Triple-Negative Breast Cancer (TNBC) Cell Metastasis is Controlled by the Opposing Actions of 5alpha-dihydroprogesterone (5alphap) and 3alpha-dihydroprogesterone (3alphahp) on Expression of Matrix Metalloproteinases and Tissue Inhibitors of Metalloproteinases

Background Triple receptor-negative breast cancers (TNBC) comprise up to 20% of all breast cancers and are the most aggressive mammary cancer subtype, marked by early-onset metastasis, high tumor recurrence rate, and low overall survival. Because TNBC cells lack expression of the estrogen receptor (ER), progesterone receptor (PR) and excess human epidermal growth factor receptor 2 (HER2), there have been no hormone-based explanations and therapeutics for TNBC. Our previous studies had shown that the endogenously produced hormones, 5αdihydroprogesterone (5αP) and 3α-dihydroprogesterone (3αHP), respectively, stimulate and inhibit TNBC cell proliferation and primary tumor development. Here we examined the role of 5αP and 3αHP in controlling metastasis of TNBC (MDA-MB-231)human breast cells. The objectives were to determine the effects of 5αP and 3αHP on TNBC cell (a) adhesion, migration and invasion and (b) expression of matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs). opposing effects of these two hormones on MMP and TIMP expression. The results indicate that the relative concentrations of the cancer-promoter hormone (5αP) and the cancer-suppressor hormone (3αHP) in the breast microenvironment control TNBC cell metastasis, and suggest that blocking 5αP synthesis and increasing 3αHP concentrations in the breast microenvironment could provide the rst hormone-based therapeutic regimen to arrest TNBC progression and metastasis. metalloproteinase; metalloproteinase

3αHP inhibited expression of MMPs and stimulated expression of TIMPS.

Conclusions
Here we show for the rst time that 5αP stimulates and 3αHP suppresses the key elements of metastasis (adhesion, migration, invasion) in TNBC cells, and that these opposing actions correlate with the opposing effects of these two hormones on MMP and TIMP expression. The results indicate that the relative concentrations of the cancer-promoter hormone (5αP) and the cancer-suppressor hormone (3αHP) in the breast microenvironment control TNBC cell metastasis, and suggest that blocking 5αP synthesis and increasing 3αHP concentrations in the breast microenvironment could provide the rst hormone-based therapeutic regimen to arrest TNBC progression and metastasis.

Background
Breast cancer is the most frequently diagnosed cancer in women [1] and triple-negative breast cancer (TNBC), the subtype which tests negative for expression of the estrogen receptor (ER), progesterone receptor (PR) and excess human epidermal growth factor receptor 2 (HER2) [ (2] comprises up to 20% of all breast cancers [2,3]. TNBC is the most aggressive mammary cancer subtype, marked by early-onset metastasis, high tumor recurrence rate, and low overall survival [4,5,6]. The etiology of TNBC is not understood and the lack of receptor expression means that the current hormone-based targeted therapies which are applicable for receptor-positive breast cancers are ineffective, and the only proven method for systemic management of TNBC is chemotherapy [7,8], which provides only limited survival for the patients, along with considerable untoward side-effects. There is, therefore, an urgent need to understand the cause(s) and regulation of TNBC , in particular the progression to metastasis, which is the leading cause of death from all breast cancers.
Previous studies demonstrated that breast tissues [9] and human breast cell lines [ 10] readily convert progesterone to 5α-dihydroprogesterone (5αP) and 3α-dihydroprogesterone (3αHP) by the actions, respectively, of 5α-reductase and 3α-HSO ( Figure 1) and that tumors and tumorigenic breast cell lines produce more 5aP and less 3aHP than normal (non-tumorous) breast tissue and non-tumorigenic breast cell lines, due to up-regulation of 5α-reductase expression and down-regulation of 3α-HSO expression [9][10][11][12]. Particularly noteworthy with respect to the present studies are the prominent differences in expression between adjacent triple receptor-negative normal and tumor breast tissues [11].
In vitro studies with various breast cell lines, including MDA-MB-231 cells ( which have been characterized as human TNBC cells [13]), showed that 5aP acts as a breast cancer-promoter hormone by stimulating cell proliferation and inhibiting apoptosis and adhesion, whereas 3αHP acts as a cancersuppressor hormone by having the opposite effects [9,14]. The pro-cancer and anti-cancer actions, respectively, of 5αP and 3αHP were con rmed with in vivo xenograft studies [15] in which MDA-MB-231 cells were implanted in mammary fat pads of immunosuppressed mice treated with 5αP and/or 3αHP.
The results showed that 5αP stimulated, whereas 3αHP inhibited, initiation and subsequent growth of primary human TNBC cell tumors; in addition, 3αHP resulted in regression of tumors that were already established.
Having demonstrated that TNBC cell proliferation and the onset and growth of primary TNBC tumors are controlled by the relative concentrations of 5αP and 3αHP in the microenvironment, it was of interest to determine the effects of 5αP and 3αHP on metastasis, the process that involves the escape of activated tumor cells from the neoplasm followed by migration to, and invasion of, distant locations where they form the malignant growths which are the primary cause of cancer deaths [16]. The metastatic cascade depends on changes in adhesion properties between cells and between cells and the extracellular matrix (ECM) [17], and the changes in adhesion involve recruitment of proteases to degrade the ECM, basement membrane and cell junctions. The matrix metalloproteinases (MMPs) are the main proteases involved in remodelling the ECM [18,19] and thus have a direct role in potentiating migration and invasion of metastatic cells [19,20]. The MMPs are secreted in a latent form, bound to speci c inhibitors called the tissue inhibitors of metalloproteinases (TIMPs) [21]. Activation of MMPs results when the TIMPs are proteolytically cleaved from them. Since the TIMPs inhibit the enzymatic activity of the MMPs, the balance between the level of active MMPs and available TIMPs is a pivotal determinant of ECM turnover [22]. Consequently, factors that result in changes in expression levels of MMPs and/or TIMPs can affect adhesive, migratory and invasive properties [ 23,24] and correlate with breast cancer prognostic factors [25].
The objectives of the current studies were to determine the effects of 5αP and 3αHP on triple-negative (MDA-MB-231) breast cells in terms of (a) adhesion, migration and invasion and (b) expression of MMP-2, MMP-9, TIMP-1 and TIMP-2.Here we show for the rst time that 5αP stimulates, whereas 3αHP inhibits, receptor-negative breast cell detachment, migration and invasion , and that these opposing actions are via opposing effects of these two hormones on MMP and TIMP expression. The results provide further evidence that the hormones, 5αP and 3αHP, play important roles in stimulating or suppressing various stages leading to deregulation from the normal state to onset and growth of primary tumors and to development of metastatic lesions of TNBC. The results also indicate that therapeutics based on raising the 3αHP:5αP concentration ratio in the microenvironment have the potential to suppress TNBC progression and metastasis. 1.2 mg/ml sodium bicarbonate and 10% full calf serum as previously described [14]. Cells were grown in T-75 asks (Sarstedt) in a humidi ed incubator at 37°C and a 5% CO 2 atmosphere, and were harvested at approximately 80% con uence using 0.025% trypsin and EDTA (0.2 M) in phosphate-buffered saline (PBS) solution (pH 7.3).

Methods And Materials
Cell adhesion. Cell adhesion was determined by a detachment assay described previously [9] Cell Migration and Invasion Assays Cell migration and invasion transwell assays were carried out in modi ed Boyden chambers [26,27] using 24-well cell culture plates with inserts containing 8 μm pore polycarbonate membranes (BD Biosciences, Canaan, CT). The treatment media contained dextran-coated charcoal-treated (DCC-treated) calf serum at 2% for 5αP studies, and 5% for 3αHP studies along with double distilled ethanol at a nal concentration of 0.1% without (control) or with either 5αP (10 -6 M) or 3αHP (10 -6 M), respectively. The reason for using higher (5%) serum concentrations for the 3αHP studies, was to have higher numbers of migrating/invading cells in the controls, thereby facilitating better detection of the suppressive effects of 3αHP.
(a) Migration. Migration was de ned as the ability of cells to move through the 8 μm pores. Cells were seeded at 1x10 5 /200 μl of treatment medium in each of the upper chambers and 800 μl of the respective treatment medium was added to the bottom wells. After incubation of the plates for 6 hours (37 C; 5% CO 2 ), non-migrated cells were gently removed from the top of the lter using cotton swabs. The bottom of each lter was gently rinsed with Earl's Balanced Salt Solution (BSS) and the cells that had migrated to the lower surface of the polycarbonate membrane were xed in 2% formaldehyde and stained with Giemsa stain. The lters were detached from the bottom of the inserts, mounted on glass slides, and viewed microscopically. The number of migrated cells per well was determined by counting the cells within a standardized grid of 16 equal-sized elds, which together covered about 90% of the membrane surface. Each treatment consisted of 4 or 5 replicate wells and experiments were repeated at least three times.
(b) Invasion.Invasion was de ned as the ability of cells to penetrate the 8 μm pores of polycarbonate membranes coated with 2.5 mg/mL of matrigel (BD Biosciences, Canaan, CT), a reconstituted basement membrane extract of the Englebreth-Holm-Swarm mouse sarcoma [28]. The gel was allowed to polymerize for 1h at room temperature. Then 1x10 5 cells in 200 μl of respective treatment medium were seeded in the upper chambers and 800 μl of the same treatment medium was added to the bottom wells. Incubation was for 24 hours at 37°C, and cells having invaded the bottom of the polycarbonate membrane were scored as described above for migration. Each treatment consisted of 4 or 5 replicate wells and experiments were repeated at least three times.

MMP and TIMP expression studies
Cells were seeded in 35 mm culture dishes (Falcon) at 5x10 4 cells per dish in the respective control/treatment media and allowed to attach for 24 hours. They were then incubated for 48 hours in medium containing serum at 5% for the 3αHP studies and 2% for the 5αP studies. After 48 hours cells were harvested and RNA was isolated.
RNA Isolation & Reverse Transcription. RNA was extracted from the cells using RNeasy Mini kit (Qiagen) following the manufacturer's protocol. Complimentary DNA (cDNA) was obtained from 10 µl of RNA using Oligo(dT) (Invitrogen), SuperScript II Reverse Transcriptase (Invitrogen) in a total volume of 20 µl following the manufacturers protocol.
RT-PCR and Quanti cation. RT-PCR was performed on a GeneAmp PCR System 9600 (Perkin Elmer) in triplicate. PCR primers were purchased from Invitrogen with the sequences given in Table 1. PCR conditions were as follows: 95°C denaturing for 2 minutes followed by 30 seconds of annealing at 42°C followed by extension for 1 minute and 30 seconds at 72°C. For each reaction, a 1 µl aliquot of cDNA product was ampli ed in a 50 µl total volume using 1 U of Platinum Taq DNA Polymerase (Invitrogen), 50 mM MgCl 2 , 10 mM dNTP's, and 4 µl of each primer. The same cycling conditions and PCR reagent concentrations were used for each set of primers. PCR products were con rmed by separation on a 1.5% agarose gel and quanti cation using the Quantity One (BioRad) program.

Statistics
Results are given as mean ± SEM. Data were analyzed by unpaired Student's t-test with InStat software (GraphPad, San Diego, CA). Differences were considered signi cant if p < 0.05.

Results
Effects of 5αP and 3αHP on proliferation and cell adhesion.
Proliferation studies (Figure 2A) showed that 5αP dose-dependently stimulated, and 3αHP inhibited, [ 3 H]thymidine incorporation, con rming that the MDA-MB-231 cells used for this study were responsive to 5aP and 3αHP as previously reported [14]. The effects of 5αP and 3αHP on adhesion were determined by detachment assays ( Figure 2B) which showed that detachment is dose-dependently increased by 5αP, and decreased by 3αHP.
Effects of 5αP and 3αHP on cell migration and invasion and cells that have migrated through the pores and attached to, and spread on, the underside, which were scored as 'migrated'. 5αP resulted in signi cant increases ( Figure 3B), whereas 3αHP resulted in signi cant decreases ( Figure 3C), in number of migrated cells. Similarly, 5αP increased, whereas 3αHP decreased, the invasive activities of the cells (Figures 3D & 3E).
Effects of 5αP and 3αHP on expression of MMPs and gelatinase activity.

Discussion
Previous studies showed that proliferation of TNBC cells in vitro [14], and initiation and growth of primary TNBC tumors in vivo [15] are stimulated by 5αP and inhibited by 3αHP. The opposing effects on TNBC cell proliferation result from opposing actions of 5αP and 3αHP on breast cell mitosis and apoptosis [14]. The ndings described here show for the rst time that 5αP stimulates, whereas 3αHP suppresses, TNBC cell detachment, migration and invasion, the cardinal processes involved in cancer cell metastasis. In addition, the studies show that these opposing effects are directly correlated with the opposing effects of 5αP and 3αHP on expression and activation of MMPs, which regulate adhesion properties between cells and between cells and ECM, and thereby control the changes in migration and invasion. Corroborating our ndings are results from other studies which have shown that up-regulation of MMPs facilitates increased migration and invasion [23,[29][30][31], whereas down-regulation of MMPs by anti-cancer and/or anti-metastatic agents inhibits cell migration and invasion [24,32,33].
The relevance of the ndings that 5αP stimulates and 3αHP inhibits the deregulation of TNBC cells from a state of normalcy to one of tumorigenesis and metastasis, is enhanced by observations that TNBC tissues [9] and tumorigenic cells [14] convert more progesterone to 5αP, and less to 3αHP, than normal tissue and nontumorigenic breast cells, as a result of higher expression of 5α-reductase (SRD5A1; SRD5A2) and lower expression of 3α-HSO (AKR1C2; AKR1C3) genes [10; 11]. The resulting increases in 5αP:3αHP concentration ratios in the breast microenvironment correlate with the observed increases in TNBC cell mitogenic and metastatic activities [9,15].

Molecular mechanisms of action of 5αP and 3αHP
The opposing actions of 5αP and 3αHP can be explained, in part, by studies which have identi ed separate receptors for 5αP (5αPR) and 3αHP (3αHPR) and their interactions with cell signalling cascades. The 5αPR and 3αHPR are located on plasma membranes of breast cells [34]. They exhibit high speci city and are distinct from each other and from known receptors of other steroid hormones [34,35 ]. Binding of 5αP to its receptor results in activation of MAPK/ERC [35] which is known to upregulate MMP expression/activity [36,37] leading to decreases in adhesion plaques, vinculin expression and polymerized actin [38] and promoting increased detachment, migration and invasion [39][40]. On the other hand, 3αHP can inhibit protein kinase C (PKC) activity ( [41], and unpublished results) which, in turn, can lead to suppression of MMP activation, and the resulting inhibition of cell invasion and migration [42,43].

Conclusions
Based on previous and current ndings we propose the following model (Fig. 6) to explain how 3αHP and 5αP, can regulate the adhesive, migratory, and invasive processes integral to metastasis of triple-negative breast cancers. 3αHP and 5αP, derived from autocrine and/or paracrine sources, couple with their respective plasma membrane-based receptors, triggering suppression or activation, respectively of PKC or MAPK signaling pathways. Binding of 5αP to its receptor activates the MAPK/ERK1/2 signaling cascade resulting in increased expression of MMPs and/or decreased expression of TIMPs, leading to stimulation of proteolytic degradation of the ECM and providing enhanced ability of cells to detach, migrate and invade. Binding of 3αHP to its receptor results in down regulation of PKC activity which leads to decreased expression of MMPs-2/9 and/or increased expression of TIMPs-1/2, resulting in decreased proteolytic degradation of the surrounding ECM, thereby suppressing cell detachment, migration, and invasion. The relative concentrations of 3αHP and 5αP in the microenvironment determine whether local clusters of breast cells maintain a state of "normalcy" or proceed toward metastasis. A higher 3αHP-to-5αP concentration ratio promotes normalcy, whereas a higher 5αP-to-3αHP concentration ratio promotes mitogenesis and metastasis of TNBC. The ndings suggest (a) that tests which measure the relative concentrations of 5αP:3αHP in a patient's blood, saliva or tissue samples could serve as prognostics/diagnostics for the status of a TNBC patient, and (b) that blocking synthesis of 5αP and increasing 3αHP concentrations in the breast microenvironment can provide a hormone-based therapeutic measure to block TNBC onset, progression and metastasis.
Abbreviations TNBC: triple-negative breast cancer; ER: estrogen receptor; PR: progesterone receptor; 5alphaP: 5alphapregnane-3,20-dione (5alpha-dihydroprogesterone); 3alphaHP: 4-pregnen-3alpha-ol-20-one (3alphadihydroprogesterone); MAPK: mitogen-activated protein kinase; PKC: protein kinase C; MMP: matrix metalloproteinase; TIMP:tissue inhibitor of metalloproteinase Author's contributions JPW conceived the study, created the study design, performed data analyses, assisted in preparation of the gures and drafted the manuscript. JJ assisted in the study design, performed the migration and invasion studies, data analyses and assisted in manuscript editing. AK performed the MMP and TIMP expression studies and assisted in data analyses and manuscript preparation. Figure 1 TNBC (MDA-MB-231) cells convert progesterone to 3αHP and 5αP by the actions of 3α-hydroxysteroid oxidoreductase (3α-HSO) and 5α-reductase (5α-R) expressed by the genes AKR1C2/3 and SDR5A1/2, respectively Figure 2 Dose-dependent effect of 5αP and 3αHP on proliferation (A) and adhesion (B) of MDA-MB-231 cells. Cells were exposed to 0 (control), 10-8, 10-7, or 10-6 M of either 3αHP or 5αP. Proliferation was determined by [3H]thymidine incorporation, and adhesion was determined by a cell detachment assay as described in [9]. Results are presented as mean ± SEM, calculated as percent in relation to control from 4 separate experiments, each treatment having 4-6 replicates. *, **, ***, signi cantly different from controls at p < 0.05, p < 0.01, p < 0.001, respectively.  (control, c) or with either 5αP or 3αHP (at 10-6 M), RNA was isolated, and RT-PCR quanti cation was as described in Methods. For the gelatinase activity studies (E & F), cells were cultured for 24 hours in serumfree medium with/without 5αP or 3αHP, medium was collected, concentrated and aliquots subjected to gelatin zymography and bands quanti ed as described in Methods. Data for the MMP-2 and MMP-9 expression studies are from 4 independent experiments, with 4 replicates in each. Data for the gelatinase activity studies are from 3 replicate incubations. Each bar and line represent the mean ± SEM, standardized to 100% for controls. Representative gels are shown above each graph. *, **, signi cantly different from controls at p < 0.05, and p < 0.01, respectively.

Figure 5
Effect of 5αP and 3αHP on expression of TIMP-1 (A & B) and TIMP-2 (C & D). Cells were incubated in medium without (control, c) or with either 5αP or 3αHP (at 10-6 M) for 48 h, RNA was isolated, and RT-PCR quanti cation was as described in Methods. Each bar and line represents the mean ± SEM of expressed TIMP-1 or TIMP-2 standardized to 100% for controls. Data are from 4 independent experiments, with 4 replicates in each. Representative RT-PCR gels are shown above each gure. *, **, signi cantly different from controls at p < 0.05 and p < 0.01, respectively.

Figure 6
Summary of opposing effects of 3αHP and 5αP on expression of MMP-2/9 and TIMP-1/2 leading to opposing effects on TNBC cell metastatic actions (detachment, migration and invasion). 3αHP suppresses ( ), whereas 5αP promotes (+), metastasis. A higher 5αP-to 3αHP concentration ratio in the microenvironment promotes metastasis, whereas a higher 3αHP-to 5αP concentration ratio suppresses metastasis and promotes normalcy. (See text for details)