Gambogic acid inhibits epithelial–mesenchymal transition in breast cancer cells through upregulation of SIRT1 expression in vitro

Gambogic acid (GA) is a natural product that selectively induces apoptosis of cancer cells in vitro with high efficiency and low toxicity. Our previous data have shown inhibitory effects of GA on breast cancer cells. However, the detailed mechanisms for GA remain largely unknown. This study aimed to investigate the effect of GA on TGF‐β1‐induced tumor invasion and EMT in MDA‐MB‐231 cells, and contribution of elevated SIRT1 to antitumor effects of GA. Human breast cell MDA‐MB‐231 and MDA‐MB‐231 were incubated with TGF‐β1 (100 ng/ml) and GA. Cell viability was determined by MTT assay, while tumor invasion was determined by Boyden chamber invasion assay. The mRNA levels of SIRT1 and TGF‐β1 were measured by quantitative real‐time polymerase chain reaction (qRT‐PCR). The protein expressions of SIRT1 and EMT‐associated mesenchymal marker Vimentin were measured by Western blotting. Interleukin‐6 (IL‐6) protein content was measured by enzyme linked immunosorbent assays (ELISAs). Our results showed that, GA showed decreased proliferation in MDA‐MB‐231 cells, especially in MDA‐MB‐231 cells with TGF‐β1 incubation. GA inhibited tumor invasion and EMT in TGF‐β1‐treated MDA‐MB‐231 cells. GA increased mRNA and protein expression of SIRT1 in MDA‐MB‐231 cells with and without TGF‐β1 treatment. SIRT1 mRNA level in MDA‐MB‐231 cells was increased by TGF‐β1 incubation and decreased by GA treatment. In addition, the production of IL‐6 was increased by TGF‐β1, and decreased by GA. In conclusion, GA inhibits tumor invasion and EMT in breast cancer, potentially through upregulating SIRT1 expression. This study provides a novel antitumor effect of GA in breast cancer.


| INTRODUCTION
Gambogic acid (C 16 H 12 O 5 , Figure 1) is a flavonoid isolated from the root of Scutellaria baicalensis Georgi. It is a traditional herbal medicine generally regarded as an analgesic, antipyretic, antitumor, and antiinflammatory agent, which has been proved to possess multiple pharmacological effects in several types of cancer both in vitro and in vivo, including apoptosis and metastasis inhibition. 1,2 GA has been found to inhibit cell proliferation, induce apoptosis and decrease NF-κB levels, thereby reducing in vivo tumor mass. 3 Our previous study found that GA inhibited cell proliferation and cell cycle progression, induced apoptosis and inhibited tumor invasion and epithelial-mesenchymal transition (EMT) process in TGF-β1-treated MDA-MB-231 cells. 4,5 However, the mechanisms underlying invasion and EMT inhibition by GA remain largely unknown.
Sirtuin-1 (SIRT1) is a nicotinamide adenine dinucleotide (NAD +)-dependent histone deacetylase, with regulatory function on cell's stress adaption, cell metabolism and apoptosis and DNA damage response. 6 SIRT1 has been shown to involve aging-related diseases, such as inflammation, diabetes and neurodegenerative diseases. 7 SIRT1 participates in tumor initiation, progression and drug resistance thorough regulating critical carcinogenesis process, such as cell growth, apoptosis, senescence, and angiogenesis. 8 However, the roles of SIRT1 in breast cancer are controversial. SIRT1 overexpression is associated with poor prognosis of breast cancer, 9 while SIRT1 represses transcriptional response to estrogen receptor and inhibits proliferation in breast cancer cells. 10 The roles of SIRT1 in breast cancer are still unclear and need further investigation on different regulating system and pathways.
In this study, we aimed to investigate regulation of SIRT1 by GA in TGF-β1-induced breast cancer cells and its related mechanisms. We incubated breast cells with TGF-β1 to induce tumor invasion and EMT, and analyzed cell viability, invasion, EMT mRNA and protein expressions of SIRT1 by GA. The effect of SIRT1 on invasion and EMT by GA was verified by nicotinamide (NAM), a specific SIRT1 inhibitor. We also detected the mRNA expression and culture supernatant cytokine interleukin-6 (IL-6) to explore potential mechanisms SIRT1 in breast cancer.

| Materials
Gambogic acid (C 16 H 12 O 5 ) was isolated from the root of S. baicalensis according to a previously reported method and dissolved in DMSO.
Samples containing 99% or higher in GA were used in all experiments.
GA was dissolved in DMSO to 200 mM and stored at À20 C. Before every experiment, the stock solution of GA was diluted with basal medium to various working concentrations. atmosphere of 5% CO 2 and 95% air. The cells were digested by trypsinization when reaching confluence, and then they were sub cultured in new culture flasks at lower numbers.

| Cell viability assay
The cell viability of MDA-MB-231/MDA-MB-231 cells was determined by MTT (Sigma Chemical Co., St. Louis, MO) assay. Briefly, cells at the logarithmic growth phase were seeded to each well of 96-well culture plates (cell density: 1 Â 10 3 cell/ml; each well 100 μl cell suspension). After 24 h incubation at 37 C in a humidified atmosphere with 5% CO 2 , cells were incubated with recombinant human TGF-β1 (100 ng/ml) (R&D Systems Inc., Minneapolis, MN). The cells with or without TGF-β1 incubation were then treated with different concentrations of GA (0, 100, 200, 500 μmol/L). After treatment for 72 h, 10 μl MTT solution (concentration: 5 mg/ml) was added to each well, followed by incubation at 37 C for 4 h. After centrifugation at 1000g for 10 min, the supernatant was discarded to obtain formazan pellet, and then it was dissolved completely with 100 μl DMSO after being shaken for 10 min. An enzyme linked immunosorbent assay (ELISA) plate reader was applied to measure the optical density (OD) at 570 nm wavelength to determine the amount of pellet.

| Statistical analysis
All quantitative data were presented as mean ± SD. Statistical analysis was performed using the commercially available software (SPSS version 19.0). Student t-test (unpaired, two tailed) was used to compare the difference between two groups. A probability value of <.05 (p < .05) was considered as statistically significant difference.

| GA enhanced mRNA and protein expressions of SIRT1 in MDA-MB-231 cells induced by TGF-β1
To explore whether SIRT1 mediates the mechanism underlying the inhibitory effect of GA against TGF-β1-induced invasion and EMT, we firstly The density of each band of Vimentin protein was converted into grayscale values and normalized to that of the internal control β-actin. Data were expressed as mean ± SD and a twotailed, unpaired t-test was performed. Significant difference from the control group is denoted by "*" (p < .05). Significant difference from the TGF-β1 group is denoted by "#" (p < .05) expression of SIRT1 protein by GA, and higher SIRT1 protein in TGF-β1-induced MDA-MB-231 cells than cells without TGF-β1 incubation ( Figure 3AC,D). These data indicate the upregulatory role in SIRT1 expression in breast cancer cells with decreased invasion and EMT by GA.

| Effect of specific SIRT-1 inhibitor on GAexerted inhibition on invasion and EMT induced by TGF-β1
To determine whether SIRT1 mediated the inhibitory effect of GA on TGF-β1-induced tumor invasion and EMT in MDA-MB-231 cells, we further explored whether nicotinamide, a specific inhibitor of SIRT1, reverse this inhibitory effect of GA against invasion and EMT. We The SIRT1 protein was detected by western blot using anti-SIRT1 antibody. Each band was converted into grayscale values and normalized to that of the internal control β-actin. Data were expressed as mean ± SD from three independent experiments, and a two-tailed, unpaired t-test was performed. Significant difference from the control group is denoted by "*" (p < .05). Significant difference from the TGF-β1 + NaSH group is denoted by "#" (p < .05) measure the mRNA expression of TGF-β1, and found TGF-β1-induced cells showed higher TGF-β1 mRNA expression compared with control cells ( Figure 5A). GA treatment significantly decreased TGF-β1 mRNA expression in TGF-β1-induced cells, and this decrease could be abolished by nicotinamide. This indicates that SIRT1 might lie in upstream of TGF-β1 and regulate its expression. We further analyzed the cytokines IL-6 in culture media from breast cancer cells by ELISA over a period of 24 h of TGF-β1 and GA incubation. The levels of IL-6 in culture media were significantly increased by TGF-β1, and were significantly decreased by GA in TGF-β1-induced cells ( Figure 5B). The decrease in IL-6 levels by GA was abolished by nicotinamide. Representative pictures are shown from three independent experiments. The density of each band was converted into grayscale values and normalized to that of β-actin. All data were expressed as mean ± SD and a two-tailed, unpaired t-test was performed. Significant difference from the control group is denoted by "*" (p < .05). Significant difference from the TGF-β1 group is denoted by "#" (p < .05) and promoted PARP-dependent apoptotic cell death. 14 Therefore, we focused on whether SIRT1 mediates the promoter or inhibitory role of GA against TGF-β1-induced invasion and EMT. We found that the GA not only upregulated the mRNA and protein expressions of SIRT1 in MDA-MB-231 cells in a concentration dependent manner, but also it increased SIRT1 expression more significantly in TGF-β1-induced cells compared with cells without TGF-β1. Indeed, SIRT1 activation by GA has been found in many physiopathological processes of ageassociated diseases, including endothelial cell senescence and oxidative damage of cardiomyocytes. [15][16][17] Our study has firstly found upregulation of SIRT1 by GA and indicates that the SIRT1 might demonstrate higher sensitivity to GA regulation in breast cancer cells.
To explore whether SIRT1 mediate the inhibition of invasion and  18 EMT can enable epithelial cancer cells to acquire highgrade malignancy and metastasis potential. 19,20 EMT is also associated with aging in cancers, 21,22 which is an aging disease with nearly 65% of cancer patients ≥65 years old. 23 As a protein regulating the pathology for aging, SIRT1 might participate EMT process in cancer cells.
Indeed, the regulatory effect of SIRT1 on EMT is controversial. SIRT1 could induce EMT in gastric cancer cells and osteosarcoma cells. 24,25 However, decreased SIRT1 expression promotes invasion and metastases via EMT in breast cancer and lung cancer, 26,27 which is in accordance with our results. The discrepancy on EMT by SIRT1 may be caused by different cancer types and various dependences of cancer cells on aging.
We further explored the detailed mechanisms of SIRT1 underlying This suggests SIRT1 might lie in upstream of TGF-β1 and regulate its expression. Inhibition TGF-β pathway by SIRT1 has been reported in renal fibrosis, 28 tissue fibrosis, 29 and AGEs-induced rat glomerular mesangial cells. 30 However, this regulation has not been reported in breast cancer cells, and was confirmed in our study by the result that decreased TGF-β1 expression can be reversed by SIRT1 inhibitor. In our study, culture supernatant IL-6 level was also increased by TGF-β1 and decreased by GA. In mouse embryonic fibroblasts oxidative stressinduced inhibition of SIRT1 can stimulates the secretion of IL-6. 31 Whether IL-6 production was enhanced directly by SIRT1 in breast cancer cells need further study. Cytokines, including IL-6, IL-8 and TGF-β, are frequently elevated in breast cancers and constitute tumor microenvironment for the expansion of cancer stem cells (CSCs). 32 Furthermore, in breast cancer cells, activation of IL-6 leads to expansion of CSCs population, and highly enriched CSCs which display an EMT phenotype secreting over 100-fold more IL-6 than parental cells, thus making an IL-6 inflammatory positive feedback loop. 33 Therefore, GA and SIRT1 may block this positive feedback loop and reverse EMT phenotype, and make a therapeutic strategy for advanced stage breast cancer, which deserve further investigation.
In conclusion, our present study demonstrated that GA can inhibit TGF-β1-induced tumor invasion and EMT phenotype in MDA-MB-231 cells through upregulating SIRT1 expression. GA also decreased F I G U R E 5 Effect of GA and specific SIRT-1 inhibitor on TGF-β1 mRNA expression and cytokine level of IL-6 in culture media. MDA-MB-231 cells were incubated with TGF-β1 (100 ng/ml), TGF-β1 + GA (200 μmol/L) or TGF-β1 + GA + nicotinamide (NAM, 100 μmol/L) for 24 h. Cells treated with DMEM media served as control. (A) qRT-PCR shows TGF-β1 mRNA expression was increased in MDA-MB-231 cells by TGF-β1, and was decreased by GA. Nicotinamide reversed the inhibitory effects on invasion and EMT by GA. (B) Western blot shows IL-6 in culture media was increased in MDA-MB-231 cells by TGF-β1, and was decreased by GA. Nicotinamide increased the IL-6 level in culture media of MDA-MB-231 cells incubated with TGF-β1 and GA. All data were expressed as mean ± SD and a two-tailed, unpaired t-test was performed. Significant difference from the control group is denoted by "*" (p < .05). Significant difference from the TGF-β1 group is denoted by "#" (p < .05) TGF-β1 mRNA expression and culture supernatant IL-6 level in MDA-MB-231 cells. Moreover, inhibition of SIRT1 reversed GA-elicited inhibition of invasion, EMT, TGF-β1 mRNA expression and culture supernatant IL-6 level in MDA-MB-231 cells induced by TGF-β1.
These findings suggest that the inhibition of invasion, EMT, TGF-β1 mRNA and IL-6 secretion by GA is mediated by SIRT1. Our results provide SIRT1-TGF-β1-IL-6 as a novel molecular mechanism underlying GA-mediated anticancer effect in breast cancer.