Progression of benign tumors to invasive, metastatic cancer is accompanied by the epithelial-to-mesenchymal transition (EMT), characterized by loss of the cell-adhesion protein E-cadherin. Although silencing mutations and transcriptional repression of the E-cadherin gene have been widely studied, not much is known about posttranslational regulation of E-cadherin in tumors. We show that E-cadherin is tightly coexpressed with the secretory pathway Ca2+-ATPase isoform 2, SPCA2 (ATP2C2), in breast tumors. Loss of SPCA2 impairs surface expression of E-cadherin and elicits mesenchymal gene expression through disruption of cell adhesion in tumorspheres and downstream Hippo-YAP signaling. Conversely, ectopic expression of SPCA2 in triple-negative breast cancer elevates baseline Ca2+ and YAP phosphorylation, enhances posttranslational expression of E-cadherin, and suppresses mesenchymal gene expression. Thus, loss of SPCA2 phenocopies loss of E-cadherin in the Hippo signaling pathway and EMT–MET transitions, consistent with a functional role for SPCA2 in E-cadherin biogenesis. Furthermore, we show that SPCA2 suppresses invasive phenotypes, including cell migration in vitro and tumor metastasis in vivo. Based on these findings, we propose that SPCA2 functions as a key regulator of EMT and may be a potential therapeutic target for treatment of metastatic cancer.

Implications:

Posttranslational control of E-cadherin and the Hippo pathway by calcium signaling regulates EMT in breast cancer cells.

This article is featured in Highlights of This Issue, p. 1595

Mortality from breast cancer is invariably due to tumor metastasis, which requires a key step known as epithelial-to-mesenchymal transition (EMT) during which cancer cells lose their polarity and cell–cell contacts, thereby acquiring the ability to migrate (1, 2). The process of EMT coincides with the loss of epithelial markers; the most critical of these is the Ca2+-binding cell-adhesion protein E-cadherin (3, 4). As a structural component of the adherens junction in polarized epithelial cells, E-cadherin is a membrane-spanning protein that mediates homophilic cell–cell contacts through its extracellular domains and communicates these interactions via its intracellular domains to a network of signaling proteins to control cell polarity, organ size, and epithelial integrity (5–7). Contact inhibition of growth is transmitted by E-cadherin to the Hippo signaling pathway, resulting in nuclear exclusion and inactivation of the transcriptional coactivator YAP (Yes-activated protein) and its paralog, TAZ (transcriptional activator with PDZ binding motif; refs. 6, 8). Loss of Hippo signaling or overexpression of YAP/TAZ promotes metastatic phenotypes, including EMT (9). A better understanding of the molecular pathways that downregulate E-cadherin and control the Hippo pathway may be pivotal to discovering new therapeutic targets for treatment of metastatic cancers.

E-cadherin expression is regulated at multiple levels, including genetic, epigenetic, transcriptional, and posttranslational. At the transcriptional level, CDH1, the gene encoding E-cadherin is repressed by a set of nuclear factors, including ZEB1 and SNAI2/SLUG, that reprogram the cancer cell to acquire malignant invasive phenotypes (10). Expression of these factors in tumors inversely correlates with E-cadherin and prognosticates a more aggressive, invasive phenotype with poor clinical prognosis. In contrast to the wealth of information on CDH1 regulation, much less is known of oncogenic control of E-cadherin biogenesis despite some evidence that posttranslational downregulation of E-cadherin precedes repression of transcription upon induction of EMT (11).

In this study, we uncover a novel functional link between E-cadherin and SPCA2, one of two secretory pathway Ca2+-ATPases that transport Ca2+ ions into the Golgi lumen for protein sorting, trafficking, and quality control (12–14). Duplication of the SPCA gene (ATP2C) occurred during tetrapod evolution (15) and paved the way for acquisition of distinct functions. Although they have similar ion transport characteristics and appear redundant at first glance, we show that SPCA2, but not SPCA1, is an epithelial signature gene (ESG) required for maximal cell surface expression of E-cadherin, and to maintain contact inhibition of EMT through the Hippo pathway. Although both secretory pathway Ca2+ pumps supply the Golgi and downstream compartments with Ca2+, SPCA2 preferentially interacts with plasma membrane Ca2+ channels to elicit Ca2+ influx (16, 17). We show that ectopic expression of SPCA2 elevates baseline Ca2+ levels and YAP phosphorylation, resulting in suppression of mesenchymal phenotypes in triple receptor-negative breast cancer (TNBC) lines, including cell migration in vitro and tumor metastasis in vivo. Our findings reveal SPCA2 as a novel regulator of E-cadherin biogenesis and EMT in breast cancer cells, and a potential therapeutic target in the treatment of metastatic cancers.

Hierarchical clustering

Cancer Cell Line Encyclopedia (CCLE) and The Cancer Genome Atlas (TCGA) invasive breast carcinoma project datasets were accessed through cBioPortal (http://www.cbioportal.org/; ref. 18). Hierarchical clustering to analyze gene expression data was performed with statistiXL software, and results were displayed as a dendrogram and correlation coefficient matrix heat map.

Cell culture

Cells used in this study were MCF7 (ATCC HTB-22), MDA-MB-231 (ATCC HTB-26), MCF10A (ATCC CRL-10317), SK-BR-3, BT-483, MDA-MB-415, and AU-565 (gift from Dr. Saraswathi Sukumar, Johns Hopkins Medical School, MD). MDA-MB-231 tet-Ecad cells were a gift from the Barry Gumbiner, Seattle Children's Hospital (6). All cell lines were grown for less than a month or no more than 6 passages, consistent with AACR guidelines. We did not see visible contamination by mycoplasma; however, no testing was performed. SK-BR-3 and AU-565 were grown in RPMI 1640 medium containing antibiotic and 10% FBS. MDA-MB-231 tet-Ecad and MCF7 cells were cultured in monolayer with DMEM containing antibiotic, antimycotic, and 10% serum. MCF10A monolayer cells were cultured in DMEM/F12 containing 5% horse serum, 20 ng/mL EGF, 0.5 mg/mL hydrocortisone, 100 ng/mL cholera toxin, 10 μg/mL insulin, and 1X penicillin/streptomycin. Note that 50 ng/mL of EGF (Sigma-Aldrich, cat#E9644) was added to the culture for 6 hours to induce EMT. Cells were cultured with 5% CO2 and at 37°C in a humidified incubator.

Lentiviral transfection

FUGW overexpression constructs and pLK0.1 shRNA and lentiviral construction of both SPCA isoforms were packaged and transfected according to previous methods using pCMV-Δ8.9 and PMDG at a ratio using 9:8:1 in HEK293T cells. A mixture of two shRNA constructs for SPCA2 gave the same results as individual constructs and have been previously validated, as reported (16, 19). Virus was collected after 48 hours and concentrated with Lenti-X Concentrator (Clontech, Cat#631231; ref. 16). After 48 hours of transfection, cells were selected with puromycin at varying concentrations according to kill curves performed for each specific cell line (1–2.5 mg/mL). Experiments were performed within 5 passages to ensure maintained knockdown.

cDNA synthesis and quantitative PCR

Note that 1 μg of RNA was collected and used for cDNA synthesis (Applied Biosystems, cat#4387406). The qPCR mastermix was made with EagleTaq Universal Mastermix (Roche, cat#07260296190), Taqman probe as specified, and 50 ng of cDNA. Probes were as follows (Applied Biosystems): GAPDH (Hs02758991_g1), SPCA2 (Hs00939492_m1), SPCA1 (Hs00995930_m1), E-cadherin (Hs01023894_m1 or HS01023895_m1), N-cadherin (Hs00983056_m1), Zeb1 (Hs00232783_m1), Snail1 (Hs00195591_m1), and Vimentin (Hs00958111_m1).

Western blotting

Note that 35 to 50 μg of lysates were utilized for protein analysis. GAPDH (Sigma-Aldrich, cat#G9295) was used as the loading control for Western blotting. E-cadherin (BD Transduction Laboratories, cat#610181), ZO-1 (H-300; Santa Cruz Biotechnology, cat#sc-10804), pYAP (Ser127, Cell Signaling Technology; #4911), and EGFR antibody (Abcam, cat# ab2430) were analyzed for total protein. Blots were analyzed using ImageJ.

Immunofluorescence

Cells were cultured on glass coverslips and were rinsed with PBS and preextracted with 1X PHEM buffer, 8% sucrose, and 0.025% saponin. Cells were fixed with 4% paraformaldehyde for 30 minutes and were rinsed and washed with PBS 3 times for 5 minutes each. After blocking in 1% BSA, cells were incubated with primary antibody overnight in 4°C. Cells were rinsed with 0.2% BSA 3 times for 5 minutes and were then incubated with a fluorescent secondary antibody in 1% BSA and 0.025% saponin buffer for 30 minutes at room temperature. Coverslips were washed and mounted onto slides with mounting media (Dako, cat#S3023). Immunofluorescent staining was analyzed using ImageJ.

Flow cytometry

Cells were rinsed with PBS and trypsinized and washed once with PBS. Resuspended cells were fixed in 4% paraformaldehyde at room temperature for 15 minutes and then washed 3 times in PBS. Fixed cells were permeabilized in a PBS buffer containing 0.1% saponin and 1% BSA for 15 minutes at room temperature. Cells were incubated with primary antibodies for 30 minutes at room temperature in the dark in the same saponin/BSA buffer and then washed 3 times. If needed, cells were incubated with secondary antibodies in the saponin/BSA buffer for 30 minutes in the dark and washed once. Cells were washed once and resuspended in a flow buffer (1 mmol/L EDTA, 0.1% Na+ azide, 1% BSA, and PBS). Cells were run using BD FACSDiva on a BD LSR II flow cytometer and were analyzed with FlowJo. Antibodies used were 1:500 FITC Mouse Anti–E-cadherin (BD Biosciences, cat#612130), 1:500 DYKDDDDK Tag Polyclonal Antibody (ThermoFisher Scientific, cat#PA1-984B), 1:1,000 Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, and Alexa Fluor 633 (ThermoFisher Scientific, cat#A-21070).

Ca2+-Dependent Fura2 fluorescence

Live imaging of Ca2+ in MDA-MB-231 cells was performed as described (20). Fura2-AM (Invitrogen; 1 μg/mL) was added to cells in imaging buffer (20 mmol/L Hepes, 126 mmol/L NaCl, 4.5 mmol/L KCl, 2 mmol/L MgCl2, 10 mmol/L glucose at pH7.4). Cells were excited at 340 and 380 nm, and Fura emission was captured at 505 nm. To show store-independent Ca2+ entry (SICE), cells were briefly washed in nominally Ca2+-free buffer followed by addition of Ca2+ (2 mmol/L) as described in the legend to Fig. 5.

Tumorsphere suspension assay

MCF7 cells (1,000 cells) were seeded and cultured in serum-free suspension media (DMEM, 31% l-glutamine, 1% penicillin/streptomycin, 2% B27 Supplement, 20 ng/mL EGF, and 20 ng/mL FGFb) in 6-well culture plates according to the method from Iglesias and colleagues (21). Culture plates were treated with poly (2-hydroxethyl methacrylate) (Sigma-Aldrich, cat# P3932) prior to culturing to prevent adherence. Tumorspheres were counted and measured using ImageJ.

YAP localization

MCF-7 cells were infected with scramble “Control” or “shSPCA2” lentivirus for 48 hours, and then selected with puromycin (2.5 mg/mL) for 48 hours. Following, cells were seeded either sparse or confluent on coverslips, as previously described (22). Briefly, adherent cells were serum starved for 24 hours before fixation and subsequent immunofluorescence. MCF7 cells, due to an aberrant mutation in PI3-K that alters YAP trafficking, were treated with wortmannin (10 mmol/L) for 4 hours to restore normal YAP trafficking prior to fixation (22). YAP intracellular localization was analyzed with anti-YAP antibody (63.7; Santa Cruz Biotechnology, cat#SC-101199) using confocal microscopy. Where indicated, MDA-MB-231 cells expressing SPCA2R were treated with YAP inhibitor Verteporfin (10 μmol/L; TOCRIS, Cat# 5305) for 24 hours and investigated for changes in EMT markers.

Cell migration

Wound-healing assay.

MDA-MB-231 cells were plated in a 12-well culture plate. Upon reaching approximately 70% confluency, cells were maintained in 0.5% FBS containing DMEM media for 16 hours. Upon reaching confluency, using a pipette tip, horizontal scratches were made across the diameter of the well as described (23). The monolayer was washed couple of times with PBS to remove any floating cells generated by the scratching process, and fresh media (0.5% FBS containing DMEM) were added. Using an inverted microscope (Zeiss Axio observer), photos of each scratch were captured immediately at 2.5X magnification and used as a reference point for the 8-, 16-, and 24-hour time point to determine percentage of scratch coverage.

Boyden chamber assay.

MDA-MB-231 cells were maintained in DMEM media without FBS for 4 hours. Note that 3 × 104 cells of each condition were plated in 100 μL DMEM media without FBS in 6.5 mm Transwell chambers with 8.0 μm Pore Polyester Membrane Insert (Corning, Cat# 3464), and transferred to a 24-well dish containing 500 μL fresh media (0.5% FBS containing DMEM). After 24 hours, cells were fixed with 4% paraformaldehyde for 30 minutes, then 0.5% crystal violet (Sigma-Aldrich, Cat# V5265) was added for 1 hour, and washed with PBS. Using a microscope (Nikon, Eclipse TS100), photos of membrane insert were captured at 10X magnification. ImageJ software was used to quantify migration.

Mouse xenografts

Animal protocols described in this study were approved by the Institutional Care and Use Committee at the Johns Hopkins Medical Institutions. MDA-MB-231-luc-D3H2LN cell line expressing the luciferase gene was obtained from Cellosaurus. MDA-MB-231-luc-D3H2LN cells stably infected with SPCA2 overexpression or GFP-containing virus. Cells were suspended in matrigel:DMEM media without FBS (1:1 volume). Six- to 8-week-old female NOD scid gamma (NSG) mice (The Jackson Laboratory) were anesthetized, and approximately 0.7 × 106 cells were injected into the mammary fat pad. Each group had 6 mice each. Animals were imaged every week to track tumor metastases using IVIS Lumina XR (Perkin-Elmer), after i.p. injection of D-luciferin-K+ (Perkin-Elmer, 150 mg/kg). Metastasis was represented as mean total flux (p/sec) ± SEM, which was calculated using Living Image Software version 4.3.1 (Perkin-Elmer). Five weeks after injection, progression of metastasis between the GFP and SPCA2 overexpression groups was compared, and the t test was used to determine the significance. Mice weights were measured every week.

SPCA2 is an epithelial marker tightly linked to E-cadherin expression

Gene expression correlations offer unique insight into disease pathophysiology, based on the premise that genes expressed together are likely to function together in a common pathway (24, 25). Both SPCA isoforms have been implicated in breast cancer oncogenesis and formation of breast microcalcifications linked to malignancy (16, 19, 26). Previously, we have evaluated subtype-specific gene expression in breast tumors: whereas high SPCA1 is associated with basal subtypes, SPCA2 was found to be significantly elevated in luminal A/B and HER2+ subtypes (19). To further distinguish between their tumor expression pattern and gain insight into isoform-specific functions, we compared the expression of SPCA1 (ATP2C1) and SPCA2 (ATP2C2) genes in breast cancer patient specimens from TCGA (n = 526) with a “benchmark” gene set, consisting of 18 ESGs and 13 mesenchymal signature genes (MSG), described previously (24). As expected, we observed a sharp dichotomy between ESG and MSG, with the epithelial and mesenchymal genes separating as two tightly distinct and mutually correlated clusters (Fig. 1A). SPCA1 and SPCA2 showed differential clustering with MSG and ESG, respectively, despite their high identity (64%) in amino acid sequence and similar Ca2+ transport characteristics (27, 28). Unexpectedly, we observed a strong coexpression correlation between SPCA2 and E-cadherin (Fig. 1A; Supplementary Fig. S1A, Pearson coefficient 0.44; P = 2.52 × 10−26). In contrast, there was no significant correlation between SPCA1 and E-cadherin expression (Supplementary Fig. S1B).

Figure 1.

SPCA2 is an epithelial marker coexpressed with E-cadherin. A, Hierarchical clustering and heat map of the correlation coefficient matrix of SPCA isoforms with ESG and MSG in breast cancer samples from TCGA (low correlation, blue; high correlation, red; n = 526; ref. 49). SPCA1 and SPCA2 cluster with MSG and ESG, respectively. Note the strong correlation of SPCA2 with CDH1 (red asterisk). B and C, Breast cancer cell lines from CCLE grouped as mesenchymal or epithelial to show Z-scores for SPCA2 (B) and SPCA1 (C) mRNA (24, 50). Note the significant upregulation of SPCA2 and, conversely, significant downregulation of SPCA1 in epithelial-like breast cancer cell lines; ***, P < 0.001, Student t test. D, MCF10A cells were cultured in both sparse and confluent conditions, and DIC images were taken with 10× and 20× objectives on a Zeiss Axio Observer A.1 microscope. E, qPCR analysis of SPCA2, SPCA1, E-cadherin, and N-cadherin expression in cDNA extracted from sparse cells or after 5 or 7 days after confluence; **, P < 0.01 and ***, P < 0.001; two-tailed Welch t test, n = 3. F–H, Box and whisker with scatter plots of the EMT score across 633 breast cancer samples from the TCGA dataset grouped into low-, intermediate-, and high-expression categories for SPCA2 (F), E-cadherin (G), and SPCA1 (H; ***, P < 0.001; ref. 30). Notably, tumors with low SPCA2 or E-cadherin expression had significantly higher EMT score than those with high SPCA2 expression. See Supplementary Fig. S1.

Figure 1.

SPCA2 is an epithelial marker coexpressed with E-cadherin. A, Hierarchical clustering and heat map of the correlation coefficient matrix of SPCA isoforms with ESG and MSG in breast cancer samples from TCGA (low correlation, blue; high correlation, red; n = 526; ref. 49). SPCA1 and SPCA2 cluster with MSG and ESG, respectively. Note the strong correlation of SPCA2 with CDH1 (red asterisk). B and C, Breast cancer cell lines from CCLE grouped as mesenchymal or epithelial to show Z-scores for SPCA2 (B) and SPCA1 (C) mRNA (24, 50). Note the significant upregulation of SPCA2 and, conversely, significant downregulation of SPCA1 in epithelial-like breast cancer cell lines; ***, P < 0.001, Student t test. D, MCF10A cells were cultured in both sparse and confluent conditions, and DIC images were taken with 10× and 20× objectives on a Zeiss Axio Observer A.1 microscope. E, qPCR analysis of SPCA2, SPCA1, E-cadherin, and N-cadherin expression in cDNA extracted from sparse cells or after 5 or 7 days after confluence; **, P < 0.01 and ***, P < 0.001; two-tailed Welch t test, n = 3. F–H, Box and whisker with scatter plots of the EMT score across 633 breast cancer samples from the TCGA dataset grouped into low-, intermediate-, and high-expression categories for SPCA2 (F), E-cadherin (G), and SPCA1 (H; ***, P < 0.001; ref. 30). Notably, tumors with low SPCA2 or E-cadherin expression had significantly higher EMT score than those with high SPCA2 expression. See Supplementary Fig. S1.

Close modal

To corroborate these findings, we evaluated SPCA gene expression in breast cancer cell lines of the Broad Institute's CCLE, previously classified into epithelial and mesenchymal subtypes based on their consensus signatures (24). In agreement with Fig. 1A, there was upregulation (1.89-fold change) of SPCA2 (Fig. 1B) and, conversely, downregulation (0.61-fold change) of SPCA1 (Fig. 1C) in epithelial-like breast cancer cell lines based on Z-score which shows the number of standard deviations above or below the mean (Z-score = 0). To determine if these expression patterns extended to nonmalignant mammary epithelial cells, we examined cell density–related changes in gene expression in sparse and confluent cultures of MCF10A. Unlike cancer cells, sparsely seeded MCF10A cells ceased to proliferate upon reaching confluence (Fig. 1D), due to contact inhibition and mesenchymal-to-epithelial transition (MET). Within 5 days after confluence, we observed upregulation of SPCA2 by 5-fold in MCF10A cultures, whereas SPCA1 levels decreased by 2-fold (Fig. 1E). These changes preceded the so-called “cadherin switch” that accompanies MET, characterized by increase in E-cadherin expression and reciprocal decrease of N-cadherin which we observed at 7 days after confluence (Fig. 1E). These findings were replicated in the breast cancer cell line MDA-MB-231 (Supplementary Fig. S1C–S1F). Similarly, cell density–dependent upregulation of SPCA2, but not SPCA1, has been observed in the human colon cancer line HCT116 (29). This distinctive transcriptional signature of Golgi calcium pumps suggests that the two isoforms play different cellular roles related to EMT–MET transitions.

Tumor EMT scores quantitatively capture the molecular features of cancer cells exhibited during EMT (25). Tumors with high EMT score have high potential for invasion and metastasis. We analyzed 633 breast cancer samples of all subtypes from TCGA to probe the link between EMT score and SPCA2 gene expression (30). Using Z-score of ±0.5 as cutoff for SPCA2 expression, breast cancer samples were stratified into low, intermediate, and high expression categories. Tumors with low SPCA2 expression had significantly higher EMT score than those with high SPCA2 expression (Fig. 1F, P < 0.001), as was seen for E-cadherin expression (Fig. 1G, P < 0.001) but not for SPCA1 expression (Fig. 1H). Consistent with these data, there was negative correlation between tumor EMT score and E-cadherin expression (Supplementary Fig. S1G; corr. −0.25, P < 0.0001) as well as for SPCA2 expression (Supplementary Fig. S1H; corr. −0.18, P = 0.0009), but no significant correlation with SPCA1 expression (Supplementary Fig. S1I). Taken together, the robust correlation in gene expression between SPCA2 and E-cadherin points to a potential EMT-related functional link in breast cancer cells.

SPCA2 regulates E-cadherin biogenesis

Within the secretory pathway, Ca2+ is important for sorting and trafficking of cell surface proteins (31, 32). Therefore, we asked if SPCA2 was required for E-cadherin biogenesis using two complementary approaches: gene knockdown in the high-SPCA2–expressing breast cancer cell line MCF7 and ectopic gene expression in the low-SPCA2–expressing line MDA-MB-231. Previously, we showed high levels of SPCA2 in ER+, luminal subtype MCF7 cell line, which we confirmed in this study (see ahead, Fig. 3A). Targeted knockdown of SPCA2 in MCF7 cells did not alter E-cadherin mRNA (Fig. 2A). However, total E-cadherin protein was significantly reduced (53 ± 7.6% of control; P < 0.05; Fig. 2B and C), suggesting an upstream requirement for SPCA2 in E-cadherin expression. These observations were replicated in MCF10A cells (Supplementary Fig. S2A–S2D). In contrast, there was no significant change in E-cadherin protein expression upon SPCA1 knockdown (Supplementary Fig. S2E and S2F) as previously reported in keratinocytes (30). Furthermore, there was no reduction of EGFR, a surface protein not found at the adherens junction (Supplementary Fig. S2G and S2H), or of zona occludens-1 (ZO-1), a tight junction protein, in shSPCA2-treated cells (Supplementary Fig. S2I and S2J). Thus, the selective defect in E-cadherin biogenesis points to a functional link with SPCA2 and is consistent with the results of the coexpression analysis (Fig. 1).

Figure 2.

Loss of SPCA2 impairs E-cadherin biogenesis. A, Knockdown of SPCA2 in MCF7 cells was confirmed by qPCR. There was no change in E-cadherin transcript; **, P < 0.01, Student t test. B, Western blot of cell lysates probed with antibody against E-cadherin and β-actin. C, E-Cadherin protein expression from densitometric analysis of Western blots, normalized against β-actin; *, P < 0.05, Student t test, n = 3. D, Immunofluorescence of surface E-cadherin in nonpermeabilized MCF7 cells treated with both control and shSPCA2. Note that the antibody recognizes an extracellular epitope. E, Total mean fluorescence was quantified; *, P < 0.05, Student t test, control n = 276, shSPCA2 n = 47. F, Immunofluorescence of E-cadherin in permeabilized MCF7 cells treated with both control and shSPCA2. Note that the antibody recognizes a cytoplasmic epitope (G). Total mean fluorescence from (F) was quantified; ***, P < 0.001, Student t test, control n = 91, shSPCA2 n = 44. H, Immunofluorescence of total E-cadherin in permeabilized MCF7 shSPCA2 cells expressing vector control, or FLAG-tagged SPCA2R (shRNA-resistant SPCA2), with quantification (I); *, P < 0.05 and **, P < 0.01, Student t test, vector n = 34, SPCA2R n = 44. Scale bars are 20 μm. J, Flow cytometry analysis of MCF7 shSPCA2 cells transfected with empty vector or SPCA2R showing percentages of cells in each quadrant (Q1–Q4). K, Number of cells with high E-cadherin fluorescence found in the pink-shaded area, representing 0.78% and 5.07% of total cells in Vector- or SPCA2R-transfected cells, respectively. L, Histogram showing shift in total E-cadherin fluorescence in shSPCA2 cells expressing SPCA2R; *, P < 0.05 and ***, P < 0.001, Student t test, n = 3 for each condition. See Supplementary Fig. S2.

Figure 2.

Loss of SPCA2 impairs E-cadherin biogenesis. A, Knockdown of SPCA2 in MCF7 cells was confirmed by qPCR. There was no change in E-cadherin transcript; **, P < 0.01, Student t test. B, Western blot of cell lysates probed with antibody against E-cadherin and β-actin. C, E-Cadherin protein expression from densitometric analysis of Western blots, normalized against β-actin; *, P < 0.05, Student t test, n = 3. D, Immunofluorescence of surface E-cadherin in nonpermeabilized MCF7 cells treated with both control and shSPCA2. Note that the antibody recognizes an extracellular epitope. E, Total mean fluorescence was quantified; *, P < 0.05, Student t test, control n = 276, shSPCA2 n = 47. F, Immunofluorescence of E-cadherin in permeabilized MCF7 cells treated with both control and shSPCA2. Note that the antibody recognizes a cytoplasmic epitope (G). Total mean fluorescence from (F) was quantified; ***, P < 0.001, Student t test, control n = 91, shSPCA2 n = 44. H, Immunofluorescence of total E-cadherin in permeabilized MCF7 shSPCA2 cells expressing vector control, or FLAG-tagged SPCA2R (shRNA-resistant SPCA2), with quantification (I); *, P < 0.05 and **, P < 0.01, Student t test, vector n = 34, SPCA2R n = 44. Scale bars are 20 μm. J, Flow cytometry analysis of MCF7 shSPCA2 cells transfected with empty vector or SPCA2R showing percentages of cells in each quadrant (Q1–Q4). K, Number of cells with high E-cadherin fluorescence found in the pink-shaded area, representing 0.78% and 5.07% of total cells in Vector- or SPCA2R-transfected cells, respectively. L, Histogram showing shift in total E-cadherin fluorescence in shSPCA2 cells expressing SPCA2R; *, P < 0.05 and ***, P < 0.001, Student t test, n = 3 for each condition. See Supplementary Fig. S2.

Close modal

Treatment of nonpermeabilized MCF7 cells with antibody directed against an external epitope of E-cadherin revealed a significant reduction in surface expression of E-cadherin in SPCA2-depleted cells (Fig. 2D and E). Total E-cadherin level in SPCA2-depleted MCF7 cells, as determined by quantitative immunofluorescence microscopy (Fig. 2F and G), was also diminished but could be largely restored by transfection of a FLAG-tagged and shRNA-resistant construct SPCA2R, relative to vector-transfected or untreated controls (Fig. 2H and I). This was confirmed by flow cytometry analysis: SPCA2R increased E-cadherin expression in knockdown cells relative to the empty vector control, as seen by forward scatter in the upper right quadrant (Fig. 2J), increased number of cells with high E-cadherin (Fig. 2K), and shift in the histogram of cell population with E-cadherin fluorescence (Fig. 2L).

Because loss of SPCA2 was sufficient to downregulate expression of E-cadherin in MCF7, an epithelial-subtype tumor cell line, we asked if ectopic expression of SPCA2 in mesenchymal-subtype tumor cells conferred a reciprocal increase in E-cadherin expression. Evaluation of a panel of established breast cancer cell lines revealed very low SPCA2 transcript levels in the TNBC cell lines MDA-MB-231 and Hs578T relative to MCF10A (Fig. 3A); in contrast, SPCA1 was expressed at levels similar to MCF10A or significantly higher in these cell lines (Supplementary Fig. S3). Lentiviral-mediated transfection with SPCA2R in the TNBC lines increased transcripts to levels approaching endogenous SPCA2 expression in MCF7 cells, although this was cell-type specific (Fig. 3A, see also Supplementary Fig. S6A and S6G). We confirmed ectopic expression of FLAG-tagged SPCA2R in MDA-MB-231 cells by Western blotting and localization to intracellular vesicular compartments by confocal microscopy (Fig. 3B). By following doxycycline-induced E-cadherin protein in Tet-Ecad-MDA-MB-231 cells, we focused only on posttranslational effects of SPCA2. E-cadherin levels were significantly increased by SPCA2R relative to vector control, as shown by both confocal microscopy and FACS analysis (Fig. 3C and D) of immunofluorescent cells. Although the bulk of E-cadherin remained within the cell, suggesting that additional factors may be necessary to restore efficient surface trafficking of E-cadherin in TNBC cells, some E-cadherin reached the cell boundary as seen in the enlarged images (arrow heads, Fig. 3C). Together, these data point to a hitherto unrecognized, posttranslational role for SPCA2 in E-cadherin expression.

Figure 3.

SPCA2 increases E-cadherin expression in TNBC cells. A, A panel of breast cancer cell lines classified as luminal, HER2+, or TNBC subtypes as indicated was evaluated for SPCA2 expression by qPCR. Results are normalized to MCF-10A. SPCA2 transcripts were lowest in two TNBC lines, MDA-MB-231 and Hs578T, and could be significantly increased by lentiviral-mediated transfection of SPCA2R. B, SPCA2R expression in transfected MDA-MB-231 cells was confirmed by Western blotting (left) and confocal immunofluorescence microscopy (right) using anti-FLAG antibody. C, MDA-MB-231 Tet-E-cadherin cells expressing either vector or SPCA2R were treated with 2 μg/mL of doxycycline, stained for E-cadherin and imaged using confocal microscopy. D, Normalized fluorescence values for each condition (top) and histograms of cells separated by flow cytometry (bottom); *, P < 0.05, Student t test, vector n = 3, SPCA2R n = 3. Flow cytometry histogram analysis of three biological replicates of each condition. Vector n = 9,468 cells, SPCA2R n = 9,258 cells. See Supplementary Fig. S3.

Figure 3.

SPCA2 increases E-cadherin expression in TNBC cells. A, A panel of breast cancer cell lines classified as luminal, HER2+, or TNBC subtypes as indicated was evaluated for SPCA2 expression by qPCR. Results are normalized to MCF-10A. SPCA2 transcripts were lowest in two TNBC lines, MDA-MB-231 and Hs578T, and could be significantly increased by lentiviral-mediated transfection of SPCA2R. B, SPCA2R expression in transfected MDA-MB-231 cells was confirmed by Western blotting (left) and confocal immunofluorescence microscopy (right) using anti-FLAG antibody. C, MDA-MB-231 Tet-E-cadherin cells expressing either vector or SPCA2R were treated with 2 μg/mL of doxycycline, stained for E-cadherin and imaged using confocal microscopy. D, Normalized fluorescence values for each condition (top) and histograms of cells separated by flow cytometry (bottom); *, P < 0.05, Student t test, vector n = 3, SPCA2R n = 3. Flow cytometry histogram analysis of three biological replicates of each condition. Vector n = 9,468 cells, SPCA2R n = 9,258 cells. See Supplementary Fig. S3.

Close modal

SPCA2 phenocopies E-cadherin in tumorsphere formation

Next, we asked if SPCA2 regulated cellular phenotypes downstream from E-cadherin. The ability of breast cancer cells to form free-floating spheroids known as tumorspheres in suspension culture has been shown to depend on surface expression of E-cadherin (21, 33). Thus, knockdown of E-cadherin in MCF7 abolished tumorsphere formation (21). Therefore, we asked if SPCA2 depletion phenocopied loss of E-cadherin in tumorsphere formation.

Depletion of SPCA2 in MCF7 cells (confirmed in Fig. 2A) resulted in nearly complete loss of tumorspheres, strongly attenuating both number and size relative to scrambled shRNA control (Fig. 4A). These results confirm and extend our previous finding that SPCA2 knockdown significantly inhibits formation of mammospheres derived from normal mammary epithelial cells embedded in extracellular matrix (20). In contrast, SPCA1 knockdown had no significant effect on tumorsphere formation (Fig. 4A). Surface expression of E-cadherin and tumorsphere formation could be rescued by reintroduction of SPCA2R in shSPCA2-treated MCF7 cells (Fig. 4B–D). Thus, tumorsphere formation closely parallels expression of both SPCA2 and E-cadherin (21).

Figure 4.

SPCA2 is required for tumorsphere formation. MCF-7 cells treated with scramble shRNA (control), shSPCA1, or shSPCA2 constructs were cultured to form tumorspheres. A, Representative images of control and shSPCA2 cells are shown in the top and bottom plots, respectively. Scale bars are 50 μm. The numbers of tumorspheres (top plot) and their diameters (bottom plots) were quantified; **, P < 0.01 and ***, P < 0.001, Student t test, n = 3. B, Three representative images of MCF7 cells (untreated control), or shSPCA2-transfected cells rescued with empty vector (GFP) or SPCA2R, and cultured as tumorspheres. Scale bars are 50 μm. C, Immunofluorescence image showing that FLAG-tagged SPCA2R (green) effectively rescues surface E-cadherin expression (red) in permeabilized MCF7 shSPCA2 cells grown in 2D culture. D, Tumorsphere area from the experiment in B was quantified; ***, P < 0.001, Student t test; for tumorsphere area, n = 31 for vector, n = 45 for SPCA2R. See Supplementary Fig. S4.

Figure 4.

SPCA2 is required for tumorsphere formation. MCF-7 cells treated with scramble shRNA (control), shSPCA1, or shSPCA2 constructs were cultured to form tumorspheres. A, Representative images of control and shSPCA2 cells are shown in the top and bottom plots, respectively. Scale bars are 50 μm. The numbers of tumorspheres (top plot) and their diameters (bottom plots) were quantified; **, P < 0.01 and ***, P < 0.001, Student t test, n = 3. B, Three representative images of MCF7 cells (untreated control), or shSPCA2-transfected cells rescued with empty vector (GFP) or SPCA2R, and cultured as tumorspheres. Scale bars are 50 μm. C, Immunofluorescence image showing that FLAG-tagged SPCA2R (green) effectively rescues surface E-cadherin expression (red) in permeabilized MCF7 shSPCA2 cells grown in 2D culture. D, Tumorsphere area from the experiment in B was quantified; ***, P < 0.001, Student t test; for tumorsphere area, n = 31 for vector, n = 45 for SPCA2R. See Supplementary Fig. S4.

Close modal

Tumorsphere formation is commonly used to indicate the presence of cancer stem cells. Interestingly, despite the inability to form tumorspheres, depletion or overexpression of SPCA2 in MCF7 and MDA-MB-231, respectively, resulted in corresponding increase and decrease in the expression of genes associated with cancer stem cells, including OCT4, NANOG, and SOX2 (Supplementary Fig. S4). These findings suggested that SPCA2 is an upstream regulator of stemness and post EMT phenotypes.

SPCA2 is an upstream regulator of YAP and Hippo pathway signaling

E-cadherin is required for cell-adhesion and contact inhibition of proliferation (34). By activating the Hippo tumor-suppressor signaling pathway, E-cadherin controls phosphorylation, inactivation, and nuclear exclusion of YAP, resulting in inhibition of cell proliferation in normal human mammary epithelial and breast cancer cells (Fig. 5A; refs. 6, 8). Given the early postconfluency induction of SPCA2 gene expression ahead of the cadherin switch (Fig. 1E), together with a role for SPCA2 in efficient surface expression of E-cadherin (Fig. 2), we postulated that SPCA2 is an upstream regulator of the Hippo pathway. We tested this hypothesis both by knockdown of SPCA2 in MCF7 cells and by introduction of SPCA2 in MDA-MB-231 cells. First, MCF7 cells were cultured under sparse and confluent conditions and immunostained for YAP localization. As previously described for control cells, nuclear occupancy of YAP in sparsely growing cells largely reverted to the cytoplasm upon reaching confluency (Fig. 5B, top plot; ref. 6). In contrast, in cells depleted of SPCA2, YAP remained predominantly in the nucleus under both sparse and confluent culture conditions (Fig. 5B, bottom plot; Fig. 5C–E). Again, loss of SPCA2 phenocopies loss of E-cadherin and downstream components of the Hippo signaling pathway (6), consistent with a functional role for SPCA2 in E-cadherin biogenesis.

Figure 5.

SPCA2 functions upstream of Hippo-YAP signaling. A, Schematic showing the role of Hippo-YAP signaling in control of EMT. Cell–cell contacts mediated by E-cadherin maintain nuclear exclusion of YAP, suppressing expression of EMT genes. B, Immunofluorescence staining of YAP in MCF7 cells treated with scramble (control) or shSPCA2 and cultured under sparse or confluent conditions. Note the distribution of YAP localization between nuclei (stained with DAPI, blue) and cytoplasm. C, Subcellular YAP distribution in MCF7 cells grown to confluence and stained as in B. D, Colocalization of YAP with DAPI staining in confluent MCF7 cells using Mander's coefficient; **, P < 0.01, Student t test. E, Schematic of YAP localization in response to cell density, EGF treatment, and SPCA2 knockdown. F, Live cell Ca2+ imaging traces using Fura2-AM–treated MDA-MB-231 cells with or without SPCA2R (Control n = 44 cells, SPCA2R n = 30 cells). Baseline readings in calcium-free conditions were established for 50 cycles (72 seconds), and then cells were recorded under 2 mmol/L Ca2+ conditions for 250 cycles (6 minutes and 20 seconds). Baseline ratios (G) and the area under the curve (H) were plotted between both Control (n = 44) and SPCA2R (n = 30) samples; ****, P < 0.0001, Student t test. Images representing time points at 0 and 450 seconds of Control cells and cells expressing SPCA2R (I) showing calcium influx. J, Western blot of cell lysates showing phospho-YAP and total YAP protein in Ca2+-free media and following addition of Ca2+ (2 mmol/L) or thapsigargin (1 μmol/L). K, Quantification of pYAP after normalization to GAPDH levels. Fold change (n = 3 independent replicates; Student t test) is shown. See Supplementary Movies S1 and S2. ns, not significant.

Figure 5.

SPCA2 functions upstream of Hippo-YAP signaling. A, Schematic showing the role of Hippo-YAP signaling in control of EMT. Cell–cell contacts mediated by E-cadherin maintain nuclear exclusion of YAP, suppressing expression of EMT genes. B, Immunofluorescence staining of YAP in MCF7 cells treated with scramble (control) or shSPCA2 and cultured under sparse or confluent conditions. Note the distribution of YAP localization between nuclei (stained with DAPI, blue) and cytoplasm. C, Subcellular YAP distribution in MCF7 cells grown to confluence and stained as in B. D, Colocalization of YAP with DAPI staining in confluent MCF7 cells using Mander's coefficient; **, P < 0.01, Student t test. E, Schematic of YAP localization in response to cell density, EGF treatment, and SPCA2 knockdown. F, Live cell Ca2+ imaging traces using Fura2-AM–treated MDA-MB-231 cells with or without SPCA2R (Control n = 44 cells, SPCA2R n = 30 cells). Baseline readings in calcium-free conditions were established for 50 cycles (72 seconds), and then cells were recorded under 2 mmol/L Ca2+ conditions for 250 cycles (6 minutes and 20 seconds). Baseline ratios (G) and the area under the curve (H) were plotted between both Control (n = 44) and SPCA2R (n = 30) samples; ****, P < 0.0001, Student t test. Images representing time points at 0 and 450 seconds of Control cells and cells expressing SPCA2R (I) showing calcium influx. J, Western blot of cell lysates showing phospho-YAP and total YAP protein in Ca2+-free media and following addition of Ca2+ (2 mmol/L) or thapsigargin (1 μmol/L). K, Quantification of pYAP after normalization to GAPDH levels. Fold change (n = 3 independent replicates; Student t test) is shown. See Supplementary Movies S1 and S2. ns, not significant.

Close modal

One surprising finding was the selective requirement for SPCA2, but not SPCA1, for E-cadherin expression. Previously, we showed an isoform-selective role for SPCA2 in activating the Ca2+ channel Orai1 to elicit robust Ca2+ influx in MCF7 cells (16). Recent findings in a glioblastoma model have revealed that Ca2+ is a crucial intracellular cue that regulates the Hippo pathway by activation of a protein kinase C beta II that leads to YAP phosphorylation and nuclear exclusion via upstream kinases Lats1/2 (35). Here, we used ectopic SPCA2 expression in MDA-MB-231 cells to elevate resting levels of Ca2+ (Fig. 5F and G). Addition of Ca2+ (2 mmol/L) to cells briefly incubated in Ca2+-free medium elicited immediate and sustained SICE that was significantly larger in SPCA2-transfected cells (Fig. 5F, H, and I; Supplementary Movies S1 and S2). In parallel, we observed increased YAP phosphorylation under baseline conditions in SPCA2-transfected MDA-MB-231 cells, relative to the vector control (Fig. 5J and K). YAP phosphorylation further increased following Ca2+ influx from the media (addition of 2 mmol/L Ca2+) or internal stores (1 μmol/L thapsigargin treatment; Fig. 5J and K), as expected. Thus, store-independent Ca2+ influx elicited by SPCA2 expression increases YAP phosphorylation associated with nuclear exclusion, extending the emerging observations from glioblastoma cells (35), and maintains Hippo pathway signaling downstream of E-cadherin in breast cancer cells.

SPCA2 antagonizes EMT

Nuclear YAP and the related coactivator TAZ interact with the TEAD family of transcriptional activators to elicit EMT by turning on mesenchymal markers such as N-cadherin and vimentin and promote malignant transformation (36). Furthermore, YAP/TAZ interact with multiple members of the EMT-associated family of transcription factors, including SNAI1, SLUG/SNAI2, and ZEB1 (37). A proteomic analysis of MCF7 cells showed that E-cadherin knockdown resulted in elevated expression of the EMT-associated transcription factors SLUG/SNAI2 and ZEB1, downstream of YAP (38). Interestingly, expression of YAP/TAZ was negatively correlated with SPCA2 expression in patient breast cancer tissue from all cancer subtypes (Supplementary Fig. S5A and S5B; P = 3.23 × 10−7, n = 295; data from GSE31448 microarray dataset; ref. 39). Specifically, high YAP/TAZ expression is correlated with SPCA2 downregulation in basal-like breast cancer (Supplementary Fig. S5A). No correlation was observed between YAP/TAZ expression and that of SPCA1 (Supplementary Fig. S5C; P = 0.178, n = 295). To test the hypothesis that SPCA2 regulated EMT–MET transitions (Fig. 6A), we elicited EMT in MCF7 cells by treatment with EGF. Although SPCA1 transcript levels remained unchanged, SPCA2 levels decreased substantially upon EGF treatment (Fig. 6B). Confirming the induction of EMT, cells treated with EGF showed enhanced expression of most EMT-associated genes including ZEB1, N-cadherin, SLUG/SNAI2, and fibronectin, as expected (Fig. 6C–G, blue bars). Thus, SPCA2 gene expression correlates with EMT.

Figure 6.

SPCA2 regulates gene expression in EMT. A, EMT–MET transitions may be regulated by SPCA2 as shown. B, Effect of EGF (20 ng/mL) on mRNA levels of SPCA1 and SPCA2 in MCF7 cells determined by qPCR; *, P < 0.05, Student t test. Gene expression changes accompanying EMT–MET transitions were evaluated following knockdown of SPCA2 in MCF7 (C–G) or upon expression of SPCA2R in MDA-MB-231 cells (H–M). MCF7 cells (scramble control or SPCA2 knockdown) were treated with 20 ng/mL EGF where indicated, and transcript levels of the following mesenchymal markers were determined by qPCR: Zeb1 (C), N-cadherin (D), SNAI2 (E), fibronectin (F), and vimentin (G); *, P < 0.05 and **, P < 0.01, Student t test. Effect of SPCA2R expression in MDA-MB-231 cells, relative to vector control, on a panel of epithelial (H) and mesenchymal (I–M) gene markers was quantified by qPCR; *, P < 0.05 and **, P < 0.01, Student t test. N, Representative confocal microscope images showing immunofluorescence staining of vimentin in MDA-MB-231 cells expressing SPCA2R, relative to vector control (40× magnification). O, Vimentin mean fluorescence intensity (MFI) was quantified by ImageJ software from 6 images each condition. Each data point represents one image; **, P < 0.01, Student t test, n = 95 for GFP control and n = 99 for SPCA2 overexpression. P, Vimentin protein expression was detected using Western blotting, and GAPDH was used as a loading control. For each condition, 30 μg of protein lysates was loaded. See Supplementary Figs. S5 and S6. ns, not significant.

Figure 6.

SPCA2 regulates gene expression in EMT. A, EMT–MET transitions may be regulated by SPCA2 as shown. B, Effect of EGF (20 ng/mL) on mRNA levels of SPCA1 and SPCA2 in MCF7 cells determined by qPCR; *, P < 0.05, Student t test. Gene expression changes accompanying EMT–MET transitions were evaluated following knockdown of SPCA2 in MCF7 (C–G) or upon expression of SPCA2R in MDA-MB-231 cells (H–M). MCF7 cells (scramble control or SPCA2 knockdown) were treated with 20 ng/mL EGF where indicated, and transcript levels of the following mesenchymal markers were determined by qPCR: Zeb1 (C), N-cadherin (D), SNAI2 (E), fibronectin (F), and vimentin (G); *, P < 0.05 and **, P < 0.01, Student t test. Effect of SPCA2R expression in MDA-MB-231 cells, relative to vector control, on a panel of epithelial (H) and mesenchymal (I–M) gene markers was quantified by qPCR; *, P < 0.05 and **, P < 0.01, Student t test. N, Representative confocal microscope images showing immunofluorescence staining of vimentin in MDA-MB-231 cells expressing SPCA2R, relative to vector control (40× magnification). O, Vimentin mean fluorescence intensity (MFI) was quantified by ImageJ software from 6 images each condition. Each data point represents one image; **, P < 0.01, Student t test, n = 95 for GFP control and n = 99 for SPCA2 overexpression. P, Vimentin protein expression was detected using Western blotting, and GAPDH was used as a loading control. For each condition, 30 μg of protein lysates was loaded. See Supplementary Figs. S5 and S6. ns, not significant.

Close modal

Next, we sought to determine if loss of SPCA2 is a driver of EMT, and conversely, if SPCA2 promotes MET (Fig. 6A). Strikingly, SPCA2 knockdown was sufficient to induce transcription of EMT-associated genes even in the absence of EGF, with further elevation of transcript in most genes upon EGF treatment (Fig. 6C–G; gray bars). Introduction of SPCA2R in MDA-MB-231 cells had no effect on transcript levels of E-cadherin (Fig. 6H) consistent with a posttranslational role for SPCA2 in E-cadherin expression. However, SPCA2R suppressed transcription of a range of mesenchymal markers, including N-cadherin, SNAI1, SNAI2/SLUG, Vimentin, and ZEB1 (Fig. 6I–M). Similar results were observed upon ectopic expression of SPCA2R in MCF10A (Supplementary Fig. S6A–S6F) and Hs578T (Supplementary Fig. S6G–S6L). We confirmed SPCA2R-mediated loss in expression of vimentin in MDA-MB-231 cells by confocal microscopy (Fig. 6N and O) and by Western blotting (Fig. 6P). These results demonstrate a cause and effect relationship between SPCA2 and EMT, pointing to a novel role for SPCA2 as driver rather than a passenger in the gene expression changes that accompany EMT–MET programs.

Finally, we investigated if gene expression changes elicited by SPCA2R in MDA-MB-231 cells were downstream of the Hippo signaling pathway. Addition of YAP inhibitor Verteporfin (40) to SPCA2R-transfected MDA-MB-231 cells rescued the expression of SNAI1, Vimentin, and ZEB1, key architects of EMT (Supplementary Fig. S6M–S6O). There was no significant rescue in the expression of CDH2 and SNAI2 (Supplementary Fig. S6P and S6Q), indicating that additional or redundant pathways may be involved. These results show that the coactivator protein YAP plays a role in SPCA2-mediated EMT.

SPCA2 suppresses cancer cell migration and tumor metastasis

Cell motility and migration underlie metastatic progression and are key features of EMT. We assessed the effects of SPCA2R expression on migration and invasion in the highly metastatic MDA-MB-231 cells, which exhibit very low endogenous levels of SPCA2 (Fig. 3A). As shown in Fig. 7A and B, wound-closure assays revealed that SPCA2R inhibited in vitro migration of MDA-MB-231 at all time points tested (P < 0.05). Similarly, expression of SPCA2R inhibited cell migration in Boyden chambers by 40% (Supplementary Fig. S7A–S7C).

Figure 7.

SPCA2 suppresses migration in vitro and metastasis in vivo. A, Representative images of wound healing in MDA-MB-231 at the indicated times in control and SPCA2R-transfected cells (2.5× magnification). B, Quantification of cell migration expressed as percentage of control. SPCA2 significantly reduced migration at 8, 16, and 24 hours; *, P < 0.05; **, P < 0.01; ***, P < 0.01; Student t test, n = 3 for each condition. C, Bioluminescent images of NSG mice following engraftment with MDA-MB-231 Luc + GFP and MDA-MB-231 Luc + SPCA2R at indicated times and luminescence scales. D and E, Quantification of bioluminescence in luciferase-labeled MDA-MB-231 Luc. Tumor growth and metastasis at week 5 were represented as mean total flux (p/sec) ± SEM. SPCA2 significantly reduced metastasis compared with control group; **, P < 0.01; Student t test, n = 6 for each condition. See Supplementary Fig. S7.

Figure 7.

SPCA2 suppresses migration in vitro and metastasis in vivo. A, Representative images of wound healing in MDA-MB-231 at the indicated times in control and SPCA2R-transfected cells (2.5× magnification). B, Quantification of cell migration expressed as percentage of control. SPCA2 significantly reduced migration at 8, 16, and 24 hours; *, P < 0.05; **, P < 0.01; ***, P < 0.01; Student t test, n = 3 for each condition. C, Bioluminescent images of NSG mice following engraftment with MDA-MB-231 Luc + GFP and MDA-MB-231 Luc + SPCA2R at indicated times and luminescence scales. D and E, Quantification of bioluminescence in luciferase-labeled MDA-MB-231 Luc. Tumor growth and metastasis at week 5 were represented as mean total flux (p/sec) ± SEM. SPCA2 significantly reduced metastasis compared with control group; **, P < 0.01; Student t test, n = 6 for each condition. See Supplementary Fig. S7.

Close modal

Next, we investigated if SPCA2 has antimetastatic effect in vivo using bioluminescent model MDA-MB-231 Luc. We injected MDA-MB-231 Luc + GFP and MDA-MB-231 Luc + SPCA2R cells in the mammary fat pad of NSG mice and monitored metastasis using whole-body bioluminescence imaging at weekly intervals (Fig. 7C; Supplementary Fig. S7D). Although there was no significant difference in the primary tumor size (Fig. 7D), tumors from cells transfected with SPCA2R showed significantly less metastasis compared with the control group at week 5 (Fig. 7E; P < 0.01) with 3 of 6 mice showing a lack of lung metastasis. We suggest that the SPCA2-mediated suppression of mesenchymal genes demonstrated in vitro (Fig. 6; Supplementary Fig. S6) confers inhibitory effects on tumor migration and metastasis.

We show an isoform-specific posttranslational role for SPCA2 in E-cadherin biogenesis that has functional consequences on cell adhesion and the EMT in the development of breast cancer metastasis. The specific molecular details of this mechanism will require further investigation and may involve a chaperone-like function of SPCA2 as previously described for Orai1 (20), isoform-specific differences in ion transport, Ca2+/Mn2+ selectivity or regulation that could alter the secretory pathway ionic milieu (41), or a new requirement for cytoplasmic Ca2+ signaling in E-cadherin trafficking or stability. It is unclear whether the secretory pathway plays a role in supplying Ca2+ to the extracellular/lumenal cadherin repeats known to mediate their structural rigidity and extended conformation (42). Interestingly, defective desmosomal adhesion is the underlying cause in Hailey–Hailey disease, an ulcerative skin disorder resulting from loss of one functional copy of the housekeeping isoform SPCA1 (43). Biogenesis of desmosomal components, including cadherins (desmoglein and desmocollin) and desmoplakins, was delayed upon knockdown of SPCA1 in keratinocytes (30). In light of our new data showing that loss of SPCA2 leads to defective biogenesis of E-cadherin, a component of the adherens junction, we propose that both secretory pathway Ca2+-ATPase isoforms play similar but nonredundant roles in the establishment of cell adhesion.

Given the developmental and pathologic implications of the EMT program, intensive efforts have been directed toward understanding the underlying molecular mechanisms and regulatory networks. Downregulation of E-cadherin is a defining mechanism in EMT. Most studies in the literature have focused primarily on silencing mutations in E-cadherin and transcriptional repression (3). However, a significant percentage of invasive tumors have normal E-cadherin gene and transcription, which may point to posttranslational downregulation of E-cadherin, although much less is known about these mechanisms (44). Here, we show that loss of SPCA2 essentially phenocopies loss of E-cadherin in tumorsphere formation, Hippo pathway signaling, and in gene regulation and tumor phenotypes associated with EMT. Previously, dysregulation of membrane trafficking of E-cadherin has been implicated in EMT associated with viral oncogenesis, such as Rous sarcoma and Kaposi's sarcoma (45, 46). This study showing SPCA2-mediated regulation of E-cadherin expression offers a new mechanism to regulate adherens junction stability in nonviral oncogenesis.

YAP/TAZ are essential for initiation and growth of most solid tumors, and their activation is critical for cancer stem cell attributes and metastasis (47). Importantly, YAP is a sensor of the structural and mechanical cues affecting tumor microenvironment and mediates contact inhibition downstream from E-cadherin (4, 8). The Hippo/YAP pathway presents a therapeutic vulnerability of cancer cells that can be exploited. Thus, our finding that SPCA2 is an upstream regulator of this pathway is an important advance that provides a mechanistic explanation for the role of SPCA2 in EMT–MET transitions via downregulation of E-cadherin. Furthermore, SICE mediated by SPCA2 demonstrates, for the first time, a role for Ca2+ in YAP phosphorylation in breast cancer cells that complements recent observations on YAP phosphorylation by store-operated Ca2+ entry in glioblastoma cells (35) and provides additional avenues for therapeutic intervention.

Downregulation of SPCA2 is associated with basal and claudin-low breast cancer subtypes (19) that are characterized by hormone receptor negativity, and a high degree of invasion and metastasis. Therefore, low SPCA2 may be a useful prognostic marker of the malignant invasive phenotype. In contrast, high SPCA2 is characteristic of luminal A/B and HER2+ subtypes where it contributes to the formation of breast microcalcifications associated with tumorigenesis (19). We propose a model to summarize the role of SPCA2 in breast cancer. During the early stages of breast oncogenesis, proliferation and dysplasia of epithelial cells result in carcinoma in situ. Tumor cells in this stage express high levels of SPCA2, elevated cytoplasmic Ca2+, and rapid cell proliferation (16). Biogenesis and trafficking of E-cadherin is normal, cell–cell contacts are maintained, and EMT genes are repressed. Induction of EMT in a subset of breast cancer cells is associated with downregulation of SPCA2 and other ESG, resulting in disrupted E-cadherin trafficking and enhanced mesenchymal gene expression that facilitates intravasation of tumor cells and escapes into the blood stream. Subsequently, extravasation of a subset of these breast cancer cells from the blood stream into tissues could be followed by MET, which reinstates expression of SPCA2 and other epithelial genes such as E-cadherin (48) leading to secondary tumors at metastatic sites.

No potential conflicts of interest were disclosed.

Conception and design: D.K. Dang, M.R. Makena, J.P. Llongueras, R. Rao

Development of methodology: D.K. Dang, M.R. Makena, J.P. Llongueras, R. Rao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.K. Dang, M.R. Makena, M. Ko, M. Bandral

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.K. Dang, M.R. Makena, J.P. Llongueras, H. Prasad, M. Ko, R. Rao

Writing, review, and/or revision of the manuscript: D.K. Dang, M.R. Makena, H. Prasad, R. Rao

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.K. Dang, M.R. Makena, M. Bandral, R. Rao

Study supervision: M.R. Makena, R. Rao

This work was supported by a grant from the NIH (DK103078) to R. Rao. H. Prasad was a Fulbright Fellow supported by the International Fulbright Science and Technology Award, J.P. Llongueras was supported in part by the NIH PREP scholar program at Johns Hopkins, and M. Bandral was supported by the Khorana Program for Scholars.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Kalluri
R
,
Weinberg
RA
. 
The basics of epithelial-mesenchymal transition
.
J Clin Invest
2009
;
119
:
1420
8
.
2.
Vega
S
,
Morales
AV
,
Ocaña
OH
,
Valdés
F
,
Fabregat
I
,
Nieto
MA
. 
Snail blocks the cell cycle and confers resistance to cell death
.
Genes Dev
2004
;
18
:
1131
43
.
3.
Ashaie
MA
,
Chowdhury
EH
. 
Cadherins: the superfamily critically involved in breast cancer
.
Curr Pharm Des
2016
;
22
:
616
38
.
4.
Mendonsa
AM
,
Na
TY
,
Gumbiner
BM
. 
E-cadherin in contact inhibition and cancer
.
Oncogene
2018
;
37
:
4769
80
.
5.
Kovacs
EM
,
Ali
RG
,
McCormack
AJ
,
Yap
AS
. 
E-cadherin homophilic ligation directly signals through Rac and phosphatidylinositol 3-kinase to regulate adhesive contacts
.
J Biol Chem
2002
;
277
:
6708
18
.
6.
Kim
NG
,
Koh
E
,
Chen
X
,
Gumbiner
BM
. 
E-cadherin mediates contact inhibition of proliferation through Hippo signaling-pathway components
.
Proc Natl Acad Sci U S A
2011
;
108
:
11930
5
.
7.
Gumbiner
BM
. 
Regulation of cadherin-mediated adhesion in morphogenesis
.
Nat Rev Mol Cell Biol
2005
;
6
:
622
34
.
8.
Gumbiner
BM
,
Kim
NG
. 
The Hippo-YAP signaling pathway and contact inhibition of growth
.
J Cell Sci
2014
;
127
(
Pt 4
):
709
17
.
9.
Janse van Rensburg
HJ
,
Yang
X
. 
The roles of the Hippo pathway in cancer metastasis
.
Cell Signal
2016
;
28
:
1761
72
.
10.
Dubois
C
,
Vanden Abeele
F
,
Lehen'kyi
V
,
Gkika
D
,
Guarmit
B
,
Lepage
G
, et al
Remodeling of channel-forming ORAI proteins determines an oncogenic switch in prostate cancer
.
Cancer Cell
2014
;
26
:
19
32
.
11.
Janda
E
,
Nevolo
M
,
Lehmann
K
,
Downward
J
,
Beug
H
,
Grieco
M
. 
Raf plus TGFbeta-dependent EMT is initiated by endocytosis and lysosomal degradation of E-cadherin
.
Oncogene
2006
;
25
:
7117
30
.
12.
Dang
D
,
Rao
R
. 
Calcium-ATPases: gene disorders and dysregulation in cancer
.
Biochim Biophys Acta
2016
;
1863
(
6 Pt B
):
1344
50
.
13.
Kienzle
C
,
von Blume
J
. 
Secretory cargo sorting at the trans-Golgi network
.
Trends Cell Biol
2014
;
24
:
584
93
.
14.
Cross
BM
,
Breitwieser
GE
,
Reinhardt
TA
,
Rao
R
. 
Cellular calcium dynamics in lactation and breast cancer: from physiology to pathology
.
Am J Physiol Cell Physiol
2014
;
306
:
C515
26
.
15.
Pestov
NB
,
Dmitriev
RI
,
Kostina
MB
,
Korneenko
TV
,
Shakhparonov
MI
,
Modyanov
NN
, et al
Structural evolution and tissue-specific expression of tetrapod-specific second isoform of secretory pathway Ca2+-ATPase
.
Biochem Biophys Res Commun
2012
;
417
:
1298
303
.
16.
Feng
M
,
Grice
DM
,
Faddy
HM
,
Nguyen
N
,
Leitch
S
,
Wang
Y
, et al
Store-independent activation of Orai1 by SPCA2 in mammary tumors
.
Cell
2010
;
143
:
84
98
.
17.
Smaardijk
S
,
Chen
J
,
Wuytack
F
,
Vangheluwe
P
. 
SPCA2 couples Ca(2+) influx via Orai1 to Ca(2+) uptake into the Golgi/secretory pathway
.
Tissue Cell
2017
;
49
(
2 Pt A
):
141
9
.
18.
Cerami
E
,
Gao
J
,
Dogrusoz
U
,
Gross
BE
,
Sumer
SO
,
Aksoy
BA
, et al
The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data
.
Cancer Discov
2012
;
2
:
401
4
.
19.
Dang
D
,
Prasad
H
,
Rao
R
. 
Secretory pathway Ca(2+) -ATPases promote in vitro microcalcifications in breast cancer cells
.
Mol Carcinog
2017
;
56
:
2474
85
.
20.
Cross
BM
,
Hack
A
,
Reinhardt
TA
,
Rao
R
. 
SPCA2 regulates Orai1 trafficking and store independent Ca2+ entry in a model of lactation
.
PLoS One
2013
;
8
:
e67348
.
21.
Manuel Iglesias
J
,
Beloqui
I
,
Garcia-Garcia
F
,
Leis
O
,
Vazquez-Martin
A
,
Eguiara
A
, et al
Mammosphere formation in breast carcinoma cell lines depends upon expression of E-cadherin
.
PLoS One
2013
;
8
:
e77281
.
22.
Sun
Y
,
Cui
X
,
Wang
J
,
Wu
S
,
Bai
Y
,
Wang
Y
, et al
Stromal interaction molecule 1 (STIM1) silencing inhibits tumor growth and promotes cell cycle arrest and apoptosis in hypopharyngeal carcinoma
.
Med Oncol
2015
;
32
:
150
.
23.
Xu
L
,
Deng
X
. 
Suppression of cancer cell migration and invasion by protein phosphatase 2A through dephosphorylation of mu- and m-calpains
.
J Biol Chem
2006
;
281
:
35567
75
.
24.
Kohn
KW
,
Zeeberg
BM
,
Reinhold
WC
,
Pommier
Y
. 
Gene expression correlations in human cancer cell lines define molecular interaction networks for epithelial phenotype
.
PLoS One
2014
;
9
:
e99269
.
25.
Tan
TZ
,
Miow
QH
,
Miki
Y
,
Noda
T
,
Mori
S
,
Huang
RY
, et al
Epithelial-mesenchymal transition spectrum quantification and its efficacy in deciphering survival and drug responses of cancer patients
.
EMBO Mol Med
2014
;
6
:
1279
93
.
26.
Grice
DM
,
Vetter
I
,
Faddy
HM
,
Kenny
PA
,
Roberts-Thomson
SJ
,
Monteith
GR
. 
Golgi calcium pump secretory pathway calcium ATPase 1 (SPCA1) is a key regulator of insulin-like growth factor receptor (IGF1R) processing in the basal-like breast cancer cell line MDA-MB-231
.
J Biol Chem
2010
;
285
:
37458
66
.
27.
Xiang
M
,
Mohamalawari
D
,
Rao
R
. 
A novel isoform of the secretory pathway Ca2+,Mn(2+)-ATPase, hSPCA2, has unusual properties and is expressed in the brain
.
J Biol Chem
2005
;
280
:
11608
14
.
28.
Vanoevelen
J
,
Dode
L
,
Van Baelen
K
,
Fairclough
RJ
,
Missiaen
L
,
Raeymaekers
L
, et al
The secretory pathway Ca2+/Mn2+-ATPase 2 is a Golgi-localized pump with high affinity for Ca2+ ions
.
J Biol Chem
2005
;
280
:
22800
8
.
29.
Jenkins
J
,
Papkovsky
DB
,
Dmitriev
RI
. 
The Ca2+/Mn2+-transporting SPCA2 pump is regulated by oxygen and cell density in colon cancer cells
.
Biochem J
2016
;
473
:
2507
18
.
30.
Raiko
L
,
Siljamäki
E
,
Mahoney
MG
,
Putaala
H
,
Suominen
E
,
Peltonen
J
, et al
Hailey-Hailey disease and tight junctions: claudins 1 and 4 are regulated by ATP2C1 gene encoding Ca(2+)/Mn(2+) ATPase SPCA1 in cultured keratinocytes
.
Exp Dermatol
2012
;
21
:
586
91
.
31.
Deng
Y
,
Pakdel
M
,
Blank
B
,
Sundberg
EL
,
Burd
CG
,
von Blume
J
. 
Activity of the SPCA1 calcium pump couples sphingomyelin synthesis to sorting of secretory proteins in the trans-Golgi network
.
Dev Cell
2018
;
47
:
464
78
.
e8
.
32.
Durr
G
,
Strayle
J
,
Plemper
R
,
Elbs
S
,
Klee
SK
,
Catty
P
, et al
The medial-Golgi ion pump Pmr1 supplies the yeast secretory pathway with Ca2+ and Mn2+ required for glycosylation, sorting, and endoplasmic reticulum-associated protein degradation
.
Mol Biol Cell
1998
;
9
:
1149
62
.
33.
Powan
P
,
Luanpitpong
S
,
He
X
,
Rojanasakul
Y
,
Chanvorachote
P
. 
Detachment-induced E-cadherin expression promotes 3D tumor spheroid formation but inhibits tumor formation and metastasis of lung cancer cells
.
Am J Physiol Cell Physiol
2017
;
313
:
C556
66
.
34.
Ozawa
M
. 
The N-cadherin cytoplasmic domain confers anchorage-independent growth and the loss of contact inhibition
.
Sci Rep
2015
;
5
:
15368
.
35.
Liu
Z
,
Wei
Y
,
Zhang
L
,
Yee
PP
,
Johnson
M
,
Zhang
X
, et al
Induction of store-operated calcium entry (SOCE) suppresses glioblastoma growth by inhibiting the Hippo pathway transcriptional coactivators YAP/TAZ
.
Oncogene
2019
;
38
:
120
39
.
36.
Zhang
H
,
Liu
CY
,
Zha
ZY
,
Zhao
B
,
Yao
J
,
Zhao
S
, et al
TEAD transcription factors mediate the function of TAZ in cell growth and epithelial-mesenchymal transition
.
J Biol Chem
2009
;
284
:
13355
62
.
37.
Tang
Y
,
Weiss
SJ
. 
Snail/Slug-YAP/TAZ complexes cooperatively regulate mesenchymal stem cell function and bone formation
.
Cell Cycle
2017
;
16
:
399
405
.
38.
Vergara
D
,
Simeone
P
,
Latorre
D
,
Cascione
F
,
Leporatti
S
,
Trerotola
M
, et al
Proteomics analysis of E-cadherin knockdown in epithelial breast cancer cells
.
J Biotechnol
2015
;
202
:
3
11
.
39.
Sabatier
R
,
Finetti
P
,
Adelaide
J
,
Guille
A
,
Borg
JP
,
Chaffanet
M
, et al
Down-regulation of ECRG4, a candidate tumor suppressor gene, in human breast cancer
.
PLoS One
2011
;
6
:
e27656
.
40.
Gibault
F
,
Bailly
F
,
Corvaisier
M
,
Coevoet
M
,
Huet
G
,
Melnyk
P
, et al
Molecular features of the YAP inhibitor verteporfin: synthesis of hexasubstituted dipyrrins as potential inhibitors of YAP/TAZ, the downstream effectors of the hippo pathway
.
ChemMedChem
2017
;
12
:
954
61
.
41.
Dode
L
,
Andersen
JP
,
Vanoevelen
J
,
Raeymaekers
L
,
Missiaen
L
,
Vilsen
B
, et al
Dissection of the functional differences between human secretory pathway Ca2+/Mn2+-ATPase (SPCA) 1 and 2 isoenzymes by steady-state and transient kinetic analyses
.
J Biol Chem
2006
;
281
:
3182
9
.
42.
van Roy
F
,
Berx
G
. 
The cell-cell adhesion molecule E-cadherin
.
Cell Mol Life Sci
2008
;
65
:
3756
88
.
43.
Hu
Z
,
Bonifas
JM
,
Beech
J
,
Bench
G
,
Shigihara
T
,
Ogawa
H
, et al
Mutations in ATP2C1, encoding a calcium pump, cause Hailey-Hailey disease
.
Nat Genet
2000
;
24
:
61
5
.
44.
Thiery
JP
. 
Epithelial-mesenchymal transitions in development and pathologies
.
Curr Opin Cell Biol
2003
;
15
:
740
6
.
45.
Palacios
F
,
Tushir
JS
,
Fujita
Y
,
D'Souza-Schorey
C
. 
Lysosomal targeting of E-cadherin: a unique mechanism for the down-regulation of cell-cell adhesion during epithelial to mesenchymal transitions
.
Mol Cell Biol
2005
;
25
:
389
402
.
46.
Qian
LW
,
Greene
W
,
Ye
F
,
Gao
SJ
. 
Kaposi's sarcoma-associated herpesvirus disrupts adherens junctions and increases endothelial permeability by inducing degradation of VE-cadherin
.
J Virol
2008
;
82
:
11902
12
.
47.
Zanconato
F
,
Cordenonsi
M
,
Piccolo
S
. 
YAP/TAZ at the roots of cancer
.
Cancer Cell
2016
;
29
:
783
803
.
48.
Wells
A
,
Yates
C
,
Shepard
CR
. 
E-cadherin as an indicator of mesenchymal to epithelial reverting transitions during the metastatic seeding of disseminated carcinomas
.
Clin Exp Metastasis
2008
;
25
:
621
8
.
49.
Cancer Genome Atlas Network
. 
Comprehensive molecular portraits of human breast tumours
.
Nature
2012
;
490
:
61
70
.
50.
Barretina
J
,
Caponigro
G
,
Stransky
N
,
Venkatesan
K
,
Margolin
AA
,
Kim
S
, et al
The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity
.
Nature
2012
;
483
:
603
7
.