Orai3 is an estrogen receptor α-regulated Ca²⁺ channel that promotes tumorigenesis

Abstract

Store-operated Ca²⁺ entry (SOCE) encoded by Orai1 proteins is a ubiquitous Ca²⁺-selective conductance involved in cellular proliferation and migration. We recently described up-regulation of Orai3 channels that selectively mediate SOCE in estrogen receptor α-expressing (ERα⁺) breast cancer cells. However, the connection between ERα and Orai3 and the role of Orai3 in tumorigenesis remain unknown. Here, we show that ERα knockdown decreases Orai3 mRNA (by ∼63%) and protein (by ∼44%) with no effect on Orai1. ERα knockdown decreases Orai3-mediated SOCE (by ∼43%) and the corresponding Ca²⁺ release-activated Ca²⁺ (CRAC) current (by ∼42%) in ERα⁺ MCF7 cells. The abrogation of SOCE in MCF7 cells on ERα knockdown can be rescued by ectopic expression of Orai3. ERα activation increased Orai3 expression and SOCE in MCF7 cells. Epidermal growth factor (EGF) and thrombin stimulate Ca²⁺ influx into MCF7 cells through Orai3. Orai3 knockdown inhibited SOCE-dependent phosphorylation of extracellular signal-regulated kinase (ERK1/2; by ∼44%) and focal adhesion kinase (FAK; by ∼46%) as well as transcriptional activity of nuclear factor for activated T cells (NFAT; by ∼49%). Significantly, Orai3 knockdown selectively decreased anchorage-independent growth (by ∼58%) and Matrigel invasion (by ∼44%) of ERα⁺ MCF7 cells with no effect on ERα⁻ MDA-MB231 cells. Moreover, Orai3 knockdown inhibited ERα⁺ cell tumorigenesis in immunodeficient mice (∼66% reduction in tumor volume). These data establish Orai3 as an ERα-regulated channel and a potential selective therapeutic target for ERα⁺ breast cancers.—Motiani, R. K., Zhang, X., Harmon, K. E., Keller, R. S., Matrougui, K., Bennett, J. A., Trebak, M. Orai3 is an estrogen receptor α-regulated Ca²⁺ channel that promotes tumorigenesis.

Breast cancer is the most widespread cancer in women and accounts for >40,000 deaths annually in the United States alone (1). In over two-thirds of breast tumors, estrogen, a steroidal hormone secreted by ovaries, plays a very important role in tumor development (1–3). Estrogen mediates most of its effects via estrogen receptors (ERs). Two types of ERs are expressed in breast tissues, namely, ERα and ERβ. Among these, ERα is the most prominent ER known to contribute to breast cancer progression (3). ERα contributes to breast cancer development largely by regulating expression and function of various oncogenes (1–4). The inhibition of this important signaling pathway using ER modulators, such as tamoxifen, has shown a clear therapeutic benefit. Nonetheless, most of the breast cancer therapies currently available are effective only in a proportion of ER-positive (ER⁺) breast cancers, and new therapeutic targets are constantly needed.

Store-operated Ca²⁺ entry (SOCE) is the most widespread agonist-evoked Ca²⁺ entry pathway in nonexcitable cells (5, 6). SOCE is defined as Ca²⁺ influx into the cell, via plasma membrane Ca²⁺-permeable channels, as a direct outcome of intracellular Ca²⁺ store depletion (5–12). These store-operated Ca²⁺ (SOC) channels conduct a highly Ca²⁺-selective current, called Ca²⁺ release-activated Ca²⁺ (CRAC) (13). After >2 decades since the concept of SOCE was introduced by Putney (5), stromal interacting molecule 1 (STIM1) and Orai1 proteins were identified as the molecular players mediating SOCE (14–17). STIM1 acts as a Ca²⁺ sensor residing in the endoplasmic reticulum, which senses depletion of Ca²⁺ from endoplasmic reticulum, oligomerizes and translocates to subplasmalemmal puncta, where it activates Orai1 channels located in the plasma membrane (10, 18). Orai1-mediated SOCE regulates many important cell functions, including proliferation, migration (10, 19–25), and downstream signaling, which are important contributors to tumor development and metastasis (9, 26–29). In addition, a study by Feng et al. (30) discovered a novel store-independent mechanism of Orai1 activation by the secretory pathway Ca²⁺-ATPase (SPCA2) in breast cancer; store-independent interaction of SPCA2 with Orai1 was shown to elicit constitutive Ca²⁺ entry that promotes tumorigenesis.

Mammals possess 3 Orai proteins (Orai1-3) encoded by independent genes. As discussed earlier, Orai3 is a unique channel whose expression is restricted to mammals (31). Orai3/Orai1 heteromultimers constitute the store-independent arachidonate-regulated Ca²⁺ (ARC) channel (32). However, the role of Orai2 and Orai3 in native SOCE pathways, their physiological function, and their pathological contribution have remained largely unexplored. Indeed, studies using either knockout mice or knockdown strategies in vitro have clearly established Orai1 as the native SOC channel mediating CRAC currents in a large number of cell types (33). Results from our group provide one exception; our study demonstrated that SOCE in ER⁺ breast tumor cells is mediated by Orai3 instead of the canonical Orai1 pathway (34). Orai3 is highly expressed and selectively involved in mediating SOCE in 5 ER⁺ breast cancer cell lines picked randomly based solely on their positivity for the ER (34). In contrast, 5 randomly picked ER-negative (ER⁻) cell lines mediate SOCE through the canonical Orai1 pathway. A subsequent study from an independent group reported up-regulation of Orai3 in ∼79% of human breast cancer samples (35). The same study implicated Orai3 in cell cycle regulation of MCF7 breast cancer cells, as Orai3 knockdown caused cell cycle arrest at the G₁ phase (35). However, the potential selective regulation of Orai3 by ERα and the role of Orai3 in ER⁺ breast cancer development in vitro and tumorigenesis in vivo using animal models remain unexplored.

Here, we show that knockdown of ERα decreases Orai3 expression, SOCE, and CRAC currents mediated by Orai3 in ER⁺ MCF7 cells. Control experiments show that neither the expression of Orai1 in ER⁺ MCF7 cells nor SOCE amplitude in ER⁻ MDA-MB 231 cells was affected on transfection with silencing RNA (siRNA) targeting ERα. Ectopic expression of Orai3 in MCF7 cells that were first depleted of ERα by siRNA rescues SOCE. Activation of ERα with 17β-estradiol increases Orai3 expression and SOCE in response to thapsigargin and epidermal growth factor (EGF). Orai3 knockdown inhibited SOCE-dependent phosphorylation of extracellular signal-regulated kinase (ERK) and focal adhesion kinase (FAK), as well as nuclear factor for activated T cells (NFAT) transcriptional activity. Orai3 knockdown selectively inhibits anchorage-independent tumor cell colony formation and Matrigel invasion of ER⁺ MCF7 cells in vitro. Furthermore, Orai3 knockdown significantly inhibits in vivo tumor size, weight, and growth rate in severe combined immunodeficient (SCID) mice. We conclude that Orai3 is a functional SOC channel regulated by ERα and contributes to breast cancer development. Therefore, Orai3 might serve as a selective therapeutic target for treatment of ER⁺ breast cancers.

MATERIALS AND METHODS

Reagents

2-Aminoethoxydiphenyl borate (2-APB) and thapsigargin were obtained from Calbiochem (Gibbstown, NJ, USA), and fura-2 AM was purchased from Molecular Probes (Eugene, OR, USA). Cs³⁺-1,2-bis-(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) was from Invitrogen (Carlsbad, CA, USA). Na-methanesulfonate and Cs-methanesulfonate were from Sigma-Aldrich (St. Louis, MO, USA). siRNAs were purchased from Dharmacon (Lafayette, CO, USA). Viral cassettes encoding nontargeting short-hairpin RNA (shNT) and short-hairpin RNA (shRNA) against Orai3 (shOrai3) in pGIPZ vector were obtained from Open Biosystems (Lafayette, CO, USA). MCF7 and MDA-MB231 cell lines were from American Type Culture Collection (Manassas, VA, USA). Primers were synthesized by Integrated DNA Technologies (Coralville, IA, USA). Matrigel Boyden chambers and Matrigel for in vivo studies were purchased from BD Biosciences (San Jose, CA, USA). Estrogen tablets were procured from Innovative Research of America (Sarasota, FL, USA). The transfection kit for breast cancer cells (VCA-1003) was from Lonza (Basel, Switzerland). All other chemicals were from Fisher (Hampton, NH, USA).

Cell culture

MCF7 cells were cultured in phenol red-free MEM, whereas MDA-MB231 cells were cultured in normal DMEM as described before (34).

Quantitative PCR

Quantitative or real-time PCRs were performed using standard protocols, as reported previously (19). Briefly, total RNA was extracted from cells using a Qiagen RNeasy Mini Kit following the manufacturer's protocol. RNA (1 μg) was reverse transcribed to cDNA using oligo(dT) primers (Invitrogen) and SuperScript III reverse transcriptase (Invitrogen). Real-time PCR analysis was performed using a Bio-Rad iCycler and iCycler iQ Optical System Software (Bio-Rad Laboratories, Hercules, CA, USA). PCR reactions were performed using SYBR Green Supermix (Bio-Rad). The PCR protocol started with 5 min at 94°C, followed by 45 cycles of 30 s at 94°C, 30 s at 54.3°C, and 45 s at 72°C. The primer sequences for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), ERα, Orai1, and Orai3 are provided in Supplemental Table S1. Quantification of PCR was measured as sample fluorescence crossed a predetermined threshold value that was above background. Expressions of ERα, Orai1, and Orai3 were compared to those of the housekeeping gene GAPDH and were measured using comparative threshold cycle values, as described previously (19).

Cell transfections with siRNAs

MCF7 and MDA-MB231 cells were transfected as described before (34) using Amaxa nucleofector device II (Lonza AG, Basel, Switzerland). In brief, the proprietary VCA-1003 transfection reagent (Lonza) was used as per the manufacturer's instructions for transfections. As a marker of cell transfection, 0.5 μg of green fluorescent protein (GFP) was cotransfected with siRNA for identification of successfully transfected cells during experiments. All siRNA sequences used in this study are provided in Supplemental Table S2.

Lentiviral particle generation and infection

For achieving stable knockdown, human specific shOrai3 and shNT were purchased from Applied Biosystems (Foster City, CA, USA), cloned in the lentiviral vector pGIPZ, and viral particles were generated in our laboratory using standard protocols. Briefly, PolyJet was used as a transfection reagent (SignaGen, Gaithersburg, MD, USA) to transfect HEK293FT cells (Invitrogen). The lentiviral constructs pCMV-VSVG, pCMV-dR8.2, and pGIPZ-shOrai3/shNT were cotransfected into a flask of 95% confluent HEK293FT cells. Cell culture medium with viral particles was collected at 48 and 72 h after transfection and was concentrated by centrifugation using an Amicon Ultra-15 filter (Millipore, Billerica, MA, USA). These concentrated viral particles were then used to infect cells seeded at 50% confluence, and stable knockdown was confirmed by performing Western blot analysis and Ca²⁺ measurements 3 wk postinfection.

Ca²⁺ measurements

Cells were cultured on 30-mm glass coverslips for performing Ca²⁺ imaging. Coverslips with cells attached were mounted in a Teflon chamber and incubated at 37°C for 45 min in culture medium containing 4 μM fura-2 AM. After incubation, cells were washed 3 times and bathed in HEPES-buffered saline solution (140 mM NaCl, 1.13 mM MgCl₂, 4.7 mM KCl, 2 mM CaCl₂, 10 mM d-glucose, and 10 mM HEPES; pH 7.4) for ≥5 min before Ca²⁺ measurements were made. A digital fluorescence imaging system (InCyt Im2; Intracellular Imaging, Cincinnati, OH, USA) was used, and fluorescence images of several cells were recorded and analyzed. Fura-2 dye was excited alternately at 340 and 380 nm, and emission fluorescence was collected at 510 nm. The 340/380-nm ratio images were obtained on a pixel-by-pixel basis. Figures showing Ca²⁺ traces are an average from several cells attached on one coverslip and are representative of several independent recordings, as mentioned in the figure legends.

Whole-cell current measurements

Whole-cell patch-clamp recordings were carried out using an Axopatch 200B and Digidata 1440A (Axon Instruments, Foster City, CA, USA), as previously published (20, 34, 36). Clampfit 10.1 software (Axon Instruments) was used for data analysis. Pipettes were pulled from borosilicate glass capillaries (World Precision Instruments, Sarasota, FL, USA) with a P-97 flaming/brown micropipette puller (Sutter Instrument Co., Novato, CA, USA) and polished with DMF1000 (World Precision Instruments) to a resistance of 2–4 MΩ when filled with the following pipette solutions: 145 mM Cs-methanesulfonate, 20 mM Cs-BAPTA, 8 mM MgCl₂, and 10 mM HEPES (pH adjusted to 7.2 with CsOH). Immediately before the experiments, cells were washed with bath solution: 135 mM Na-methanesulfonate, 10 mM CsCl, 1.2 mM MgSO₄, 10 mM HEPES, 20 mM CaCl₂, and 10 mM glucose (pH was adjusted to 7.4 with NaOH). Only cells with tight seals (>16 GΩ) were selected for break-in. Cells were maintained at a 0-mV holding potential during experiments and subjected to voltage ramps every 2 s from +150 to −140 mV lasting 290 ms. “Reverse” ramps were designed to inhibit Na⁺ channels potentially expressed in these cells. High MgCl₂ (8 mM) was included in the patch pipette to inhibit TRPM7 currents.

NFAT luciferase activity assays

MCF7 cells were transfected with siRNA against Orai3 (siOrai3) or nontargeting control siRNA (siControl). After 2 d, cells were cotransfected with pIL-2-Luc (a luciferase reporter plasmid with NFAT response element) and Renilla luciferase plasmid (minTK pRL). Along with these plasmids, siOrai3 or siControl were also transfected into the cells a second time for maintaining Orai3 knockdown for 5 d. At 24 h after transfection, cells were treated with 2 μM thapsigargin for 15 min. NFAT activity was measured 24 h after treatment in cell lysates by firefly luciferase assay using the dual-luciferase reporter assay kit (Promega, Madison, WI, USA), and normalized to Renilla luciferase activity.

Western blot analysis

Cells were lysed using RIPA buffer (50 mM Tris-HCl, pH 8; 150 mM NaCl; 1% Triton X-100; 0.5% sodium deoxycholate; 0.1% SDS; and 0.2 mM EDTA). Proteins in denaturing conditions were subjected to SDS-PAGE. Proteins from gels were then electrotransferred onto polyvinylidene fluoride membranes. After blocking with 5% nonfat dry milk (NFDM) dissolved in Tris-buffered saline containing 0.1% Tween 20 (TTBS) for overnight at 4°C, blots were given three washes with TTBS for 5 min each and probed overnight at 4°C, with specific primary antibodies in TTBS containing 2% NFDM. The primary antibodies used were ERα (1:250; Cell Signaling, Beverly, MA, USA), hOrai1 (1:1000; Alomone, Jerusalem, Israel), hOrai3-NT (1:500; ProSci, Loveland, CO, USA), PCNA (1:1000; Abcam, Cambridge, MA, USA) β-actin (1:2000; Sigma-Aldrich), polyclonal anti-phospho-ERK1/2, T202/Y204 (1:1000; Cell Signaling), monoclonal anti-ERK2 (1:500; BD Biosciences); polyclonal anti-phospho-FAK, Y397 (1:1000; Invitrogen), and polyclonal anti-FAK (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA). The next day, membranes were washed with TTBS (3 washes of 5 min each) and were incubated for 1 h at room temperature with a horseradish peroxidase-conjugated anti-mouse (1:5000; Jackson Laboratories, Bar Harbor, ME, USA) for ERα and actin or anti-rabbit IgG for Orai1 and Orai3 (1:10,000; Jackson Laboratories) in TTBS containing 2% NFDM. Detection was performed using the enhanced chemiluminescence reagent (ECL Western blotting detection reagents; Amersham Biosciences, Piscataway, NJ, USA). Quantification of bands was achieved by densitometry using ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA).

Cell proliferation and Matrigel invasion assay

MCF7 cells were transfected with either siControl or siOrai3. At 72 h post-transfection, cell viability was evaluated, and proliferation was quantified by cell counting on the Coulter counter (Beckman Coulter, Fullerton, CA, USA).

Invasion assay was performed using Matrigel-coated Invasion Chambers (BD Biosciences) in accordance with the manufacturer's protocol. In brief, after rehydration of Matrigel inserts, cells (33×10⁴ cells/chamber in case of MCF7 cells and 25×10⁴ cells/chamber in case of MDA MB-231 cells) transfected with either siControl or siOrai3 were plated in the top chamber in serum-free medium at 72 h post-transfection. Cells were allowed to invade overnight toward the bottom chamber containing complete medium with 10% FBS. Next day, the bottom of the top insert was fixed and stained with DAPI (Vectashield; Vector Laboratories, Burlingame, CA, USA) for staining the nuclei of invaded cells. Twelve random pictures of each insert were taken with a Leica DM IRB microscope (×20 view; Leica Microsystems, Wetzlar, Germany), and the number of invaded cells was counted manually. All invasion assays were performed in triplicates, and ≥3 independent experiments/transfections were performed for both cell lines.

Soft agar colony formation

Soft agar colony formation assay was performed in 6-well plates. Three wells of each plate were dedicated to one condition for running experiments in triplicates. The base layer of wells was prepared with complete growth medium in 0.6% agar and was solidified at 4°C. After solidification, 1 × 10⁴ MCF7 cells infected with either shNT or shOrai3 were suspended in 1 ml of complete growth medium in 0.3% agar and were poured over base layer. In case of experiments with MDA-MB231 cells, 0.5 × 10⁴ cells were seeded in the top growth agar layer. These plates were kept at 4°C for few minutes until the top growth agar layer solidified. On the top of growth agar layer, 1 ml of complete growth medium was added and was replaced with fresh medium 1×/wk. The cells were cultured at 37°C for 4 wk. After 4 wk, cell colonies were stained with Crystal violet, and 12 random pictures of each well were taken with a Leica DM IRB microscope (×10 view). Total colonies were then counted (>50 μm in diameter for MCF7 and >25 μm in diameter for MDA-MB231) and statistically analyzed by paired t test as described below. Assays were repeated in 3 independent infections performed in triplicates for each cell line.

In vivo tumor development

Animal protocols were approved by the Institutional Animal Care and Use Committee at the Albany Medical College Animal Resource Facility, which is licensed by U. S. Department of Agriculture and New York State Department of Public Health, Division of Laboratories and Research. Four- to 6-week-old female SCID mice (Taconic Farms, Germantown, NY, USA) were acclimated at the animal facility for 2 wk before any intervention. Animals were anesthetized using i.p. injection of pentobarbital sodium. First, an estradiol tablet (SE-121, 0.72 mg/pellet, 60-d release) was implanted subcutaneously on the right flank of each mouse by making a small incision that was subsequently closed by a wound clip. Next, 2.5–4 × 10⁶ MCF7 cells infected with lentiviral particles encoding either shNT or shOrai3 in a volume of 50 μl along with 50 μl Matrigel were injected into the mammary fat pads of 10 mice/condition. Mice were examined every 2 d for tumor incidence. Tumor volume was assessed using Vernier calipers, and tumor volume was calculated using the formula V = W² × L/2, where V is tumor volume, W is tumor width, and L is tumor length.

The differences in tumor volume between shNT and shOrai3 groups were confirmed by ultrasound on the tumors on d 45 postinjection. Ultrasound scans were performed on mice that were anesthetized with 2% isoflurane using a small-animal anesthesia machine (Visual Sonics, Toronto, ON, Canada). The fur on the thorax was removed, and the transducer with gel was placed on the site of injection. Images were obtained using the Vevo770 software (Visual Sonics). Animals were finally euthanized on d 50 postinjection, and pictures of mice bearing tumors as well as isolated tumors were taken. Tumor weight was measured using a digital scale.

Statistics

Statistical analysis was performed using the paired t test. For tumor growth curves (in vivo studies in Fig. 7), 1-way ANOVA was performed. Values of P < 0.05 were considered significant.

Figure 7.

Orai3 knockdown inhibits in vitro anchorage-independent growth and invasion in MCF7 cells. A) Western blot analysis showing decrease in Orai3 expression at 96 h after siOrai3 transfection. Densitometry data from 4 independent experiments are quantified in bar graph. B) Representative DAPI-stained fields showing decrease in MCF7 cell invasion on Orai3 knockdown in in vitro Matrigel invasion assays. Bar graph shows statistical analysis from 3 different transfections performed in triplicates, representing 12 fields/membrane. C) Western blot analysis showing decrease in Orai3 protein on infection with shOrai3-encoding lentiviruses compared to shNT-encoding lentiviruses. Densitometry data from 4 independent experiments are quantified in bar graph. D) Images from soft agar growth assay 4 wk after seeding of MCF7cells infected with shNT and shOrai3. Bar graph shows quantification of data obtained from 3 independent transfections performed in triplicate, representing 12 fields/condition. *P < 0.05, **P < 0.01.

RESULTS

ERα selectively regulates Orai3 expression

We have recently reported that Orai3 is upregulated in ER⁺ breast cancer cells in comparison to “normal” breast epithelial cells or ER⁻ breast cancer cells. In addition, we also showed that in contrast to ER⁻ breast cancer cells, ER⁺ breast cancer cells use Orai3 for mediating SOCE instead of Orai1 (34). However, the potential connection between the presence of ER and function of Orai3 as a Ca²⁺-selective channel is unknown. To address this question, we characterized siRNA sequences targeting the dominant and major isoform of ER involved with breast cancer development, ERα. Four independent siRNA sequences targeting ERα were transfected into ER⁺ MCF7 cells. As shown in Supplemental Fig. S1A, all four siRNAs significantly decreased ERα mRNA levels three d post-transfection, compared to control nontargeting siRNA. Western blotting on MCF7 cell lysates harvested 4 d post-transfection (with either control nontargeting siRNA or different siRNA against ERα) showed a significant decrease in ERα protein levels (Supplemental Fig. S1B). As evident from these data, ERα siRNA 2 and 3 were the most efficient at decreasing ERα expression (Supplemental Fig. S1). Therefore, for the remainder of the study, we used ERα siRNA 2 and 3.

Interestingly, transfection of MCF7 cells with ERα siRNA caused a significant decrease of both ERα and Orai3 mRNA levels, with no effect on Orai1 levels as measured using quantitative PCR (Fig. 1A; see also Supplemental Fig. S2), suggesting that ERα might be regulating Orai3 mRNA expression. Furthermore, ERα knockdown resulted in significant decrease in ERα and Orai3 protein levels, with no effect on Orai1 proteins (Fig. 1B, C).

Figure 1.

ERα selectively regulates Orai3 expression. A) quantitative PCR (qPCR) data showing significant decrease in the transcript levels of ERα and Orai3 after transfection of MCF7 cells with siRNA targeting ERα (siERα; 72 h post-transfection) compared to siControl. Note that siERα does not affect Orai1 mRNA levels (n=4 independent transfections performed in duplicates). B) Western blot was performed with 100 μg of protein samples obtained from cells transfected with either siControl or siERα. Detection for ERα, Orai1, and Orai3 proteins was performed 96 h post-transfection using specific antibodies; β-actin loading control is also shown. C) Statistical quantification of Western blot data. *P < 0.05, **P < 0.01.

ERα regulates SOCE and CRAC currents mediated by Orai3

Interestingly, ERα knockdown using siRNA 2 (Fig. 2A, B) caused a significant decrease in SOCE by ∼43% in MCF7 cells (Fig. 2C, D). SOCE was activated by passive store depletion using thapsigargin, a pharmacological blocker of the sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase (SERCA) pump. Similar results were obtained with an additional ERα siRNA (siRNA 3), thus ruling out any nonspecific effects of siRNA (Supplemental Fig. S3A, B). The pharmacological compound 2-APB at 30 μM was shown in HEK293-overexpression systems to inhibit Orai1-mediated SOCE, while potentiating Orai3-mediated SOCE (37–41). In a previous study, we showed that native SOCE in MCF7 cells is potentiated by 2-APB and mediated by Orai3, independently of Orai1 (34). Therefore, we sought to determine the effect of ERα knockdown on 2-APB-mediated potentiation of SOCE in MCF7 cells. As shown in Fig. 2E, F, ERα knockdown resulted not only in a reduction of SOCE in response to thapsigargin, but also in a reduction (by ∼42%) of the subsequent potentiation of this SOCE by 2-APB. Control experiments conducted under the same ERα-knockdown and Ca²⁺-imaging conditions in ER⁻ MDA-MB231 cells clearly showed that thapsigargin-mediated SOCE in MDA-MB231 cells was not affected (Supplemental Fig. 3C, D). We performed additional Ca²⁺-imaging experiments in MCF7 cells, where Orai3 was ectopically expressed after ERα knockdown. Interestingly, Orai3 introduction into MCF7 cells depleted of ERα rescued SOCE activated by thapsigargin (Fig. 2G, H).

Figure 2.

ERα regulates Orai3-mediated SOCE in MCF7 cells. A) Western blot analysis showing that siERα was able to significantly reduce ERα expression. B) Statistical analysis of densitometry data from 4 independent transfections. C) Representative trace showing thapsigargin-activated SOCE in MCF7 cells measured using the standard Ca²⁺ off/Ca²⁺ on protocol simultaneously in several cells transfected with either siControl (n=17) or siERα (n=18) performed on the same day. D) Statistical analysis of Ca²⁺ imaging data from several independent runs (each representing several cells as shown in C). Note that in this figure and subsequent figures, the numbers in bar graphs (i.e., n=x, y) represent the total number of independent runs (x) and the total number of cells from all runs (y). E, F) Representative traces depicting the effect of siERα and siControl on 2-APB-mediated potentiation of SOCE (E) and statistical analysis on several independent experiments (F). G) Thapsigargin-activated SOCE was measured in MCF7 cells after 3 independent transfections with siControl, siERα, and siERα and eYFP-Orai3 cDNA plasmid. H) Statistical analyses on several independent runs similar to G. All data are derived from 3 or 4 independent transfections. *P < 0.05, **P < 0.01.

We previously demonstrated that Orai3 encodes a functional CRAC channel in MCF7 cells and that application of 2-APB to these cells amplifies these Orai3-mediated currents and causes a change in Orai3 Ca²⁺ selectivity, as evidenced by left shift in reversal potentials and manifestation of outwardly rectifying currents (34). We performed whole-cell patch-clamp measurements of CRAC currents activated by store depletion through dialysis with the fast Ca²⁺ buffer BAPTA (20 mM) in MCF7 cells transfected either with control siRNA or ERα siRNA. Dialysis with 20 mM BAPTA through the patch pipette in control siRNA-transfected MCF7 cells resulted in development of CRAC currents measured in Ca²⁺-containing bath solutions (20 mM Ca²⁺), which were amplified in divalent-free (DVF) solutions and further potentiated with 2-APB addition to the bath solution (Fig. 3A). The same experiments in MCF7 cells transfected with ERα siRNA showed a significant reduction in CRAC currents as well as 2-APB-mediated potentiation of these currents (Fig. 3B). The I/V relationships of CRAC currents measured in DVF solutions and their potentiation with 2-APB are shown in Fig. 3C, D, respectively, for both control siRNA- and ERα siRNA-transfected MCF7 cells (see statistics in Fig. 3E). The I/V relationships clearly show that CRAC currents potentiated by 2-APB have a left shift in reversal potential and an outward component, consistent with the biophysical properties of Orai3 (41).

Figure 3.

ERα regulates Orai3-mediated CRAC currents in MCF7 cells. A, B) SiERα causes abrogation of CRAC currents measured in Ca²⁺-containing and DVF solutions, as well as 2-APB-mediated potentiation of these currents in comparison to siControl; data represented were taken from each ramp at −100 mV. C, D) I/V relationships of monovalent CRAC in MCF7 cells transfected with siControl or siERα before (C) and after (D) addition of 2-APB. I/V relationships are taken from traces in A and B where indicated by solid and shaded plus symbols and asterisks. E) Statistical analysis on several recordings (n=5–6) of monovalent CRAC currents and 2APB-mediated potentiation of these currents measured at −100 mV. **P < 0.01.

ERα is a well-established transcription factor that regulates the expression of genes associated with tumorigenesis (1–3). The pathway involved in this regulation is initiated on ERα activation by its ligand, estradiol. Treatment of MCF7 cells with 100 nM of 17β-estradiol overnight resulted in ∼3-fold increase in Orai3 mRNA levels, suggesting that Orai3 expression is regulated by ERα (Fig. 4A). Increased Orai3 levels on activation of ERα resulted in increase in the magnitude and rate of Orai3-mediated SOCE (Fig. 4B). Figure 4B–D shows that overnight treatment of MCF7 cells with 100 nM 17β-estradiol resulted in an increase in SOCE, as well as 2-APB-mediated potentiation of SOCE, in comparison to control vehicle-treated MCF7 cells.

Figure 4.

ERα activation increases Orai3 expression and SOCE mediated by Orai3. A) Treatment with 100 nM 17β-estradiol increases Orai3 mRNA expression by 3-fold; data represent 4 independent experiments. B) Representative Ca²⁺ imaging traces showing that 17β-estradiol treatment increases SOCE, as well as 2-APB-mediated potentiation in MCF7 cells. C, D) Quantification of Orai3-mediated SOCE (C) and 2-APB-mediated SOCE potentiation (D) in MCF7 cells with or without 17β-estradiol treatment from 7 independent runs originating from 3 independent 17β-estradiol treatments. *P < 0.05, **P < 0.01.

Breast cancer growth factors activate Ca²⁺ influx in ER⁺ cells through Orai3

Our data so far have relied on the pharmacological activation of SOCE by thapsigargin. Therefore, we sought to determine whether physiological agonists that act through phospholipase C (PLC)-coupled receptors and are pertinent to breast cancer development could also mediate SOCE via Orai3 in ER⁺ MCF7 cells. We tested two growth factors with established roles in breast cancer growth: EGF (42), which acts through a tyrosine kinase receptor; and thrombin (43, 44), which activates G-protein-coupled receptors. Both growth factors caused Ca²⁺ release from internal stores (assayed in nominally Ca²⁺-free solutions) and induced Ca²⁺ influx across the plasma membrane (measured when 2 mM Ca²⁺ was subsequently replenished to the bath solution) in MCF7 cells (Fig. 5B, C). A trace representing thapsigargin-activated SOCE is shown in Fig. 5A for comparative purposes; the Ca²⁺ influx signal activated by EGF and thrombin is relatively smaller compared to that activated by thapsigargin. ER⁻ MDA-MB231 displayed increased Ca²⁺ release and influx signals only in response to thapsigargin and thrombin but not to EGF, consistent with the reported lack of EGF receptors in these cells (Fig. 5E–G); again, thapsigargin-activated Ca²⁺ influx was more robust than that activated by thrombin in MDA-MB231cells. Interestingly, the Ca²⁺ influx pathway activated by EGF and thrombin in MCF7 cells was pharmacologically identical to that activated by thapsigargin in the same cells, namely potentiation with 30 μM 2-APB (Fig. 5A–C), while thapsigargin- and thrombin-activated Ca²⁺ influx in MDA-MB231 cells were both inhibited by the same concentration of 2-APB (Fig. 5E, G), suggesting that EGF- and thrombin-activated Ca²⁺ influx in MCF7 cells likely occurs via Orai3-mediated SOCE, while thrombin-activated Ca²⁺ entry in ER⁻ MDA-MB231 cells occurs through Orai1. Indeed, knockdown of Orai3 in MCF7 cells (Fig. 5D) and Orai1 in MDA-MB231 cells (Fig. 5H) using previously described siRNAs (34) abrogated thrombin-activated Ca²⁺ entry in these cells. Furthermore, treatment of MCF7 cells with 100 nM 17β-estradiol overnight resulted in an increase in EGF-activated Ca²⁺ entry (Supplemental Fig. S4), strongly suggesting that EGF-mediated SOCE occurs via ERα-regulated Orai3 channels.

Figure 5.

EGF and thrombin activate Ca²⁺ influx through Orai3. A–D) Ca²⁺ release and subsequent Ca²⁺ influx measured simultaneously in several MCF7 cells, in response to thapsigargin (A; n=32), EGF (B; n=30) and thrombin (C; n=29) and showing that 2-APB potentiated the Ca²⁺ influx in response to all three stimuli. D) MCF7 cells transfected with siOrai3 or siControl show that siOrai3 significantly abrogated Ca²⁺ influx in response to thrombin in comparison with siControl. E–H) Ca²⁺ release and subsequent Ca²⁺ influx measured simultaneously in several MDA-MB231 cells, in response to thapsigargin (E; n=56), EGF (F; n=39), and thrombin (G; n=19) and showing that 2-APB inhibited the Ca²⁺ influx in response to thapsigargin and thrombin (EGF caused no perceptible rise in either Ca²⁺ release or Ca²⁺ entry. Ionomycin (Iono, 10 μM) was added at the end of the recording as a control). H) MDA-MB231 cells transfected with siOrai1 or siControl show that siOrai1 significantly abrogated Ca²⁺ influx in response to thrombin in comparison with siControl. **P < 0.01, ***P < 0.001.

Orai3 is required for SOCE-dependent NFAT activity and ERK1/2 and FAK kinase phosphorylation in MCF7 cells

Next, we sought to determine whether Orai3 knockdown has any effect on Ca²⁺-dependent downstream effectors that control cell proliferation (e.g., NFAT, ERK1/2) and migration (ERK1/2, FAK). SOCE was activated by thapsigargin (2 μM for 15 min) in MCF7 cells, and NFAT activity was measured using a luciferase reporter assay. Figure 6A shows a significant increase in thapsigargin-induced NFAT activity in cells treated with siControl, and this increase was significantly inhibited in cells transfected with siOrai3 (Fig. 6A). Similarly, thapsigargin-induced phosphorylation of ERK1/2 (Fig. 6B, C) and FAK (Fig. 6D, E) was increased in cells transfected with siControl, and this phosphorylation was significantly inhibited when MCF7 cells were transfected with siOrai3.

Figure 6.

Orai3 knockdown inhibits SOCE-mediated NFAT activity and ERK1/2 and FAK phosphorylation in MCF7. A) NFAT activity assay was performed as described in Materials and Methods. Thapsigargin (2 μM for 15 min)-mediated NFAT activity was increased by siControl treatment, and this increase was significantly inhibited by siOrai3 treatment. Data are quantified from 3 independent experiments that were run in triplicates. B) PhosphoERK1/2 levels were measured in cell lysates containing proteases and phosphatases inhibitors before (time 0) and after thapsigargin stimulation for 5 and 15 min; ERK2 was used as a loading control. C) Ratio of densitometry of phosphoERK1/2 over ERK2; quantification of data obtained from 3 independent experiments. D) PhosphoFAK was also measured before (time 0) and after thapsigargin stimulation for 5 and 15 min, with total FAK used as loading control. E) Ratio of densitometry of phosphoFAK over total FAK; Quantification of data obtained from 3 independent experiments. *P < 0.05, **P < 0.01.

Orai3 is required for breast cancer invasion and colony formation in vitro

Orai3 has been shown to inhibit cell proliferation of MCF7 cells and cause cell cycle arrest at the G₁ phase (35). Orai3 knockdown using previously characterized siRNA (34) caused a significant decrease in Orai3 protein levels by ∼52% (Fig. 7A), resulting in ∼40% reduction in MCF7 cell proliferation at 72 and 96 h post-siRNA transfection (Supplemental Fig. S5A) and significant reduction in protein expression of the proliferation marker PCNA (Supplemental Fig. S5B, C). These data confirmed earlier work showing that Orai3 plays an important role in regulating breast cancer cell proliferation. We next sought to determine the role of Orai3 in breast cancer development in vitro and in vivo. Tumor cell invasion in vitro was determined using Boyden chamber invasion assays, as described in Materials and Methods. We seeded MCF7 cells 3 d post-transfection with either siControl or siOrai3 in Matrigel-coated top chambers; invasion was allowed to occur overnight, and cells that were able to invade the Matrigel were stained with DAPI and counted. Figure 7B shows representative images demonstrating that Orai3 knockdown resulted in significant decrease in tumor cell invasion in comparison to cells transfected with control siRNA. Quantification of data from 3 independent transfections performed in triplicate is represented in Fig. 7B. Similar studies in ER⁻ MDA-MB231 cells showed no change in invasion on Orai3 knockdown (Supplemental Fig. S6).

To evaluate the contribution of Orai3 in anchorage-independent colony formation, the use of siRNA transfection of MCF7 cells is not practical, since siRNA-induced protein knockdown lasts for days, while these cells require weeks to form colonies in soft agar. Therefore, we generated lentiviral particles encoding two different shRNA sequences: shNT (control shRNA) and shOrai3 to achieve stable Orai3 knockdown. As shown in Supplemental Fig. S7A, lentiviral particles encoding shOrai3 or shNT were able to infect MCF7 cells with >90% efficiency, as evidenced by GFP fluorescence encoded by the viral particles (see Materials and Methods). Further, Western blot analysis and Ca²⁺-imaging experiments confirmed the decrease in both expression and function of Orai3 in shOrai3-infected cells (Supplemental Fig. S7B–D). ShOrai3 was specific to Orai3, as protein expression of its closest homologue, Orai1, was not affected by shOrai3 (Supplemental Fig. S7B). Anchorage-independent soft agar colony assays on MCF7 cells were performed on stable infection with lentiviral particles encoding either shNT or shOrai3, and images of tumor cell colonies were obtained 4 wk after seeding cells on soft agar. Cell colonies of >50 μm in size were counted from both shNT and shOrai3 conditions. As represented in Fig. 7C, Orai3 knockdown with shRNA-encoding lentiviruses significantly reduced Orai3 protein levels (Fig. 7C) and the number and size of MCF7 cell colonies formed on soft agar (Fig. 7D). Data are from 3 independent experiments performed in triplicate and are quantified in Fig. 7D. Significantly, there was no effect of shOrai3 on the ability of ER⁻ MDA-MB231 cells to form colonies on soft agar (Supplemental Fig. S8).

Orai3 is involved in breast cancer development in vivo

To determine the contribution of Orai3 to ER⁺ breast tumor development in vivo, we performed orthotopic injections of MCF7 cells in mammary fat pads of SCID mice. MCF7 cells injected into mice were first infected with lentiviral particles encoding either shNT or shOrai3 (Supplemental Fig. S7). Mice were implanted with a 60-d slow-release estradiol pellet to provide ER⁺ tumor cells a continuous supply of estrogen for tumor development. Tumor volumes were determined using the Vernier caliper, as described in Materials and Methods. We also performed ultrasound imaging on the tumors at d 45 postinjection to confirm caliper readings. As shown in Fig. 8A, B, the size of tumors at d 50 postinjection was dramatically reduced on Orai3 knockdown. Representative ultrasound images taken at d 45 postinjection visibly ascertain the smaller tumor size of the shOrai3 group (Fig. 8C). Furthermore, while all mice (10/10) injected with MCF7 cells infected with lentiviral particles encoding shNT developed tumors, only 60% of mice (6/10) injected with MCF7 cells infected with lentiviral particles encoding shOrai3 developed tumors (Fig. 8D). Interestingly, shNT-group mice started developing tumors 10 d after cell injections, whereas shOrai3-group mice took ≥20 d to develop tumors (Fig. 8D). Figure 8D, showing tumor volume as a function of time, shows that the rate of tumor growth was significantly lower in the shOrai3 group in comparison to the shNT group. Tumor volumes at d 50 postinjection were 651 ± 123 mm³ for shNT vs. 225 ± 101 mm³ for shOrai3. Animals were euthanized 50 d postinjection, and tumors were isolated and weighed (Fig. 8E), showing that tumors in the shOrai3 group weighed significantly less than those in the shNT group; the shOrai3 group developed on average smaller tumors (190±77 mg) than the shNT group (512±55 mg) group (Fig. 8E). Stable knockdown of Orai3 in solid tumors was confirmed by Western blots at d 50 postinoculation (Fig. 8F, G).

Figure 8.

Orai3 knockdown inhibits breast tumor growth in vivo. A) Images of tumors in SCID mice injected with MCF7 cells infected with lentiviral particles encoding either shNT or shOrai3 at d 50 postinjection. B) Representative image of tumors isolated from shNT and shOrai3 groups at d 50 postinjection. C) Representative ultrasound scans on SCID mice injected with either shNT- or shOrai3- MCF7 cells in the mammary fat pad at d 45 postinjection; tumors are encircled in red. D) Tumor growth curve measured as increase in tumor volume as a function of time (d) postinjection of MCF7 cells (infected with either shNT- or shOrai3-encoding lentiviruses) into mammary fat pads of SCID mice. E) Scatterblot showing tumor weight at d 50 from all injected mice. F) Stable Orai3 knockdown was evaluated by Western blot on total tumor lysates from shNT and shOrai3 group at d 50 postinoculation. G) Densitometry data normalized to actin were quantified from 3 independent experiments. *P < 0.05, **P < 0.01.

DISCUSSION

Our previous studies have shown that, unlike ER⁻ breast cancer cell lines, which mediate SOCE and CRAC currents through the canonical Orai1 pathway, 5 ER⁺ breast cancer cell lines picked at random based solely on their positivity for ER showed up-regulated Orai3 proteins and selectively mediated SOCE and CRAC currents via Orai3 (34). An independent group subsequently reported up-regulation of Orai3 in ∼79% of human breast cancer samples and implicated Orai3 in MCF7 cell proliferation by showing that Orai3 knockdown caused cell cycle arrest at the G₁ phase (35). Here, we used two independent siRNA sequences to knock down ERα. We show that Orai3 expression is specifically regulated by ERα, since down-regulation of ERα levels caused a significant decrease of Orai3 mRNA and protein levels with no effect on Orai1 levels. Although we achieved ∼82% knockdown of ERα, Orai3 protein levels, SOCE magnitude, and CRAC currents in MCF7 were decreased by only ∼44, ∼43, and ∼42%, respectively. This likely reflects the existence of additional regulatory mechanisms that control Orai3 expression in MCF7 cells. Significantly, the decrease in SOCE in MCF7 cells on ERα knockdown can be rescued by ectopic expression of human Orai3 into these cells. Furthermore, activation of ERα by 17β-estradiol led to a 3-fold increase in Orai3 mRNA levels, while SOCE showed a relatively smaller increase (∼23%). This can be explained by the fact that protein levels of the Ca²⁺ sensor STIM1, which is required for the function of all Orai channels (45), are likely limiting. Indeed, on 17β-estradiol stimulation, the potentiation of Orai3-mediated SOCE by 2-APB, which does not depend on STIM1 (41), is substantially increased (∼2.3 fold). The decrease and increase of Orai3 mRNA on ERα knockdown and ERα stimulation, respectively, suggests that ERα likely regulates Orai3 at the transcriptional level. However, it remains to be determined whether this regulation is a direct (i.e., through binding of ERα to ERα-responsive elements on the Orai3 promoter) or an indirect one. Although the sequence and location of Orai3 promoter remain unknown, there are 2 putative ERα binding sites in the 3′UTR and 6 in the intronic sequence of Orai3 gene. Additional studies are required to determine whether these putative ERα-binding sequences are involved in directly controlling Orai3 transcriptional activity. It is worth mentioning, a recent study by Floukaris et al. showing that androgen receptor knockdown in the prostate cell line LNCaP caused a decrease in Orai1 protein levels and associated store depletion-activated currents, suggesting the regulation of Orai1 by androgen receptors (29). These data suggest that sex hormones play differential roles in regulating Orai channel isoform expression and that increased expression of Orai3 relative to Orai1 might be a characteristic of females. The functional outcome of such differences is an important question that warrants further investigations.

EGF and thrombin are breast cancer growth factors with established roles in cancer proliferation and invasion that are known to mediate their effects partially through initiation of Ca²⁺ signaling. However, the precise Ca²⁺ entry channel activated by these agonists in ERα⁺ breast cancer cells remained unknown. We show data supporting a role for Orai3 in mediating Ca²⁺ entry in response to EGF and thrombin. Thrombin activates Ca²⁺ entry in MCF7 cells via Orai3, and EGF-mediated Ca²⁺ entry is further increased on ERα stimulation with 17β-estradiol. Furthermore, Orai3 knockdown inhibits activity of downstream Ca²⁺-dependent proliferative and migratory pathways. Indeed, Orai3 knockdown inhibits SOCE-mediated NFAT activity, as well as ERK1/2 and FAK phosphorylation in MCF7 cells, suggesting that Orai3 is required for providing the Ca²⁺ signal necessary for MCF7 cell proliferation and migration. Indeed, we further show that Orai3 plays a significant role in tumor invasion and colony formation in vitro and tumorigenesis in vivo using SCID mice. Orai3 knockdown was achieved using two independent siRNA sequences. SiRNA against Orai3 achieved ∼54% knockdown of Orai3 protein levels and affected in vitro invasion of MCF7 cells by ∼44%. The use of lentiviruses encoding shRNA against Orai3 allowed for stable knockdown of Orai3 over a period of weeks and achieved greater levels of Orai3 protein knockdown (∼68%); interestingly, Orai3 knockdown with shRNA caused a substantial reduction in anchorage-independent growth of MCF7 cells (by ∼58%). Most important, Orai3 knockdown with shRNA in MCF7 cells led to a significant reduction in ER-dependent in vivo tumor growth in SCID mice. Orai3 knockdown with lentiviruses encoding shRNA caused an average reduction of tumor volume by ∼66% and tumor weight by ∼63%. Furthermore, 4 of 10 SCID mice (40%) injected with MCF7 cells stably infected with lentiviruses encoding shRNA against Orai3 did not develop any tumors. In contrast, 100% of SCID mice injected with MCF7 cells stably infected with lentiviruses encoding shNT developed tumors. On the basis of our data, we would like to propose that channel inhibitors capable of achieving selectivity for Orai3 over Orai1 (and Orai2) should be seriously considered for selective targeting of ER⁺ breast tumors.