneaud
N.
,
Poulat
F.
,
Mattéi
M. G.
,
Jay
P.
,
Berta
P.
(
2001
).
SOX7 transcription factor: sequence, chromosomal localisation, expression, transactivation and interference with Wnt signalling.
Nucleic Acids Res.
29
,
4274
4283
.
Tsai
M. T.
,
Cheng
C. J.
,
Lin
Y. C.
,
Chen
C. C.
,
Wu
A. R.
,
Wu
M. T.
,
Hsu
C. C.
,
Yang
R. B.
(
2009
).
Isolation and characterization of a secreted, cell-surface glycoprotein SCUBE2 from humans.
Biochem. J.
422
,
119
128
.
van de Vijver
M. J.
,
He
Y. D.
,
van't Veer
L. J.
,
Dai
H.
,
Hart
A. A.
,
Voskuil
D. W.
,
Schreiber
G. J.
,
Peterse
J. L.
,
Roberts
C.
,
Marton
M. J.
 et al. (
2002
).
A gene-expression signature as a predictor of survival in breast cancer.
N. Engl. J. Med.
347
,
1999
2009
.
van Roy
F.
,
Berx
G.
(
2008
).
The cell-cell adhesion molecule E-cadherin.
Cell. Mol. Life Sci.
65
,
3756
3788
.
van 't Veer
L. J.
,
Dai
H.
,
van de Vijver
M. J.
,
He
Y. D.
,
Hart
A. A.
,
Mao
M.
,
Peterse
H. L.
,
van der Kooy
K.
,
Marton
M. J.
,
Witteveen
A. T.
 et al. (
2002
).
Gene expression profiling predicts clinical outcome of breast cancer.
Nature
415
,
530
536
.
Vleminckx
K.
,
Vakaet
L.
 Jr
,
Mareel
M.
,
Fiers
W.
,
van Roy
F.
(
1991
).
Genetic manipulation of E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role.
Cell
66
,
107
119
.
Wang
Y.
,
Klijn
J. G.
,
Zhang
Y.
,
Sieuwerts
A. M.
,
Look
M. P.
,
Yang
F.
,
Talantov
D.
,
Timmermans
M.
,
Meijer-van Gelder
M. E.
,
Yu
J.
 et al. (
2005
).
Gene-expression profiles to predict distant metastasis of lymph-node-negative primary breast cancer.
Lancet
365
,
671
679
.
Wang
G.
,
Hu
X.
,
Lu
C.
,
Su
C.
,
Luo
S.
,
Luo
Z. W.
(
2008
).
Promoter-hypermethylation associated defective expression of E-cadherin in primary non-small cell lung cancer.
Lung Cancer
62
,
162
172
.
Woods
I. G.
,
Talbot
W. S.
(
2005
).
The you gene encodes an EGF-CUB protein essential for Hedgehog signaling in zebrafish.
PLoS Biol.
3
,
e66
.
Wu
B. T.
,
Su
Y. H.
,
Tsai
M. T.
,
Wasserman
S. M.
,
Topper
J. N.
,
Yang
R. B.
(
2004
).
A novel secreted, cell-surface glycoprotein containing multiple epidermal growth factor-like repeats and one CUB domain is highly expressed in primary osteoblasts and bones.
J. Biol. Chem.
279
,
37485
37490
.
Yang
R. B.
,
Ng
C. K.
,
Wasserman
S. M.
,
Colman
S. D.
,
Shenoy
S.
,
Mehraban
F.
,
Komuves
L. G.
,
Tomlinson
J. E.
,
Topper
J. N.
(
2002
).
Identification of a novel family of cell-surface proteins expressed in human vascular endothelium.
J. Biol. Chem.
277
,
46364
46373
.
Yoshiura
K.
,
Kanai
Y.
,
Ochiai
A.
,
Shimoyama
Y.
,
Sugimura
T.
,
Hirohashi
S.
(
1995
).
Silencing of the E-cadherin invasion-suppressor gene by CpG methylation in human carcinomas.
Proc. Natl. Acad. Sci. USA
92
,
7416
7419
.
Zajchowski
D. A.
,
Bartholdi
M. F.
,
Gong
Y.
,
Webster
L.
,
Liu
H. L.
,
Munishkin
A.
,
Beauheim
C.
,
Harvey
S.
,
Ethier
S. P.
,
Johnson
P. H.
(
2001
).
Identification of gene expression profiles that predict the aggressive behavior of breast cancer cells.
Cancer Res.
61
,
5168
5178
.

Competing interests

The authors declare no competing interests.

Supplementary information