Suppression of Lipogenesis via Reactive Oxygen Species–AMPK Signaling for Treating Malignant and Proliferative Diseases

Sterol regulatory element-binding protein 1 functions as a biomarker and potential therapeutic target for malignant and highly proliferative diseases

To assess lipogenesis in malignant and highly proliferative diseases, we used NSCLC and RA as representative models for analysis. By comparing the expression of critical factors and enzymes associated with lipogenesis in different cell types from these two model diseases, we found that sterol regulatory element-binding protein 1 (SREBP1), a key positive regulator of sterol synthesis, was highly expressed in both NSCLC cells and MH7A human synovial fibroblasts following TNF-α stimulation but was poorly expressed in normal BEAS-2B lung epithelial cells or unstimulated human synovial fibroblasts (Fig. 1A, B). Among these NSCLC cell lines, H1975 and H1650 are G-R, HCC827 is gefitinib sensitive, and A549 and H2228 harbor wild-type EGFR that is not applicable for gefitinib treatment. Thus, we hypothesized that suppressing SREBP1 and lipogenesis can inhibit the progression of malignant and proliferative diseases. Metformin (25) and BBR (48), which have previously been identified to be effective suppressors of SREBP1, were used to confirm our hypothesis, and their inhibitory effects were compared. Interestingly, although metformin has been marketed globally, BBR showed much stronger cytotoxic effects on G-R NSCLC H1975 cells than did metformin (Fig. 1C); thus, we selected BBR as the representative drug in the following experiments.

OPEN IN VIEWER

FIG. 1.

SREBP1 was upregulated in malignant and proliferative diseases, and BBR was an effective inhibitor of lipogenesis. (A) mRNA and SREBP1 expression level in lung cancer cells and normal lung epithelial cells. (B) SREBP1 expression in the human fibroblast cell line MH7A with or without TNF-α stimulation. (C) The cytotoxic effect of metformin and BBR on H1975 cells. (D) BBR decreased SREBP1 levels at 24 h in both NSCLC cells and human synovial fibroblasts. (E) Liquid chromatography/mass spectrometry results indicated that the key cellular intermediate palmitic acid was significantly inhibited by BBR. (F) BBR effectively inhibited the growth of the NSCLC cell lines H1975 and H1650. (G) MH7A human fibroblasts and primary RASFs stimulated with TNF-α were more sensitive to BBR than were normal cells. ^*/# p < 0.05, ^**/## p < 0.01, ^***/### p < 0.001 vs control. BBR, berberine; NSCLC, nonsmall-cell lung cancer; RASFs, rheumatoid arthritis synovial fibroblasts; SREBP1, sterol regulatory element-binding protein 1; TNF, tumor necrosis factor.

First, the inhibitory effect of BBR on lipogenesis was validated. G-R NSCLC cell lines H1975 and H1650 with the highest SREBP1 expression were selected for further studies. BBR significantly suppressed the expression of SREBP1 in the H1975 (3 μM) and H1650 (6 μM), the human fibroblast cell line MH7A (40 μM), and synovial fibroblasts (80 μM) from RA patients (RASFs) (Fig. 1D). We found that BBR substantially reduced the triglyceride, cholesterol, and other lipid metabolite levels in H1975 cells (Supplementary Fig. S1A–C; Supplementary Data are available online at www.liebertpub.com/ars). Notably, the intracellular level of palmitic acid, which is a critical intermediate metabolite of lipogenesis (35), was substantially decreased in the BBR-treated H1975 cells, by ∼100-fold, based on the same cell numbers for analysis (Fig. 1E). These results indicate that BBR is a potent suppressor of SREBP1 and lipogenesis. By assessing the cytotoxic effects of BBR, we found that BBR significantly inhibited the growth of H1975 NSCLC cells but had a much weaker cytotoxic effect on normal BEAS-2B lung epithelial cells (Fig. 1F). For synovial fibroblasts, TNF-α was added to induce inflammatory responses, and we observed that after treatment with TNF-α, the two types of human fibroblasts under study (MH7A cells and RASFs) showed increased sensitivity to BBR (Table 1; Fig. 1G).

Furthermore, we used both an adjuvant-induced arthritis (AIA) model in rats and a G-R NSCLC xenograft model of cancer in mice to verify the effects of BBR on suppressing malignant carcinoma and proliferative conditions in vivo. In the AIA model, the levels of SREBP1, TNF-α, and interleukin (IL)-β were all significantly increased (Supplementary Fig. S1D), and an increase in hind paw volume was considered a major indicator of arthritis. Methotrexate (MTX), the most commonly prescribed drug worldwide for treating RA, and metformin, a drug marketed for treating diabetes and cancer, were used as positive controls (33). Notably, a significant regression in edema of the hind paw was observed in AIA animals treated with 25 mg/kg BBR compared to that of the vehicle-treated AIA rats, and this regression was similar to the effects of MTX and metformin (Fig. 2A and Supplementary Fig. S1E). However, 25 mg/kg metformin showed no inhibitory effect, indicating that BBR is more effective than metformin. Moreover, bone damage induced by RA was effectively blocked by BBR at 25 mg/kg concentration (Fig. 2B). In the mouse G-R NSCLC xenograft model, at 10 mg/kg BBR successfully suppressed tumor growth (Fig. 2C), and this was almost as efficacious as the positive control drug afatinib. Both tumor volume and weight in BBR-treated animals were significantly decreased, to one half of that of the control animals. We further analyzed EGFR and SREBP1 expression in tumor xenograft samples (Fig. 2D), as well as the levels of lipid metabolites, and found that expression of EGFR and SREBP1 was significantly decreased, as well as the cellular and plasma levels of the lipid metabolites palmitic acid and mevalonic acid were substantially reduced by BBR treatment (Fig. 2E).

OPEN IN VIEWER

FIG. 2.

Suppression of lipogenesis by BBR selectively inhibited the progression of arthritis and tumor growth in vivo . (A, B) AIA in rats. BBR (25 mg/kg body weight) significantly inhibited AIA progression and suppressed bone damage of the ankle joint. (C) Gefitinib-resistant NSCLC xenograft model in mice. BBR substantially reduced the tumor volume and weight, demonstrating a comparable potency to that of the positive control afatinib. (D) Western blot and immunohistochemistry analyses detected the expression of EGFR and SREBP1 in tumor samples. (E) The levels of intermediate metabolites in plasma and tumors were decreased dramatically by BBR treatment. *p < 0.05, **p < 0.01, ***p < 0.001 vs control. AIA, adjuvant-induced arthritis.

Taken together, the results showed that BBR effectively inhibited the progression of both arthritis and growth of G-R NSCLC in vivo. Importantly, the level of SREBP1, a key transcription factor for regulating genes associated with de novo lipogenesis, was increased in both G-R NSCLC and RA cells, as well as in the in vivo model, and BBR significantly decreased SREBP1 expression and the levels of its downstream lipid metabolites, indicating that suppressing SREBP1 and lipid metabolism can substantially inhibit both malignant and proliferative diseases.

Inhibition of lipogenesis selectively suppresses the growth of human synovial fibroblasts and G-R NSCLC cells

To determine whether targeting lipogenesis indiscriminately affects normal cells, fibroblasts, and malignant cells, we compared the cytotoxicity toward the normal and abnormal cells in both animal models. First, the cytotoxic effect of BBR was compared among different types of lung cancer cells (Table 2 and Supplementary Fig. S2). Intriguingly, BBR showed a strong cytotoxic effect on G-R NSCLC cells but not on normal cells: the IC₅₀ value of BBR toward BEAS-2B normal lung epithelial cells was ∼10-fold higher than that toward H1975 G-R NSCLC cells containing a double mutation (L858R and T790M) in EGFR. These results suggest that BBR may be specifically cytotoxic toward G-R NSCLC cells. To determine why BBR is selective and has the strongest effect on G-R NSCLC cells, we hypothesized that BBR inhibits EGFR signaling by suppressing lipogenesis. First, EGFR expression was investigated. We found that the level of EGFR was in accordance with that of SREBP1, which was significantly downregulated by BBR (Fig. 3A). To further clarify whether downregulation of EGFR is mediated through suppression of lipogenesis, we used siRNA to knock down SREBP1. Knockdown of SREBP1 significantly reduced the protein level of EGFR (Fig. 3B). Moreover, a combination of BBR and gefitinib showed interactive effects on H1975 cells (Fig. 3C), which indicated that BBR could be combined to use with gefitinib in G-R NSCLC cells.

OPEN IN VIEWER

FIG. 3.

Suppression of lipogenesis significantly inhibited EGFR signaling and inflammatory cytokines. (A) BBR decreased the EGFR level. (B) Knockdown of SREBP1 caused EGFR degradation. (C) Combination treatment with BBR and gefitinib had synergistic effects. (D, E) The effect of BBR on various chemokines in MH7A cells and RASFs was analyzed. ^*/# p < 0.05, ^**/## p < 0.01, ^***/### p < 0.001 vs control.

To assess the effect of BBR on RA, we used MH7A cells and RASFs stimulated with TNF-α to induce generation of the proinflammatory cytokines IL-1β, IL-6, IL-8, IL-25, and IL-33. The results demonstrated that BBR effectively attenuated the upregulation of inflammatory cytokines induced by TNF-α (Fig. 3D, E, and Supplementary Fig. S3).

Next, the essential role of suppression of lipogenesis in the anticancer or antiarthritic effects mediated by BBR was examined. A critical intermediate metabolite of lipogenesis, palmitic acid, was applied with BBR to the cell culture, and the results showed that coadministration of palmitic acid significantly relieved the cytotoxic effect of BBR on G-R NSCLC cells and RASFs (Fig. 4A–D). On the basis of these results, we concluded that BBR effectively inhibited lipogenesis and thus selectively suppressed the growth of G-R NSCLC cells and the inflammation mediated by human synovial fibroblasts.

OPEN IN VIEWER

FIG. 4.

Inhibition of lipogenes effectively suppressed cell growth and caused cell cycle arrest. (A–D) Addition of palmitic acid largely counteracted the BBR-mediated suppression in NSCLC cells and RASFs. (E, F) BBR induced cell cycle arrest in G2 in H1975 cells and RASFs. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001 vs control.

The mitochondrial respiratory chain is crucially involved in the suppressive effects of BBR on both human fibroblasts and malignant cells

Since BBR effectively suppressed lipogenesis, to further explore the candidate targets for regulating lipogenesis, we first determined the effect of BBR on apoptosis and the cell cycle. BBR did not significantly increase the percentages of apoptotic cells in the H1975 cell population (Supplementary Fig. S4A), while it caused significant G2 arrest in both G-R NSCLC cells and RASFs (Fig. 4E, F, and Supplementary Fig. S4B, C), indicating that BBR suppressed the growth of G-R NSCLC carcinoma and the proliferation of human synovial fibroblasts through induction of cell cycle arrest rather than apoptosis. Since lipogenesis is directly involved in cell cycle progression (18), these results are consistent with and are a further confirmation of the suppressive effect of BBR on lipogenesis.

To identify the potential targets of BBR, we performed iTRAQ proteomic analysis. The results demonstrated that 107 genes showed a more than twofold change in expression following BBR treatment (3 μM). Of these, 61 genes were downregulated, while 46 genes were upregulated (Table 3). Bioinformatic analysis of these genes indicated that 15 genes were associated with the mitochondrial respiratory chain, which was the center of the connection network generated from the data and was significantly suppressed by BBR (Fig. 5A and Supplementary Fig. S5). Enzyme activity assays of complex I in the mitochondrial respiratory chain also demonstrated that BBR significantly inhibited the activity of complex I (Fig. 5B).

OPEN IN VIEWER

FIG. 5.

BBR targeted the cellular mitochondrial respiration chain. (A) The targets of BBR in H1975 cells were identified by proteomic analysis. (B) BBR inhibited the activity of complex I. (C) The OCR of H1975 was suppressed by BBR. (D) H1975 NSCLC cells showed a higher OCR than that of the normal BEAS-2B cells. (E) RASFs showed a higher OCR than that of normal fibroblasts from metabolic arthritis patients. OCR, oxygen consumption rate. ***p < 0.001.

Since the mitochondrial respiratory chain involves oxygen consumption and energy production, to clarify whether the suppression of the mitochondrial respiratory chain is the major cause of selectivity of BBR on different types of cells, we compared the oxygen consumption rate (OCR) of the cells. OCR is a critical indicator of mitochondrial function and directly reflects oxidative phosphorylation (OXPHOS) of the mitochondria, which are the major source of ATP for eukaryotic cells. First, compared with the control H1975 cells, treatment with BBR (3 μM) dramatically decreased the OCR and extracellular acidification rate (ECR) of the cells (Fig. 5C and Supplementary Fig. S6A). Moreover, the basal OCRs of BEAS-2B (normal lung epithelial cells) and H1975 cells were compared (Fig. 5D). Interestingly, H1975 cells had a higher OCR than did BEAS-2B cells, which is consistent with the cytotoxic effects of BBR. Similarly, a much higher OCR was also observed in RASFs than in non-RA fibroblasts (primary human fibroblast like synoviocytes [HFLS] cells derived from normal synovial tissues) (Fig. 5E). Higher rates of oxygen consumption by the cells indicated increased sensitivity of the cells to BBR treatment.

Mitochondria are the major organelles involved in energy and ROS generation. Dysfunction of mitochondria will lead to energy deprivation and ROS accumulation (15). Therefore, we first examined whether BBR caused ATP reduction. As expected, the cellular ATP level in BBR-treated H1975 cells was significantly decreased (Supplementary Fig. S6B). It is well established that nutrient deprivation or high levels of stress can increase mitochondrial fission and/or decrease fusion (45). Second, the cellular ROS level was analyzed and compared. BBR notably enhanced the ROS generation within 3 h (Supplementary Fig. S7A). Collectively, our findings show that BBR inhibits lipogenesis mainly by targeting AMPK pathway and thus suppresses the growth of NSCLC and the inflammatory responses elicited by RA fibroblasts.

ROS–AMPK activation signaling is required for the inhibitory effects of BBR on both human synovial fibroblasts and malignant cells by suppressing lipogenesis

AMPK is an energy sensor and regulates energy homeostasis of cells by detecting changes in the AMP/ATP ratio. As expected, the deprivation of ATP by BBR directly activated AMPK and its direct downstream pathway, which involves Acetyl-CoA carboxylase, even at 3 h (Fig. 6A, B). Moreover, treatment with the AMPK inhibitor compound C (10 μM) significantly blocked the inhibitory effect of BBR (Fig. 6C–F), whereas overexpression of AMPK (to simulate the effect of AMPK activation) significantly suppressed cellular viability; a dominant negative mutant of AMPK was used as a control (Fig. 6G). Since BBR could effectively induce AMPK activation and ROS generation, we explored whether there are direct associations involved. Cells were cotreated with BBR (3 μM) and ROS blocker N-acetylcysteine (NAC, 10 mM). Interestingly, the activation of AMPK induced by BBR was mostly counteracted by NAC, which indicated that the activation of AMPK was largely dependent on the stimulation of ROS increasing (Supplementary Fig. S7B). These results suggested that ROS increasing and AMPK activation are essential requirements for the anticancer and antiarthritic effects of BBR.

OPEN IN VIEWER

FIG. 6.

The anticancer and antiarthritic effects of BBR that occur via the activation of AMPK. In H1975 (A) and RASFs (B), BBR significantly activated AMPK and acetyl-CoA carboxylase at 3 h. (C–F) The AMPK inhibitor compound C significantly counteracted the effect of BBR on H1975 cells and RASFs. (G) Overexpression of AMPK decreased cellular viability. ^*/# p < 0.05, ^**/## p < 0.01, ^***/### p < 0.001 vs control.

Materials

BBR and metformin powder were purchased from Sigma-Aldrich (St. Louis, MO). Propidium iodide (PI) and annexin V/PI staining dye were purchased from BD Biosciences (San Jose, CA). Cell lysis buffer (RIPA) and primary antibodies against actin, total EGFR, and FAS were purchased from Cell Signaling Technology (Danvers, MA). Fluorescein-conjugated goat anti-rabbit and mouse antibodies were purchased from Odyssey (Lincoln, NE).

Cell lines and cell culture

BEAS-2B, A549, HCC827, H820, H1650, H1975, and H2228 cells were purchased from ATCC (Manassas, VA) and cultivated in RPMI 1640 medium supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA), 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen). The BEAS-2B culture flasks were precoated with a mixture of 0.01 mg/ml fibronectin, 0.03 mg/ml bovine collagen type I, and 0.01 mg/ml bovine serum albumin dissolved in bronchial epithelial basal medium (BEBM) medium (Lonza, Allendale, NJ). MH7A, a human RA fibroblast line, was purchased from ATCC and cultured in Dulbecco's modified Eagle's medium (DMEM). RASFs are primary RA fibroblasts isolated from RA patients and cultured in DMEM supplemented with 10% fetal bovine serum. HFLS cells were purchased from Cell Application and cultured in Human Synoviocyte Growth Medium. All the cells were cultivated at 37°C in a 5% CO₂ incubator. In the cotreatment experiments, the cells were pretreated with compound C for 1 h and then cotreated with BBR.

3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide assay

Cells were seeded in a 96-well microplate at 5000 cells/well and cultured overnight for cell adhesion. Various concentrations of BBR were added with the vehicle control, dimethyl sulfoxide (DMSO), and the microplates were incubated for 24 or 48 h. Each dose was assayed in triplicate. 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (5 mg/ml, 10 μl) solution was added to every well, and the plate was incubated for 4 h. Then, 100 μl of solution (10% sodium dodecyl sulfate [SDS] and 0.1 mM HCl) was added to each well, and the plate was further incubated at 37°C for another 4 h to dissolve the formazan crystals. Finally, the absorbance of the plate was measured at 570 nm (absorbance) and 650 nm (reference) using a microplate reader (Tecan, Männedorf, Switzerland). The cell viability was calculated as the percentage change in the absorbance of the treated cells divided by the absorbance of the untreated cells.

Assessment of apoptosis via annexin V/PI staining

After treatment, the cells were harvested, washed with phosphate-buffered saline (PBS), and resuspended in 1 × binding buffer. Then, the cells were double stained with annexin V/PI for 15 min at room temperature in the dark. The apoptotic cells were quantified by flow cytometry (BD FACSAria III; BD, Franklin Lakes, NJ).

Cell cycle assessment by PI staining

Cells were seeded into six-well plates with 1 × 10⁵ cells/well and cultured in an incubator overnight for cell adhesion. After treatment with BBR, the cells were trypsinized and centrifuged. The cells were washed with PBS and centrifuged to remove the PBS. The pellet was resuspended and fixed in 70% ethanol for 30 min at 4°C. After centrifugation, the cells were stained with PI for 30 min and then analyzed by flow cytometer.

ROS detection

After treatment with different concentrations of BBR for 3 h, cells were stained with DCFDA for 15 min and loaded onto flow cytometer.

Western blot analysis

Cells were lysed in RIPA lysis buffer with protease and phosphatase inhibitors to extract the total protein. The concentration of the total protein extract was determined with a Bio-Rad DCTM Protein Assay Kit (Bio-Rad, Hercules, CA). Then, 30 μg of total protein lysate was loaded onto a 10% SDS-polyacrylamide gel electrophoresis gel; the separated proteins were transferred to a nitrocellulose membrane. Membranes were blocked with 5% milk without fat in TBST for 1 h at room temperature. Primary antibodies (Santa Cruz Biotechnology; 1:500 dilution; Invitrogen and Cell Signaling Technology; 1:1000 dilution) were incubated overnight at 4°C. After the membrane was washed three times with TBST (5 min/wash), a secondary fluorescent antibody (1:10,000 dilutions) was added to the membrane at room temperature for 1 h. Actin was used as the loading control and for normalization. The signal intensity of the membranes was detected with the LI-COR Odyssey Scanner (Belfast, ME).

RNA extraction

RNA was extracted from NSCLC cells with an RNA extraction kit (Invitrogen) according to the manufacturer's instructions. After the medium was removed, 0.3 or 0.6 ml lysis buffer with 1% 2-mercaptoethanol was added to the cells. The total cell lysate was transferred to an RNase-free tube and vortexed until the cell pellet was dispersed. One volume of 70% ethanol was added to the cell lysate and then mixed thoroughly to disperse any visible precipitate. The mixture was transferred to a spin cartridge and centrifuged at 12,000 g for 15 s at room temperature to discard the flow-through. The spin cartridge was sequentially washed with wash buffer I and wash buffer II. Finally, RNA was eluted with 50 μl of RNase-free water. The RNA concentration was determined using Nano2000, and RNA quality was assessed by electrophoresis.

cDNA synthesis

The synthesis of first-strand cDNA was carried out following the instructions of the cDNA synthesis kit (Roche, Basel, Switzerland). One microgram of DNase-treated RNA was used for each cDNA synthesis. First, RNA sample was incubated with 2 μl of random primers and 1 μl of oligo (dT) primer provided with the kit at 65°C for 10 min and then cooled on ice for 2 min. The reaction mixture, which contained reaction buffer, RNase inhibitor, and reverse transcriptase, was added to the tube and then incubated at 25°C for 10 min and 55°C for 30 min. The reaction was terminated by heating at 70°C for 15 min. The final product (cDNA) was stored at −80°C until further use.

Polymerase chain reaction and real-time polymerase chain reaction

Polymerase chain reaction (PCR) was performed following the instructions of the FastTaq DNA Polymerase Kit (Roche). Each PCR consisted of 1 × PCR buffer, 200 μM dNTP, 200 nM forward primer, 200 nM reverse primer, 0.2 μl Taq, and 50 ng cDNA template in a total volume of 25 μl. The PCR running protocol is as follows: 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, annealing for 30 s, 72°C polymerization for 1 min, and a final extension at 72°C for 10 min. Real-time PCR was performed using FastStart Universal SYBR Green Master. The reaction mixture preparation consist of 10 μl of SYBR Master Mix, 1 μl forward and reverse primers, 0.2 μl template, and water to a final volume of 20 μl. The PCR protocol was 94°C for 10 min, followed by 40 cycles at 94°C for 10 s and 60°C for 30 s.

Transfection

siRNA transfections were performed with Lipofectamine™ LTX Reagent (Invitrogen). 1.5 × 10⁵cells were seeded to each well of six-well plate 1 day before transfection. The transfection complexes were prepared as follows: 100 μM siRNA was diluted in 100 μl Opti-MEM^® I Reduced Serum Medium (Invitrogen) and mixed gently. Next, 4 μl LTX Reagent was directly added to each diluted solution sample and incubated for 5–15 min at room temperature. Finally, the transfection mixture was added to the well, mixed gently by rocking the plate back and forth.

Automated liquid chromatography-mass spectrometry/mass spectrometry analysis of the peptides

After collection, protein lysate was trypsin digested, and then the peptide was analyzed using ultrahigh-performance liquid chromatography with an Agilent 1290 Infinity system (Agilent, Santa Clara, CA) as previously reported (23). 0.1% formic acid in Milli-Q water (A) and 0.1% formic acid in acetonitrile (B) consist of mobile phase. Chromatography was performed on an ACQUITY UPLC^® BEH C18 column (2.1 × 100 mm, 1.7 μm) and performed on Waters Mass spectrometry (Milford, MA). After acquiring raw data, the results were analyzed by Agilent Bioconfirm protein deconvolution software (Agilent).

OCR assay

Initially, 3000–5000 cells were suspended in 80 μl medium and seeded in each well of a Seahorse plate from B–G; wells A and H contained 80 μl medium without cells. Then, the plate was incubated at 37°C for overnight adhesion. BBR was added to wells E–G for 24 h, while wells A–D contained DMSO, as the control. After 24 h, the culture medium in all wells of the plate was removed and replaced by 80 μl assay medium. Oligomycin (a complex V inhibitor) at 10 μM, carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP) (a protonophore) at 0.5 μM, and antimycin A and rotenone (inhibitors of complex III and I) at 0.5 μM were added to the kit pack according to the user manual. After calibration, the mitochondrial OCR of cells was detected by an Extracellular Flux Analyzer (Seahorse Bioscience, MA).

Xenograft model

The xenograft model was established as previously described (23). Briefly, 1 × 10⁶ H1975 cells/100 μl were mixed with 50 μl Matrigel and then implanted into the right forelimb of each nude mice. After the diameter of tumors reached 5 mm³, the mice were to be treated with BBR (10 mg/kg) and afatinib (10 mg/kg) 5 days/week through oral administration. The tumor volume was calculated by the following equation: volume = (width² × length)/2.

Statistical analyses

All data are expressed as mean ± standard error of the mean of three individual experiments. Differences between groups were determined using a one-way analysis of variance using GraphPad Prism 5 (GraphPad Software, La Jolla, CA), followed by Bonferroni's test to compare all pairs. Student's t test was used to compare two groups. The level of significance was set at p < 0.05 for all tests.