Molecular mechanisms underlying progesterone-enhanced breast cancer cell migration

Progesterone (P4) was demonstrated to inhibit migration in vascular smooth muscle cells (VSMCs), but to enhance migration in T47D breast cancer cells. To investigate the mechanism responsible for this switch in P4 action, we examined the signaling pathway responsible for the P4-induced migration enhancement in breast cancer cell lines, T47D and MCF-7. Here, we demonstrated that P4 activated the cSrc/AKT signaling pathway, subsequently inducing RSK1 activation, which in turn increased phosphorylation of p27 at T198 and formation of the p27pT198-RhoA complex in the cytosol, thereby preventing RhoA degradation, and eventually enhanced migration in T47D cells. These findings were confirmed in the P4-treated MCF-7. Comparing the P4-induced molecular events in between breast cancer cells and VSMCs, we found that P4 increased p27 phosphorylation at T198 in breast cancer cells through RSK1 activation, while P4 increased p27 phosphorlation at Ser10 in VSMCs through KIS activation. P27pT198 formed the complex with RhoA and prevented RhoA degradation in T47D cells, whereas p-p27Ser10 formed the complex with RhoA and caused RhoA degradation in VSMCs. The results of this study highlight the molecular mechanism underlying P4-enhanced breast cancer cell migration, and suggest that RSK1 activation is responsible for the P4-induced migration enhancement in breast cancer cells.

Values shown in parentheses represent the quantified results adjusted with G3PDH (A) or with G3PDH and cadherin for the cytosol and the membrane, respectively (D) and expressed as ratio over its own control. cadherin and G3PDH were used as a membranous and cytosolic protein marker, respectively, to confirm the purities of isolation and to verify equivalent sample loading. The gels have been run in the same experimental conditions and the cropped blots were shown. The entire gel pictures of 1C were shown in the Supplemental Fig. 1. In (B,E), values represent the means± s.e.mean. (n = 3). * P < 0.05 different from control group. # P < 0.05 different from P4-treated group. Con, control; IB, immunoblotting; IP, immunoprecipitation; siRNA, small interfering RNA. in T47D cells. Treatment with P4 (50 nM) increased the levels of p27, RSK1 and p27pT198 protein ( Fig. 2A) in T47D cells. Pre-treatment with BI-D1870 (31 nM), a RSK1 inhibitor, prevented the P4-increased level of p27pT198 protein (Fig. 2B), membrane translocation of RhoA (Fig. 2C), and migration ( Fig. 2D), suggesting that RSK1 activation might play an important role in the P4-enhanced migration in T47D cells. To confirm the involvement of RSK1 in the P4-induced migration enhancement in T47D cells, the expression of RSK1 was knocked-down by antisense oligonucleotide. As shown in Fig. 2E, the P4-induced migration enhancement in T47D cells was abolished by pre-treatment the cell with the RSK1 antisense oligonucleotide, but not with the scramble oligonucleotide.

Involvement of cSrc activation in the P4-induced up-regulation of p27 protein and migration enhancement in T47D cells.
Previously, we demonstrated that P4 might inhibit RASMCs migration through up-regulation of p27 mediated by activating the cSrc/AKT/ERK/p38 signaling pathway 16 . Accordingly, we examined whether this signaling pathway is also involved in the P4-induced up-regulation of p27 and migration enhancement in T47D cells. Treatment with P4 (50 nM) for 5-15 min activated cSrc (Fig. 3A), AKT and ERK1/2 (Fig. 3B), and induced phosphorylation of p27 at T198 (Fig. 3C). Pre-treatment with PP2 (200 nM), a cSrc inhibitor, abolished activations of AKT (Fig. 3B), ERK1/2 (Fig. 3B) and Iκ Bα (Fig. 3D, left panel), and increases of the nucleus translocation of NFκ B (Fig. 3D, right panel). The increased levels of RSK1, p27 and p27pT198 protein (Fig. 3E) and migration enhancement (Fig. 3F) in T47D cells induced by P4 treatment were also abolished by pre-treatment with PP2. These data suggest that P4-induced increases in the levels of RSK1, p27 and p27pT198 protein were mediated through a cSrc-mediated pathway.
Role of AKT in the P4-induced phosphorylation of p27 at T198 and migration enhancement in T47D cells. It has been indicated that RSK1 is an effector of both Ras/MEK/MAPK and PI3K/PDK1/AKT pathways. Therefore, we further examined whether activations of PI3K and ERK are involved in the P4-induced RSK1 activation. Treatment with P4 for 15 min did not affect the levels of RSK1 and p27 protein, but increased the levels of p-RSK1 and p27pT198, and these effects were prevented by pre-transfection with dn-AKT, but not by pre-treatment with U0126 (1 μ M), an ERK inhibitor (Fig. 4A). However, treatment with P4 for 6 h increased the Values shown in parentheses represent the quantified results adjusted with G3PDH (A,B) or cadherin (C) and expressed as ratio over its own control. Caherin and G3PDH were used as a membrane and cytosolic protein marker, respectively, to confirm the purities of isolation and to verify equivalent sample loading. In (B-E), values represent the means± s.e.mean. (n = 3). * P < 0.05 different from control group. # P < 0.05 different from P4-treated group. AS, antisense; BI, BI-D1870; Con, control; SC, scramble.
Scientific RepoRts | 6:31509 | DOI: 10.1038/srep31509 levels of both total and phosphorylated RSK1 and p27 protein. As shown in Fig. 4B, pre-treatment with U0126, but not with dn-AKT, significantly reduced P4-induced increases in the levels of RSK1 and p27 protein, whereas pre-treatment with dn-AKT, but not with U0126, significantly reduced P4-induced increases in the levels of p-RSK1 and p27pT198 protein. These findings suggest that P4 induced RSK1 activation mainly through activating the AKT-mediated pathway and increases in the levels of RSK1 and p27 protein mainly through activating the ERK1/2-mediated pathway. To further confirm this notion, the following experiments were conducted. Treatment with P4 for 20 min induced Iκ Bα activation ( Fig. 4C) and nuclear translocation of NFκ B (Fig. 4D), and these effects were prevented by pre-treatment with U0126, but not with dn-AKT. Moreover, the P4-induced migration enhancement was completely abolished by pre-transfection with dn-AKT (Fig. 4E) and reduced for only 25% by pre-treatment with U0126 (Fig. 4F).
Requirement of new protein synthesis in the P4-induced RSK1 activation, phosphorylation of p27 at T198, and migration enhancement in T47D cells. We further examined whether increases of protein syntheses are required for the P4-induced migration enhancement. Pre-treatment with, cycloheximide (4 μ M), an inhibitor of protein biosynthesis, prevented the P4-increased the levels of RSK1 and p27 protein (Fig. 5A), but did not significantly affect the P4-induced migration enhancement (Fig. 5B), suggesting that new protein synthesis is not essential for the P4-induced migration enhancement.

Involvement of the cSrc/AKT/RSK1-mediated signaling pathway in the P4-induced p27 phosphoryation at T198 and migration enhancement in MCF-7 cells.
To examine whether the signaling pathway involved in the P4-induced migration enhancement in T47D cells was also observed in other breast cancer cell lines, the effect of P4 on the migration of MCF-7 breast cancer cells was studied. As shown in Fig. 6A, treatment with P4 (50 nM) for 5 min induced cSrc activation in MCF-7 cells. Pre-treatment with PP2   (200 nM) abolished P4-induced activations in AKT and ERK1/2 (Fig. 6B), whereas pre-treatment with wortmannin (100 nM), a PI3K inhibitor, abolished P4-induced increases of the levels of p-RSK1 and p27pT198 protein (Fig. 6C). The P4-induced phosphorylation of p27 at T198 in MCF-7 was abolished by pre-treatment with BI-D1870 (31 nM). Moreover, the P4-induced migration enhancement in MCF-7 was abolished by pre-treatment with PP2 (Fig. 7A), wortmannin (Fig. 7B) or BI-D1870 (Fig. 7C).

Discussion
P4 was previously shown to promote T47D movement via actin cytoskeleton remodeling mediated by the RhoA/ROCK-2 cascade-triggered moesin activation 12 . However, the signaling pathway involved in the cSrc/ PI3K-mediated activation of the RhoA/ROCK2 cascade is still unclear. In the present study, we further investigated the molecular mechanism underlying the P4-induced migration enhancement in breast cancer cells. Our results showed that P4 stimulated the cSrc/AKT signaling pathway to active RSK1, which in turn induced p27 phosphorylation at T198, subsequently increasing formation of the p27-RhoA complex and causing membrane translocation of RhoA, and finally enhanced migration in breast cancer cells. To our knowledge, this is the first demonstration that the RSK1-induced activation of RSK1 and phosphorylation of p27 at T198 are responsible for the P4-induced migration enhancement in breast cancer cell lines.
P27, an inhibitor of cyclin/cyclin-dependent kinase (CDK) complexes, can bind with cyclin-CDK complexes and cause inactivation of specific CDK enzymes required for cell division, and thereby arrest the cell cycle in the G1 phase 17,18 . It has been shown that targeted disruption of the CDK-inhibitory domain of mouse p27 gene results in enhanced growth of mice, multiple organ hyperplasia, and a predisposition to tumors [19][20][21] . Therefore, p27 has been considered as a tumor suppressor because of its function as a negative regulator of the cell cycle 22,23 . In addition to the function in regulating cell cycle activity, p27 has also been indicated to play additional roles outside of the nucleus 24 and to be involved in regulating cancer cell differentiation 25,26 and apoptosis 27 and cell migration 28 .
The activity of p27 is controlled by its concentration, distribution among different cellular complexes, and subcellular localization 29,30 . P27 contains nuclear export signal (NES) 31 and nuclear localization signal (NLS) 32 . In normal cells, p27 is largely localized in the nucleus and regulates cell cycle activity. In contrast, many human cancers exhibit cytoplasmic p27 localization 22 . It has been demonstrated that AKT and serum and glucocorticoid-regulated kinase 1 (SGK1) can phosphorylate p27NLS at T157 25,[33][34][35] , and their constitutive activation localizes p27 to the cytoplasm by impairing its import into the nucleus 25 . In RASMCs, we previously demonstrated that P4 induced phosphorylation of p27 at serine 10 (Ser10) in the nucleus, subsequently enhancing p27 export from the nucleus to the cytosol, which in turn increased formation of the p27-RhoA complex, and finally caused p27 degradation through the ubiquitin-proteasome pathway 13 . In the present study, our data showed that P4 increased the level of p27pT198 protein ( Fig. 2A) and formation of the p27-RhoA complex (Fig. 1C) in T47D, but did not affect the level of KIS protein (data not shown) in T47D cells. Threonine T198 in the c-terminal of p27 is a phosphorylation site, which controls several important aspects of p27 function including binding to cyclin-CDK complexes, subcellular localization and protein stability 15 . It was previously reported that phosphorylation of p27 at T198 increases formation of the p27-RhoA complex 14 and prevents ubiquitin-dependent degradation of p27 15 . Our data also showed that knock-down of p27 or RSK1 expression, blockade of the RSK1 activity, or interruption of the RhoA-mediated pathway (Fig. 1E) abolished the P4-induced enhanced migration in T47D cells.
RSK1, an effector of both Ras/MEK/MAPK and PI3K/PDK1 pathways, can induce p27 phosphorylation at T198 to increase formation of the RhoA-p27 complex, p27 stability and cell motility 14 . In the present study, we showed that P4 activated RSK1 in T47D cells (Fig. 4A). Our data suggest that the P4-induced activation of RSK1 through a direct activation of the cSrc/AKT-mediated pathway. Pre-treatment with the RSK1 inhibitor, BI-D1870, prevented the P4-induced phosphorylation of p27 at T198 and migration enhancement in both T47D and MCF-7 cells, suggesting that RSK1 activation plays an important role in the P4-induced migration enhancement in breast cancer cells. The involvement of RSK1 in the P4-induced migration enhancement in breast cancer cells was further confirmed by the evidence showing that knock-down of RSK1 expression abolished the P4-induced migration enhancement. P4 also induced up-regulations of RSK1 and p27 through activating the cSrc/ERK1/2/ NFκ B-mediated pathway. However, pre-treatment with cycloheximide, an inhibitor of protein synthesis, which blocked the P4-induced up-regulations of RSK1 and p27 (Fig. 5A), did not cause any significant reduction in the P4-induced migration enhancement in T47D cells (Fig. 5B), suggesting that new protein synthesis might not be required for the P4-induced migration enhancement.
P4-induced activation of cSrc is mediated through a non-genomic action of P4 36 . This rapid action of P4 in the cytosol and consequent activation of cSrc has also been implicated in the regulation of cell cycle and migration 37,38 . Further indication that cSrc activation is involved in P4-induced activations of RSK1 and migration enhancement was evidenced by failure of these effects of P4 to occur in the PP2-treated T47D and MCF-7 cells. Taken together, these results suggest that activation of the cSrc-mediated signaling pathway might be involved in the P4-induced migration enhancement in breast cancer cells.
In conclusion, this study provides evidence that P4 induced RSK1 activation mediated by the cSrc/AKT signaling pathway, subsequently causing phosphorylation of p27 at T198, which in turn increased formation of the p27-RhoA complex and RhoA activation, and finally enhanced migration in breast cancer cells. P4 also increased the levels of p27 and RSK1 protein through activating the cSrc/ERK1/2 signaling pathway in T47D cells. Although increased expressions of RSK1 and p27 are not required for the P4-induced migration enhancement in T47D cells, it might have some contributions in the P4-regulated T47D cell migration. Based on the results from the present study, we propose a model of the molecular mechanism underlying the P4-induced migration enhancement in breast cancer cells as shown in Fig. 8. The findings from the present study suggest that activation of RSK1 is the key factor responsible for the P4-induced migration enhancement in breast cancer cells.

Methods
Cell Cultures. T47D and MCF-7 cell lines established from the pleural effusion of a ductal carcinoma of the breast were purchased from the American Type Culture Collection/Bioresource Collection and Research Center (BCRC) (Taiwan). These cells have performed STR-PCR profile at BCRC. The cells were grown in DMEM (GIBCO, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; GIBCO), penicillin (100 U/mL), and streptomycin (100 μ g/mL, GIBCO) in a humidified 37 °C and 5% CO 2 incubator, and the media was switched to DMEM supplemented with 10% charcoal-stripped fetal bovine serum at 24 h prior to and during P4 or vehicle (0.05% DMSO) treatment. For the experiments, T47D and MCF-7 cells were treated with either P4 (50 nM) in 0.05% DMSO (Sigma-Aldrich, St. Louis, MO) for the indicated times or 0.05% DMSO (control).
Wound healing assay. Wound healing assay was performed as previously described 16,39 . Briefly, cells were grown in 24-well plates. White tips for 10 μ L micropipette were used to scalp a "wound" in a cell monolayer. The images captured at the beginning were compared to the images after 8 h in a humidified 37 °C, CO 2 incubator. The cells migrating to close the wound were counted to quantify the migration rate of the cells.

Western blot analysis.
To determine the protein levels in breast cancer cells, total proteins were extracted and Western blot analyses were performed as described previously 16 . Electrophoresis was carried out using a 12% SDS-polyacrylamide gel (50 μ g protein per lane). Separated proteins were transferred onto PVDF membranes, treated with 1% BSA/0.02% NaNO 3 to block the non-specific IgGs, incubated for 1 h with specific primary antibody (0.2 μ g/mL), and then incubated with second antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) linked to horse radish peroxidase (1:10,000) for 1 h. Subsequently, the blot was developed using the ECL (enhanced chemiluminescence) system (GE, Healthcare, NJ). The intensity of each band was quantified by densitometry analysis using Image Pro-Plus 4.5 Software.

Subcellular fractionation.
To examine the membrane translocation of RhoA, breast cancer cells were washed with ice-cold phosphate-buffered saline (PBS) and re-suspended in hypotonic buffer (20 mM Tris-HCl, pH 7.5, 5 mM EGTA, 2 mM EDTA, 1 mM NaVO 3 , 1 mM DTT, 10 mM NaF, 10 mM Na 2 H 2 P 2 O 7 ) plus protease inhibitor cocktail (Sigma-Aldrich). After being frozen at − 80 °C for 2 h, the cell lysates were disrupted by passing through a 3/10 mL BD insulin syringe and 30-gauge needle for 120 times, and then incubated on ice with repeated vortex mixing for 6 times (5 min each time). The supernatant was collected as the cytosolic fraction after being centrifuged at 14,000 g for 10 min at 4 °C. Pellets were washed with ice-cold PBS for 2 times, and then lysed in the extract buffer (50 mM HEPES, pH 7.0, 250 mM NaCl, 2.5 mM EDTA, 1% NP-40, 5% Glycerol, 1 mM NaVO 3 , 10 mM NaF, 10 mM Na 2 H 2 P 2 O 7 ) plus protease inhibitor cocktail (Sigma-Aldrich) on ice with repeated vortex mixing for 9 times (5 min in each), and then centrifuged at 14,000 g for 30 min at 4 °C. The supernatant was collected as the particulate (membrane) fraction.

Nuclear extraction.
To examine the effect of P4 on nuclear translocation of NFκ B (p65), the NE-PER nuclear and cytoplasmic extraction reagents (Thermo Fisher Scientific, Rockford, IL) were used and the extraction was performed as previously described 38 . Cell transfection. For transient transfection of the indicated constructs into T47D cells, jetPEI-T47D transfection reagent (Polyplus Transfection, Bioparc, France) was used and the transfection was performed as previously described 40 .
Small interfering RNA (siRNA) knockdown assay. Expression of p27 was knocked-down in T47D with at least three independent small interfering RNAs (siRNAs). The target sequences of p27 mRNA were selected to suppress p27 gene expression. Non-target sequences of each siRNA (NT-siRNA) were used as controls 41 . After BLAST analysis to verify that there were no significant sequence homologies with other human genes, the selected sequences were inserted into BglII/HindIII-digested pSUPER vectors to generate the pSUPER-Si p27 (p27-Si) and pSUPER-non-target p27 (p27-NT). The cells were transfected with either combined three anti-sense siRNAs or non-target RNA (control). Three different anti-sense siRNAs targeted against different parts of the p27 sequence are listed below: Si-1: 5′ -GATCCCCGGAAGCGACCTGCAACCGATTCAAGAGATCGGTTGCAGGTCGCTTCCTTTTTA-3′ ; Si-2: 5′ -GATCCCCACCCGGGAGAAAGATGTCATTCAAGAGATGACATCTTTCTCCCGGGTTTTTTA-3′ ; Si-3: 5′ -GATCCCCGAAGCGACCTGCAACCGACTTCAAGAGAGTCGGTTGCAGGTCGCTTCTTTTTTA, and NT: 5′ -GATCCCCATGGACAGACGCA CGCACGTTCAAGAGACGTGCGTGCGTCTGTCCATTTTTTA-3′ . All constructs were confirmed by DNA sequence analysis. The transfection protocol has been previously described 16 . Briefly, 2 × 10 5 cells were washed twice with PBS and mixed with 0.5 μ g of plasmid, and then one electric pulse was applied for 20 ms under a fixed voltage of 1.3 kV on a pipette-type MP-100 microporator (Digital Bio, Seoul, Korea). Antisense oligonucleotide. The antisense oligonucleotide (AS) sequence of RSK1: 5′-ATGCCGCTC GCCCAGCTCAA-3′ is complementary to the region of the sense template of RSK1. The scramble oligonucleotide (Sc) sequence of RSK1: 5′-GCCCACCCTACGAGCTACGT-3' was used for the control. Briefly, a final concentration of 100 nM RSK1 antisense oligonucleotide or non-targeting RSK1 scramble oligonucleotide (control) was added in the culture medium for 24 h. The oligonucleotide-treated T47D cells were rendered for quiescent by incubation for 21 h in phenol red free/charcoal-stripped DMEM containing 0.04% FBS. The phenol red free/ charcoal-stripped DMEM supplemented with 10% FBS and 0.05% DMSO with or without 50 nM P4 was added to the cell and the culture was conducted for the wound healing assay.
Statistical analysis. All data were expressed as the mean value± s.e.mean. Comparisons were subjected to one way analysis of variance (ANOVA) followed by Fisher's least significant difference test. Significance was accepted at P < 0.05.