ROS-p53-cyclophilin-D signaling mediates salinomycin-induced glioma cell necrosis

Salinomycin, sanglifehrin A (SfA), cyclosporine A (CsA), n-acetyl cysteine (NAC), temozolomide (TMZ) and pyrrolidinedithiocarbamate (PDTC) were purchased from Sigma (St. Louis, MO). Necrostatin-1 (Nec-1) was purchased from Cayman Chemical (Beijing, China). Antibodies against tubulin and Cyp-D were purchased from Santa Cruz Biotech (Santa Cruz, CA), antibodies for p53 (regular and specific sites of phosphorylation) were purchased from Cell Signaling Technology (Danvers, MA).

U87MG, U251MG and EFC-2 glioma cells were maintained in dulbecco’s modified Eagle’s medium (DMEM, Sigma, St. Louis, MO), supplemented with a 10 % fetal bovine serum (FBS, Sigma), penicillin/streptomycin (1:100; Sigma) and in a CO₂ incubator at 37 °C.

Tissues from whole brains of post-natal (P1–P2) mice were triturated, and then cells were placed on poly-d-lysine pre-coated cell culture flasks in DMEM containing 15 % FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cultures were maintained at 37 °C in a humidified atmosphere of 5 % CO₂/95 % filtered air. After reaching a confluent monolayer of glial cells (10–14 days), microglia were separated from astrocytes by shaking off for 5 h at 100 rpm. The enriched astrocytes were >96 % positive for glial fibrillary acidic protein (GFAP).

The cell viability was measured by the 3-[4,5-dimethylthylthiazol-2-yl]-2,5 diphenyltetrazolium bromide (MTT) (Sigma, St. Louis, MO) method as reported [28]. Briefly, cells were seeded in 96-well plates with 70–80 % confluence. After indicated treatment/s, MTT tetrazolium salt (0.25 mg/ml) was added to each well for 2 h at 37 °C. Afterwards, 200 μl of DMSO was added to dissolve formazan crystals. The absorbance of each well was observed by a plate reader at a test wavelength of 490 nm. The value of each treatment group was expressed as percentage change of that of control group.

As reported [28], the number of dead glioma cells (trypan blue positive) after treatment was recorded, and the percentage (%) of dead cells was calculated by the number of the trypan blue stained cells divided by the total cell number, which was automatically tested by a handheld automated cell counter (Merck Millipore, Shanghai, China).

As reported [28], U87MG cells (5× 10³) were suspended in 1 ml of DMEM containing 0.1 % agar (Sigma, St. Louis, MO), 10 % FBS and with indicated treatments or the vehicle control. The cell suspension was then added on top of a pre-solidified 100 mm culture dish. The medium was replaced every two days. After 10 days of incubation, colonies were photographed at 4×. The number of large colonies (>50 μm in diameter) was manually counted and recorded, and the number was expressed as percentage change of the control group.

As reported [28], the glioma cell apoptosis and necrosis were determined by the Annexin V In Situ Cell Apoptosis Detection Kit (Roche Molecular Biochemicals, Indianapolis, IN, USA) according to the manufacturer’s protocols. Briefly, after indicated treatment/s, glioma cells were stained with Annexin V and propidium iodide (PI) (Molecular Probes). The cell apoptosis percentage was reflected by Annexin V^+/+/PI^−/− plus Annexin V^+/+/PI^+/+ percentage detected by fluorescence-activated cell sorting (FACS) (BD, Shanghai, China). Annexin V^−/−/PI^+/+ percentage was utilized as an indicator of cell necrosis. The time point for FACS assay was selected based on pre-experiment results.

The cytosol proteins of approximately 2 × 10⁶ glioma cells were extracted in cell lysis buffer containing 25 mm HEPES, 5 mm MgCl₂, 5 mm EDTA, 5 mm dithiothreitol and 0.05 % phenylmethylsulfonyl fluoride, (pH 7.5). Twenty μg of cytosolic extracts were added to caspase assay buffer (312.5 mM HEPES, pH 7.5, 31.25 % sucrose, 0.3125 % CHAPS) with benzyloxycarbonyl-DEVD-7-amido-4-(trifluoromethyl)-coumarin as substrate (Calbiochem, Darmstadt, Germany). After 2 h of incubation at 37 °C, the release of 7-amido-4-(trifluoromethyl) coumarin (AFC) was quantified, using a Fluoroskan system set to an excitation value of 355 nm and emission value of 525 nm. The value of each result was normalized, and was expressed as fold change vs. that of vehicle control group.

Cells were washed with ice-cold PBS and then lysed using lysis buffer containing 1 % Nonidet P-40, 1 % deoxycholate, 0.1 % sodium dodecyl sulfate, 150 mmol/L sodium chloride and 10 mmol/L Tris–HCl (pH, 7.4). The lysates were collected and centrifuged. The concentration of the extracted protein was measured by bicinchoninic acid assay kit (Sigma). The extracted protein was boiled for 5 min in loading buffer. Samples were separated on 10 % SDS-polycrylamide gel, and after electro-blotting onto polyvinylidene fluoride (PVDF) membranes (Millipore, Shanghai, China), the membranes were blocked with blocking solution [10 % (w/v) milk in Tris-buffered solution plus Tween-20 (TBST), incubated overnight at 4 °C with primary antibodies, and then incubated with HRP-conjugated anti-rabbit/mouse second antibodies. The detection was performed by Super-signal West Pico Enhanced Chemiluminescent (ECL) Substrate according to the manufacturer’s protocol. The blot intensity was quantified and normalized as previously reported [28]. For detecting proteins in the mitochondria, intact mitochondria of cultured 2.0 × 10⁷ glioma cells were isolated using the “Mitochondria Isolation Kit for Cultured Cells” (Pierce, Rockford, IL).

After treatment, 600 μg of cell lysates from mitochondrial fractions of glioma cells were pre-cleared. The supernatant was then rotated overnight with 2 μg of anti-Cyp-D (Santa Cruz Biotech). Next, the lysates were centrifuged for 5 min at 4 °C in a micro-centrifuge to remove nonspecific aggregates. The protein A/G PLUS-agarose (35 μl, Sigma) was then added to the supernatants for 4 h at 4 °C. Pellets were washed six times with PBS, resuspended in lysis buffer, and then assayed by Western blots. The time point for IP was based on previous publications and results from pre-experiments.

The SiRNA duplexes against human cyclophilin-D were purchased from Santa Cruz. Lipofectamine^TM 2000 was applied to transfect RNAi (100 nM) into cultured glioma cells according to manufacturer’s protocol. The same amount of scramble non-sense siRNA (control siRNA, Santa Cruz) was transfected into control cells. After 48 h, the Cyp-D expression and the equal loading in transfected cells were examined by Western blots. Only successfully transfected cells, demonstrated by markedly reduced level of Cyp-D, were used for further experiments. The transfection was repeated one more round if necessary.

Similar to previous reported, the MMP of glioma cells was measured through JC-10 dye (Invitrogen, Carlsbad, CA) [29]. The JC-10 dye exhibits two staining spectra. In normally resting cells, the dye forms aggregates in the mitochondrial membrane, exhibiting orange fluorescence. When the membrane potential is decreasing, the monomeric JC-10 will form in the cytosol, exhibiting the green fluorescence. Thus, the intensity of green fluorescence can be used as indicator of MMP loss. Briefly, after treatment, glioma cells were stained with 5.0 μg/ml of JC-10 for 5 min at room temperature under dark. Cells were then washed twice with warm PBS, and resuspended in fresh culture medium and read immediately on a microplate reader with an excitation filter of 485 nm. The OD value was used as an indicator of MMP loss.

The wt-Cyp-D plasmid (pSuper-puromycin-GFP- Cyp-D) and the empty vector (pSuper-puromycin) were gifts from Dr. Wang [29]. Lipofectamine^TM and PLUS reagent (Invitrogen, Carlsbad, CA) were applied to transfect Cyp-D plasmid or the vector (1 μg/ml) into cells according the manufacturer’s protocol. The stable clones were selected by puromycin (0.25 μg/ml medium). The puromycin-containing medium was refreshed every 3 days, until single resistant colony can be formed (4 weeks). The Cyp-D expression in the stable clones was always detected by Western blots to confirm the transfection efficiency. Only colonies with over-expressed Cyp-D were utilized for experiments.

P53 shRNA containing lentiviral particles were added to cultured glioma cells at the dose of 20 μl/ml medium, the infection took 48 h, afterwards, puromycin (0.25 μg/ml) was added to select the stable clones. The puromycin-containing medium was refreshed every 3 days, until single resistant colony can be formed (4 weeks). P53 expression in the stable colony was always detected by Western blots to confirm the infection efficiency. Control cells were infected with same concentration of scramble shRNA containing lentiviral particles.

Intracellular ROS generation was measured by flow cytometry using dichlorofluorescin (DCF) oxidation assay [30]. DCFH-DA enters passively into cells and is cleaved by nonspecific cellular esterase and oxidized in the presence of ROS. Briefly, 3*10⁵ glioma cells were plated in 60-mm culture plates overnight. After treatment, cells were incubated with DCFH-DA (5 μM) for 1 h at 37 °C. Thereafter, cells were washed with PBS and kept in 1 ml PBS, ROS fluorescence was analyzed using flow cytometer. The value of treatment group was expressed as fold change to that of untreated control group.

The SOD activity was measured using a SOD activity assay kit (BioVision, Mountain View, CA) according to the manufacturer’s protocol. Briefly, cells were seeded onto 100-mm² dishes at a density of 5 × 10⁶ cells/dish. After treatment, the cells were washed with PBS, scraped from the plates into 1 mL of ice-cold PBS (containing 0.05 mM of EDTA), and homogenized. The homogenate was centrifuged at 4000 × g for 30 min at 4 °C. The resulting supernatant (20 μl) was added with 200 μl of water soluble tetrazolium working solution and 20 μL of enzyme working solution to a 96-well plate. After incubating the plate at 37 °C for 20 min, the absorbance at 450 nm was read using a microplate reader.

Nude mice were purchased from Suzhou University Institute of Biological Science. U251 cells (10⁶ cells in 100 μL of saline/Matrigel (BD Pharmigen San Jose, Ca), 1:1 v/v) were injected subcutaneously into the right flank of 4-weeks-old female mice. Treatments were started after tumor reached approximately 200 mm³ (around 4 weeks after inoculation). Salinomycin (5.0 mg/kg) or with CsA (5.0 mg/kg) were administered to mice (10 per group) once daily for two weeks, intraperitoneally. The size of the tumors was measured by caliper every week, and tumor volumes were calculated using the following formula: π/6 × width ²× length. Control mice received vehicle only with the same schedule. All animal procedures were performed according to the Animal Experimentation guidelines upon approval of the experimental protocol by Soochow University and Shanghai Jiao Tong University. The study was approved by Soochow University and Shanghai Jiao Tong University review boards. All investigations were conducted according to the principles expressed in the Declaration of Helsinki.

The data presented in this study were means ± standard deviation (SD). Statistical differences were analyzed by one-way ANOVA followed by multiple comparisons performed with post hoc Bonferroni test (SPSS version 18). Values of p < 0.05 were considered statistically significant. The difference between two groups was tested using paired-samples t test.

We first tested the potential role of salinomycin in cultured glioma cells. MTT cell viability assay results in Fig. 1a demonstrated that salinomycin dose-dependently inhibited U87MG cell survival. Meanwhile, the anti-survival effect of salinomycin was time-dependent (Fig. 1b). The viability of U87MG cells started to decrease 48 h after salinomycin (5 μM) treatment in U87MG cells (Fig. 1b). Using the trypan blue staining assay, we found that salinomycin treatment caused significant cytotoxicity of U87MG cell death, and the effect again was dose-dependently (Fig. 1c). To further confirm the potential role of salinomycin on proliferation of glioma cells, clonogenicity assay was performed, and results showed that salinomycin (5 μM) significantly reduced the number of survival colonies of U87MG cells (Fig. 1d). Figure 1e demonstrated that salinomycin inhibited the viability of U251MG and EFC-2 glioma cells. We also tested the potential cytotoxicity of salinomycin against primary cultured mouse astrocytes, and found that salinomycin (5–10 μM) only slightly inhibited the survival of primary astrocytes (Fig. 1f and g).

The results above confirmed the cytotoxic effect of salinomycin in glioma cells; we then explored the role of apoptosis and necrosis in it. By using two different assays including Annexin V-FACS assay and caspase-3 activity assay, our results demonstrated that salinomycin induced moderate cell apoptosis (less than 15 %) in U87MG cells (Fig. 2a and b), which was blocked by the apoptosis inhibitor zVADfmk (Fig. 2a and b). Meanwhile, the Western blot results in Fig. 2c showed caspase-3 cleavage and cytochrome C release in salinomycin-treated U87MG cells, further supporting apoptosis induction. Note that less than 10 % cells were apoptotic after 5 μM of salinomycin stimulation in U87MG cells (Fig. 2a). Further, zVADfmk only slightly inhibited salinomycin-induced viability reduction and cell death in cultured glioma cells (Fig. 2d-f). On the other hand, we noticed a large proportion of necrotic U87MG cells after same salinomycin stimulation (PI positive and Annexin V negative, which was largely inhibited by necrosis inhibitor Nec-1 (Fig. 2g). More importantly, combination of Nec-1 and zVADfmk almost completely blocked salinomycin-induced cytotoxicity in U87MG cell (Fig. 2h). Together, these results suggest that salinomycin induces some apoptosis, but mainly necrosis in cultured glioma cells.

A number of recent studies have shown that the mitochondrial protein Cyp-D is required for stresses-induced necrotic, but not apoptotic cell death [16‐19, 23, 31]. Thus, we tested the possible involvement of Cyp-D in salinomycin’s cytotoxicity in glioma cells. Results demonstrated that two Cyp-D inhibitors, sanglifehrin A (SfA) [32] and cyclosporin A (CSA) [23, 31], significantly inhibited salinomycin-induced viability reduction in both U87MG cells (Fig. 3a) and in U251MG cells (Additional file 1: Figure S2A). Further, siRNA-depleting Cyp-D (Fig. 3b) also alleviated salinomycin’s cytotoxicity in U87MG cells (Fig. 3c). Meanwhile, salinomycin-induced U87MG cell death was also inhibited by SfA, CsA or Cyp-D depletion (Fig. 3d). These results suggest that Cyp-D is important for salinomycin-induced cell death of glioma cells. To further support our hypothesis, we tested salinomycin’s effect in Cyp-D over-expressing glioma cells. As shown in Fig. 3e, we exogenously introduced wt-Cyp-D in U87MG cells. These cells were hyper-sensitive to salinomycin, as more cell death were achieved in Cyp-D over-expressing cells (Fig. 3f). We also noticed some spontaneous cell death in Cyp-D over-expressing cells (Fig. 3f). Significantly, salinomycin-induced cell apoptosis, shown by the Annexin V percentage, was not affected by Cyp-D depletion or Cyp-D over-expression (Fig. 3g), while cell necrosis induced by salinomycin was inhibited by Cyp-D siRNA, but was aggravated by Cyp-D over-expression (Fig. 3h). Together, these results indicated that the mitochondrial protein Cyp-D is required for salinomycin induced necrosis in cultured glioma cells.

Previous studies have shown that Cyp-D-mediated necrotic cell death is through binding with p53 in mitochondria [19, 27, 33]. Since we have shown that Cyp-D is required for salinomycin’s cytotoxicity, we thus tested the role of p53 in the process. As demonstrated, p53-shRNA stable knockdown (Fig. 4a) significantly inhibited salinomycin-induced U87MG cell necrosis, while apoptosis was not affected (Fig. 4b). We noticed p53 mitochondrial translocation in salinomycin-treated U87MG cells (Fig. 4c), which formed a complex with the local protein Cyp-D (Fig. 4d). Similar results were obtained in U251MG cells (Additional file 1: Figure S2B). The mitochondrial complexation between Cyp-D and p53 was blocked by Cyp-D inhibitor CsA or p53-shRNA depletion (Fig. 4d). Our data suggested that the complexation was required for mPTP opening, as both CsA and p53 shRNA largely inhibited salinomycin-induced MMP loss (Fig. 4e). We failed to see p53 phosphorylation at Ser 15 or Ser 20, nor p53 upregulation by salinomycin, while temozolomide (TMZ) induced p53 phosphorylation and upregulation (Fig. 4c and f). Based on these data, we suggest that salinomycin induces p53 mitochondrial translocation to interact with Cyp-D, which is required for mPTP opening and programmed necrosis.

Next we tested the possible upstream signal for salinomycin-induced mPTP opening and necrosis. FACS results in Fig. 5a showed ROS production by salinomycin in U87MG cells, which are consistent with other studies [12, 34, 35]. We found that salinomycin decreased the activity of SOD in glioma cells, which could be one reason for ROS increase (Additional file 2: Figure S1). N-acetyl cysteine (NAC) and pyrrolidinedithiocarbamate (PDTC), two well known anti-oxidants inhibited ROS production by salinomycin. Significantly, both NAC and PDTC suppressed p53 mitochondrial translocation (Fig. 5b, input) and following Cyp-D association (Fig. 5b). Further, salinomycin-induced MMP loss, an indicator of mPTP opening, was also inhibited by these two anti-oxidants (Fig. 5c). These results indicate that salinomycin-induced p53 mitochondrial translocation, p53-Cyp-D complexation, and the subsequent mPTP opening might be dependent on ROS production. Based on these results, we predicted that ROS might also be important for salinomycin-induced necrosis. As a matter of fact, salinomycin-induced U87MG cell necrosis was alleviated by NAC or PDTC (Fig. 5d), which could explain that NAC inhibited salinomycin-induced viability loss in both U251 MG cells and EFC-2 cells (Fig. 5e). Interestingly, using the cells above, ROS results in Fig. 5f showed that salinomycin-induced ROS was inhibited by Cyp-D depletion, but was aggravated by Cyp-D over-expression. Thus, salinomycin-induced mPTP opening might play a vital role in ROS regulation in glioma cells. To evaluate the anti-tumor activity of salinomycin in vivo, we generated subcutaneous xenografts by inoculating the U251 cells to nude mice. Treatment with salinomycin elicited a marked inhibitory effect on U251 xenograft development (Fig. 5g). Such an effect by salinomycin was significantly inhibited by co-administration with CsA (Fig. 5g). Note that CsA alone failed to affect U251 growth in vivo (Fig. 5g). Further, the mice body weight was not significantly affected by salinomycin or with CsA, indicating the relative safety of this regimen (Additional file 3: Figure S3).