2-deoxy-D-Glucose Synergizes with Doxorubicin or L-Buthionine Sulfoximine to Reduce Adhesion and Migration of Breast Cancer Cells

Breast cancer is the most common cancer in women worldwide. It is the major cause of death from cancer among women globally (Mu et al., 2012). In 2014, an estimated 232,670 new cases of invasive breast cancer will be diagnosed in U.S. women. In addition to invasive breast cancer, an estimated 62,570 cases of carcinoma in situ (CIS) will be diagnosed (ACS, 2014). Cancer metastasis represents the final, most advanced stage of malignancy and is the leading cause of death by cancer (Pani et al., 2010; Mu et al., 2012). Metastasis is a multistep process that includes a wide array of cellular changes involving alterations in cell structure by affecting cytoskeleton dynamics and expression of adhesion molecules (Mu et al., 2012; Lee and Kang, 2013; Meng and Yue, 2014).


Introduction
Breast cancer is the most common cancer in women worldwide. It is the major cause of death from cancer among women globally (Mu et al., 2012). In 2014, an estimated 232,670 new cases of invasive breast cancer will be diagnosed in U.S. women. In addition to invasive breast cancer, an estimated 62,570 cases of carcinoma in situ (CIS) will be diagnosed (ACS, 2014).
Cancer metastasis represents the final, most advanced stage of malignancy and is the leading cause of death by cancer (Pani et al., 2010;Mu et al., 2012). Metastasis is a multistep process that includes a wide array of cellular changes involving alterations in cell structure by affecting cytoskeleton dynamics and expression of adhesion molecules (Mu et al., 2012;Lee and Kang, 2013;Meng and Yue, 2014).
Reactive oxygen species (ROS) are highly reactive molecules that are constantly produced in all aerobic organisms, mostly as a consequence of aerobic respiration (Ortega et al., 2010;Nourazarian et al., 2014). The term covers several types of chemical species, including free radicals such as superoxide (O 2-) or hydroxyl (OH·), and non-radicals such as hydrogen peroxide (H 2 O 2 ) (Ortega et al., 2010). It has been reported that ROS contribute, in different ways, to carcinogenesis and to malignant progression of tumor cells (Pani et al., 2010;Nourazarian et al., 2014). The levels of ROS determine their effects on metastasis; where high levels suppress metastasis and destroy cancer cell directly or through the activation of apoptosis (Zhang et al., 2009), while the moderate levels induce cancer cell proliferation, migration, and invasion (Zhang et al., 2009). In addition, ROS play a vital role in regulating actin dynamics, which is sensitive to oxidative modification (Barth et al., 2009).
Previous studies suggest that cancer cells produce more ROS than normal cells (Schumacker, 2006). 8-hydroxy-2-deoxyguanosine (8-OH-DG), which is the major oxidized modified DNA damage, has been reported to be at higher contents in human renal cell carcinoma than in non-tumorous tissue (Toyokuni et al., 1995;Nourazarian et al., 2014). Also, high levels of H 2 O 2 and O 2have been reported in several human carcinoma cell lines, including malignant melanoma, colon carcinoma, pancreatic carcinoma, neuroblastoma, breast carcinoma, and ovarian carcinoma (Toyokuni et al., 1995).
The level of ROS is determined by a balance between ROS generation and detoxifying system (Nakashima et al., 1984), where the imbalance between the oxidants and antioxidants could lead to oxidative stress (Sies, 1997). In order for cells to protect themselves against the damaging effects of ROS, they have evolved enzymatic antioxidants such as super oxide dismutase (SOD), catalase (CAT), heme oxygenase-1 (HO-1) and glutathione peroxidase (GPx) (Steinbrenner and Sies, 2009), and non enzymatic antioxidation mechanisms such as glutathione (GSH) (Wu and Cederbaum, 2003). However, many studies revealed that tumor cells have lower Mn-SOD, CuZn-SOD, and CAT activities when compared to their normal cells counterparts (Oberley and Buettner, 1979;Gurudath et al., 2012). In addition, it has been proposed that GPx activity is highly variable in different tumor tissues (Toyokuni et al., 1995;Gurudath et al., 2012).
Accordingly, cancer cells should have mechanisms other than the enzymatic antioxidant to resist oxidative stress. It has been reported that the metabolism of glucose and cellular pools of glutathione are involved in the protection of cancer cells against exogenous exposure to the hydroperoxide toxicity (Averill-Bates and Przybytkowski, 1994;Hegde et al., 2013).
Glucose metabolism is thought to be involved in cellular sensitivity to oxidative stress mediated by hydroperoxide, presumably via the formation of pyruvate and NADPH (Lee et al., 1998). In this scenario, pyruvate reacts directly with H 2 O 2 resulting in decarboxylation of pyruvate to acetic acid and reduction of H 2 O 2 to H2O (Ahmad et al., 2005;. NADPH is the source of the reducing equivalents for the glutathione/glutathione peroxidase/glutathione reductase system (Coleman et al., 2008).
Recently, it was discovered that the pharmacologic inhibition of glycolysis with 2-deoxy-D-glucose (2DG; a structural analogue of glucose) induced cytotoxicity in MCF-7 breast carcinoma cell line via metabolic oxidative stress (Lin et al., 2003;Lee and Kang, 2013). We have also shown that the combination of 2DG and doxorubicin (DOX) resulted in a significant cell killing in rapidly dividing T47D cells compared to 2DG or DOX alone (Ahmad et al., 2010). In addition, BSO sensitized these cells to the cytotoxicity of the combination of 2DG and DOX (Ahmad et al., 2010).
DOX is an effective therapy that has been used to treat cancers for more than 30 years (Takanashi and Bachur, 1976). However, the chronic side effects of DOX, the most detrimental is cardiomyopathy, limit the amount of this drug that can be given to patients (Takanashi and Bachur, 1976;Bertazzoli et al., 1979;Siveski-Iliskovic et al., 1995;Gille and Nohl, 1997;Zhou et al., 2001;Wold et al., 2005). DOX-causing cardiomyopathy is believed to be redox generated (Wold et al., 2005). Therefore, decreasing the amount of DOX given to patients using combination therapy such as 2DG or BSO which inhibit critical aspects of thiol-mediated hydroperoxide metabolism, leading to increased steady-state levels of ROS, might suppress vital steps during metastasis which could lead to highly desirable outcomes of therapy.
The current work aimed at studying the effects of sublethal metabolic oxidative stress mediated by combination therapy (2DG/BSO, 2DG/DOX, or BSO/DOX) on actin cytoskeleton, focal adhesion and cell migration in T47D human breast cancer cells.

Drug treatment
2DG, BSO and DOX were obtained from Sigma Aldrich. Drugs were added to cells at a final concentration of 20 mM, 1.0 mM and 0.1 µM, respectively. Stock solutions of 2DG and BSO were dissolved in phosphate buffered saline (PBS, pH 7.5), while DOX was dissolved in normal saline (0.9% NaCl). The required volume of each drug was added directly to complete cell culture medium to achieve the desired final concentration. All experiments were performed 24h post treatment.

Glutathione assay
Total glutathione (GSH+GSSG) content in control and treated T47D cells was measured according to Anderson (Anderson, 1985). The method depends on the colorimetric reaction of 5, 5-dithio-bis-2-nitrobenzoic acid (DTNB) with GSH to form 5-thio-2nitrobenzoic acid (TNB). The rate of TNB formation is proportional to the sum of GSH and GSSG present. GSSG is reduced to GSH by glutathione reductase (GR) and NADPH.The rate of TNB formation was determined at 412 nm, every 10 sec for 2.5 min using Jasco-V-530 spectrophotometer.The amount of GSSG in the sample was determined according to (Griffith, 1980). The reduced GSH was determined by subtracting the GSSG content from the total GSH. All Asian Pacific Journal of Cancer Prevention, Vol 16, 2015 3215 DOI:http://dx.doi.org/10.7314/APJCP.2015.16.8.3213 2-Deoxy-D-Glucose Synergizes with Doxorubicin or L-BSO to Reduce Adhesion and Migration of Breast Cancer Cells biochemical determinations were normalized to the protein content using the Bradford method.

Lipid peroxidation assay
Lipid peroxidation was determined by measuring thiobarbituric acid-reactive species (TBARS) as described by Ohkawa et al. (1979) with modifications. Cells were washed with ice-cold PBS, scraped into cold PBS and centrifuged for 5 min at 15,000 rpm. Cell pellet was homogenized in 250 μL 1% sodium dodecyl sulfate (SDS; BDH, England), and 1 ml of 20% trichloroacetic acid (TCA; Sigma Aldrich, USA). 2 ml of 0.8% thiobarbituric acid (TBA; ACROS, USA) and 250 μL of 1% phosphoric acid were added to the homogenate. The mixture was then boiled for 1 h, cooled on ice and then extracted with 1 ml n-butanol. The mixture was centrifuged briefly, and the absorbance of the upper phase was then measured at 530 nm with Jasco-V-530 spectrophotometer (Gavino et al., 1981;Lykkesfeldt, 2007). The values are expressed in nmole/mg protein malondialdehyde [MDA; end product of lipid peroxidation (Nourazarian et al., 2014).

Cell proliferation assay (MTT assay)
Proliferation assay is widely used as an indicator of cell metabolic viability (Coleman et al., 2008). Cell proliferation assay (MTT assay) was performed according to the method of Mosmann (Mosmann, 1983) with modifications. Briefly, 5510 4 viable cells were seeded in the wells of a 96-well tissue culture plate containing growth media supplemented with FBS. Cells were kept in a humified 5% CO 2 incubator at 37 o C for 24h. Freshly prepared MTT salt (5 mg/ml in PBS) (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) was added to each well to give a final concentration of 0.5 μg/μl. The plates were incubated for 4 h and the formation of formazan crystals was checked using an inverted microscope. Equal volume of 1:1 (200 μl) mixture of dimethyl sulfoxide (DMSO; TEDIA Co., USA) and isopropanol was added to each well and incubated for 30-45 min. Cell viability was evaluated by measuring the absorbance at 570 nm using a Jasco-V-530 spectrophotometer. The percentage of survival was calculated as the proportion between absorbance of treated and control cells.

Cell fixation and immunostaining
T47D cells were plated on sterile glass cover slips at density of 1510 4 . Twenty four h post drug treatment, cells were fixed with 4% paraformaldehyde (Sigma Aldrich, USA) in cytoskeleton buffer with sucrose (CBS) [10 mM MES, pH 6.1, 138 mM KCl, 3 mM MgCl 2 , 10 mM EGTA, 0.32 mM sucrose] for 45 min at room temperature. Cells were then washed three times five min each with 0.1% Triton X-100 in PBS.
To visualze actin cytoskeleton, F-actin was stained with fluorescent-conjugated phalloidin (Invitrogen, USA) in CBS for 1h. To visualize adhesion structures, cells were incubated with anti-paxillin antibody (BD Pharmingen, USA) [1:250 in 2% bovine serum albumin (BSA) in CBS] for 1 h and then with fluorescent-conjugated goat anti-mouse IgG (1:400 in 2% BSA in CBS) for 1h. Finally, cells were washed three times five min each with 0.1% Triton X-100 in PBS and mounted with Prolong Gold Antifade with DAPI (Invitrogen, USA). Images were captured using either a 40× NA 0.65 or 100× NA 1.3 objectives on an inverted Nikon microscope with a CCD camera.

Cell adhesion assay
Cells were plated on ice-cold freshly prepared collagen I (10μg/ml) (Sigma Aldrich, USA) -precoated 12 well tissue culture plate at a final concentration of 5510 4 /well (Estornes et al., 2007). Each well was treated with 150 μl collagen I, left for 1 h at 37oC incubator and then washed three times with PBS. Cells were incubated at 37 o C on collagen I for 1 h and washed twice with PBS. Adherent cells were fixed with 4% paraformaldehyde for 30 min at room temperature, stained with 1% methylene blue (Gurr Certistain, England), diluted in 1 % Borax for 30 min at room temperature. After solubilization with 1% SDS, absorbance at 630 nm was measured.

Invadopodia assay
Ethanol-flamed sterile 18 mm glass coverslips were placed on the wells of 12-well tissue-culture plate and were coated with 50 μg/ml poly-L-lysine for 20 min at room temperature. The coverslips were covered with 0.5% glutaraldehyde for 15 min, and then were coated with 37oC-preheated 0.2% gelatin (Sigma Aldrich, USA) and Alexa Fluor 488 or 568-gelatin (Invitrogen, USA) mixture at a 8:1 ratio for 10 min at room temperature. The residual reactive groups in the gelatin matrix were quenched with 5 mg/ml sodium borohydride for 15 min at room temperature. Cells were plated at a concentration of 2×104/coverslip and incubated at 37°C for 6h. Finally, cells were stained for F-actin with fluorescent phalloidin (Artym et al., 2009).

Wound healing assay
T47D cells were grown on 6-well tissue culture plate to about 80% confluency. Cultures were wounded by a heat polished glass pipette (~ 30-50-μm tip) and overlayered with dimethyl polysiloxane (Sigma Aldrich, USA) to reduce evaporation while allowing gas exchange. Detailed observation on the behavior of live cells was monitored and the effects of the different drug treatments were assessed by measuring the time and the distance migrated by cells to close the wound. Live cell migration in woundhealing assay was followed using a CCD camera operated by Nikon ECLIPSE Ti widefield microscope equipped with a 37°C stage and 5% CO 2 .

Migration assay
Control and treated T47D cells (1510 6 ) were seeded into the upper compartment of a 12-well chemotaxis chamber (Neuroprobe, Gaithersburg, MD). Both the upper and lower compartments were filled with HGDMEM:F12 containing 0.35% BSA and were physically separated by a polycarbonate membrane (8-µm pore size) precoated for 4 h with 100 µg/ml collagen I. Cells were then incubated for 6h at 37 o C in 5% CO 2 humidified conditions. Cells were fixed with 4% paraformaldehyde, and stained with 1% borax and 1% methylene blue. Cells of the upper surface of the filter were removed with a cotton swab, and those underneath were quantified (Estornes et al., 2007;Mu et al., 2012).

Statistical analysis
Statistical analysis was performed using STATISITCA 7 analysis program (StatSoft Inc., Ok, USA). To determine differences between 3 or more means, one-way ANOVA with Fisher's LSD for multiple comparisons post-tests were performed. Results are presented as mean±standard error of the mean (SEM). All statistical analysis was performed at p<0.05 level of significance.

Combination therapy causes disruption in intracellular thiol pool
The changes in intracellular total glutathione (TGSH; GSH+GSSH) levels were measured in T47D human breast cancer cells treated with 2DG, DOX, and BSO individually or in combination for 24h. Figure 1A shows that individual drug treatment caused a significant decrease in total glutathione compared to untreated control cells, and that two drug combinations (2DG/BSO and BSO/DOX) enhanced TGSH depletion compared to individual drug treatment ( Figure 1A). In addition, drug combinations caused a significant increase in % GSSG (oxidized glutathione) compared to control cells and to individual drug treatment (except for 2DG/DOX) ( Figure  1B). We have shown previously that the effects of these drugs on intracellular glutathion pool was mediated by disruption in thiol metabolism and thus oxidative stress (Ahmad et al., 2010). Lipid peroxidation is the hallmark of oxidative stress (Wu and Cederbaum, 2003), thus malondialdehyde (MDA) level (secondary product of lipid peroxidation) was measured. As shown in Figure   1C, MDA levels were significantly elevated in treated cells. In addition, all of the drug combinations (2DG/BSO, 2DG/DOX or BSO/DOX) caused a significant increase in MDA levels when compared to individual drug treatment ( Figure 1C).

Drug combination causes sub-lethal T47D cytotoxicity
T47D cells were treated with the different drugs individually or in combination and cell viability was measured using the MTT-based cell viability assay ( Figure  2A). Our data show that individual drug treatment had a sub-lethal cytotoxity effect (around 20% reduction in cell viability) on T47D cells, which was more profound with drug combination (around 30% reduction in cell viability) (Figure 2A). BSO/DOX drug combination was the only combination that caused a significant increase in cytotoxity when compared to individual drug treatment (BSO or DOX alone) (Figure 2A).

Heme oxygenase-1 is induced in treated T47D breast cancer cells
To study the effects of TGSH depletion on the oxidative status of treated T47D breast cancer cells, the expression of three major antioxidant enzymes (CAT, HO-1, and MnSOD) was studied by immunoblotting ( Figure  2B, C). Drug treatment (individually or in combinations) induced the expression of HO-1 as compared to control,   Figure 2C). On the other hand, only 2DG/BSO drug combination induced the expression of MnSOD significanlty. The expression of catalase was reduced in all drug combinations as compared to control but not to individual drug treatment ( Figure 2C).

Drug-treated T47D cells are characterized by a smaller cell area
To investigate the effect of drug treatment on the morphology of T47D cells, we measured cell length, width, length to width ratio (L/W ratio) and area of control and treated cells (Table 1). Cell width of treated cells decreased significantly when compared to control cells (Table 1). This in turn caused a decrease in cell area in   B) Data are expressed as mean±SEM, n≥600 cells in each experiment, three independent experiments. a p<0.05; a 2 p<0.01; a 3 p<0.001 versus control, b p<0.05 2DG/BSO versus 2DG alone, c p<0.05 2DG/BSO versus BSO alone, e 2 p<0.01 2DG/DOX versus DOX alone, f p<0.05 BSO/DOX versus BSO alone, g p<0.05 BSO/DOX versus DOX alone. Scale bar: 10 µm drug-treated cells when compared to control cells (Table  1), while the length and L/W ratio were not affected (Table1).

Drug combination affects mitotic parameters
Since mitosis and cytokinesis are actin-dependent processes, we wanted to investigate the effects of the drugs on certain mitotic parameters such as mitotic index (no. of mitotic cells/total no. of cells × 100%) ( Figure 3A, B), percentage of multinucleation (no. of cells having two or more nuclei/total no. of cells × 100%) ( Figure 3A, B), and percentage of micronucleation (no. of cells having fragments or whole chromosomes lagging behind in anaphase/total no. of cells × 100%) ( Figure 3A, B). Drug treatment caused a significant reduction in the percentage of mitotic cells as compared to control cells ( Figure  3B). Except for 2DG/DOX drug combination, the other two combinations (2DG/BSO and BSO/DOX) caused a significant decrease in the percentage of mitotic cells as compared to individual drug treatment.

Drug-treated cells exhibit changes in actin cytoskeleton
To determine if T47D-treated cells show changes in F-actin organization, T47D cells were treated with the different drugs for 24h, fixed and stained with fluorescentphalloidin. Cells were observed and divided into three categories according to the structure of their actin cytoskeleton ( Figure 4A) as described previously (Zebda et al., 2000;Sidani et al., 2007;Meng and Yue, 2014). Our results show that all of the drugs, used individually or in combination, significantly caused enhanced F-actin aggregates formation as compared to control cells accompanied by a significant decrease in normal and prominent F-actin ( Figure 4B). Cells treated with the drug combination BSO/DOX showed the most enhanced actin aggregates formation ( Figure 4B).

Disruption of thiol metabolism reduces T47D cell adhesion to collagen I
Changes in cell morphology and actin cytoskeleton structure suggested that drug treatment might have affected cell adhesion. We next investigated the effect of disrupting thiol metabolism on T47D cell adhesion ( Figure  5). Control and drug-treated cells were stained with antipaxillin antibody ( Figure 5A) and the size and number of focal adhesions at cell periphery were measured per unit area (30 μm 2 ) ( Figure 5A, B). Our results indicate that individual drug treatment decreased the total area occupied by focal adhesion as compared to control cells ( Figure  5B). When drugs were used in combinations, the redcution in the total area occupied by focal adhesions decreased even more when compared to individual drug treatment ( Figure 5B). In addition, control and treated cells were seeded onto collagen I-coated dishes, and adherent cells were quantified after 1 h (adhesion assay; (Estornes et al., 2007). We found that the percentage of adherent cells was less in drug-tretaed cells compared to control cells ( Figure  5C). All drug combinations enhanced this effect on cell adhesion when compared to individual drug treatment.    DOI:http://dx.doi.org/10.7314/APJCP.2015.16.8.3213 2-Deoxy-D-Glucose Synergizes with Doxorubicin or L-BSO to Reduce Adhesion and Migration of Breast Cancer Cells To examine the effects of drug treatment on T47D cells ability to degrade the extracellular matrix (ECM), cells were grown on fluorescent gelatin and the proteolytic activity (dark non-fluorescent areas; Figure 6A) was measured ( Figure 6B). Drug treatment reduced the degradative ability of cells as compared to control and all drug combinations reduced this ability when compared to individual drug treatments ( Figure 6B).

Metabolic oxidative stress decreases the rate of migration of T47D breast cancer cells
Since disruption of thiol metabolism affected actin organization and cell adhesion, we next analyzed the effect of drug treatment on the migration of T47D breast cancer cells. The wound healing assay measures cell directed migration as a response to clearing of cells in a monolayer (Montanez et al., 2007). The migration rate of cells treated individually decreased significantly when compared to the control (p<0.01) ( Figure 7A). Treating T47D breast cancer cells with the combination therapy enhanced this reduction ( Figure 7B). In addition, we measured the number of control and treated cells migrating across type I collagen-coated 8 µm filters (migration assay). Treating T47D cells with BSO or the combination therapy reduced T47D cell migration by almost 80% compared to control cells ( Figure 7B).

Discussion
Breast cancer has become the second leading malignancy among women in both developing and developed countries (Li et al., 2013). Cancer cell metastasis is the most sever stage in cancer and is the main cause of morbidity in cancer patients (Gupta and Massague, 2006). A major challenge in cancer therapy is to discover means to inhibit tumor cell migration from primary site to other body organs (i.e. inhibit metastasis) (Hayot et al., 2006;Hegde et al., 2013).
As in most other solid tumors, breast cancer metastatic phase rather than the primary lesion is responsible for the vast majority of patients' morbidity and mortality (Dong et al., 2009). Therefore, metastasis continues to be the main obstacle to the effective treatment of breast cancer, and there is an urgent need to develop novel therapeutics to inhibit this process.
Recently, it has been noted that metastatic cells have an increased dependency on glycolytic pathways while down-regulating pathways involved in oxidative phosphorylation (Sottnik et al., 2011;Jin and Wei, 2014). This is a further enhancement of the Warburg effect which states that tumor cells become dependent on glycolysis and decrease their oxygen utilization as they become more aggressive (Zhao et al., 2013;Jin and Wei, 2014). These metabolic differences arise as a result of a fundamental defect in the tumor mitochondrial electron transport chain which leads to increased steady state levels of pro-oxidants (Nakashima et al., 1984). Previous studies suggest that breast cancer patients are exposed to conditions of oxidative stress (Noda and Wakasugi, 2001). As a compensatory mechanism in response to this increase in ROS, cancer cells exhibit an increase in glucose cells were grown into monolayers and wounds were made with a sterile tip. The wound area was measured at 0 hr and later at 6 h, and migration rate was expressed as µm/h. Three independent experiments each performed in triplicate. (B) Cells were serumstarved, seeded on collagen I-precoated filters and then subjected to migration assay. Cell migration is expressed as percent of control cells.  The degradation area of gelatin (µm 2 ) was quantified in NIS-Elements AR 4.10.00 and divided by number of cells in the same field and expressed as degradation area (µm 2 )/cell. n≥25 cells in each experiment, three independent experiments. a 2 p<0.01 versus control, b p<0.05 2DG/BSO versus 2DG alone, d 3 p<0.001 2DG/ DOX versus 2DG alone, e 3 p<0.001 2DG/DOX versus DOX alone, f p<0.05 BSO/DOX versus BSO alone, g p<0.05 BSO/ DOX versus DOX alone. Scale bar: 10 µm metabolism by forming pyruvate and NADPH which are believed to function in hydroperoxide detoxification (Nakashima et al., 1984). Accordingly, the increased dependency of metastatic cancer cells on glucose metabolism is an attractive target for treatment strategies since these pathways are common to tumor cells.
Several studies have shown that glucose deprivation can induce cytotoxicity in transformed human cell types via metabolic oxidative stress (Lee et al., 1998;Lin et al., 2003). In our previous study, we demonstrated that the inhibition of glycolysis using 2DG induced cytotoxicity in T47D breast cancer cells (Ahmad et al., 2010). This is consistent with previous studies in which it was shown that 2DG induced growth inhibition and cytotoxicity in MDA-MB231 breast cancer cells (Andringa et al., 2006), and FaDu squamous cell carcinoma (Simons et al., 2007).
On the other hand, when combining 2DG with DOX a significant cell killing in rapidly dividing T47D cells compared with 2DG or DOX alone was observed (Ahmad et al., 2010). This is also consistent with previous studies, suggesting that 2DG may potentially increase the efficacy of standard chemotherapeutic drugs (Maschek et al., 2004). Furthermore, we previously reported that BSO, a potent inhibitor of glutathione biosynthesis, sensitized T47D cells to apoptosis induced by 2DG and DOX, which suggests that inhibition of GSH synthesis further enhanced the oxidative stress induced by these drugs (Ahmad et al., 2010).
The role of ROS in controlling vital cellular processes has emerged in the last decade (Moldovan et al., 2006). These molecules function as signaling molecules in many physiological processes including cell growth, gene expression, and cell migration (Moldovan et al., 2006;Shinohara et al., 2007). The role of ROS in cell migration depends on the type of the ROS, dose, production site and cell type (Luanpitpong et al., 2010).
In the current study we validated our previous findings in terms of being able to induce metabolic oxidative stress, and our results were consistent with what has been published previously (Ahmad et al., 2010) (Figure 1 and 2). Accordingly, we hypothesized that 2DG or DOX can act synergistically with inhibitors of hydroperoxide detoxification (i.e. BSO) to produce oxidative stress and to revert highly metastatic cancer cells to a less aggressive phenotype, leading to inhibition of metastasis.
Metastasis is a multistep process characterized by cellular deformation involving the formation of protrusions and new adhesions to surfaces as well as cellular contractility, which is required for rear retraction and cellular transmigration (Menhofer et al., 2014). In all of these processes, actin cytoskeleton plays a pivotal role and undergoes a constant polymerization and depolymerization to form protrusions and stress fibers (Sakthivel et al., 2012;Tahtamouni et al., 2013).
ROS have a key regulatory effect on actin cytoskeleton, which has a dynamic structure that participates in cell motility. It has been reported that oxidation of β\γ (non-muscle forms of actin) leads to complete loss of polymerization (Lassing et al., 2007). Another study reported that the increase in the ROS superoxide (O 2 -) causes carbonylation of actin, leading to loss of lamellopdia formation (Barth et al., 2009). In addition, ROS regulate actin reorganization via Rac1-mediated ROS production from NADPH oxidase, which is responsible for electron transfer to molecular oxygen leading to superoxide production (Alexandrova et al., 2006). The prevention of superoxide formation from NADPH oxidase reduces cell migration in hypoxic endothelial cell (EC) (Moldovan et al., 2006). Therefore, we first examined alterations in actin cytoskeleton organization and cell shape changes. In T47D cells, actin cytoskeleton exists in three different categories: normal F-actin, F-actin aggregates, and prominent F-actin (stress fibers) ( Figure 4A). Our results show that all of the drugs used individually or in combination, significantly enhanced F-actin aggregates formation as compared to control cells accompanied by a significant decrease in normal and prominent F-actin ( Figure 4B). The combination of BSO/DOX elicited the most enhanced actin aggregate formation ( Figure 4B). F-actin aggregate formation might indicate enhanced myosin II activities (Wiggan et al., 2012).
Actin cytoskeleton is also involved in regulating the cell cycle, during which it undergoes drastic changes and remodeling (Heng and Koh, 2010). Thus, the regulation of actin cytoskeleton and the cell cycle progression appears to be connected (Heng and Koh, 2010). Our result show that disruption of actin cytoskeleton dynamics was accompined with a delay in entry to mitosis ( Figure 3B).
ROS regulate cell migration via regulating many proteins such as integrins, small Rho GTPase family, focal adhesion-forming proteins, and extracellular matrix degrading enzymes such as metalloproteinase (MMP) (Luanpitpong et al., 2010). An important observation of our study is that individual or combination treatments induced morphological changes in T47D cells (Table  1). T47D cells, when grown on tissue culture dishes, display typical epithelioid characteristics; cells are flat and polygonal in shape and are very adhesive to the plastic substratum (Shiu and Paterson, 1984). We found that treated cells became smaller and round compared to control cells (Table 1). Consequently, cell adhesion and invasion were also dramatically impaired ( Figure 5, 6, and 7).
Our data show that the percentage of adherent cells was less in drug-treated cells compared to control cells ( Figure 5C). All drug combinations enhanced this effect on cell adhesion when compared to individual drug treatment ( Figure 5C). Focal adhesions are sites of large macromoleuclar assemblies containing integrins with linkages to cytoplasmic actin bundles and collagen (Chen et al., 2003;Zaidel-Bar et al., 2004). Since our drug treatment increased F-actin aggregate formation with a reduction in stress fiber formation, this in turn reduced focal adhesion formation ( Figure 5A). Focal adhesions provide a tension force needed for the forward movement of cell body.
In addition, drug treatment reduced the degradative ability of cells to ECM as compared to control, and all drug combinations reduced this ability when compared to individual drug treatments ( Figure 6). Different reports have shown that successful tumor cell metastasis requires 2-Deoxy-D-Glucose Synergizes with Doxorubicin or L-BSO to Reduce Adhesion and Migration of Breast Cancer Cells cells to adhere to blood vessels of host organ (Korb et al., 2004). Thus, impairing tumor cell ability to form stable adhesion to vessel wall would help in preventing tumor ability to migrate, invade, and metastasize. Our data presented in Figure 5 indicate that drug treatment decreased T47D cell adhesion, leading to decreased migration rate (Figure 7) (Tahtamouni et al., 2013).
The results presented here indicate that the antimetastatic properties of the different drugs used in the current study, may be mediated by impaired actin polymerization and cytoskeletal control. Our findings may contribute to the understanding of the relationship between metabolic oxidative stress and breast cancer cell metastasis progression. In addition, our findings provide a strong support for the potential therapeutic use of sublethal metabolic oxidative stress mediated by glucose and glutathione depletion.