Background
Cancer metastasis and burden of secondary tumors are the most common causes of mortality for many patients, accounting for nearly 90% of cancer-related deaths [
1]. One key aspect of metastasis is the invasive capacity of the cells, which is mainly driven by cell motility [
2]. Cancer cell motility, in turn, is heavily dependent on changes in tumor cell morphology caused by dynamic modifications in the polymerization of actin leading to rearrangements of the cytoskeleton [
3]. The changes in cellular morphology and their impact on motility are associated with changes among epithelial and mesenchymal phenotypes, a process known as epithelial-to-mesenchymal transition (EMT). The transition from an epithelial to a more mesenchymal state is linked to morphological modifications, loss of tight junctions, remodeling of the cytoskeleton, and acquisition of migratory and invasive capacities [
4]. Such migratory and invasive capacities are commonly assessed by a variety of experimental approaches which have been amply described [
5].
The rationale for classical anti-cancer therapy has long been to target cell proliferation at the primary site without, unfortunately, discriminating cancer cells from normal cycling cells [
6]. This approach was improved by the introduction of targeted therapies and immunotherapies that brought about reduced toxicities as they target cancer cells while sparing normal cells [
7,
8]. However, to treat cancer more effectively, we should further focus on preventing the formation and growth of metastatic carcinoma cells. We should consider that inhibition of migration, associated with the process of metastasis, might be as important as inhibition of cell proliferation. Agents that negatively influence
both mechanisms might provide a novel tool to fight cancer, in particular if they inhibit cell proliferation at the sites of metastasis while preventing migration of such cells to new niches.
Previous work in our laboratory has shown that the prototypical member of the family of antiprogestins, mifepristone (MF), can efficiently inhibit the growth of cancer cells of ovarian, breast, prostate, and glial origin, all known for their high metastatic potential [
9]. We demonstrated that the anti-cancer effect of MF does not require the presence of progesterone receptors [
9], involves cell cycle arrest at the G1 phase of the cell cycle associated with the inhibition of cyclin-dependent kinase Cdk2 [
10,
11], and triggers cellular stress and autophagy, making it useful in combination therapies with proteasome inhibitors and autophagy blockers [
12]. Furthermore, we provided evidence that MF interferes with the adhesive capacity of cancer cells by altering fibrillar actin (F-actin) distribution and promoting the formation of membrane ruffling [
13], represented by sheet-like membrane protrusions devoid of adhesive properties [
14].
In this work, we demonstrated that MF inhibits migration and invasion via standard approaches, and validated and enhanced a method for visualizing migratory and invasive cells upon double fluorescence cytochemical staining. Labeling migratory and invasive cells with a fluorescent probe linked to Phalloidin allowed us to observe the changes in F-actin distribution while cells are migrating and invading. Furthermore, the simultaneous labeling of the nucleus with a second fluorescent agent capable of binding DNA added further detail to visualize the position of the nucleus relative to the cytoplasm while cells are migrating and invading. The double fluorescence cytochemical labeling increases the value of two well-known approaches—the wound healing and Boyden chamber assays—to study cell movement by simple staining, increasing the level of detail of cell morphology in motion.
Discussion
In this work we focused on studying whether MF operating at cytostatic doses, while altering the morphology of the cytoskeleton and reducing cellular adhesion, also negatively impacts the migratory and invasive capacities of the cancer cells. Furthermore, we validated these approaches using an enhanced version of standard migration and invasion assays adding an extra, yet simple, visualization step involving double fluorescence staining of the cytoskeleton and the DNA, upon cytochemical reactions.
The use of fluorescent labeling of the cells to assess their migration and invasion capacities is not novel. However, these methods usually rely on the in vivo labeling of the cells with vital fluorescent stains that are retained by the cells for a number of hours [
5,
21], adding the confounding factor that another chemical is added to the cells while assessing mobility. In our work, we stained the cells with fluorescent stains post-fixation, providing certainty that the staining did not influence migration/invasion. Another advantage of the method is that it does not involve an immune reaction: it is based on the affinity of Phalloidin to F-actin, and of SYTOX® Green or DAPI to DNA, making the approach easy to perform, robust, and straightforward. Finally, the double staining provides extra-information on the morphology of the cytoskeleton while cells are migrating under different conditions. The addition of fluorescence labeling uncovered varying migration patterns between different cancer cell lines. Certain cell lines were found to undergo single-cell migration, while others were found to undergo collective-cell migration (rev. in [
17‐
19]). Notably, staining F-actin and DNA allowed the distinction of these two types of migratory patterns, as well as the recognition of cell lines with a mixed pattern of migration.
Moreover, in the assays involving movement through a membrane, the double staining allows for detailed visualization of how cellular compartments move through the 8 μm membrane pores, giving precise information on the localization of the nucleus relative to the movement of the cytoplasm. This is important as the positioning and movement of the nuclei are essential for the process of migration and, hence, invasion [
22]. For instance, whereas in bi-dimensional models the nucleus is often positioned in the back of the cell, during tri-dimensional migration the nucleus may be positioned in the front or back of the cell depending on cellular type (rev. in [
23]) (Additional file
2: Figure S2). During metastasis, the nucleus leading edge is the main barrier when migrating through tight spaces, such as when cancer cells pass through an endothelial cell layer. Supporting this concept, it was demonstrated with leucocytes that the localization of their nuclear lobes at the forefront of the cells act as a ‘drilling’ device by bending endothelial actin [
24].
Finally, another feature that was enhanced using dual fluorescence along tri-dimensional migration (Boyden chamber assay) was the capacity to detect modifications in the distribution of F-actin among the cytoplasm and the nucleus. This was evident in SKOV-3 and LNCaP cells in which treatment with MF increased the number of cells showing yellow fluorescence in the nucleus, signifying an overlapping of SYTOX® Green Nucleic Acid Stain binding DNA (green) with that of AlexaFluor®594 Phalloidin binding F-actin (red) (Additional file
3: Figure S3). This feature is important to point out as F-actin can be visualized in the cytoplasm and the nucleus under particular conditions as a result of its movement via transport molecules (rev. in [
25]). For instance, nuclear actin has been shown to be important in chromatin remodeling and organization [
26], and during cell death [
27], highlighting the relevance of its location, in particular, when exposing cells to cytotoxic agents.
Expectations suggest that the impairment in the adhesive capacity of cells exposed to MF should negatively impact their dissemination capacity, because both de-adhesion and adhesion are critical to cancer cell metastasis, since they reflect the initial detachment from the primary tumor site and the re-attachment leading to re-growth at a secondary location. In addition, cell migration is characterized by cyclic detachments of the rear of the cells and attachments to the front, which propel the cell forward. The turnover rate of these adhesions and de-adhesions is critical for migration, and a relationship has been shown between cell de-adhesion and rate of migration [
28]. We previously demonstrated that when cancer cells are exposed to cytostatic concentrations of MF, they suffer significant decrease of their ability to attach to various extracellular matrix proteins while suffering major alterations in the distribution of F-actin associated with an increase in the formation of actin ruffles with no adhesion capacity [
13]. These results were recently confirmed in two ovarian cancer cells lines in which MF caused a decrease in the visualization of stress actin fibers, with a concomitant increase in cortical actin and reduced adhesion to extracellular matrix proteins [
29]. MF also retained the anti-adhesive properties against human melanoma cells when given in combination with doxycycline, aspirin, and lysine [
30]. Furthermore, it was demonstrated that the major metabolite of MF, a mono-demethylated derivative termed metapristone [
31], is capable of diminishing adhesion of HT-29 colorectal adenocarcinoma cells to human extracellular matrix and to umbilical vein endothelial cells (HUVECs) [
32].
The effect of MF reducing cellular adhesion and changing cellular morphology, from their original epithelial shape towards a spindle-like appearance, suggests that the speed of migration may also be positively affected, likely involving EMT [
33]; however, contrary to the expectation, in all cancer cells studied, treatment with MF significantly attenuated migration of cells in wound healing assays. Firstly, we validated the results of the wound healing assay by showing that cells transiting M phase seem to only be found away from the wound, but not at the wound site, as demonstrated by the expression of the mitotic marker pHH3 (Additional file
4: Figure S4) [
34,
35]. As mitoses are usually coupled to cell proliferation [
36], labeling with pHH3 could prevent the need to use serum starvation or mitotic poisons to block cell division without killing the cells and, at the same time, allows closure of the wound to be attributed to migration and not cell proliferation [
21,
37]. Secondly, we validated the negative action of MF on migration in a tri-dimensional (Boyden chamber) assay. MF was capable of diminishing cell migration through the 8 μm pore polycarbonate membrane as early as 6 h in each cell line studied. Supporting our data with ovarian, breast, glial, and prostate cancer cell lines, it was reported very recently that MF inhibited migration induced by progesterone in human astrocytoma cells [
38], that both MF and its metabolite metapristone inhibited the chemotactic migration and mobility in SKOV-3 and IGROV-1 ovarian cancer cell lines facilitated by activation of the chemokine SDF-1/CXCR4 [
29,
39], and that MF inhibited migration and invasion of endometrial carcinoma cells [
40].
MF may have the simultaneous ability to inhibit cell growth and migration via a common mechanism: increase in expression of cyclin-dependent kinase inhibitor p21
cip1. We have shown that MF and MF-related compounds block growth of cancer cells inhibiting the activity of cyclin-dependent kinase Cdk2 as a consequence of p21
cip1 upregulation [
9‐
11]. In support of this concept, it was found that inducing p21
cip1 expression inhibits vascular smooth muscle cell proliferation and migration [
41]. More studies need to be done addressing the role of p21
cip1 in cancer-cell migration; however, the possibility of MF carrying out its various effects through cell-cycle inhibitors should not be overlooked.
Another in vitro aspect studied was cellular invasion, which involves the penetration through tissue barriers, including the basement membrane and stroma, and involves adhesion, proteolytic degradation of the extracellular matrix, and migration [
28]. We clearly showed previously that MF treatment significantly reduced adherence [
13], and now demonstrate that it also attenuates migration and invasion of four highly metastatic cancer cell lines. Supporting our results, in astrocytoma cells, MF blocked the acceleration in the invasive capacity of the cells triggered by progesterone [
38]. In addition, in human gastric adenocarcinoma cells, MF inhibited adhesion, migration, and invasion [
42]. Also, in two melanoma cell lines, metapristone, the demethylated metabolite of MF [
31], significantly impaired invasion through the complex extracellular matrix contained in Matrigel® [
43].