Suppression of adenine nucleotide translocase-2 by vector-based siRNA in human breast cancer cells induces apoptosis and inhibits tumor growth in vitro and in vivo

The human metastatic breast carcinoma cell lines MCF7 and MDA-MB-231 as well as ovarian cancer cell lines SK-OV-3 and SNU8 were used throughout the study. These cells were provided by the Korean Cell Line Bank, Seoul, Korea and were cultured in DMEM supplemented with 10% FBS, 100 u/ml penicillin and 100 μg/ml streptomycin. MCF-10A, a line of healthy epithelial cells from human mammary gland (CRL-10317) provided by the American Type Culture Collection (Manassas, VA, USA) was used as a control. This cell line was maintained in the DMEM/Ham's Nutrient Mixture F-12 (1:1) with the addition of epidermal growth factor (20 ng/ml), cholera enterotoxin (100 ng/ml), insulin (10 μg/ml) and hydrocortisone (500 ng/ml) in the presence of 5% horse serum.

ANT2 siRNA-1, siRNA-2 and siRNA-3 were synthesized by Bioneer (Daejeon, Korea), and pSilencer™ 3.1-H1 puro plasmids for DNA vector-based siRNA synthesis were purchased from Ambion (Austin, TX, US). The oligonucleotide pairs of ANT2 siRNA-1, siRNA-2, and siRNA-3 are complementary to exon 2 or exon 4 (Genbank accession number NM001152), and the sequences of ANT2 siRNA-1, siRNA-2 and siRNA-3 are 5'-GCAGAUCACUGCAGAUAAGTT-3', 5'-CTGACATCATGTACACAGG-3' and 5'-GATTGCTCGTGATGAAGGA-3', respectively. The oligonucleotide pairs were designed to contain a terminal BamHI or HindIII restriction site for subcloning into the BamHI or HindIII site of pSilencer™ 3.1-H1 puro vector to generate pSilencer™ 3.1-H1 puro ANT2 siRNA vectors (shRNAs). These vectors produce a shRNA with a TTCAAGAGA linker sequence that forms looped structures. This linker is processed with Dicer to generate an ANT2-specific siRNA. A negative scrambled siRNA (Ambion) control with no significant homology to mouse or human gene sequences was designed to detect nonspecific effects.

For transfection, cells were plated on either six-well plates (2 × 10⁵ cells per well) or 100 mm dishes (2 × 10⁶ cells) and were allowed to adhere for 24 hours. Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) was used for the transfections. pSilencer™ 3.1-H1 puro ANT2 siRNA vectors or pSilencer™ 3.1-H1 puro scramble siRNA vector were transfected into the cells. Transfected cells were then cultured for 4 hours and the culture media were replaced with fresh media supplemented with 10% FBS. The cells were harvested at 24–48 hours after transfection.

After 48 hours of transfection, the cells were collected and total RNA was extracted using Trizol (Invitrogen) according to the manufacturer's instructions. For RT-PCR analysis, 5 μg total RNA was reverse-transcribed using RT-PCR kits (Promega, Madison, WI, USA). PCR was used to amplify target cDNA with the following conditions: 35 cycles of 94°C for 1 minute, 55°C for 1 minute and 72°C for 2 minutes. The PCR products were analyzed using standard agarose gel electrophoresis.

The primers used for RT-PCR are ANT1 forward, 5'-ACAGATTGTGTGGTTT-3' and reverse, 5'-TTTTGTGCATTAAGTGGTCTTT-3' ; ANT2 forward, 5'-CCGCAGCGCCGGAGTCAAA-3' and reverse, 5'-AGTCTGTCAAGAATGCTCAA-3' ; ANT3 forward, 5'-AACCAAGAGAACCACGTAGAA-3' and reverse, 5'-CTTAGAACAGACTTGGCTC-3' ; TNF-receptor I forward, 5'-CTGCCTCAGCTGCTCCAAA-3' and reverse, 5'-CGGTCCACTGTGCAAGAAGAG-3' ; and β-actin forward, 5'-GGAAATCGTGCGTGACATTAAGG-3' and reverse, 5'-GGCTTTTAGGATGGCAAG GGA C-3'.

Anti-ANT, anti-Bcl-xL, anti-Bax, anti-α-tubulin as well as anti-caspase-3 antibodies were obtained from Santa Cruz Biotech (Santa Cruz, CA, USA) and polyclonal anti-ANT3 antibody was donated by Dr HH Schmid (University of Minnesota, MN, USA).

For western blot analyses, cells were harvested after 48 hours of transfection and were lysed with lysis buffer (5 mM/l ethylenediamine tetraacetic acid; 300 mM/l NaCl; 0.1% NP-40; 0.5 mM/l NaF; 0.5 mM/l Na₃VO₄; 0.5 mM/l phenylmethylsulfonyl fluoride; and 10 μg/ml each of aprotinin, pepstatin and leupeptin; Sigma, St Louis, MO, USA). After centrifugation at 15,000 × g for 30 minutes, the concentrations of supernatant proteins were analyzed by Bradford reagent (Bio-Rad, Hercules, CA, USA).

For the analysis of protein contents, 50 μg total proteins was electrophoresed in 10% SDS-PAGE gel, transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA) and were then incubated with the respective antibodies indicated above. Immunoblots were visualized using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech, Uppsala, Sweden).

Approximately 2 × 10⁵/ml MDA-MB-231 cells were transfected with respective pSilencer™ 3.1-H1 puro ANT2 siRNA-1, siRNA-2 and siRNA-3 vectors as well as with pSilencer™ 3.1-H1 puro scramble siRNA vector for the indicated times. The transfected cells were harvested, washed twice with PBS and were then incubated for 15 minutes at room temperature with a solution of annexin V conjugated with fluorescence isothiocyanate (2.5 μg/ml) and propidium iodide (PI) (5 μg/ml) (BD Pharmingen, San Diego, CA, USA) for flow cytometry (Epics XL; Coulter, Marseille, France) to detect the levels of apoptosis. Genomic DNA was extracted using genomic DNA extraction kits (G-DEX™IIc; Invitrogen, Seoul, Korea) and was subjected to electrophoresis in 2% agarose gels for DNA fragmentation analysis.

ATP assays were conducted using CellTiter-Glo™ Luminescent Cell Viability assay kits (Promega) that quantify ATP levels in viable cells. This bioluminescence assay utilizes luciferase, which induces light emission during the interaction between ATP and luciferin. Lyophilized enzyme/substrate mixtures (250 μl) were transferred to opaque 96-well microplates containing cell lysates. The plates were incubated at room temperature for 10 minutes to stabilize luminescence signals and then the stabilized signals were quantified with an Orion Luminometer (Berthold Detection Systems, Oak Ridge, TN, USA).

The cells transfected with pSilencer™ 3.1-H1 puro ANT2 siRNA or pSilencer™ 3.1-H1 puro scramble siRNA vector were trypsinized, counted, centrifuged and fixed in ethanol for 3 hours. These cells were then washed twice in PBS and centrifuged. Pellets were resuspended with a solution containing RNase (0.02 mg/ml) (Sigma), incubated at 37°C for 30 minutes and were stained with PI (0.02 mg/ml) (Sigma). The cells were analyzed by flow cytometry (Epics XL; Coulter).

To measure mitochondrial membrane potential disruption, the cells transfected with pSilencer™ 3.1-H1 puro ANT2 siRNA or pSilencer™ 3.1-H1 puro scramble siRNA vector were harvested, washed twice with PBS and were incubated with 20 nM 3,3'-diethyloxacarbocyanine (Molecular Probes, Eugene, OR, USA) for 15 minutes at 37°C. Mitochondrial membrane potential values were determined by flow cytometry (Epix XL; Coulter).

MDA-MB-231 cells (1.5 × 10³) were cultured with the culture media of pSilencer™ 3.1-H1 puro ANT2 siRNA-1/siRNA-2/siRNA-3-transfected cells for 24 hours and were then harvested. Cell death was determined by annexin V–fluorescence isothiocyanate/PI staining followed by flow cytometry analysis (Epics XL; Coulter).

Respective pSilencer™ 3.1-H1 puro ANT2 siRNA-1, siRNA-2 and siRNA-3 vectors as well as pSilencer™ 3.1-H1 puro scramble siRNA vector were transfected into MDA-MB-231 cells for the indicated times. Six hours before harvesting, the cells were treated with brefeldin A (10 μg/ml), washed twice with PBS, fixed with 2% paraformaldehyde, permeabilized with buffer (1% BSA, 0.1% saponine, 0.1% sodium azide in PBS) and were then stained with phenylethylene-conjugated anti-TNFα, anti-IFNγ, anti-IL-12 as well as anti-mouse IgG antibodies for 1 hour at 4°C (BD Pharmingen). Surface staining was performed with phenylethylene-conjugated anti-TNF-receptor 1 as well as anti-mouse IgG antibodies and then analyzed by flow cytometry (Epics XL; Coulter).

For tumor challenges, we established tumor models in 6-week-old to 8-week-old Balb/c nude mice by subcutaneously injecting 5 × 10⁶ MDA-MB-231 cells into the right flanks. Treatments were started from 2 weeks after tumor inoculation when tumor volumes were <100 mm³. Intratumoral injections of PBS, pSilencer™ 3.1-H1 puro scramble siRNA vector (100 μg) or respective pSilencer™ 3.1-H1 puro ANT2 siRNA-1, siRNA-2 and siRNA-3 supplemented with Lipofectamine 2000 (200 μl) were performed three times per day for 5 days. Tumor sizes were measured using a caliper every week until 56 days after tumor challenges, and tumor volumes calculated using the formula: m₁² × m₂ × 0.5236 (where m₁ represents the shortest axis and m₂ the longest axis)

Data were analyzed using the Student's t test. P < 0.05 was considered statistically significant.

To explore the functions of ANT2 in cancer, we first examined its mRNA expression levels in human cancer cell lines obtained from various organs (Figure 1a). RT-PCR showed that ANT2 mRNA was expressed in various human cancer cells originated from the stomach, lung, liver, ovary and breast. ANT2 mRNA was dramatically overexpressed in human breast cancer cell lines (MCF7, MDA-MB-231 and SK-BR-3) and was overtly increased in ovarian cancer cell lines (SK-OV-3 and SNU8), however, suggesting that overexpression of ANT2 could be a unique feature of breast cancer and ovarian cancer that may play a role in cancer development.

To expand the knowledge of ANT mRNA expressions among breast cancer cell lines, the MCF7 and MDA-MB-231 cell lines, as well as the non-neoplastic mammary epithelial cell line MCF10A, the relative expressions of three ANT isoforms from these cell lines were investigated at mRNA and protein levels (Figure 1b). In terms of mRNA expressions, ANT2 is highly expressed in both breast cancer cell lines while ANT1 and ANT3 expressions are not detectable. In contrast, the non-neoplastic MCF10A cell line barely expresses ANT2 as well as ANT1 but highly transcribes ANT3. Serendipitously, the transcriptional patterns of ANT isoforms in ovarian cancer cell line SK-OV-3 are identical to breast cancer cell lines (Additional File 1a), emphasizing the significance of ANT2 overexpression in breast cancer cells and ovarian cancer cells.

ANT2 protein levels were also tested from the breast cell lines by western blotting (Figure 1b). Owing to the absence of ANT1-specific and ANT2-specific antibodies, however, the protein levels of ANT2 were indirectly detected by subtracting the amount of ANT3 from total ANT protein levels. The total protein levels of ANT in respective breast cell lines are about the same, but ANT3 protein is not detectable from the breast cancer cell lines. On the other hand, ANT3 protein is dominantly expressed in the non-neoplastic MCF10A cell line where ANT3 mRNA is highly expressed. This result suggests that the breast cancer cell lines have more ANT1 and ANT2 proteins than ANT3 proteins. No ANT1 mRNA is detectable from both breast cancer cell lines, however, indicating the majority of ANT proteins in breast cancer cell lines might be ANT2 proteins. Collectively, these arrays of results suggest that the higher expression of ANT2 in breast cancer and ovarian cancer cell lines is a unique feature that may play a key role in cancer development.

Owing to the unique feature of overt ANT2 overexpression in breast cancer and ovarian cancer cell lines, we hypothesized that the downregulation of ANT2 in these cell lines may induce serious physiological effects that result in cell growth inhibition or cell death. To evaluate the functions of ANT2 in cancer development or the possible utility of ANT2 downregulation for breast cancer therapy, therefore, ANT2-specific knockdown experiments were performed with human breast cancer cell lines. Initially, respective siRNA-1, siRNA-2 and siRNA-3 oligonucleotides directed at exon 2 or exon 4 of ANT2 mRNA were synthesized to determine whether ANT2 siRNA suppresses ANT2 expression. At the same time, DNA oligonucleotides representing the siRNA duplex were cloned into pSilencer™ 3.1H1 puro vector to produce a high-level silencing effect based on a DNA vector system. The synthesized shRNA derived from DNA templates was composed of two identical 21-nucleotide sequence motifs in an inverted orientation, separated by a 9 base pair nonhomologous spacer (shRNA).

To confirm the silencing efficacies of ANT2 siRNA-1, siRNA-2 and siRNA-3 as well as pSilencer™ 3.1H1-ANT2 siRNA (ANT2 shRNA-1, siRNA-2 and siRNA-3), the cell lines MCF7 and MDA-MB-231 were transfected with the above samples as well as scramble siRNA and pSilencer™ 3.1H1-scramble siRNA (scramble shRNA) as negative controls. After culturing for 48 hours, their efficacies in extinguishing ANT2 mRNA expression were evaluated by RT-PCR. As shown in Figure 1c, the treatment with ANT2 shRNA was much more effective than the treatment with ANT2 siRNA in terms of downregulating the ANT2 mRNA level. Treatment with ANT2 shRNA for 48 hours decreased ANT2 mRNA expressions in MCF7 and MDA-MB-231 cells by over 90% compared with the cells transfected with control vectors, while ANT2 siRNA treatment resulted in markedly less reduction. The total protein levels of ANT were also downregulated by ANT2 shRNA, indirectly indicating that the knockdown of ANT2 reduces the protein levels of ANT2 (Figure 1c). These results suggest that ANT2 shRNA is superior to ANT2 siRNA for downregulating ANT2 mRNA as well as protein levels and can be used to target ANT2 for breast cancer therapy.

Taken together, we have uniquely identified the higher expression of ANT2 mRNA from human breast cancer cell lines MCF7, MDA-231 and SK-BR-3 as well as ovarian cancer cell lines SK-OV-3 and SNU8. To investigate the possible utility of ANT2 downregulation as a cancer therapy, ANT2 shRNA systems were adopted and the actual downregulation of ANT2 mRNA as well as protein mediated by ANT2 shRNA was confirmed from human breast cancer cell lines. As a result, we established a shRNA system to investigate the therapeutic value of ANT2 downregulation in breast cancer in vitro and in vivo.

To explore the potential of ANT2 shRNA as a treatment for human breast cancer, the diverse phenotypic changes of respective cancer cell lines affected by ANT2 shRNA were investigated. Firstly, the cell survival and proliferation rate were investigated using MDA-MB-231. After transfecting MDA-MB-231 cells with ANT2 shRNA-1 for 24 hours, the transfected cells became less confluent as compared with the cells transfected with scramble shRNA. Many ANT2 shRNA transfected cells even became rounded and detached from culture plates, providing one line of evidence about cell death or cell cycle arrest (Figure 2a). In addition, the proliferation rate of ANT2 shRNA-1 transfected MDA-MB-231 cells was also obviously reduced (Figure 2b). To understand the effects of ANT2 shRNA in MDA-MB-231 cells more specifically, the levels of intracellular ATP and the cell cycle status in transfected cells were investigated. Intracellular ATP levels were significantly reduced by up to 50% in the cells transfected with ANT2 shRNA-1 (Figure 2c), suggesting the death of ANT2 shRNA transfected cells is ascribed to the reduction of ATP synthesis. Cell cycle analysis indicated that the G₁ populations of ANT2 shRNA-treated MDA-MB-231 cells were 12% (24 hours) and 14.6% (48 hours), while scramble shRNA-treated cells displayed 0.8% (24 hours) and 7.6% (48 hours) of the G₁ populations (Figure 2d). Additional analysis gating on the sub-G₁ population also showed similar results after 24 or 48 hours of transfection, indicating ANT2 shRNA induces G₁ arrest in breast cancer cells.

We finally tested whether the knockdown of ANT2 shRNA induces cell death in the breast cancer cell lines MCF7 and MDA-MB-231 in vitro. Based on annexin V and PI staining, both ANT2 siRNA and ANT2 shRNA increased early apoptotic cells (AV⁺PI^-), intermediate apoptotic cells (AV⁺PI⁺) as well as late apoptotic cells (AV^-PI^-) compared with control groups after 48 hours of transfection (Figure 2e). ANT2 shRNA was more effective than ANT2 siRNA, however, in terms of inducing cell death. For example, ANT2 siRNA induced around 25–35% of cell death, while ANT2 shRNA generated about 50–60% of apoptotic cells. These results coincide with our previous observation regarding the higher efficacy to suppress ANT2 expression by ANT2 shRNA than ANT2 siRNA as shown in Figure 1c, suggesting increased ANT2 downregulation induces more cell death. Additionally, the ovarian cancer cell lines SK-OV-3 and SNU8 transfected with ANT2 siRNA or shRNA demonstrated similar results to those obtained from the breast cancer cell lines (Additional File 1b). These results imply that the physiological effects of ANT2 downregulation could be consistent in both cancers.

Taken together, these results suggest that the downregulation of ANT2 efficiently inhibits cell growth and proliferation by reducing ATP levels as well as inducing G₁ arrest, which ultimately leads to the apoptosis of human breast cancer cells in vitro.

To confirm apoptosis was responsible for the cell death induced by ANT2 shRNA in breast cancer cell lines, we examined the presence of genomic DNA fragmentation. As we expected, DNA laddering – a characteristic of apoptotic cell death – was abundantly observed from ANT2 shRNA-1 transfected MDA-MB-231 cells (Figure 3a), proving the death of transfected cells mediated by ANT2 shRNA is apoptotic.

It has been reported that Bcl-2 family molecules can physically interact with ANT and that this interaction regulates ANT-mediated mitochondrial permeability transition pore formation [7]. To understand the mechanisms of apoptosis induced by ANT2 shRNA, we examined the changes of Bcl-2 family protein levels during the ANT2 shRNA-1-induced, shRNA-2-induced and shRNA-3-induced apoptosis in MDA-MB-231 cells by western blotting. The transfection of ANT2 shRNA-1, shRNA-2 and shRNA-3 resulted in the upregulation of Bax (proapoptotic) and the downregulation of Bcl-xL (antiapoptotic) (Figure 3b). In addition, caspase-3 activation – defined by the appearance of cleaved caspase 3 – was observed from ANT2 shRNA-treated cells. These results clearly demonstrate that ANT2 shRNA induces the apoptosis via the Bcl-2 family.

We assumed the apoptosis triggered by ANT2 shRNA must be involved in the mitochondria-mediated apoptotic pathway because ANT2 protein is abundant in mitochondrial membranes. The potentials of mitochondrial membranes were measured by 3,3'-diethyloxacarbocyanine staining to prove our hypothesis. As expected, the cells transfected with ANT2 shRNA-1 displayed about 10-fold shifted curves to the left after 24 and 48 hours of transfection, indicating that mitochondrial membrane potentials are disrupted in apoptotic cells transfected with ANT2 shRNA (Figure 3c).

Taken together, these results implicate that ANT2 shRNA changes the Bcl-2 family balance in mitochondrial membranes, favoring a proapoptotic pore-forming status, and thereby causes the disruption of mitochondrial membrane potentials resulting in cell apoptosis.

Before transfecting MDA-MB-231 cells with ANT2 shRNA-1, shRNA-2 and shRNA-3, we also assessed transfection efficiencies using GFP (Green Fluorescence Protein)-expressing vector. Although the transfection efficiency was only about 30%, the proportion of cells killed by ANT2 shRNA treatment exceeded 60%, raising the possible existence of bystander effect. To confirm the presence of a bystander effect in MDA-MB-231 cells, we cultured nontransfected cells with the supernatants obtained from ANT2 shRNA-1, shRNA-2 and shRNA-3 transfected cells and then evaluated cell death by annexin V–PI staining.

As shown in Figure 4, nontransfected cells cultured with the supernatants from ANT2 shRNA-1, shRNA-2 and shRNA-3 transfected cells committed to cell death defined by AV^-PI⁺. The absence of DNA laddering and the lack of an AV⁺ population lead us to consider that necrosis might be the major mechanism of cell death induced by this bystander effect. These results therefore suggest that cytotoxic agents must be secreted from ANT2 shRNA-transfected cells to the media.

To determine the cause of this bystander effect, we examined the levels of proinflammatory cytokines such as TNFα, IFNγ and IL-12 from MDA-MB-231 cells treated with ANT2 shRNA or scramble shRNA by intracellular fluorescent-activated cell sorter. TNFα expression was elevated eight-fold in the cells treated with ANT2 shRNA-1 but the levels of IFNγ and IL-12 were unchanged (Figure 5a).

In addition, we evaluated TNF family death receptor I expression by RT-PCR and flow cytometry. Both transcription and surface expression of TNF-receptor I were upregulated by the transfection with ANT2 shRNA (Figure 5b). Additionally, the bystander effect was partially neutralized by anti-TNFα antibody (Figure 5c). These results suggest that the observed bystander effect might be caused by TNFα secretion and TNF-receptor I expression.

We finally evaluated the antitumor effects of ANT2 shRNA in vivo using a nude mouse tumor xenograft model. Tumor sizes were measured twice per week and the observations lasted over 56 days after tumor challenge with MDA-MB-231 cell lines. Tumor growth was significantly inhibited in ANT2 shRNA-1-treated, shRNA-2-treated and shRNA-3-treated mice as compared with PBS-treated mice or scramble shRNA-treated mice (P < 0.05) (Figure 6a), suggesting intratumoral ANT2 shRNA treatment has a strong antitumor effect in vivo. One week after the final intratumoral injection of respective ANT2 shRNA-1, shRNA-2 and shRNA-3 (day 32), increased apoptotic cell death was detected by Terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling assay (Figure 6b) and the suppression of ANT2 was detected by RT-PCR (data not shown) from the tumor tissues obtained from ANT2 shRNA-treated mice. As a result, these data indicate that the injection of ANT2 shRNA induces tumor regressions associated with apoptosis in vivo.