Biochemical and Anti-Triple Negative Metastatic Breast Tumor Cell Properties of Psammaplins

Breast tumors reprogram their cellular metabolism, nutrient uptake, and utilization-associated biochemical processes. These processes become further transformed as genetically predisposed metastatic breast tumor cells colonize specific organs. Breast tumor cells often metastasize to the brain, bone, lung and liver. Massagué and colleagues isolated organotropic subclones and established organ-specific gene signatures associated with lung-, bone-, and brain-specific metastatic triple-negative breast cancer (TNBC) MDA-MB-231 cells. Using these genetically characterized metastatic subclones specific to lung (LM4175), bone (BoM1833), and brain (BrM-2a), we evaluated marine natural products for the ability to differentially suppress metastatic breast cancer cells in a target organ-dependent manner. Psammaplin-based histone deacetylase (HDAC) inhibitors were found to differentially inhibit HDAC activity, induce activation of hypoxia-inducible factor-1 (HIF-1), and disrupt organotropic metastatic TNBC subclone growth. Further, psammaplins distinctly suppressed the outgrowth of BoM1833 tumor spheroids in 3D-culture systems. Similar results were observed with the prototypical HDAC inhibitor trichostatin A (TSA). These organotropic tumor cell-based studies suggest the potential application of HDAC inhibitors that may yield new directions for anti-metastatic breast tumor research and drug discovery.


Introduction
Prevention and improved therapies have produced a steady decline in cancer rates in developed countries [1]. In spite of this success, systemic metastasis-associated disease relapse accounts for over 90% of cancer mortality. The five-year survival rate is 27% among the 162,000 American women with

Differential HDAC Inhibition by Psammaplin Analogues
The effects of 1-5 on HDAC activity were determined in a human melanoma MDA-MB-435 cell-based assay. Test compounds were added at specified concentrations to exponentially grown cells plated in 96-well plates. After 30 min incubation, the HDAC activity was determined using a commercial kit (HDAC-Glo™, Promega, Madison, WI, USA) and normalized to that of the dimethyl sulfoxide (DMSO) solvent control. The protypical HDAC inhibitors trichostatin A (TSA, 1 nM) and suberanilohydroxamic acid (SAHA, vorinostat, Zolinza TM , 100 nM) were included as positive controls. In the MDA-MB-435 cell-based assay, TSA and SAHA inhibited HDAC by 52 ± 8% and 60 ± 5%, respectively (average ± SE, n = 6). The IC 50 values for 1-5 to inhibit HDAC are summarized in Table 1. The most potent compound (i.e., 2) inhibited HDAC with an IC 50 of 0.019 µM, while the least active (i.e., 5) had an IC 50 of 0.948 µM. The HDAC inhibitory activities of these compounds mirrored those observed in the T47D cell-based reporter assay (2 is the most potent compound and was 5 the least potent). cells transfected with pHRE-luc for HIF-1 activity. Test compounds were added at the increasing concentrations of 0.1, 0.3, 1, 3, 10, and 30 µM, as specified. The positive control 1,10-phenanthroline (1,10-phen) was used at 10 µM. Data shown are average ± standard deviation (n = 3). (C) Similar to described in (B) except that the pHRE-luc transfected T47D cells were exposed to test compounds in the presence of 10 µM 1,10-phen, and the data were normalized to the positive control (1,10-phen).

Effects of Psammaplin A on HIF-1 Target Gene Expression
Around 100 genes have been identified as HIF-1 target genes that encode proteins involved in various aspects of cellular physiology [27]. While most of these genes are regulated in a cell type-specific manner, some are induced upon HIF-1 activation in most cell types [28]. Based on availability and potency, compound 4 (a stereoisomer of 2) was selected for follow-up studies. The effects of 4 on the expression of HIF-1 target genes cyclin dependent kinase inhibitor 1A (CDKN1A) and vascular endothelial growth factor A (VEGFA) were examined by quantitative real time RT-PCR ( Figure 2). The HIF-1 activator 1,10-phen (10 µM) and the pan-HDAC inhibitor TSA (0.1 and 1 µM, respectively) were included as positive controls. In T47D cells, 4 and TSA each increased the levels of CDKN1A mRNA in a concentration-dependent manner (2.9-fold for 4 at 10 µM and 7.2-fold for TSA at 1 µM, Figure 2A). In contrast, neither 4 nor TSA exerted greater than 20% effect on the levels of VEGF mRNA ( Figure 2B). The gene VEGF encodes vascular endothelial growth factor A (a potent angiogenic factor) and agents that inhibit VEGF are in clinical use for cancer [29,30]. The expression of cellular and secreted VEGF proteins was examined in T47D cells by ELISA assay. As anticipated, the positive control 1,10-phen induced VEGF expression at the levels of mRNA ( Figure 2B), cellular protein ( Figure 2C), and secreted protein ( Figure 2D). None of the HDAC inhibitors examined (4 and TSA) increased VEGF protein levels at the concentrations tested ( Figure 2C,D). the presence of 10 μM 1,10-phen, and the data were normalized to the positive control (1,10-phen). (D) Similar to described in (C) except that hypoxic exposure (1% O2: 5% CO2: 94% N2, 16 h) was applied in place of 1,10-phen. (E) As described in (B) except that T47D cells were transfected with the pGL3-control construct. An asterisk "*" indicates p < 0.05 when compared to the controls ("Media" for B, D, and E; and "1,10-phen" for C).

Effects of Psammaplin A on HIF-1 Target Gene Expression
Around 100 genes have been identified as HIF-1 target genes that encode proteins involved in various aspects of cellular physiology [27]. While most of these genes are regulated in a cell typespecific manner, some are induced upon HIF-1 activation in most cell types [28]. Based on availability and potency, compound 4 (a stereoisomer of 2) was selected for follow-up studies. The effects of 4 on the expression of HIF-1 target genes cyclin dependent kinase inhibitor 1A (CDKN1A) and vascular endothelial growth factor A (VEGFA) were examined by quantitative real time RT-PCR ( Figure 2). The HIF-1 activator 1,10-phen (10 μM) and the pan-HDAC inhibitor TSA (0.1 and 1 μM, respectively) were included as positive controls. In T47D cells, 4 and TSA each increased the levels of CDKN1A mRNA in a concentration-dependent manner (2.9-fold for 4 at 10 μM and 7.2-fold for TSA at 1 μM, Figure 2A). In contrast, neither 4 nor TSA exerted greater than 20% effect on the levels of VEGF mRNA ( Figure 2B). The gene VEGF encodes vascular endothelial growth factor A (a potent angiogenic factor) and agents that inhibit VEGF are in clinical use for cancer [29,30]. The expression of cellular and secreted VEGF proteins was examined in T47D cells by ELISA assay. As anticipated, the positive control 1,10-phen induced VEGF expression at the levels of mRNA ( Figure 2B), cellular protein ( Figure 2C), and secreted protein ( Figure 2D). None of the HDAC inhibitors examined (4 and TSA) increased VEGF protein levels at the concentrations tested ( Figure 2C,D).

A.
B.

Psammaplins Suppress Cell Proliferation/Viability in a Cell Line-Dependent Manner
In the T47D cell-based reporter assay, psammaplins regulated HIF-1 activity in a biphasic manner ( Figure 1). To discern if cytostatic/cytotoxic effects contributed to the drop in HIF-1 activity at higher concentrations, the effects of psammaplins on cell proliferation/viability were examined in a panel of established human breast cancer cell lines. The protein synthesis inhibitor cycloheximide (CHX, 10 μM) was used as a positive control and the pan-HDAC inhibitor TSA was included for comparison. Following 48 h of compound treatment, all compounds affected cell proliferation/viability to a certain extent. Among the psammaplins, the potency rank of 2 and 4 > 3 > 1 > 5 mirrored that observed in the HDAC assay (Table 1)

Psammaplins Suppress Cell Proliferation/Viability in a Cell Line-Dependent Manner
In the T47D cell-based reporter assay, psammaplins regulated HIF-1 activity in a biphasic manner ( Figure 1). To discern if cytostatic/cytotoxic effects contributed to the drop in HIF-1 activity at higher concentrations, the effects of psammaplins on cell proliferation/viability were examined in a panel of established human breast cancer cell lines. The protein synthesis inhibitor cycloheximide (CHX, 10 µM) was used as a positive control and the pan-HDAC inhibitor TSA was included for comparison. Following 48 h of compound treatment, all compounds affected cell proliferation/viability to a certain extent. Among the psammaplins, the potency rank of 2 and 4 > 3 > 1 > 5 mirrored that observed in the HDAC assay (Table 1)   The effects of psammaplins on the colony-forming ability of single cells were assessed in a clonogenic assay. Cells seeded at low density were exposed to test compounds at the specified concentrations for 24 h. The conditioned media were replaced with growth media and the colonies formed from surviving single cells in 10 days. While the cell lines differ in their colony-forming abilities, the positive control paclitaxel blocked colony formation in all cell lines (Figure 4). Less pronounced colony-suppressing activity was observed with the HDAC inhibitors.

Psammaplin A and TSA Inhibit Tumor Cell Invasion
In order to form metastatic lesions, metastasis-initiating tumor cells must invade and intravasate into the lymphatic vasculature and/or blood vessels. Psammaplins were evaluated in a Cultrex ® 3-D cell invasion assay that monitors the invasion and migration of tumor cells grown as spheroids, which closely model in vivo pathophysiological conditions. The bone metastatic BoM subclone displayed the most aggressive behavior (a network of extensive projections from the spheroid, T 96 , Media, Figure 5).

Discussion
The overwhelming majority of all targeted therapies approved for the treatment of breast cancer specifically target either estrogen or progesterone hormone receptors (e.g., tamoxifen, anastrozole, letrozole) or target tumors that overexpress the proto-oncogenic receptor tyrosine-protein kinase known as human epidermal growth factor receptor 2 (HER2) (e.g., trastuzumab, pertuzumab) [31]. Human TNBCs represent a major treatment challenge because of the lack of estrogen receptor (ER),

Figure 4. Effects of 1-5 and TSA (trichostatin A) on colony formation.
Cells plated at low density were exposed to compounds (24 h) at specified concentrations (1 μM for paclitaxel and TSA, and 10 μM for 1-5). After a period of ten days, the cells were fixed and stained.

Discussion
The overwhelming majority of all targeted therapies approved for the treatment of breast cancer specifically target either estrogen or progesterone hormone receptors (e.g., tamoxifen, anastrozole, letrozole) or target tumors that overexpress the proto-oncogenic receptor tyrosine-protein kinase known as human epidermal growth factor receptor 2 (HER2) (e.g., trastuzumab, pertuzumab) [31]. Human TNBCs represent a major treatment challenge because of the lack of estrogen receptor (ER),

Discussion
The overwhelming majority of all targeted therapies approved for the treatment of breast cancer specifically target either estrogen or progesterone hormone receptors (e.g., tamoxifen, anastrozole, letrozole) or target tumors that overexpress the proto-oncogenic receptor tyrosine-protein kinase known as human epidermal growth factor receptor 2 (HER2) (e.g., trastuzumab, pertuzumab) [31]. Human TNBCs represent a major treatment challenge because of the lack of estrogen receptor (ER), progesterone receptor (PR), and HER2. Genetic studies identified at least half a dozen distinct TNBC subtypes with unique gene expression profiles that appear to respond differentially to NVP-BEZ235 (PI3K/mTOR inhibitor), dasatinib (abl/src inhibitor), and the androgen receptor antagonist bicalutamide [32]. Mooberry and coworkers recently demonstrated that certain DNA damaging agents, such as the Tolypocladium sp. fungi hybrid polyketide-shikimate-nonribosomal peptide synthetase metabolite, maximiscin, exhibits a differential pattern of triple-negative breast tumor cell line selectivity [33]. It is clear that small molecules selectively target various subtypes of primarily tumor biopsy-derived TNBCs can be identified through targeted screening efforts.
To discover potential new agents that differentially inhibit patient-derived TNBC metastases or specifically target organotropic metastatic TNBCs, we initiated a screening campaign using the TNBC MDA-MB-231 derived organotropic clones as in vitro models. Our initial screening efforts indicate that natural product-based HDAC inhibitors (i.e., psammaplins and trichostatin A) differentially suppress the proliferation and three-dimensional invasive growth potential of TNBC cells that are oncogenically predisposed to colonize specific organs. These psammaplins (1)(2)(3)(4) were isolated from the sponge Dendrilla lacunosa for their HIF inducing activity. The most potent two compounds (E,Z)-psammaplin A (2) and (E,E)-psammaplin A (4) are stereo isomers, while the structurally related compounds psammaplin E (1) and (E,E)-psammaplin K (3) are less active. In T47D cell-based reporter assays, psammaplins displayed biphasic effects (induction at lower concentrations and inhibition at higher concentrations). This is not surprising because hypoxia and chemical hypoxia impose cellular stress. The extent of damage caused by severe hypoxia or high concentrations of chemicals that activate hypoxic signaling may induce senescence or cell death. Either of these conditions will result in decreased reporter gene expression. In comparison to the common HIF stimuli (e.g., hypoxia, iron chelator, etc.), (E,E)-psammaplin A (4) and the HDAC inhibitor standard TSA are unique in the aspect that they induced the expression of cyclin dependent kinase inhibitor 1A (CDKN1A), but not the expression of vascular endothelial growth factor A (VEGFA). While both are HIF-1 target genes, CDKN1A stalls cell cycle progression and VEGFA promotes tumor angiogenesis. Further study is required to resolve the mechanisms behind the differential effects exerted by HDAC inhibitors on HIF-1 target gene expression. Clinically, the natural product HDAC inhibitor romidepsin (Istodax TM ) and SAHA (vorinostat, Zolinza TM ) are used for the treatment of cutaneous T-cell lymphoma [34,35]. However, HDAC inhibitors have not found general utility in the treatment of genetically diverse primary solid tumors.
Recent studies suggest that psammaplins induce Sirtuin 1-dependent autophagic cell death in human breast tumor cell lines and xenografts [36,37]. Psammaplin A decreased SIRT1 enzyme expression and activity in breast tumor cells, increased the acetylation of the SIRT1 target p53, and produced an overall increase in autophagy-related protein expression, including that of the p53-induced protein, DRAM (damage-regulated autophagy modulator). It is possible that psammaplins stall tumor progression by inducing the expression of proliferation inhibiting and cell death promoting genes, without stimulating the expression of survival genes (i.e., VEGFA). Among the three MDA-MB-231 derived organotropic clones examined, the slowest growing BrM subclone was least affected by psammaplins. The most aggressive BoM cells invaded extracellular matrix in a 3D spheroid invasion assay and (E,E)-psammaplin A (4) inhibited this invasion. Compound 4 and the standard HDAC inhibitor TSA were tested at the IC 50 values determined in a 48 h proliferation/viability. Because the invasion assay was conducted over the course of 96 h, it is also possible that target proteins other than histones (e.g., tubulins) were affected by HDAC inhibitors. Increased acetylation of these non-histone target proteins may directly contribute to the blockade of cell invasion.
The concept of natural product-derived HDAC inhibitors as potential antimetastatic agents that target TNBCs is further supported by the work of Lu, Wang, and coworkers with garlic diallyl trisulfide that acts as a natural HDAC inhibitor [38]. Garlic diallyl trisulfide was found to suppress MDA-MB-231 metastatic potential in embryonic zebrafish, xenografts, and orthotopic tumor models, and inhibited MDA-MB-231 migration and angiogenesis in vitro. Although these are only the first efforts to identify natural product-derived agents with potential to selectively suppress organotropic metastatic TNBCs, they provide vital new prospects for the repurposing of clinically approved chemotherapeutic drugs and other natural product-based agents in the treatment or chemoprevention of organ-specific TNBC metastases.

Sponge Material, Extract Preparation, and Bioassay-Guided Isolation
The sponge material was part of the NCI Open Repository Collection.   C-7). The structure of 2 was confirmed by comparison with previously published 1 H-NMR and 13 C-NMR data [40].  C-7). The structure of 3 was confirmed by comparison with previously published 1 H-NMR and 13 C-NMR data [40].

T47D Cell-Based Reporter Assay
To monitor HIF-1 activity, T47D cells were transfected with the pHRE3-TK-Luc construct and the cell-based luciferase reporter assay was performed as described [20]. Cells were exposed to test compounds in the absence and presence of 1,10-phenanthroline (10 µM) or hypoxic conditions (1% O 2 : 5% CO 2 : 94% N 2 ) for 16 h. The cells were lysed and luciferase activity determined with a commercial kit (Promega, Madison, WI, USA). For the cell-based control reporter assay, T47D cells were transfected with the pGL3-control construct (Promega, Madison, WI, USA), exposed to test compounds for 16 h, and the luciferase reporter assay performed as described [21].

MDA-MB-435 Cell-Based HDAC Assay
Human melanoma MDA-MB-435 cells (ATCC, Manassas, VA, USA) were maintained in RPMI 1640 medium supplemented with 10% FBS, 100 units/mL penicillin, and 100 µg/mL streptomycin. Exponentially grown cells were seeded at the density of 5000 cells/well into 96-well plates (Corning, Corning, NY, USA) and incubated overnight. Compounds dissolved in DMSO were added to achieve the specified final concentrations (total volume: 100 µL, DMSO: 0.5%). The incubation continued for 30 min at 37 • C and the HDAC activity determined using a commercial luminescent assay (HDAC-Glo™, Promega, Madison, WI, USA). The HDAC inhibitors trichostatin A (TSA, 1 nM) and SAHA (100 nM) were used as positive controls and the data presented as percentage inhibition of the solvent control.

Quantitative Real-Time RT-PCR and ELISA Assay
The effects of test samples on HIF-1 target gene expression were assessed in T47D cells.
To determine the levels of CDKN1A and VEGF mRNA, quantitative real-time RT-PCR was performed as described [21,43]. To determine the levels of cellular and secreted VEGF proteins, T47D cells were exposed to compounds as described [44], the levels of VEGF proteins in the conditioned medium and cell lysate samples determined by ELISA [21], the amount of proteins in the cell lysate samples quantified using a micro BCA assay kit (Thermo Fisher Scientific, Rockford, IL, USA), and the level of VEGF proteins normalized to that of cellular proteins.

Cell Proliferation/Viability and Clonogenic Survival Assays
The cell proliferation/viability assay (48 h exposure) was performed as described, using the sulforhodamine B method [22]. The data are presented as '% Inhibition' of the media control.
For the clonogenic assay, exponentially grown cells were seeded at the density of 1000 cells/well into 6-well plates (Cellstar ® , Greiner Bio-One GmbH, Kremsmünster, Austria) and incubated at 37 • C for 4 h to allow the cells to adhere. Compound addition was similar as above. After 24 h, the compound-containing conditioned media were replaced with fresh medium containing FBS (10%) and antibiotics. The incubation continued for another 10 days with a change of fresh medium every 5 days, the cells were fixed with methanol and stained with crystal violet (1 mg/mL in 20% ethanol), and the images were acquired with a Kodak digital camera.

3D Cell Invasion Assay
The 3D spheroid cell invasion assay was performed using a commercial kit, following the manufacturer's instructions (Cultrex ® 3D Spheroid Cell Invasion Assay, Trevigen, Gaithersburg, MD, USA). Briefly, exponentially grown MDA-MB-231, LM, BoM, and BrM cells were trypsinized, collected, and resuspended in serum-free DMEM/F12 media. For spheroid formation, 3000 cells were added in a volume of 50 µL serum-free DMEM/F12 media with 1× spheroid formation solution/well into a pre-cooled 96-well plate. The plate was centrifuged at 100× g for 5 min at 4 • C, incubated at 37 • C for 24 h, and the spheroids imaged using an Axiovert 40 CFL microscope (Zeiss, Oberkochen, Germany). The plate was placed on ice, the invasion mix added in a volume of 50 µL/well, centrifuged at 350× g for 5 min at 4 • C, and incubated at 37 • C for 1 h. Test compounds and controls were diluted to two times the final concentrations in DMEM/F12 media supplemented with FBS (10%) and antibiotics, and added in a volume of 100 µL/well. The incubation continued for another 4 days at 37 • C and the cells/spheroids imaged (T 96 ).

Statistical Analysis
GraphPad Prism 6 was applied to analyzed data. Data were compared by one-way ANOVA followed by Bonferroni post-hoc analyses. Differences were considered statistically significant when p < 0.05.