Characterization of inorganic phosphate transport in the triple-negative breast cancer cell line, MDA-MB-231

Background Recent studies demonstrate that interstitial inorganic phosphate is significantly elevated in the breast cancer microenvironment as compared to normal tissue. In addition it has been shown that breast cancer cells express high levels of the NaPi-IIb carrier (SLC34A2), suggesting that this carrier may play a role in breast cancer progression. However, the biochemical behavior of inorganic phosphate (Pi) transporter in this cancer type remains elusive. Methods In this work, we characterize the kinetic parameters of Pi transport in the aggressive human breast cancer cell line, MDA-MB-231, and correlated Pi transport with cell migration and adhesion. Results We determined the influence of sodium concentration, pH, metabolic inhibitors, as well as the affinity for inorganic phosphate in Pi transport. We observed that the inorganic phosphate is dependent on sodium transport (K0,5 value = 21.98 mM for NaCl). Furthermore, the transport is modulated by different pH values and increasing concentrations of Pi, following the Michaelis-Menten kinetics (K0,5 = 0.08 mM Pi). PFA, monensin, furosemide and ouabain inhibited Pi transport, cell migration and adhesion. Conclusions Taken together, these results showed that the uptake of Pi in MDA-MB-231 cells is modulated by sodium and by regulatory mechanisms of intracellular sodium gradient. General Significance: Pi transport might be regarded as a potential target for therapy against tumor progression.


Introduction
triple-negative molecular subtype) is a cellular model with a high metastatic capacity and, therefore, is considered a more aggressive strain compared to MCF-7 (ER+, PR+ and HER is classified as a tumorigenic, but non-metastatic cell line) [17,18].
MDA-MB-231 cells exhibit increased migratory capacity upon increasing Pi concentrations in the growth medium. There are two solute carrier families of Pi transport in mammals: SLC20 and SLC34 and both protein families transport Pi using the electrochemical gradient for Na + [19]. It has been previously described that NaPi-IIb (SLC34A2) is up-regulated in ovarian carcinomas and benign tumors compared to normal ovary tissues [20]. Another study showed that the upregulated expression of SLC34A2 in hepatocellular carcinoma cell lines, and the knockdown of this Pi transporter decrease cell proliferation, migration and invasion as well as the epithelial-mesenchymal transition [21]. In lung cancer cells, SLC34A2 was also necessary for proliferation and tumorigenesis [22]. Overexpression of the NaPi-IIb transporter has been described in breast cancer tumors as opposed to normal tissues and has been proposed as a novel diagnostic marker and a therapeutic target [23]. Thus, the aim of this study was to characterize the Pi transport kinetics in the MDA-MB-231 breast cancer cell line by investigating whether such biochemical features may suggest interference tactics using inhibitors that could impact upon migration and invasive capacity.

Materials
Reagents were bought from E. Merck (Darmstadt, Germany) and Sigma Chemical Co. (St. Louis, MO, USA). Radioactive inorganic phosphate ( 32 P i ) used was from Instituto de Pesquisas Energéticas e Nucleares (IPEN). In this work, we used distilled water through a Milli-Q system of resins to prepare all solutions (Millipore Corp., Bedford, MA, USA).

Cell culture
MDA-MB-231, T47D and MCF-7 cells were grown at 37˚C in Iscoves Modified Dulbeco's Medium (IMDM-LCG Biotechnology, Brazil) supplemented with sodium bicarbonate, 10% of foetal bovine serum (FBS) (Cripion Biotechnology, Brazil), 100 U/mL penicillin and streptomycin (Thermo Fisher, Brazil). 67NR and 4T1 cell lines, which originated from a spontaneous mammary carcinoma arising in a BALB/c mouse [24], were purchased from Karmanos Cancer Institute (Detroit, MI, USA). Cells were maintained in high glucose Dulbecco's modified Eagle medium (DMEM), supplemented with L-GlutaMax, 10 mM sodium carbonate, Hepes Buffer and 10% FBS and maintained at 37˚C in a humidified atmosphere of 5% CO 2 . For the experiments, cells were harvested from the culture medium, washed two times with buffer comprising of 116 mM NaCl, 5.4 mM KCl, 5.5 mM glucose, 0.8 mM MgCl 2 and 50 mM HEPES (pH 7.2). Hank's EDTA solution was used to isolated cells from dishes. Cell number was estimated by counting in a Neubauer chamber. The protein concentration was measured with the Bradford methodology [25].

Pi transport assay in MDA-MB-231
Living MDA-MB-231 cells (10 4 cells) were incubated at 37˚C for 1 hour in a reaction mixture (0.5 mL) containing 116 mM NaCl or choline cloride, 5.4 mM KCl, 5.5 mM glucose, 50 mM HEPES (pH 7.2), 0.8 mM MgCl 2 , 0.1 mM KH 2 PO 4 and 2.5 μCi/nmol 32 P i [26]. The reaction was stopped with 0.5 ml of an ice-cold PBS buffer (pH 7.2). Cells were washed with the same cold buffer at 4˚C and disrupted with 0.25 mL Hank's solution (5.37 mM KCl, 0.44 mM KH 2 PO 4 , 136.8 mM NaCl, 0.33 mM NaH 2 PO 4 , 5.03 mM D-glucose, 4.16 mM sodium bicarbonate, 6.35 mM EDTA, pH 7.2) and 0.25 ml 0.1% SDS. These lysed cells with the internalized Pi were moved to a filter paper and, then, to a scintillation liquid. Blank values of uptake were obtained as previously described [26].

Real-time PCR analysis
Total RNA was purified from MDA-MB-231 cells using TRIzol Reagent (Invitrogen, Thermo Fisher Scientific, Massachusetts, USA) as described by manufacturer's manual. After treatment of RNA with DNase I, a first-strand synthesis kit (Invitrogen) was used to generate full-length cDNA from 1 μg of total RNA. qPCR was carried in StepONE Plus Real Time PCR System (Applied Biosystems, Massachusetts, USA), using a FastStart Master SYBR Green I Kit (from Roche, Mannheim, Germany). The primers for amplification are shown in Table 1. Gene expression data were normalized to an endogenous reference β-actin (ACTB) as previously described [27], and according to the manufacturer's instructions.

Migration assay
Tumor cell migration was assayed in a 48-well Boyden chamber (Neuro Probe Microchemotaxis System, Gaithersburg, MD) using 12-μm polycarbonate filters, as previously described [28]. 28 μl of IMDM containing 2.5% FBS was added at the bottom of the chamber as the chemoattractant. At the upper part, 5 x 10 4 cells/well, pretreated with ouabain, furosemide, monensin or phosphonoformic acid (PFA) for 1h, was added in 50 μL of medium in the absence of FBS. Controls were performed by pretreating the cells, for 1 h, in the specific drug diluents: 1% DMSO or 1% ethanol in 50 μL of medium in the absence of FBS., After migration for 1.5h, filters were removed from the chambers, fixed and stained with Panoptic kit (Interlab, São Paulo, Brazil). Tumor cells that had migrated through the membrane were counted in a LX400 light microscopy under a 100 x objective (Labomed Inc, Los Angeles, CA) on at least five random fields. The results are representative of three independent experiments performed in triplicate. Adhesion assay 96-well tissue culture plates were performed pre-coated with 32 μg/mL ECM gel (from mouse sarcoma) diluted in phosphate buffered saline (PBS), 100 μL per well for overnight in 4˚C and non-specific binding sites in the wells were blocked with 1 mg/mL BSA diluted in PBS for 2 h at room temperature. Cells grown at 80% confluency in IMDM with 10% FBS were pretreated with the phosphate transport inhibitors for 1h, then the cells were trypsinized, suspended in serum-free medium at the concentration of 2 x 10 4 cells/mL and applied 100 μL for each well incubated under routine condition as above for 3h. After incubation, non-adherent cells were removed by carefully washing twice with PBS, fixed with 3% paraformaldehyde for 10 min. Cleaned with PBS two times, stained with 100 μL 0.5% crystal violet for 5 min, washed two times and lysed with 100 μL ethanol of 1% acetic acid solution read at A570. Results were expressed 100% as control [29].

Statistical analysis
All experiments were performed, at least, three times in triplicate. The experiments were represented with values of mean ± SE. We used nonlinear regression analysis of the data to the Michaelis-Menten equation (K 0.5 and V max values). Significant differences: p<0.05. Statistical analyses were performed using Prism 6.0 software (GraphPad Software, San Diego, USA).

Pi transport in different lineages of breast cancer cells
In an initial approach, we assayed the Pi transport in three different human breast cancer cell lines. When we compared them, the most aggressive, MDA-MB-231, had the higher P i transport activity (Fig 1A). In addition, we employed two isogenic murine breast cancer cells: 67NR, is a low-aggressive, nonmetastatic, cell line and 4T1, is a highly aggressive, metastatic, cell line [24]. As seen in the human cell lines, 4T1 cells presented higher 32 P i transport than 67NR (Fig 1B). Based on these initial results, we opted to concentrate our study specifically on the MDA-MB-231-line. We next analyzed the gene expression pattern of Pi transporters. As shown in Fig 2, MDA-MB-231 cells exhibited the highest NaPi-IIb gene expression levels when compared to the other NaPi transporters (Fig 2).

Pi uptake and kinetic parameters
The Pi uptake in MDA-MB-231 was evaluated at different times (0 to 60 minutes) and these cells showed a linear uptake for up to 60 minutes ( Fig 3A). We did not measure longer times of incorporation because it could lead to saturation. Then, all experiments were carried out at 60 minutes. NaP i -II (SLC34) is a member of the sodium-dependent P i transporter family and transports the monovalent form of P i . Fig 3B shows  To evaluate the affinity of the transport for P i , we performed the Pi uptake in the P i concentration range of 5-500 μM (Fig 3C). The transporter followed a Michaelis-Menten kinetic and the parameter values were calculated, presenting an apparent K 0.5 = 84.9 ± 10.4 μM and a V max = 71.67 ± 3 nmol x h -1 x mg protein -1 .  pH values ranging from 6.4 to 9.2 were tested, and the transport was higher at pH 7.3-7.7, as shown in Fig 4A. Fig 4B shows the cell viability during the pH experiments. Thus, all the following experiments were performed at pH 7.2, the same as the mean pH of the cell culture medium.
In order to evaluate the inhibition mechanism of Pi transport, we tested some inhibitors: bafilomycin A 1 (100 nM), SCH28080 (100 μM), valinomycin (100 μM), ouabain (1 mM), Furosemide (1 mM), monensin (100 μM) and PFA (5 mM) ( Table 2). With the purpose of verifying the influence of H + gradient on the Pi uptake, we tested the vacuolar H + ATPase inhibitor bafilomycin A 1 and SCH28080, a H + ,K + -ATPase inhibitor, that were not able to inhibit the  P i transport in this cell, as well as valinomycin, a K + ionophore. Only furosemide, ouabain, monensin and PFA inhibited the Pi transport. Taken together, these data suggest the importance of Na + gradient to P i transport, possibly by the involvement of an Na + ,K + -ATPase and/ or Na + -ATPase. The influence of PFA in P i transport was also verified. Increasing concentrations of this inhibitor were tested and Fig 5 shows the dose-response to PFA. These inhibitors do not affect the viability of MDA-MB-231 (Table 2).

Cell migration and adhesion
In order to evaluate the possible impact of Pi transport inhibition on tumor cell properties, we further tested the effect of furosemide, ouabain, monensin and PFA on cell migration and adhesion. All inhibitors were able to inhibit the migration and adhesion of MDA-MB-231 cells (Fig 6). Pi transport inhibitors decreased cell adhesion by approximately 50%. In the case of migration, the inhibition was also significant, especially in the presence of furosemide, ouabain and PFA. These results indicate that Pi uptake is important for pro-tumoral processes such as cell motility, migration and adhesion. None of the inhibitors affected cell viability ( Table 2).

Discussion
Inorganic phosphate (Pi) has been recently proposed as a key component of the "growth rate hypothesis", in which tumors exhibit high phosphorus concentration due to the requirement of a high amount of ribosomes necessary to produce proteins that support the accelerated proliferation of cancer cells [12]. Consistent with the GRH, it was reported that some tumors have higher phosphorous and RNA content than normal tissue, and that phosphorous in RNA have an important contribution to the total biomass of phosphorous in malignant rather than normal tissues [12]. More recently, it was also demonstrated that interstitial inorganic phosphate, which is elevated in tumor microenvironment, could be a new marker of tumor progression [13]. In this study, the authors propose that measurement of interstitial inorganic phosphate would be more sensitive and specific for tumor detection, as compared to blood Pi concentration measurements [13].
In this work we evaluated the biochemical behavior of Pi transport in an aggressive breast cancer cell line, MDA-MB-231 in an attempt to understand the importance of Pi to cancer cells. Initially, we compared the Pi transport of MCF-7, T47D and MDA-MB-231 cells. The high level of Pi transport observed in MDA-MB-231 cell line may reflect aspects of the metabolism associated to the energy demands linked to its metastatic phenotype. Conceivably, the extra Pi incorporated by MDA-MB-231 cells could be utilized as a substrate in ATP biosynthesis, thus enabling cells to sustain their typical proliferation and migration patterns. Indeed, the increased expression of Pi transporters in MDA-MB-231 cells is compatible with such a view [18,30]. Other parameters have been reported that stress the distinct metabolic identities of these cell lines. For example, MCF-7, T47D and MDA-MB-231 cells differ in lactate secretion, 2-NBDG uptake, expression of LDH and sensitivity to histone deacetylase inhibitors [31]. As a matter of fact these differences should not be surprising in view of the diverse phylogenies of MCF-7, T47D and MDA-MB-231 cell lines. Regarding the expression of Pi transporters, NaPi-IIb has been shown to be highly expressed in breast cancer [23]. We observed that the NaPi-IIb had a higher expression compared to other inorganic phosphate transporters in the MDA-MB-231 cell line (Fig 2).
Because a high expression of Pi transporters has been described in breast cancer compared to normal tissue, we have sought to identify the biochemical profile of the Pi transporter (NaPi-IIb) in kinetic terms and to correlate the transport of inorganic phosphate with the tumor processes [23]. According to Forster et al. [6] and Takeda [4], the Pi transporter in breast cells would be sodium dependent and with high affinity for inorganic phosphate. Partially, we correlated this high affinity of the phosphate transporter in breast cancer with the high energy requirement typical of cancer cells [30].
Pi is a tryptotic acid and, thus, has different physiological forms according to the pH range in which it is found [1]. Thus, when the Pi transport levels at different pH ranges were tested, we observed higher levels of transport at neutral pH than at more alkaline pH. With this result, we verified that the sodium-dependent inorganic phosphate carrier has a higher affinity for Pi in the diprotic form (H 2 PO 4 ) found in the blood (pH 7.2).
Previous studies have demonstrated that in mammary glands NaPi-IIb was prevalently expressed as the carrier of Pi, a sodium-dependent carrier capable of transporting the Pi molecule. Furthermore, we demonstrated that the Pi transporter in tumor cells was also sodium dependent. In addition, ouabain (1 mM), furosemide (1 mM) and monensin (100 μM) were able to inhibit the transport of Pi. Furosemide inhibits the Na + -ATPase and ouabain inhibits the Na + ,K + -ATPase. As a result the Na + ion accumulates in the cytosol, leading to indirect inhibition of Pi transport. Monensin is a Na + ionophore, undoing the gradient of this ion across the cytoplasmic membrane. Collectively these results suggest that the inhibitors may deregulate the intracellular sodium gradient and consequently affect sodium-dependent Pi transport (Fig 7).
Phosphonoformic acid is a competitive antagonist of the Pi transporter belonging to the NaPi-II family [32]. As expected, Pi transport into breast cancer cells in the presence of this compound was inhibited in a dose dependent manner, further evidence of the existence of the Na + -dependent Pi transporter on the membrane of these cells.
It was recently demonstrated in other cell types [26,33,34] that some inhibitors of Pi transport, such as FCCP, valinomycin and SCH28080, could have metabolic effects at the mitochondrial level leading to severe ATP depletion, thus indirectly affecting the active Pi uptake. Therefore, we measured the intracellular ATP content under Pi transport conditions, in the absence or presence of monensin, furosemide, ouabain and PFA, which inhibit Pi transport in MDA-MB-231 cells. However, neither inhibitor modulated cellular ATP levels (S1 Fig). Lin et al. [35] identified a dose-dependence of inorganic phosphate for the cell migration of the MDA-MB-231 strain, suggesting the importance of Pi for the metastatic process. Based on this result, we sought to evaluate the importance of the Pi transporter for the metastatic process. Therefore, we used the inorganic phosphate transport inhibitors (monensin, ouabain and furosemide) and the Pi transport inhibitor (PFA) in the cell migration and adhesion assays. Interestingly, we noted a significant inhibition of cell migration, as well as cell adhesion with all inhibitors of Pi transport. PFA has been reported to inhibit the NaPi transport in osteoclast cells [36]. From these results it can be inferred that Pi directly (free Pi), or indirectly (conjugated as ATP) could be accessory to the glycolytic pathway and thus play a role in supplying energy for cellular motility and adhesion of breast cancer cells. Pi uptake appears to be related to the sodium gradient rather than to the Pi transporter itself since ouabain, furosemide and monensin decreased Pi transport, cell migration and adhesion. Therefore, the use of phosphonoformic acid clearly demonstrated that the cells were fully dependent on Pi transport for such functions.
In this study, we characterized the Pi transport in the triple-negative breast cancer cell line, MDA-MB-231. The Pi transport was found to be Na + -dependent, had high affinity for Pi and was more active in acid pH range; it was also inhibited by classical Pi transport inhibitor, PFA. Remarkably, the inhibition of Pi transport caused significant decrease in tumor cell migration and adhesion, suggesting a prominent role for Pi transporters in tumor progression. In this context, these proteins might be regarded as potential therapeutic targets in breast cancer.