MCF-7 cells were grown in DMEM medium containing 10% fetal calf serum and 1% penicillin/streptomycin serum as described (9). Cells were plated onto ø13-mm coverslips and used when 60–70% confluent.

The growth medium was removed and cells were rinsed once in Earle’s balanced salts solution (EBSS; Invitrogen). Calcium-green of 50 μg AM (C3012; Invitrogen) or Fura-2 AM (F1221; Invitrogen) were dissolved in 20 μl 20% pluronic acid in DMSO (0.01 g in 50 μl DMSO stock). Before the experiment, mixtures of 1 μl dye preparation in 200 μl EBSS was applied and cells were incubated for 60 min. Prior to placing the coverslip into the recording chamber, coverslips were rinsed in EBSS to remove residual dye. Data acquisition and analysis were performed via OpenLab v.3.1.7 (Improvision Ltd., Coventry, UK). A CCD camera (ORCA-AG; Hamamatsu Ltd., Japan) was used to capture the fluorescent image by using Fura-2-AM and calcium green. In the experiments performed using Fura-2, fluorescent intensities were measured with dual-sequential-wavelength excitation at 340 and 380 nm, and emission at 510 nm. Changes in Ca²⁺ concentration were expressed as ratios of 340/380. Fluorescent intensity of calcium green-1 was measured with a single wavelength excitation at 488 nm and emission at 528 nm. Changes in the Ca²⁺ concentration were expressed as ΔF/F, where F was the fluorescence intensity when cells were at rest, and ΔF was the change in fluorescence during stimulation.

Stealth siRNA (Invitrogen) was obtained from Invitrogen. MCF-7 cells were passaged onto coverslips in 500 μl Opti-MEM (Invitrogen) one day before transfection and reached about 40–50% confluence at the time of transfection. siRNA of 20 pmol (against TRPC3) or the siRNA negative control complex, with a 1:125 final dilution of Lipofectamine 2000 (Invitrogen) was used according to the manufacturer’s instructions. The knockdown effects were examined at 48 h and the results were compared with control and control without knockdown. Results were collected from 3 different batches of MCF-7 cells. Human hCOX2 plasmids were obtained from Professor R. Kulmacz (University of Texas Health Science Center at Houston). Cells were transfected with hCOX2 by Lipofectamine 2000. The effects of transfection were examined by western blot analysis at 24 and 48 h.

RT-PCR experiments followed standard protocols. Primers were designed with primer 3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) for TRPC1 (NM_003304/92 bp), TRPC3 (NM_003305/157 bp), TRPC4 (NM_016179/191 bp), TRPC5 (NM_012471/108 bp), TRPC7 (NM_020389/135 bp) and the α1C subunit (NM_000719/194 bp), α1G subunit (AH_007322/135 bp) and α1H subunit of VGCCs (NM_021098/123 bp). Antibodies against TRPC1, 3, 4 and 5 were the kind gift from Professor W.P. Schilling (Case Western University, Cleveland, OH, USA). The peptide sequence (26) used to generate the antibody against TRPC3 was RRRRLQKDIEMGMGN.

After removal of methanol, cells were treated with a Coulter DNA-Prep reagent kit (Beckman-Coulter, France). Cells were resuspended in 40 μl of a lysing and permeabilizing reagent and 400 μl of a propidium iodide solution containing RNAse. Flow cytometry analyses were performed using a Coulter Epics Elite ESP flow cytometer (Beckman-Coulter) equipped with a 488 nm argon laser running at 15 mW. The red DNA fluorescence signal was analyzed as total (area) vs. peak signal, in order to eliminate doublets and aggregates. Data were recorded for at least 10,000 events. Cell cycle distribution was analyzed with the Multicycle-AV software (Phoenix Flow Systems, San Diego, CA, USA).

Cell proliferation was determined using the tetrazolium salt reduction method (MTT). Cells were seeded at 4×10⁴ cells/well on a 24-well plate (for a given condition on three separate experiments) and allowed to start growth for 48 h. Drugs and AA were added for 24 and 48 h at the concentrations indicated in the figure legends. Cells were incubated at 37°C with the tetrazolium salt [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] and metabolically active cells reduced the dye to purple formazan. After 45 min of incubation at 37°C, the medium was discarded and MTT formazan crystals were solubilized with DMSO. Absorbance was measured in a multiwell plate spectrophotometer at 570 nm (Molecular Devices model Thermomax microplate reader, Les Ulis, France). For easier comparison between conditions, results obtained for proliferation were normalized. The means were then calculated on the daily calculated ratios. The mean of each triplicate was used to create a data point for comparing cell growth in different conditions.

Invasion and migration were analyzed in BD Falcon 24-well plates with 8-μm pore size polyethylene terephtalate membrane BD BioCoat cell culture inserts. For the invasion assay, the membrane was covered with a film of Matrigel (Becton-Dickinson). The upper compartment was seeded with 4×10⁴ viable cells in DMEM with 5% FBS ± drugs/AA and the lower compartment was filled with DMEM supplemented with 10% FBS as a chemo-attractant ± drugs/AA. After 48 h at 37°C, cells remaining on the upper side on the membrane were removed with a cotton swab and the cells that had migrated and were attached to the lower side were stained with hematoxylin for 2 min and counted on the total surface of the insert using light microscopy at x200 magnification.

EBSS for calcium imaging recording contained: NaCl 116.3 mM, NaH₂PO₄ 1 mM, KCl 5.3 mM, MgCl₂ 1 mM, CaCl₂ 1.8 mM, NaHCO₃ 26 mM, D-glucose 5.5 mM, HEPES 10 mM. The EGTA solution for calcium imaging recording contained: NaCl 116.3 mM, NaH₂PO₄ 1 mM, KCl 5.3 and 1.8 mM, NaHCO₃ 26 mM, D-glucose 5.5 mM, HEPES 10 mM, EGTA 0.2 mM. All solutions were prepared fresh on the day of experimentation.

Chemicals dissolved in ethanol or DMSO were made up as 1,000 times stock solutions. All 1X chemical solutions were made with an appropriate bath solution on the day of experimentation. Solutions such as TG, AA, ETYA, were kept in the dark and on ice and added into a syringe which led to the recording chamber when necessary. The solvent, ethanol at the same dilution, was tested alone in controls and had no effects on calcium entry evoked either by store depletion or OAG stimulation.

Expression of the TRPC3, α1H (T type) but not the α1C or α1G subunits of the voltage-gated calcium channel (VGCC) was observed in MCF-7 cells (Fig. 1).

The spatial distribution of TRPC3 protein in MCF-7 cells was determined by immunohistochemistry, as previously reported in other cell types (27). Abundant, non-discrete TRPC3 expression throughout the cell cytoplasm and at the cell surface was observed (Fig. 1B). Western blot analysis consistently demonstrated abundant expression of TRPC3 in MCF-7 cells (Fig. 1C).

Recombinant human TRPC3 can be directly activated by diacylglycerol (DAG) (28,29). In MCF-7 cells, OAG (100 μM), a synthetic DAG, induced a significant intracellular Ca²⁺ elevation in an extracellular calcium-dependent manner (Fig. 2). Recent reports have confirmed that TRPC3 could also mediate store operated calcium entry (SOCs) (30). Activation of native TRPC3 via store depletion by SERCA pump inhibitors, such as TG, has been reported in HSG (31) and epithelial cells (32). In MCF-7 cells, store depletion by TG (1 μM) and EGTA bath led to significant Ca²⁺ elevations (Fig. 2A).

The Ca²⁺ elevation induced by store depletion was almost quenched by bath application of 10 μM AA (Fig. 2A, yellow line) or 10 μM LA (Fig. 2A, blue line). 2APB, as a common antagonist of TRPC3, at the concentration of 100 μM, caused a transient calcium peak followed by decay (Fig. 2A, pink line) to a level similar to that in the presence of AA or LA. In addition, the CCE following Ca²⁺ replacement after EGTA, could be prevented by 10 μM AA and CCE was observed after removal, by washout, of the AA inhibition (Fig. 2B).

Exogenous AA can freely penetrate membranes, but will be challenged and degraded by endogenous intracellular cyclooxygenases (COX) or lipooxygenases (LOX). To exclude the possibility that the inhibition effect is due to degenerated metabolites of AA rather than AA itself, 14-eicosatetraynoic acid (ETYA, 20 μM), a competitive analogue of AA resistant to LOX and other degradative enzymes, was used to mimic the effect of AA. ETYA, like AA, inhibited CCE induced by store depletion (Fig. 2A, brown line), suggesting that PUFA interacts directly with TRPC channels.

The effect of fatty acids on ion channels is sometimes attributed to non-specific influences upon cell membranes (33,34). Polyunsaturated fatty acids may, for example, increase membrane fluidity. AA at a concentration between 25–100 μM directly affected channel proteins via a change of the lipid environment (34). We observed that application of 500 μM AA induced a massive, irreversible Ca²⁺ elevation, presumably following the collapse of cell integrity. We therefore maintained PUFA concentrations below the threshold (15 μM) for such effects. Additionally, CCE was also inhibited by the double-bonds unsaturated fatty acid, e.g. LA (5 and 10 μM) and the mono-bond unsaturated fatty acid, oleic acid (10 μM). However, the saturated fatty acid, stearic acid (20 μM), had no effect upon induced CCE.

To confirm that CCE in MCF-7 cells was mediated via TRPC3 channels, we examined Ca²⁺ entry in MCF-7 cells by knocking down TRPC3 (shealth RNAi; Invitrogen) (Fig. 2D and E). A negative control iRNA was utilized as a control to discount the non-specific effect of TRPC3 iRNA. Transfected cells gradually lost the TRPC3 channels and after 48 h, detectable TRPC3 protein was almost eliminated (Fig. 2D). In parallel to the decrease of TRPC3 protein, calcium elevation evoked by store depletion was also significantly reduced; suggesting that calcium elevation induced by store depletion was mediated via the TRPC3 channels.

Ca²⁺ entry in MCF-7 cells could be evoked by application of OAG (Fig. 2C). AA in a range from 0.1 to 20 μM was employed to study the inhibitory effect of AA on OAG induced Ca²⁺ elevation. A single, sigmoidal dose-response curve was fitted according to data (Fig. 2C) and gave an IC₅₀ of AA upon OAG induced Ca²⁺ elevation of 1.78±0.17 μM.

There are generally two principal sources in AA generation, i) cytosolic phospholipase A2 (PLA₂) releasing AA from appropriate phospholipids (35–37) and ii) fatty acid amide hydrolase (FAAH) degenerating endogenous anandamide and relatives into AA. The AA metabolism diverges down two main pathways, the COX and LOX pathways (38).

cPLA2, FAAH and COX2 were demonstrated to be present in MCF-7 breast cancer cells by western blot analyses (Fig. 3A). Accordingly, reduction of endogenous PUFA could be achieved by either inhibition of cPLA2, FAAH or overexpression of COX2. Inhibition COX2 or LOX could induce the elevation of endogenous PUFA. Consistent with our results above, reduction of endogenous PUFA by the cPLA2 antagonist (AACOCF3), FAAH antagonist (PMSF), and overexpression of COX2 significantly enhanced Ca²⁺ entry via TRPC3. Niflumic acid, the antagonist of COX2, reduces degeneration of PUFA, resulting in elevation of endogenous PUFA. Ca²⁺ entry via TRPC3 was reduced by pre-incubation with niflumic acid (Fig. 3D).

The common antagonist of TRP channels, 2-APB, was used to investigate the effects of these channels on the proliferation of MCF-7 breast cancer cells. 2-APB (50 μM) significantly reduced the S phase of the cell cycle (Fig. 4A-i), and cell proliferation (MTT test; 100 μM), whilst D609, an inhibitor of phospholipase C, had no effect (Fig. 4A-ii). 2-APB also reduced MCF-7 cell migration and invasion in a dose-dependent manner (Fig. 4A-iii and iv). The effects of increasing concentration of AA on cell migration are shown in Fig. 4B. Taken together, these results suggest that cellular AA inhibits Ca²⁺ entry via TRPC3, thus interfering with the Ca²⁺ requirement for cell proliferation and invasion.