S1 was synthesized as previously reported (19) and dissolved in dimethyl sulfoxide (DMSO). The Bcl-2 inhibitors ABT-737, ABT-199, and obatoclax mesylate (Oba) were purchased from ApexBio (Boston, MA, USA). The glycolysis inhibitor 2-DG was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Hoechst 33342, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), and fetal bovine serum (FBS) were purchased from Sigma (St. Louis, MO, USA). Anti-SIRT3, anti-Mcl-1, anti-Bcl-2, anti-lamin A/C, anti-Cox IV, anti-Tom20, anti-caspase-3, anti-PARP and anti-β-actin antibodies were purchased from Santa Cruz Biotechnology.

Human ovarian carcinoma cells (SKOV3 cells) were purchased from American Tissue Culture Collection (Rockville, MD, USA) and cultured in RPMI-1640 medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin (complete medium), and were grown in a humidified cell culture incubator equilibrated with 95% air and 5% CO₂ at 37°C.

Cells were seeded in 96-well plates at a density of 1×10⁴ cells/well in 200 μl of complete medium at 37°C with 5% CO₂. A total of 10 μl of 10 mg/ml MTT reagent in phosphate-buffered saline was added to each well and incubated for 4 h, after which any formazan crystals that formed were dissolved in 150 μl DMSO. Absorbance was recorded at a wavelength of 490 nm. The treatment was repeated in six separate wells.

All animal manipulations were performed in accordance with institutional guidelines under approved protocols. Six-week-old female athymic BALB/c nude mice were purchased from the Animal Experimental Center (Beijing, China). SKOV3 ovarian cancer cells (2×10⁶ cells), were s.c. implanted into into the upper flank. Seven days after tumor cell injection, the nude mice were randomized into 2 groups of 5 mice per group: vehicle (control) and S1 (6 mg/kg). The groups of mice were intraperitoneally administered with 0.2 ml of vehicle (DMSO in PBS, control group) or S1 in 0.2 ml of vehicle every 2 days. The body weights and tumor volumes of the control and drug-treated mice were recorded prior to the start of the experiment. Tumor volume was monitored by using external caliper measurement and calculated according to the formula tumor volume (in millimeters cubed) = D × d²/2, where D and d are the longest and shortest diameters, respectively. The mice in both the control and S1-treated groups were sacrificed at 21 days after the tumor cell injection. Tumors from each mouse treated with vehicle or S1 were carefully dissected out at the time of sacrifice, weighed, and photographed for graphical representation.

Apoptotic morphological alterations in nuclear chromatin were detected by Hoechst 33258 staining. Briefly, SKOV3 cells were cultured in 24-well plates and treated as indicated for 24 h. Cells were washed with ice-cold PBS and fixed with 4% (w/v) paraformaldehyde overnight. The plates were then incubated with 10 M Hoechst 33258 staining solution for 10 min. The cells were visualized under a fluorescence microscope (IX-71, Olympus).

Briefly, after quantitating protein in each sample with the Bio-Rad protein reagent (Bio-Rad Laboratories, Hercules, CA, USA), (40 μg protein per well) were separated by SDS-PAGE and transferred onto an Immun-Blot PVDF membrane (Bio-Rad Laboratories). After blocked in Tris-buffered saline containing 5% (w/v) non-fat dry milk at room temperature for 1 h, the membranes were incubated with specific primary antibodies overnight at 4°C. After washed with PBS-Tween-20, membranes were incubated with horseradish-peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology) at room temperature for 1 h. Membranes were then incubated in ECL reagents and images captured by Syngene Bio Imaging (Synoptics, Cambridge, UK). Densitometric quantitation of bands was performed using Syngene Bio Imaging tools.

Briefly, SKOV3 cells were plated at 8×10⁴ cells/well in 96-multiwell, clear-bottom, black-body plates and allowed to adhere overnight. The following day, different concentrations of drugs were added and incubated for 6 h. Each treatment was repeated in three wells. The determination of cell oxygen consumption and extracellular acidification rates was carried out using the fluorescent oxygen-sensitive and pH-sensitive probes Mito-Xpress and pH-Xtra (Luxcel Bioscience, Cork, Ireland), respectively, as described previously (20).

Following treatment as indicated, the medium of SKOV3 cells was collected. The glucose and lactate concentrations were then determined with the glucose assay kit and the lactate assay kit (Bioassay Systems), respectively.

Following treatment as indicated, the SKOV3 cells were lysed. After centrifugation, ATP concentrations were measured in the supernatant based on ATP-dependent luciferase activity with the ATP assay kit (Beyotime Biotechnology).

The Human Glucose Metabolism RT² Profiler™ PCR array(SABiosciences-Qiagen, Hilden, Germany) profiles the expression of 84 key genes involved in the regulation and enzymatic pathways of glucose and glycogen metabolism. Total cellular RNA was extracted from cultured cells according to the manufacturer's instructions. Single-stranded cDNA was obtained by reverse transcription of 1 μg of total RNA using the SABiosciences RT² First Strand kit. Real-time qPCRs were performed using Applied Biosystems 7300 Fast with SYBR Green Fluorophore. The reactions were carried out using RT² SYBR Green Mastermix. cDNA was used as template and cycling parameters were 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. Fluorescence intensities were analyzed using the manufacturer's software, and relative quantification was calculated using the 2^−ΔΔCt method. Change of expression of the 84 genes was shown by heat imaging. GAPDH was used as a reference gene.

Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. First-strand cDNAs were generated by reverse transcription of RNA samples with SuperScript preamplification system (Promega, Madison, MI, USA). Reverse transcribed products were used to amplify SIRT3 by qRT-PCR. qRT-PCR was performed using a FastStart Universal SYBR Green Master (Roche) on an ABI 7300 instrument. The primer sequences used for qRT-PCR were 5′-TGGAAAGCCTAGTGGAGCTTCTGGG-3′ (forward) and 5′-TGGGGGCAGCCATCATCCTATTTGT-3′ (reverse) for SIRT3, and 5′-GGGTGATGCTGGTGCTGAGTATGT-3′ (forward) and 5′-AAGAATGGGAGTTGCTGTTGAAGT-3′ (reverse) for GAPDH. The expression of each investigated gene was normalized to the housekeeping gene GAPDH. Data are presented as the fold change in gene expression relative to the control sample.

Both oxidized and reduced forms of intracellular nicotin-amide adenine dinucleotide were determined using an NADH/NAD quantification kit (BioVision Inc., Milpitas, CA, USA) as described by the manufacturer. Briefly, 2×10⁵ cells were washed with cold PBS, pelleted, and extracted with 400 μl of NADH/NAD extraction buffer by two freeze/thaw cycles. Samples were vortexed and centrifuged at 14,000 rpm for 5 min. For each sample, a 200-μl aliquot of extracted NADH/NAD supernatant was transferred to a microcentrifuge tube. Samples were heated to 60°C for 30 min to allow decomposition of all NAD⁺ while retaining NADH intact and then placed on ice. Samples were rapidly centrifuged and transferred to a multiwell-plate. Standards and a NAD cycling mix were prepared according to the manufacturer's protocol. An amount of 100 μl of the reaction mix was added into each well of NADH standards and samples, and all samples were incubated at room temperature for 5 min to convert NAD⁺ to NADH. NADH developer solution was added to each well, and plates were incubated at room temperature. The reaction was stopped by adding 10 μl of stop solution into each well and mixing well, and absorbance was measured at 450 nm.

To generate fractions enriched in nuclear and mitochondrial proteins, 5.0×10⁶ cells were harvested and washed in PBS. Cells were resuspended in 0.2 ml of buffer MgRSB (10 mM NaCl, 1.5 mM MgCl₂, 10 mM Tris-HCl at pH 7.5, 0.1 mM PMSF, 1 mM DTT) and incubated on ice for 10 min. Cells were dounced to yield free nuclei, and 0.034 ml of sucrose buffer (2 M sucrose, 35 mM EDTA, 50 mM Tris-HCl at pH 7.5) was immediately added. Free nuclei were collected by centrifugation at 1,000 g for 10 min. Mitochondria were collected by centrifugation at 11,000 g for 10 min. Nuclei were resuspended in buffer C (20 mM Tris-HCl at pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.1 mM PMSF, 0.5 mM DTT) and incubated on ice for 20 min. To generate nuclear extract, nuclei were spun down at 14,000 g, and the supernatant was collected. Mitochondria were resuspended in buffer C to be used in western blot analyses.

SKOV3 cells were cultured on coverslips at a density of 5×10⁴ cells/well in 500 μl of complete medium. After treatment, SKOV3 cells were washed with cold PBS three times and fixed in 4% (w/v) paraformaldehyde/PBS for 20 min then washed with cold PBS three times. Fixed cells were digested by protein enzyme K for 1 min and washed with PBS twice. Cells were incubated with 0.1% (v/v) Triton X-100 for 6–10 min, washed once with PBS, and then blocked for 30 min in 5% (v/v) non-immune animal serum/PBS. Cells were incubated with primary antibody (SIRT3, Tom20, dilution 1:200; Santa Cruz Biotechnology) overnight and washed three times with PBS. Cells were then incubated with secondary antibody (rabbit and mouse, dilution 1:400; Thermo, Waltham, MA, USA) for 30 min in the dark. Plates were washed three times in PBS, treated with Hoechst 33342/H₂O (1 μg/ml) for 2 min, and then washed three times with PBS. Cells were examined by Olympus FV1000 confocal laser microscope.

RNA Small interfering RNA (siRNA) sequences targeting human SIRT3 (GenBank accession no. NM_012239) and a non-target sequence were constructed by Genechem (Shanghai, China). The SIRT3-RNAi (23504) sequence was GCTTGATGGACCAGACAAA, SIRT3-RNAi (23505) sequence was GCTGTACCCTGGAAACTAC, SIRT3-RNAi (23506) sequence was ACCTCCAGCAGTACG ATCT, and SIRT3-RNAi (23507) sequence was TCGATGGGC TTGAGAGAGT. The non-target siRNA (Scramble) sequence was TTCTCCGAACGTGTCACGT. Transfections with siRNA were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Briefly, SKOV3 cells were placed into 6-well plates, and transfected the next day with 4 μg si-SIRT3 or si-Scramble, using 10 μl (1 μg/μl) Lipofectamine 2000. Cells were harvested 2 days after transfection; whole cell lysates were isolated for western blot analyses. For MTT assay, transfected cells were treated with S1 for 24 h, followed by MTT assay to determine cell viability.

Cells were seeded at 2.0×10⁵ cells/well in 6-well cell culture dishes and treated with experimental conditions, as indicated in the individual figures, during their logarithmic growth phase. After various treatments and time-points, morphological alterations were analyzed under an inverted phase contrast microscope (Olympus, Tokyo, Japan) at ×20 magnification.

Caspase-3/7 activity was measured using the Caspase-Glo 3/7 assay kit (Promega) according to the manufacturer's instructions. Each experiment was performed in quintuplicate and experiments were carried out twice.

Data are expressed as the mean ± SD. SPSS 17.0 software was used for analysis. All experiments were repeated at least three times. The statistical significance of the difference between two groups was assessed using Student's t-test, and P<0.05 was considered to be significant.

SKOV3 cells were treated with different doses of S1 for 24 h, and the survival rate was then examined using MTT assays. The viability of SKOV3 cells was shown to be decreased by S1 treatment (Fig. 1A). To initiate the treatment on palpable tumors, SKOV3 xenografts were inoculated subcutaneously into Balb/c nude mice. S1 was then injected 7 days later, and shown to significantly inhibit the growth of tumor xenografts compared with vehicle-treated controls. At sacrifice, the mean tumor volumes in animals treated with S1 were 235 mm³ compared with 443.3 mm³ for control animals (Fig. 1C), while mean tumor weights were 0.116 and 0.25 g, respectively (Fig. 1D). Notably, there was no overt gross toxicity in either vehicle- or S1-treated mouse, as determined by measuring body weights (Fig. 1E).

Ovarian tumor growth was compared between animals with and without S1 treatment. The direct measurement of tumor volume was performed at the start of tumor inoculation. After S1 treatment, a significant reduction in tumor volume and weight was observed compared with the vehicle-treated control (Fig. 1B–D). We found that 6 mg/kg of S1 caused a reduction in tumor growth of 53.7% compared with the vehicle-treated control group, while it had no adverse effects on body weight, indicating that it lacked potential toxicity in nude mice. These results indicate that S1 inhibits SKOV3 cell proliferation and retards the growth of SKOV3 xenografts in vivo.

Based on MTT results, we treated SKOV3 cells with 5 μM S1 for 6, 12 and 24 h, and detected apoptotic chromatin condensation by Hoechst 33342 staining using confocal microscopy. S1-induced apoptotic chromatin condensation was obviously seen after 12-h treatment (Fig. 2A). We next examined the apoptotic effects through detecting the expression of Bcl-2, Mcl-1, cytoplasmic cytochrome c and the activation of caspase-3 and PARP by western blotting. S1 decreased the expression of Bcl-2 and Mcl-1, and enhanced expression of cleaved caspase-3, cleaved PARP and cytoplasmic cytochrome c in SKOV3 cells (Fig. 2B and C). These results indicate that S1 efficiently induces apoptosis in SKOV3 cells.

Previous studies showed that aerobic glycolysis is important for the growth of ovarian tumors (13). To determine whether S1 affects glucose metabolism before exerting obvious cytotoxity such as the induction of apoptosis, we first analyzed its effects on oxygen consumption rate (OCR) and extracellular acidification rates (ECAR) in SKOV3 cells after short-term exposure for 6 h (Fig. 2D and E). Following the inhibition of mitochondrial respiration by antimycin A in SKOV3 cells, we observed a drop in OCR, whereas ECAR increased. We also found that S1 induced a significant decrease in OCR, whereas ECAR increased slightly. Moreover, the effects of several different Bcl-2 inhibitors, including ABT-737, ABT-199 and obatoclax mesylate (oba), were also observed. ABT-737 showed no obvious effect on either OCR or ECAR, while ABT-199 and oba induced an increase in both OCR and ECAR. These results indicate that the effects of different Bcl-2 inhibitors on glucose metabolism are specific, and S1 mainly inhibits mitochondrial respiration in SKOV3 cells before exerting obvious cytotoxity.

We next determined cellular metabolic profiles by measuring extracellular glucose and lactate levels. Compared with the control group, S1-treated cells showed a decreased glycolytic phenotype with significantly lower glucose consumption and lactate production after long-term exposure for 12 h (Fig. 2F and G). Because glycolysis and mitochondrial respiration are the two main sources of cellular ATP production, we assessed whether S1 affected ATP levels. S1 was shown to decrease ATP levels in a dose-dependent manner (Fig. 2H). Taken together, these results suggest that S1 inhibits mitochondrial respiration and glycolysis and induces apoptosis in SKOV3 cells.

To gain further insight into the regulation of glucose metabolism by S1, the expression levels of the human glucose metabolism-related genes were examined in SKOV3 cells treated by S1 (5 μM, 6-h exposure) using the RT² Human Glucose Metabolism Profiler PCR array technology.

As shown in Fig. 3, significant downregulation of gene expression indicative of glycolysis was induced by S1 including hexokinase 2 (HK-2; position C7), liver phosphofructokinase (PFKL; position E5) and pyruvate dehydrogenase kinase isozyme 2 (PDK2; position D12), consistent with the detection of S1-induced inhibition of glucose uptake and lactate secretion (as detailed in Fig. 2). Upregulation of hexose-6-phosphate dehydrogenase (H6PD; position C6), which is responsive to the pentose phosphate pathway, was observed. S1 also affected tricarboxylic acid (TCA) cycle encoding genes [upregulation of ATP citrate lyase (ACLY; position A1), aconitase 1 (ACO1; position A2), dihydrolipoamide S-acetyltransferase (DLAT; position A10), isocitrate dehydrogenase 2 (IDH2; position C10), isocitrate dehydrogenase 3α (IDH3A; position C11), malate dehydrogenase 1B (MDH1B; position D3), malate dehydrogenase 2 (MDH2; position D4), oxoglutarate α-ketoglutarate ehydrogenase (OGDH; position D5), succinate dehydrogenase complex subunit C (SDHC; position G4), succinate-CoA ligase GDP-forming β subunit (SUCLG2; position G8)]. In addition, S1-treatment changed expression of genes encoding enzymes involved in gluconeogenesis [upregulation of fructose-1,6-bisphosphatase 1 (FBP1; position B4), fructose-1,6-bisphosphatase 2 (FBP2; position B5), glucose-6-phosphatase catalytic subunit (G6PC; position B7), phosphoenolpyruvate carboxykinase 1 (PCK1; position D7), phosphoenolpyruvate carboxykinase 2 (PCK2; position D8)], together with downregulation of glucose 6 phosphatase catalytic 3 (G6PC3; position B8)]. The array results show that S1 inhibits glycolysis and activates gluconeogenesis and TCA cycle at the mRNA level in SKOV3 cells.

SIRT3, a mitochondrial stress protein, is a key regulator of cancer metabolism and apoptosis (21). Because S1 inhibited mitochondrial respiration and glycolysis and induced apoptosis in SKOV3 cells, we considered whether SIRT3 was involved in this role. By observing the gene and protein expression of SIRT3, we found that S1 upregulated SIRT3 in a time-dependent manner (Fig. 4A–C). As NAD⁺ is an important cofactor for the activation of SIRT3, we also checked whether the intracellular NAD⁺/NADH ratio was modulated by S1. We found that S1 effectively increased the intracellular NAD⁺/NADH ratio, which may contribute to the activation of SIRT3 (Fig. 4D).

Although SIRT3 has been reported to be exclusively expressed in mitochondria (14), another study showed that it may exist in the nucleus and translocate to the mitochondria under stress (22). In this study, western blot analysis and immunofluorescence studies of the protein expression of SIRT3 suggested that S1 induced the SIRT3 translocation from the nucleus to mitochondria (Fig. 4E–G). A decrease in the nuclear expression of SIRT3 was matched by an increase in its colocalization with TOM20, a mitochondrial marker. These results indicate that S1 upregulates the expression of SIRT3 and induces its translocation from the nucleus to mitochondria.

To examine the role of SIRT3 activation after S1 treatment, we first used small interfering (si)RNA to knockdown SIRT3 expression, and then determined the effects of this on S1-induced growth inhibition. We transfected SKOV3 cells with siRNA against SIRT3 or non-target sequence for 48 h. As shown in Fig. 5A, the SIRT3 protein expression was significantly decreased in siSIRT3-23507 group compared with cells transfected with control siRNA, and we adopted this effective siRNA to knockdown SIRT3 expression. Importantly, SIRT3 knockdown alleviated S1-induced growth inhibition compared with control siRNA cells (Fig. 5B and C).

Next, we assessed the role of SIRT3 in S1-induced apoptosis. After 24-h treatment with 5 μM S1, a significant decrease in caspase-3/7 activity was shown in siSIRT3-SKOV3 cells compared with cells transfected with control siRNA (Fig. 5D). To further investigate whether the activation of SIRT3 is involved in the regulation of glucose metabolism by S1, we examined cellular metabolic profiles by measuring extracellular glucose and lactate levels in the growth media of SKOV3 cells transfected with siRNA against SIRT3 or non-target sequence. The consumption of glucose and lactate secretion after 12-h treatment with S1 were considerably higher in siSIRT3-SKOV3 than control-transfected cells (Fig. 5E and F). Previous research found that a potential anticancer drug Oroxylin A could inhibit glycolysis and induce the dissociation of hexokinase II from the mitochondria in human breast carcinoma cell lines, which is related to the activation of SIRT3 (23). We also checked whether the activation of SIRT3 induced by S1 could affect the expression of HK-II in the mitochondria. As shown in Fig. 5G, compared with control transfected group, knockdown of SIRT3 increased the expression of HK-II in the mitochondria after S1 treatment. These results indicate that the activation of SIRT3 exerts a pro-apoptotic role in S1-treated SKOV3 cells, and may contribute to the regulation of glucose metabolism by S1.

In MTT assays, use of a combination of 2-DG with Bcl-2 inhibitors such as ABT737, ABT199, Oba, and S1 inhibited cell viability to a greater extent than that of cells treated only with Bcl-2 inhibitors (Fig. 6A–E). To determine the effect on apoptosis, we examined the activation of caspase-3 in SKOV3 cells treated with S1 and/or 2-DG for 24 h (Fig. 6F). 2-DG upregulated the activation of caspase 3 induced by S1, which correlated with the upregulation of SIRT3 (Fig. 6G). These results demonstrated that the manipulation of Bcl-2 inhibitors combined with the use of classic glycolysis inhibitors may be rational strategies to improve ovarian cancer therapy through glycolysis inhibition and targeting mitochondrial apoptotic machinery.