Background
High-grade serous ovarian cancer (HGSOC) is the deadliest histotype of ovarian cancer, accounting for 70–80% of all deaths from the disease [
1]. In most cases, HGSOC is diagnosed at late stages, when the disease has already metastasized to distant organs within the peritoneal cavity. HGSOC most commonly metastasizes by transcoelomic dissemination during which cells detach from the primary tumor, either as single cells or as multicellular aggregates, and use the natural flow of peritoneal fluid to reach distant tissues [
2,
3]. Organs in the peritoneal cavity are covered by a protective layer of peritoneal tissue, made up of a submesothelial stroma composed of fibroblasts and extracellular matrix (ECM), and topped with a monolayer of mesothelial cells. Mesothelial cells form tight junctions between one another and serve a protective function [
4]. Two-dimensional in vitro assays are the standard methods to study the metastatic abilities of cancer cells in vitro (i.e.. adhesion, migration, and invasion) [
5]. However, communication between mesothelial cells, fibroblasts, and HGSOC plays an important role in the adhesion and invasion of cancer cells to the peritoneal tissue. To address this process, there has been a rise in the development of organotypic models of disease composed of a minimum of two different cell lines and ECM, with the goal of mimicking microenvironments and physical components that more closely resemble those observed in patients [
6]. For the study of ovarian cancer, organotypic models mimicking both the peritoneal tissue and omentum are mostly used, as these two sites are known to often be affected by late-stage disease [
2]. Peritoneal tissue models typically contain fibroblasts embedded in collagen I, as ovarian cancer cells have the highest level of adhesion and invasion in the presence of this ECM component [
7]; such fibroblasts are topped with a monolayer of mesothelial cells. Tri-dimensional organotypic models of disease have shown, in many cancers, to provide different results than 2D assays in terms of cellular behavior or in drug responses [
8‐
12].
Treatment options for ovarian cancer have remained stagnant for the last four decades, since the establishment and acceptance of platinum-taxane combination chemotherapy in the 1980s. Although the current standard of treatment, debulking surgery followed by platinum-taxane combination chemotherapy, has a 70–80% initial response rate, most patients will relapse with a chemoresistant disease. The 5-year survival rate after relapse is less than 48% [
1]. For this reason, developing new treatment options for these patients is of upmost importance. Previous work in our laboratory has demonstrated the potential of mifepristone (MF), a synthetic steroid that acts both, as anti-glucocorticoid and anti-progestin, as a potential treatment option for ovarian cancer. We demonstrated that MF blocks the growth of ovarian cancer cells [
13] and prevents their repopulation after chemotherapy [
14] by blocking cells in the G1 phase of the cell cycle as a consequence of increasing the levels of Cdk inhibitors and thus reducing the activity of Cdk2 [
15]. We have also shown that MF has antimetastatic potential and can slow the adhesion of highly metastatic cancers, including ovarian, breast, prostate and glial, by rearranging the distribution of fibrillar actin [
16], while also interfering with their migratory and invasive capabilities [
17]. MF has also been shown to inhibit the metastatic abilities of many cancers, including breast [
18,
19], gastric adenocarcinoma [
20], and cervical cancer [
21].
In this work, we demonstrate that differences in cellular behaviours can be observed between simplistic 2D assays and more complex 3D organotypic assays. Two-dimensional adhesion, migration, and invasion assays demonstrated, in two cases of patient-derived HGSOC cell lines, that cells representing an early-stage disease showed higher metastatic potential than those representing late-stage disease. Conversely, an adhesion assay using an organotypic model demonstrated higher adhesion capacities for the cells obtained at more advanced stage, which coincided with their higher tumorigenicity in vivo. Of relevance, MF inhibited the adhesion, migration, and invasion capacities of all cell lines studied regardless of their metastatic capabilities along disease progression.
Materials and methods
Cell lines, culture conditions and treatments
The HGSOC cell lines used were established from two patients and represent disease progression: PEO1, PEO4, and PEO6 from a first patient, and PEO14 and PEO23 from a second patient. From patient 1, PEO1 was originally collected from ascites after first treatment with cisplatin, 5-fluorouracil, and chlorambucil. PEO4 was collected from ascites 10 months after the initial treatment. The patient was treated once more with cisplatin and relapsed a final time, at which point PEO6 was collected from ascites. From patient 2, PEO14 was collected from the ascites prior to any treatments (i.e. chemonaïve). PEO23 was collected from the ascites 7 months after initial treatment with cisplatin and chlorambucil [
22]. With written consent from Dr. Langdon (Edinburgh Cancer Research Center, Edinburgh, UK), PEO1, PEO4, and PEO6 cell lines were obtained from Dr. Taniguchi (Fred Hutchinson Cancer Center, University of Washington, Seattle, WA, USA). PEO14 and its longitudinally patient-matched pair PEO23 were obtained from Culture Collections, Public Health England (Porton Down, Salisbury, UK). All cells were cultured in RPMI-1640 (Mediatech, Manassas, VA, USA) supplemented with 5% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA, USA), 5% bovine serum (Life Technologies, Auckland, New Zealand), 1 mM sodium pyruvate (Corning, Corning, NY, USA), 2 mM
l-Alanyl-
l-Glutamine (Glutagro™, Corning), 10 mM HEPES (Corning), 0.01 mg/ml human insulin (Roche, Indianapolis, IN, USA), 100 IU penicillin (Mediatech), and 100 µg/ml streptomycin (Mediatech). Cell culture was carried out at 37 °C in a humidified incubator with 95% air/5% CO
2 in standard adherent plastic plates.
WI38 fibroblasts were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and were cultured in the same medium as the HGSOC cells. LP9 mesothelial cells are normal diploid (karyotype 46XX), untransformed human mesothelial cells originally described by Wu and colleagues in 1982 [
23]. They were obtained from the ascites fluid of a patient with ovarian cancer. The cells were acquired from the Coriell Institute for Medical Research (Coriell, Camden, NJ, USA) and were cultured in a 1:1 combination of F-12 containing L-glutamine (Gibco, from Thermo Fisher Scientific, Waltham, MA, USA) and Medium 199 (Corning) and supplemented with 10% FBS and 0.4 µg/ml hydrocortisone (Sigma Chemical Co., St. Louis, MO, USA).
Mifepristone (MF; Corcept Therapeutics, Menlo Park, CA, USA) was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 4655 µM, and stored at − 20 °C. During treatment, MF was diluted into culture medium to reach a final concentration of 20 µM, which was previously deemed to be cytostatic [
24]. The final concentration of DMSO in culture medium was 0.43%; therefore, vehicle treated cells were exposed to 0.43% DMSO diluted in culture medium. Cells were pre-treated with either MF or medium containing DMSO (vehicle) for 72 h prior to plating and their corresponding treatment was maintained throughout the experiment.
Cell proliferation and doubling times
The doubling times for the five HGSOC cell lines (PEO1, PEO4, PEO6, PEO14 and PEO23) were repeated twice to confirm the accuracy of the results. For each experiment, cells were plated in 6-well plates at a density of 200,000 cells per well and left to attach for 72 h. This cell density ensured exponential growth of each cell line while preventing the cells from reaching 100% confluence over the course of the experiment. After 72 h, triplicate cultures were trypsinized, pelleted by centrifugation at 500
g for 5 min, and resuspended in the appropriate growth medium. An aliquot of each cell suspension was counted using the Muse™ Cell Analyzer and the Muse™ Count & Viability Assay Kit. A Count & Viability assay was then run on the Muse™ Cell Analyzer and the total number of cells per sample was collected. These results were considered to be the 0 h time-point. The same procedure was repeated every 24 h for a total of 96 h. Doubling times were calculated as previously described [
25] using a nonlinear regression analysis on exponentially growing cells.
Wound healing assay
Each HGSOC cell type was plated in 6-well plates at a density of 200,000 cells per well. Once cells reached approximately 80% confluency, a wound was created along the center of the well, using a 1000 µl pipette tip and a ruler as a guide to ensure the wound was straight and reproducible. Multiple images were then immediately taken of each well, along the wound, using an IN480 Series inverted biological microscope (United Scope LLC, Irvine, CA, USA). Cells were incubated at 37 °C for various time points up to 36 h. At each time point, images of the wounds were once again taken of each well. Using the OMAX Toupview software (United Scope), the wound width was measured four times per image to determine the wound average. Experiments were repeated with a pre-treatment of either 0.4% DMSO or 20 µM of MF for 72 h.
Boyden chamber assay
Migration
Twenty-four hours before the start of the experiment, cell culture medium was replaced with 0.1% FBS-containing medium. The next day, cells were plated in 0.1% FBS-containing medium in the upper chamber of a 6-well Boyden Chamber (BC) plate at a density of 200,000 cells per well. Ten percent serum-containing medium was added to the lower chamber to act as a chemoattractant. Cells were incubated at 37 °C for 30 h, after which non-migratory cells left in the upper chamber were removed using a sterile cotton swab. The upper chamber was washed with phosphate-buffer saline (PBS) and a second sterile cotton swab was used to ensure that all cells had been removed. Migrated cells were fixed with 4% PFA solution. Twenty 20 × field images were taken of each insert by fluorescent microscopy using a Leica DMi8 inverted fluorescence microscope and a Leica LAS X software (Leica Microsystems Canada, Concord, ON). Cells were counted in each image and the average number of cells per 20 × field was calculated. Experiments were repeated with a pre-treatment of either 0.4% DMSO or 20 µM of MF for 72 h.
Invasion
To study the invasion of HGSOC cells, the BC was performed in the same way as described for migration. However, 24 h before cells were plated, inserts were coated with a layer of extracellular matrix (ECM) (Sigma). The stock solution of ECM (9.11 mg/ml) was thawed at 4˚C and diluted to a working concentration of 0.6 mg/ml. Wells were coated with 500 µl of diluted ECM gel and incubated at 37 °C overnight. The next morning, excess ECM gel was removed and the rest of the experiment was performed in the same manner as the migration assay.
Visualization of migrating and invading cells using cytochemical double fluorescence staining
To improve the visualization of cell migration, cells were stained with a combination of Alexa Fluor®-594 Phalloidin (Life Technologies, Carlsbad, CA, USA) and SYTOX® Green (Molecular Probes, Eugene, OR, USA) or DAPI (Life Technologies), to stain the cytoskeleton and the nucleus respectively. Cells were permeabilized with 0.1% Triton X-100 in PBS for 5 min. To reduce background staining, cells were then incubated in a PBS solution containing 1% BSA for 20 min. The stock solution of Alexa Fluor® 594 Phalloidin was diluted from its 6.6 µM solution in PBS containing 1% BSA at a 1:40 ratio (5 µl stock solution in 200 µl PBS). Cells were incubated with diluted Alexa Fluor® 594 Phalloidin for 20 min. During the last 10 min of incubation, cells were exposed to either 4 µl of 50 µM SYTOX® Green (5 mM stock solution) or 300 nM (5 mg/ml stock solution) of DAPI (Life Technologies, Carlsbad, CA, USA) in PBS containing 1% BSA. Finally, the cells were washed with PBS and stored at 4 °C.
Adhesion assay to fibronectin
Twenty-four hours before the start of the experiment, 12-well plates were coated with a layer of fibronectin (Gibco). A stock solution of fibronectin (1.0 mg/ml) was thawed at 4˚C and diluted to a working concentration of 6 µg/ml in PBS. The next morning, excess fibronectin was removed from the wells and all wells were blocked with a PBS solution containing 1% BSA. Cells were then plated at a density of 100,000 cells/well and left to incubate for 0.5, 1, or 2 h. The cells were washed twice with PBS, fixed with 100% methanol for 30 min, and stained with a filtered solution of 0.5% (w/v) crystal violet (Sigma) for 10 min. Ten 20 × images were taken using an IN480 Series inverted biological microscope (AmScope) and the average number of adherent cells per 20 × field was calculated for each well. Experiments were repeated with a pre-treatment of either 0.4% DMSO or 20 µM of MF for 72 h.
Organotypic culture model
Adhesion assay
The peritoneal tissue is the main site for EOC metastatic lesions; therefore, an organotypic model was utilized based on the model developed by Kenny et al. [
7]. To study the adhesive capacity of the five HGSOC cell lines in a tri-dimensional (3D) environment, an organotypic model composed of WI38 fibroblasts embedded in collagen I topped with a monolayer of LP9 mesothelial cells was created. Four thousand WI38 cells were suspended in a solution of 2 ml of LP9 culture medium and 30 µg of collagen I (Corning), per single-well chamber slide. The fibroblasts were incubated in a humidified environment at 37 °C and 5% CO
2 for 4 h. Three-hundred and fifty thousand LP9 mesothelial cells were plated per chamber slide and incubated in a humidified environment at 37 °C and 5% CO
2 overnight, to form a confluent monolayer. HGSOC were incubated for 45 min with 5 µM of CellTracker™ Deep Red Dye (Invitrogen) diluted in serum-free medium. Stained HGSOC cells were then plated at a density of 250,000 per chamber slide and left to adhere to the LP9 monolayer for 2, 4, or 24 h. Cells were fixed with 4% PFA and stored at 4 °C. Experiments were repeated with a pre-treatment of either 0.4% DMSO or 20 µM of MF for 72 h.
Conditioned medium
Organotypic culture models were created as previously described and incubated with conditioned medium from all 5 HGSOC cell lines for 24 h. Medium was collected from each cell culture after incubating with the cancer cells for 24 h. Organotypic models were fixed with 4% PFA and stored at 4 °C.
Immunofluorescence
Cells were permeabilized using 0.1% Triton X-100 for 5 min. To reduce background staining, the cells were then incubated with a PBS solution containing 1% BSA for 20 min. Cells were incubated with the primary antibody diluted in 0.2% BSA at 4 °C overnight (see Additional file
1: Table S1 for the source and specific concentration of each antibody). Cells were washed twice with PBS for 3 min and then incubated with the secondary antibody for 30 min. Again, cells were washed twice with PBS for 3 min and stored at 4 °C. Images were taken using a Leica DMi8 inverted fluorescence microscope and a Leica LAS X software (Leica Microsystems) or a Cytation™ 3 Cell Imaging Multi-Mode Reader with Gen5 software (Biotek, Winooski, VT, USA).
Western blot analysis
The breast cancer cells MDA-MB-231 and MCF-7 cells were used as positive controls for glucocorticoid receptor (GR) and progesterone receptor (PR), respectively, as we have shown before [
25]. These cells as well as the HGSOC cells were washed with ice-cold PBS, scraped, collected, and centrifuged. Pellets were stored at -80˚C. NP40 lysis buffer was used to extract protein lysates and 25 µg of proteins per sample were resolved in 10% gels (TGX Stain-Free™ FastCast™ Acrylamide kit, Bio-Rad) via electrophoresis. A Trans-Blot
® Turbo™ Transfer System (BioRad) was used to transfer the resolved proteins to Immuno-Blot
® PVDF membranes. Membranes were blocked in 5% milk for 1 h and incubated at 4 °C overnight with primary antibodies for E-Cadherin, EpCAM, cytokeratin 7 (CK7), CD44, Vimentin, N-Cadherin (N-Cad), PR, GR, and β-actin. Membranes were then washed and incubated with secondary antibodies for 1 h. Protein detection was performed using a ChemiDoc Imaging System (BioRad) using chemilluminescence (Clarity Western ECL Imaging System, BioRad). See Additional file
1: Table S1 for the source and specific concentration of each antibody. Ultraviolet activation of the TGX stain-free gels on a ChemiDoc MP Imaging System (BioRad) was used to control for protein loading in addition to β-actin. The original membranes, which contain detailed information from where the immunoblots images were obtained, are shown in Additional file
9: Fig. S8.
Statistical analysis
All data represent means ± s.e.m. and statistical significance was defined as p < 0.05. One-way analysis of variance (ANOVA) followed by Bonferroni’s test was used to compare the means among three different cell lines. Two-way ANOVA followed by Bonferroni’s test was used to compare the means of groups receiving different treatments over time. Unpaired student’s t-test was used when comparing the means between two different cell lines. Experiments were repeated three times in triplicates.
Discussion
In this work, we characterized the metastatic abilities of two series of HGSOC cell lines, developed at different disease stages, and studied the ability of cytostatic doses of MF to inhibit these capabilities. The majority of ovarian cancer cases are diagnosed once the disease has already metastasized, emphasizing the importance of understanding and treating this phenomenon. Metastasis can be divided into three complex processes: adhesion, migration, and invasion. Migration involves the rearrangement of the cytoskeleton leading to an elongated morphology and increased contractility [
5]. Two types of cellular migration have been described: single cell and collective cell migration. In single cell migration, each individual cell undergoes physical changes before migrating. Single cell migration is often associated with EMT, in which E-cadherin is often downregulated indicating the loss of an epithelial phenotype [
26]. Collective cell migration describes the ability of cells to migrate as clusters, in which cells retain their cell–cell junctions and migrate as a collective group. This allows for migrating cells to carry along cells with a more epithelial phenotype, all while maintaining contact with one another. Collective cell migration has been associated with a more invasive phenotype in many cancers including breast [
35], thyroid [
36] and prostate [
37], with leading cells having a more mesenchymal phenotype than those in the rear. Contrary to what we expected, our results demonstrated that cells produced at late-stage disease were less migratory than their early-stage counterparts in 2D in vitro assays. Furthermore, PEO4 and PEO23 obtained at advanced disease were observed migrating as clusters of cells, suggesting that they undergo collective cell migration. These observations combined suggest that HGSOC may not undergo what has long been believed to be the classic form of metastasis, in which invasive cells undergo EMT, developing a more mesenchymal phenotype to enhance migration and invasion, before adhering and invading a distant site [
38]. Instead, although cells undergoing single cell invasion migrate more rapidly, cells undergoing collective cell migration are more efficient, as the crosstalk between cells helps to coordinate the direction of the migration of the group [
39]. This could explain the more rapid migration rates found in cells generated at early-stage disease (i.e. PEO1 and PEO14) and the observation that these seem to be undergoing single cell migration, although they may be less efficient to metastasize into a 3D organotypic model system. Due to collective cell migration being associated with increased metastatic potential, the ability of HGSOC cells to travel also as spheroids in the peritoneum could contribute to the aggressivity and low survival rate of the disease compared to other cancers.
Our results in terms of EMT biomarkers are in agreement with the latest guidelines and definitions on EMT stating that the transition between epithelial and mesenchymal phenotypes are incomplete leading to cells in an intermediary EMT state or EMT plasticity [
40]. This phenomenon can occur in ovarian cancer metastasis that takes place in a transcoelomic manner with the cells arranged as spheroidal structures that not necessarily need to entirely disintegrate into individual cells to attach to the peritoneal cavity.
The upregulation and downregulation of various adhesion molecules is important throughout the entire process of cancer metastasis and is involved in triggering conformational changes within cancer cells as well as determining the ECM protein cells preferentially adhere to [
41]. HGSOC has been shown to have a predisposition for the peritoneal tissue, composed of a single layer of mesothelial cells atop of sub-mesothelium of ECM, composed of many proteins including collagen I and fibronectin [
42]. Although both ECM proteins play a role in ovarian cancer metastasis, studies have shown that ovarian cancer cells adhere more readily to collagen I than fibronectin [
6], as well as demonstrate an increased ability to migrate when cultured on collagen I and a decreased ability when cultured on fibronectin [
43]. This coincides with the differences observed between simple adhesion assays on fibronectin and organotypic models containing collagen I. Adhesion to fibronectin showed minimal differences in adhesion rates between cell lines (except for the only chemonaïve cell line, PEO14 that had superior adhesion when compared to the other cell lines). It was only when plated onto the more complex organotypic models that major differences between the adhesive capabilities of cell lines began to be observed. In this case, there was a clear increase in adhesion along disease progression in both cell line groups. It has been shown that single cell migration is associated with weak adhesion to ECM, while collective cell migration involves strong cell–cell junctions and interactions with underlying ECM in order for the cellular group to move forward [
44]. This could explain the reason for the differences observed in adhesion to the organotypic model, as cells with the higher adhesion rates seems to be cells with a collective cell migration pattern.
Hyaluronic acid (HA) is another major ECM protein that binds the adhesion molecule CD44 and has been implicated in ovarian cancer metastasis; increased levels have been associated with poor ovarian cancer outcome [
45,
46]. Immunofluorescent staining of CD44 in a wound healing assay showed that both PEO4 and PEO23, cells developed at late-stage disease, showed CD44 positive cells when migrating into the wound, presumably to facilitate adhesion. This pattern is consistent with a collective cell migration pattern, in which leading cells are primed to undergo migration and are the first to adhere. PEO1 and PEO14, cells produced at early-stage disease, demonstrated extremely different levels of CD44. PEO1 cells were found to be CD44 positive, while the majority of PEO14 cells were found to be CD44 negative, which is somewhat controversial as both were generated at early-stage disease with the only difference that PEO1 are chemosensitive whereas PEO14 are chemonaïve. However, when analysing the data on western blot, we observed high expression of CD44 only in PEO1 and PEO6 cells, but neither in PEO4 in the case of patient 1 nor in cells isolated from patient 2 (PEO14 and PEO23). The highly specific localization of CD44 at the border of the wound in PEO4 and PEO23 cells (Fig.
5C [ii], D [ii]) suggests a response to the stimuli provided by the scratch, while the CD44 detected by western blotting might correspond to the basal expression of this protein in proliferating cells in the absence of the specific challenge of the scratch.
In this study, we demonstrated that behaviors of HGSOC cells differ between 2D simplistic in vitro adhesion assays vs. 3D organotypic and more complex in vitro adhesion assays. Organotypic models have the goal of providing cells with an artificial tumor microenvironment, which is not present in standard in vitro assays. Cell–cell communication between mesothelial cells, fibroblasts, and ovarian cancer cells are relevant to achieve metastasis to the peritoneal tissue. For example, cancer-associated fibroblasts promote HGSOC metastasis [
47,
48], while mesothelial cells inhibit the adhesion and invasion of HGSOC cells [
49]. Therefore, the presence of both cell types is important in observing HGSOC cell behaviors more closely resembling those encountered in patients. Adhesion assays in organotypic models showed an increase in adhesion rate along disease progression in both series of cell lines (i.e. in both patients). Furthermore, through the addition of an immunofluorescence staining with calretinin to label the mesothelial and vimentin to label the fibroblasts, a displacement of mesothelial cells was observed. This is concordant with what it was found by Kenny et al. that showed that mesothelial cells are absent under the peritoneal tumor mass in wide spread ovarian cancer disease [
50]. In our results, PEO4, PEO6, and PEO14 were all observed disrupting the mesothelial cell monolayer after 24 h of incubation. Although mesothelial cell displacement was not dependent on disease progression in this case, it was associated with tumorigenicity in vivo. Our laboratory recently demonstrated that only PEO4, PEO6, and PEO14 are capable of forming tumors in vivo, while PEO1 and PEO23 show no signs of tumorigenicity even after 14 months of inoculation into the peritoneal cavity of immunosuppressed mice [
34]. The disruption of the mesothelial monolayer is an important step of HGSOC metastasis, as cancer cells can adhere more easily to the underlying ECM than to the mesothelial cells themselves. The exact mechanism for mesothelial cell displacement is still unknown, however it was suggested that it involves the removal of mesothelial cells by ovarian cancer cells using force [
51]. Of interest, when organotypic models were incubated with conditioned media from PEO4, PEO6, or PEO14 cells for 24 h but without the cancer cells, the LP9 monolayer was found to be disrupted in a similar fashion as when the cells were present and adhered. These results suggest the presence of secreted factors that could be communicating with the mesothelial cells, promoting their displacement. It has been shown that the presence of pro-inflammatory cytokines can cause structural changes to mesothelial cells. Particularly, tumour necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and hepatocyte growth factor (HGF) have all been shown to cause mesothelial cell retraction, exposure of the underlying ECM, while facilitating the adhesion of cancer cells to the mesothelial cell monolayer [
52‐
55]. Transforming growth factor-β1 (TGF-β1) has also been suggested to be involved in the progression of peritoneal metastasis by causing fibrosis of mesothelial cells, decreasing the integrity of the mesothelium, and increasing the secretion of chemokines, homing HGSOC cells to the peritoneal tissue [
56]. Furthermore, exposure to various pro-inflammatory cytokines causes variations in the level of adhesion molecules present on the surface of mesothelial cells. Interestingly, IL-1β has been shown to increase levels of CD44 while TNF-α causes an opposite effect [
52].
The formation of HGSOC spheroids is thought to serve a protective function while traveling in the peritoneal fluid, increasing cell survival and resistance to anoikis [
57,
58]. As the spheroids reach the peritoneal tissue, communication between mesothelial cells and HGSOC cells primes the environment, promoting adhesion and invasion of cancer cells [
59,
60]. PEO6 spheroids demonstrated increased adhesion capacity when exposed to adherent PEO6 conditioned media, implying the requirement of secreted factors for the spheroids to adhere to the organotypic system representing a reductionist model mimicking the peritoneal wall.
The effect of MF on the adhesion, migration, and invasion abilities of each cell line was studied in this work. It was found that regardless of the variations in metastatic abilities of each cell type, MF was able to inhibit the migration and invasion in wound healing and Boyden chamber assays, adhesion to fibronectin, and to the 3D organotypic model system. The ability of MF to inhibit these properties was not dependent on disease progression nor platinum-sensitivity, making it a good potential treatment candidate for progressive HGSOC disease. The exact mechanism behind this inhibition is still unknown; however, a previous report from our laboratory demonstrated that MF inhibits the adhesion of cancer cells representing ovarian, breast, prostate, and nervous system cancers, through the redistribution of fibrillar actin into cortical actin ruffles that are not adherent, thus diminishing the surface of cells with adhesion capacity [
16]. This is a mechanism that could also affect the ability of cells to migrate and invade, as these processes involve front to rear polarization and actin cytoskeletal rearrangement, in order for cells to move forward [
39]. When lamellipodia cannot establish themselves firmly, they tend to retract towards the center of the cell, halting the migration process [
61]. Stress fibers are essential for the adhesion of cells to the substratum, as well as the morphological changes undergone during migration. The loss of stress fibers in MF-treated cells could also be a reason for the reduction in migration capacity [
62]. Cells undergoing collective cell migration undergo similar actin cytoskeletal rearrangement as cells undergoing single cell migration, just with the addition of adherens junctions between cells [
39]. Therefore, this hypothesis for the mechanism of action of MF in the inhibition of migration and invasion would pertain to cells undergoing both single cell and collective cell migration. Finally, we demonstrate that spheroidal HGSOC structures are dissociated by MF into small clusters and isolated cells in the presence of conditioned media. This suggests that MF may target both individual cells as well as spheroidal structures floating in the peritoneal cavity and prevent their adhesion to the peritoneal wall.
It is known that MF is both an antiglucorticoid and an antiprogestin agent [
25]. We previously demonstrated that the inhibition of growth of cancer cells by MF is not dependent on the presence of PR [
25]. Likewise, the anti-adhesive, anti-migratory, and anti-invasive effect of MF may not rely on progesterone receptor (PR) expression as the levels of PR-A an PR-B isoforms are highly variable among the cell lines (Additional file
8: Fig. S7 upper panel), while MF was equally efficient inhibiting those processes. Notwistanding, a role of PR-A and PR-B in mediating the effect of MF cannot be dismissed as it was shown in breast cancer cells that they contribute to cell migration through interactions with focal adhesion kinase complexes [
63]. Other putative targets of MF, the glucocorticoid receptors (GRα and GRβ), show variable levels of expression (Additional file
8: Fig. S7 lower panel), yet MF had a consistent efficacy among all of cells in terms of its anti-adhesive, anti-migratory and anti-invasive effects suggesting their lack of involvement. This concept has however to be further tested as GRα, despite with different levels, is expressed in all cell lines, and GRβ, whereas expressed in only some cells, has been shown to bind MF and drive transcription [
64,
65].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.