A novel non-sulphamoylated 2-methoxyestradiol derivative causes detachment of breast cancer cells by rapid disassembly of focal adhesions

The MCF-7, MDA-MB-231 and BT-20 cell lines were obtained from Cellonex (Johannesburg, South Africa). Most chemical reagents including glutaraldehyde, crystal violet, bovine serum albumin, Tween-20, 4′,6-diamidino-2-phenylindole (DAPI), and Fluoromount aqueous mounting fluid were obtained from Sigma Aldrich (Darmstadt, Germany). NuPAGE LDS sample buffer, MOPS running buffer and 4–12% NuPAGE SDS-PAGE gels were obtained from Invitrogen (Johannesburg, South Africa). The lactate dehydrogenase assay kit II was obtained from Biovision (Johannesburg, South Africa). Total FAK and phospho-S732 FAK antibodies were from Abcam (Pretoria, South Africa). Secondary HRP tagged antibodies were from KPL (Gaithersburg, USA) while the polyvinylidene difluoride (PVDF) membrane was from Amersham (Johannesburg, South Africa). Clarity Western ECL substrate was from Biorad (Johannesburg, South Africa).

MCF-7 and MDA-MB-231 cells were cultured in GlutaMAX DMEM medium containing 10% fetal calf serum (FCS), while BT-20 cells were cultured in a 1:1 (v/v) DMEM:HamF12 mix containing 10% FCS, and glutamine.

To determine the effect of the compounds on cell viability we used crystal violet staining to determine cell numbers. Exponentially growing cells were seeded in 96-well tissue culture plates (5 × 10³ cells/well for MCF-7 and MDA-MB-231 cells, and 1 × 10⁴ cells/well for BT-20 cells) and incubated overnight to allow attachment. The following day cells were exposed to either DMSO (v/v%) or compounds diluted in 200 µl medium. At termination of the experiment cells were fixed with 1% (v/v) glutaraldehyde for 15 min at room temperature (RT). The glutaraldehyde was discarded and cells were stained using 0.1% (w/v) crystal violet at RT for 30 min. The 96-well plate was washed by rinsing the plate in water. Cells were permeabilised and crystal violet solubilised by adding 0.2% Triton X-100 at RT for 30 min. The absorbance of the resulting solution was measured at 570 nm. Six technical repeats were included in each experiment and at least three independent experiments were performed. Graphs represent the average of independent experiments with the error bars representing standard error of the mean. Student’s t-tests were performed to determine statistical significance.

To quantify the number of cells that round up after exposure to the compounds, cells were seeded in 24 well plates (5 × 10⁴/well) and incubated overnight to attach and spread. Cells were exposed to the compounds after having been photographed and were subsequently photographed at indicated timepoints. At least three images were captured per well and three wells per condition were used in each experiment. Photos were taken at 10× magnification using a Zeiss Primovert microscope and a Zeiss Axiocam ERc5s camera (Zeiss, Oberkochen, Germany). Graphs represent three independent experiments with error bars indicating standard error of the mean.

To quantify the cytotoxicity of our compounds on breast cancer cells the presence of lactate dehydrogenase (LDH) in the medium was analysed using the LDH cytotoxicity assay kit II from Biovision according to the manufacturer’s protocol. In short, MCF-7 and MDA-MB-231 cells were seeded in 96-well cell culture plates (5 × 10³/well) and incubated overnight before exposure to 5 µM EE-15-one for the indicated timepoints. Medium from samples (200 μl) was transferred and centrifuged at 2400g for 10 min. Afterwards, 10 μl was transferred to a clear 96-well plate. LDH reaction mix (100 μl) was added to the samples and incubated for 90 min at RT. The absorbance was read at 460 nm, with a reference wavelength of 630 nm. Three independent experiments were performed each with three technical repeats. Graphs represent the average of independent experiments with error bars showing standard error of the mean. Statistical significance was determined using the student’s t-test.

To determine the effect of compounds on initial cell adhesion and spreading on a rigid surface we performed crystal violet-based adhesion assays. MCF-7 cells were trypsinized and kept in suspension in complete medium. Cells were pre-treated in suspension with either DMSO or 5 µM EE-15-one for 2 h with continuous gentle agitation to prevent cells from aggregating. Cells were subsequently seeded onto plastic culture dishes (96 well, 4 × 10⁴ cell/well) and at indicated time points loose cells were removed while attached cells were fixed in 1% (v/v) glutaraldehyde and processed for crystal violet staining as described previously. Total cell numbers were obtained by collecting 4 × 10⁴ cells by centrifugation and fixing these in glutaraldehyde followed by staining using crystal violet. Three independent experiments were performed each with six technical repeats. Values were calculated as averages with error bars representing standard error of the mean. Statistical significance was calculated using a two-tailed student’s t-test.

Cells were seeded on glass coverslips and incubated overnight to attach and spread. Subsequently, cells were exposed to 5 µM EE-15-one or DMSO for different times before they were fixed in 2% (w/v) paraformaldehyde for 20 min, washed with 1× PBS, permeabilised with 0.2% (v/v) triton X-100 for 5 min, and washed with 1×  PBS. Coverslips were subsequently blocked with 2% (w/v) bovine serum albumin (BSA) in 1× PBS for 1 h, incubated with primary antibodies against FAK for 1 hat RT. Coverslips were washed once more with 1× PBS, and incubated with the secondary antibodies DAPI, and fluorescently conjugated phalloidin for 1 h at RT. After washing with 1× PBS, coverslips were mounted in mounting fluid. Slides were examined using a Zeiss LSM800 Meta confocal microscope furnished with a 63× magnification oil objective.

Glass slides (2 cm diameter) were sterilised in 70% ethanol and air-dried in a sterile flow cabinet. The slides were placed into a sterile six well cell culture plate after which cells were seeded on top of the slides at a density of 2.5 × 10⁴ cells per slide. The plates were incubated at 37 °C overnight to allow for attachment. The following day cells were transfected with a paxillin-GFP (Addgene) construct along with a LifeAct-RFP (Addgene) construct. Mirus T2020 transfection reagent was used according to the manufacturer’s protocol. The following day the slides were mounted in the Zeiss life imaging chamber and the microscope incubator set to 37 °C and 5% CO₂. The Zeiss LSM800 Meta confocal microscope was used to image the slides every 30 s for 2 h at 63× magnification. The images and videos were processed using ImageJ (FIJI) software [9].

For western blotting 2 × 10⁵ cells were seeded in six well cell culture plates and incubated overnight for attachment. After exposure of EE-15-one for indicated timepoints the medium was removed and collected, and cells were carefully washed with ice-cold 1× PBS. Detached cells were collected by centrifugation and added to adherent cells. Cells were lysed in hot (80 °C) NuPAGE LDS sample buffer supplemented with 2.5% (v/v) β-mercaptoethanol. Cells were scraped and heated at 80 °C for 10 min and centrifuged to remove cell debris. Proteins were separated on 4–12% NuPAGE gels in MOPS running buffer and transferred to polyvinylidene difluoride (PVDF) membranes. After blotting, membranes were blocked in 2% (w/v) BSA and subsequently incubated with anti- β-tubulin, anti-FAK or anti-pFAK primary antibody in BSA for 1 h. The membrane was subsequently washed three times in PBS-T (1× PBS containing 0.5% Tween-20) and incubated with HRP-labelled secondary antibodies for 1 h followed by three more PBS-T washes. Protein expression was analysed using chemiluminescence by incubating membranes in ECL substrate and imaging the membrane using the ChemiDoc MP system (BioRad). Analysis was performed using the ImageLab software from BioRad. Bands were quantified and normalized against β-tubulin protein bands. The graphs represent three independent experiments with error bars as standard deviation. Statistical significance was determined using a two-tailed student’s t-test.

To analyse the effect of EE-15-one (Additional file 1: Figure S1) on breast cancer cell line survival, dose response curves were generated by quantifying cell numbers using crystal violet staining and spectrophotometry. ER positive MCF-7, ER negative MDA-MB-231, and ER negative BT-20 cells were exposed to increasing concentrations of EE-15-one for 24 h after which cell numbers were quantified (Fig. 1a). Exposure to micromolar concentrations of EE-15-one led to significant reductions in cell numbers after 24 h. EC₅₀ calculations show that MCF-7 cells were most sensitive to EE-15-one with a half maximal effective response (EC₅₀) of 1.89 µM whereas MDA-MB-231 cells (EC₅₀ = 3.29 µM) and BT-20 cells (EC₅₀ = 6.86 µM) were somewhat more resistant.

To compare EE-15-one with other sulphamoylated 2ME2 derivatives that induce cell cycle arrest, dose response curves were performed in MCF-7 cells for EE-15-one and its sulphamoylated homologue, ESE-15-one (Fig. 1b). After 24 h exposure to 5 μM EE-15-one, almost 100% of MCF-7 cells were lost while the maximal cell loss induced by ESE-15-one reached only 43.3%. Therefore, EE-15-one has a significantly greater effect on cell survival than ESE-15-one after 24 h exposure.

To identify the unique structural components of EE-15-one that make it the only non-sulphamoylated 2ME2 derivative to cause significant reduction in cell numbers, closely related derivatives of EE-15-one were analysed for their ability to inhibit breast cancer cell survival. EE-one lacks the alkene group on C15 but contains the ketone group on C17 while EE-15-ol has a hydroxyl group instead of a ketone group on C17 but does contain the alkene group at C15 (Additional file 1: Figure S1). Their effect on MCF-7 cell survival was measured and compared to EE-15-one (Fig. 1c). Neither EE-15-ol nor EE-one had any significant effect on cell numbers after 24 h (vehicle control vs. all concentrations of EE-one or EE-15-ol, P > 0.05) while EE-15-one exposure led to almost complete cell loss at low micromolar concentrations. Thus, both the ketone group at C17 and the alkene group at C15 are needed for the effect of EE-15-one to be present. Our results on MCF-7 cells closely mimic those obtained in other cell lines for these compounds [7, 8].

To discover the mechanism through which EE-15-one causes the observed rapid loss of cells, we analysed the effect of EE-15-one on cell morphology. Light microscopy images of MCF-7, MDA-MB-231 and BT-20 cells showed that cells became rounded or detached within 24 h after exposure to 5 μM EE-15-one (Fig. 2a). To quantify this observation, MCF-7 and MDA-MB-231 cell morphology were analysed using light microscopy and the number of spread cells and rounded cells were quantified at different times after exposure (Fig. 2b, c). MCF-7 cells exposed to 5 µM EE-15-one rapidly lost their spread morphology with 60% being rounded after 2 h and almost all cells rounded or detached within 4 h. In contrast, neither DMSO nor the sulphamoylated version of EE-15-one, ESE-15-one, induced any significant rounding during this time. In MDA-MB-231 cells 5 µM EE-15-one exposure also resulted rapid cell rounding with almost all cells rounded or detached after 6 h (Fig. 2c). The data suggest that EE-15-one inhibits the ability of cells to remain spread on a rigid surface. Such rapid changes in cell morphology could be the result of the inhibition of cell attachment or of necrosis.

To determine if EE-15-one impacts on the adhesive ability of cells, initial adhesion and spreading were analysed in MCF-7 cells. To measure the effect of EE-15-one, suspended cells were either treated for 2 h with 5 µM EE-15-one before they were seeded and adhesion was assayed, or 5 µM EE-15-one was added to cells as they were seeded after which they were assayed. Control cells were exposed to DMSO before seeding (Fig. 3). Cells treated for 2 h with EE-15-one lost all adhesive ability with only 5% of seeded cells adhering after 3 h in contrast to 40% of cells adhering when exposed to DMSO. Interestingly, even exposure to EE-15-one at the time of plating led to significant inhibition of adhesion with no more than 15% of plated cells adhering after 3 h. This data suggests that the loss of cells observed after exposure to EE-15-one may be due to an acute inhibition of cell–substrate adhesion.

Acute loss of adhesion can be due to rapidly induced necrosis. To determine if necrosis was induced by EE-15-one exposure, MCF-7 cellular health was assessed by analysing lactate dehydrogenase (LDH) leakage from cells into the medium which acts as an indicator of a compromised cell membrane signifying necrosis or late stage apoptosis (Fig. 4). Exposure of MCF-7 cells to 5 μM EE-15-one did not lead to significant increases in LDH activity up to 6 h even though all cells were rounded or detached at this time. Only after 24 h was a small but significant increase in LDH activity (20% of positive control) measured. Similarly, LDH activity is only increased significantly after 24 h in MDA-MB-231 cells suggesting that the majority of cells were still alive even though they were rounded or detached (Additional file 2: Figure S2). Therefore, the data suggests that EE-15-one causes rapid cell rounding and detachment most likely through inhibition of cell–substrate adhesion but does not induce rapid and massive necrosis or cell death even after all cells were detached.

EE-15-one inhibits cell adhesion and induces cell detachment. Therefore, it is likely that it impacts on focal adhesion function. To determine if EE-15-one does impact on focal adhesions we visualised these along with the actin cytoskeleton. To visualise focal adhesions and the actin cytoskeleton, adherent MCF-7 cells were exposed to DMSO or 5 μM EE-15-one for 2 h before they were fixed and focal adhesions were visualised with an anti-FAK antibody (green) while the actin cytoskeleton was visualised using fluorescently labelled phalloidin (red) (Fig. 5a). DMSO-treated cells possessed multiple focal adhesions throughout the basal membrane which were connected to well-defined, thick actin stress fibres. However, EE-15-one treated cells were poorly spread with long protrusions. No clear FAK positive structures were visible and instead a general cytoplasmic staining was observed. The actin cytoskeleton was almost completely lost with staining only visible close to the cell membrane and no observable stress fibres. Since the effect of EE-15-one was rapid, real time imaging was used to investigate the dynamics of the actin cytoskeleton and the focal adhesions after EE-15-one exposure (Fig. 5b, Additional file 3: Video S3). MCF-7 cells were transfected with paxillin-GFP and actin-mRFP constructs and were filmed using confocal microscopy. Still images of co-transfected MCF-7 cells show that within 1 h after exposure to 5 μM EE-15-one, actin fibres were becoming severed and stress fibres started thinning. After 1.5 h most stress fibres were lost with only cortical actin fibres remaining. At the same time, focal adhesions rapidly reduced in size so that at 1.5 h many focal adhesions had completely disassembled. Together, these data suggest that EE-15-one induces the disassembly of focal adhesions and the loss of tension through actin stress fibres. To analyse if EE-15-one exposure leads to the inhibition of focal adhesion signalling, FAK Y397 phosphorylation, which acts as a marker for FAK activation, was quantified by western blot (Fig. 5c). Western blot analysis using antibodies targeting phosphorylated FAK, total FAK and β-tubulin as loading control were used to assess the phosphorylation status of FAK in control cells and cells exposed to 5 µM EE-15-one for 1–4 h (Fig. 5c). The results show that EE-15-one significantly reduced FAK phosphorylation even after 1 h exposure to EE-15-one. Phosphorylation remained diminished throughout the time course while total FAK levels remained constant. Together, these data suggest that EE-15-one impacts on cell adhesion by abrogating focal adhesion function along with a loss of tension via the actin cytoskeleton leading to cell rounding and detachment.

EE-15-one exposure leads to rapid cell detachment from the tissue culture surface which is dependent on the formation of functional focal adhesions. Whether EE-15-one directly impacts on focal adhesions or on the actin cytoskeleton is unclear from the previous experiments. Therefore, we tested the ability of EE-15-one to interfere with adhesion in a setting in which integrin dependent adhesion was absent. Epithelial cell adhesion in three dimensional settings such as cell spheroids does not depend on cell–substrate adhesions mediated by integrins but rather on Ca²⁺-dependent cadherin adhesions or adherens junctions [15]. Importantly, both integrin based focal adhesions and cadherin based adherens junctions link to the actin cytoskeleton. Therefore, we hypothesised that if EE-15-one is an actin disruptor, cancer cell spheroids would be affected after exposure to EE-15-one. Due to a lack of effective spheroid formation using MCF-7 and MDA-MB-231 cells, BT-20 cell spheroids were generated by growing cells in non-adhesive cell culture plates. BT-20 cells formed compact, well-defined spheroids over 4 days (Fig. 6a, control). To show that these spheroids do indeed depend on Ca²⁺ dependent adhesions, they were cultured for 24 h in the presence of the Ca²⁺ chelator EGTA (Fig. 6a, low Ca²⁺). The well-defined spheroid border was lost, and numerous single cells were spread around the remnants of the spheroid suggesting that chelating Ca²⁺ led to reduced adhesion within the spheroid. However, when spheroids were cultured with EGTA and an excess of Ca²⁺, the border of the spheroid remained well-defined and the spheroid was intact (Fig. 6a, high Ca²⁺). Therefore, BT-20 spheroids are assembled by Ca²⁺ dependent cell–cell adhesion. To test if EE-15-one inhibits cell–substrate adhesion or rather impacts on the actin cytoskeleton, BT-20 spheroids were exposed to DMSO as control, 5 μM EE-15-one, 1 μM ESE-15-one as a control for compound exposure and 500 nM paclitaxel as positive control (Fig. 6b). Images of spheroids exposed to DMSO and EE-15-one showed well-defined spheroids with a smooth spheroid border that remained similar in size over the measured timeframe. In contrast, ESE-15-one exposed spheroids lost the smooth spheroid edge and diminished in size. Spheroid volumes were quantified at the time of exposure, and 3 and 11 days after exposure. Interestingly, while BT-20 cells grown in monolayer detached significantly within 24 h after exposure to 5 μM EE-15-one (Figs. 1, 2a), exposure of spheroids to 5 μM EE-15-one had no effect on spheroid volume at any time. In contrast, the sulphamoylated derivative ESE-15-one caused a significant reduction in spheroid volume similarly to the paclitaxel positive control. Therefore, the spheroid volume data suggests that EE-15-one is affecting cells attached via cell–substrate adhesion while it does not affect cells adhering via cell–cell adhesion.