Curcumin inhibits the growth of triple‐negative breast cancer cells by silencing EZH2 and restoring DLC1 expression

Abstract Enhancer of zeste homolog 2 (EZH2), an oncogene, is a commonly up‐regulated epigenetic factor in human cancer. Hepatocellular carcinoma deletion gene 1 (DLC1) is an antioncogene that is either expressed at low levels or not expressed in many malignant tumours. Curcumin is a promising anticancer drug that has antitumour effects in many tumours, but its mechanism of action is unclear. Our research demonstrated that EZH2 was up‐regulated in breast cancer (BC) tissues and cells, whereas DLC1 was down‐regulated, and the expression of EZH2 and DLC1 was negatively correlated in BC. By analysing the characteristics of clinical cases, we found that positive expression of EZH2 and negative expression of DLC1 may be predictors of poor prognosis in patients with triple‐negative breast cancer (TNBC). Moreover, knockdown of EZH2 expression restored the expression of DLC1 and inhibited the migration, invasion and proliferation, promoted the apoptosis, and blocked the cell cycle of MDA‐MB‐231 cells. Furthermore, we found that curcumin restored the expression of DLC1 by inhibiting EZH2; it also inhibited the migration, invasion and proliferation of MDA‐MB‐231 cells, promoted their apoptosis and blocked the cell cycle. Finally, xenograft tumour models were used to demonstrate that curcumin restored DLC1 expression by inhibiting EZH2 and also inhibited the growth and promoted the apoptosis of TNBC cells. In conclusion, our results suggest that curcumin can inhibit the migration, invasion and proliferation, promote the apoptosis, block the cycle of TNBC cells and restore the expression of DLC1 by inhibiting the expression of EZH2.

as Asia, Africa and South America. 2 The increased risk of BC is closely related to several important factors, including increased age, number of first-degree relatives with BC, atypical hyperplasia and age of menarche. 3 As a molecular subtype of BC, triple-negative breast cancer (TNBC) lacks the expression of ER, PR and HER2. 4 The prognosis of patients with TNBC is worse than that of non-TNBC patients. Non-TNBC patients may benefit from the anti-HER2 antibody trastuzumab and endocrine therapy. 5,6 Studies have found that the release of upstream epigenetic regulatory factors can promote epigenetic changes, leading to abnormal silencing of antioncogenes, which is also an important mechanism for promoting cancer. 7 Drosophila zeste gene enhancer homologue 2 (EZH2), as a catalytic subunit of PRC2, is a commonly up-regulated epigenetic factor in cancer. 8 As a histone methyltransferase, EZH2 can specifically catalyse histone H3K27me3 and inhibit histone modification to control epigenetic transcriptional regulation. 9 Therefore, the up-regulation of EZH2 can promote the metastasis of cancers and play a pivotal role in the progression of cancers. 10 Hepatocellular carcinoma deletion gene 1 (DLC1), an antioncogene, is located on human chromosome 8p22 and is either expressed at low levels or not expressed in 50% of human hepatocellular carcinomas and many other human cancers (including colon cancer, lung cancer, prostate cancer and BC). [11][12][13] In addition to DLC1, 8p22 also contains other tumour suppressor genes, such as MTUS1, TUSC35 and FGL1. 14 Moreover, in addition to genomic deletion in tumours, down-regulation of DLC1 expression and promoter methylation is also common in human tumours, which makes DLC1 the most important tumour suppressor on 8p22. 11,15 A large number of studies have proven that curcumin has strong anti-inflammatory, antioxidant and antitumour properties. 16,17 In pancreatic cancer, curcumin has been known to inhibit many oncogenes, including VEGF, Akt, Erk, cytochrome c oxidase subunit II and EZH2. 18 EZH2 has been considered as a target for curcumin in pancreatic and colorectal cancer. 19,20 EZH2 is an oncogene, and DLC1 is an antioncogene, and both of them are related to epigenetics. The objective of our study was to investigate the regulatory relationship between EZH2 and DLC1 in BC and to determine whether curcumin can inhibit the growth of TNBC cells by silencing EZH2 to restore the expression of DLC1.

| Analysis of gene differential expression and correlation based on TCGA database
UALCAN (http://ualcan.path.uab.edu/) and GEPIA (http://gepia. cance r-pku.cn/), two-level online analysis tools based on TCGA, were used to analyse the differential expression and correlation of EZH2 and DLC1 in BC and a variety of other tumours, and further stratified analyses were conducted on the expression of EZH2 and DLC1 in patients with different clinical characteristics.
Tumours with grades greater than or equal to grade 2 are considered positive for antigen expression.

| RNA extraction and qRT-PCR
Total RNA was extracted by TRIzol reagent (TaKaRa), and cDNA was synthesized by reverse transcription using PrimeScript™ RT Master Mix (TaKaRa) according to the manufacturer's instructions.
The levels of GAPDH, DLC1 and EZH2 were detected by qRT-PCR using SYBR@ Premix Ex TaqTM (Roche). The results were normalized to the expression of GAPDH. The primer sequences were synthesized by Sangon ( Table 1). The relative expression levels of EZH2 and DLC1 were quantitatively calculated by the 2 (−ΔΔCT) method.

| Western blot assay and antibodies
Total proteins were extracted with RIPA lysate containing protease inhibitors. Protein was isolated by SDS-PAGE. The protein bands were transferred to PVDF membranes after electrophoresis. After blocking with 5% skimmed milk powder, the primary antibody was added, and the reaction lasted overnight at 4°C. Immunoblots Afterwards, the membrane was incubated with the secondary antibody for 1 hour. ECL detection reagent (Santa Cruz) was used to detect the signal. ImageJ software was used to analyse the grey values of the protein bands.

| Chromatin immunoprecipitation (ChIP) assay
ChIP assays were performed using a ChIP kit (Cell Signaling Technology, USA) according to the manufacturer's protocol. ChIPgrade antibody against H3K27me3 (Abcam, USA) was used in the assay. Rabbit IgG (Cell Signaling Technology, USA) was used as a negative control in the assay. Precipitated DNA samples were analysed by PCR with primer pairs specific for the promoters of the DLC1 gene. The following primers were used: DLC1, sense primer, anti-sense primer, 5'-CTGGAATACAACAACTTTGCACC-3'. The fold enrichment from the ChIP assay was calculated with reference to the IgG control after normalization to the input DNA. Each experiment was repeated three times.

| Proliferation assay
MDA-MB-231 cells (3000 cells/well) from the experimental and control groups were seeded in 96-well plates. The cell viability at 24,48 and 72 hour was measured by CCK-8 assay. The cell proliferation ability was expressed as the absorbance ratio of the experimental group and the control group at 450 nm.

| Cell apoptosis assay
The Annexin V-FITC/PI Apoptosis Detection Kit (BD) was used to detect the apoptosis of cells by flow cytometry according to the manufacturer's instructions. The cells were collected and resuspended in buffer, stained in the dark and analysed by flow cytometry.

| Cell cycle assay
A cell cycle assay kit (KeyGEN, China) was used to assess cell cycle distribution by flow cytometry according to the manufacturer's instructions. The cells were collected and fixed overnight with 500 µL of 70% cold ethanol. Finally, the cell cycle assay kit was used for staining, and the results were analysed by flow cytometry.

| Cell migration and invasion assays
A wound healing assay was used to detect cell migration. First, the cells were inoculated into 6-well plates and cultured to 80% conflu-

| Statistical analysis
Statistical analysis was performed using SPSS 17.0 (SPSS, Inc, USA) and GraphPad Prism 5 (GraphPad Software Inc, San Diego, CA, USA) in this study. The data among the groups were analysed by t test or one-way analysis of variance (ANOVA) using the mean ± standard deviation. The chi-square test was used to analyse the difference and correlation between clinicopathological parameters, and the Kaplan-Meier chart and log rank test were used for survival analysis. The difference was statistically significant with P < 0.05.

| The expression of EZH2 is up-regulated and DLC1 is down-regulated in various tumours
To explore the expression of EZH2 and DLC1 in tumours, we ana-

| EZH2 and DLC1 may be predictors of prognosis in patients with TNBC
To explore the relationship between EZH2 and DLC1 and the clinical characteristics of BC patients, we first carried out subgroup analysis based on the TCGA database. The results showed that EZH2 and DLC1 were differentially expressed in patients of different ages ( Figure 3A,G), races ( Figure 3B,H), cancer stages ( Figure 3C,I), molecular subtypes ( Figure 3D,J) and histologic subtypes ( Figure 3E,K).
The expression of DLC1 was also different between premenopausal and postmenopausal patients with BC ( Figure 3L). In addition, as shown in Table 2, the positive expression of EZH2 is associated with lymph node metastasis (P = 0.02) and tumour size (P = 0.01), while the negative expression of DLC1 is associated with lymph node metastasis (P = 0.043) and tumour size (P = 0.036). Moreover, we found that TNBC patients with positive expression of EZH2 and negative expression of DLC1 had shorter overall survival (P < 0.05, Figure 2F).

| EZH2 regulates the expression of DLC1 in MDA-MB-231 cells by mediating H3K27me3
The results of correlation analysis based on the TCGA database showed that the expression levels of EZH2 and DLC1 were negatively correlated in BC (R = −0.13, P = 2.7 × 10 −5 , Figure 4A). Moreover, our correlation analysis also showed that the expression levels of EZH2 and DLC1 were negatively cor-

| Both EZH2 and DLC1 are related to genes involved in cell proliferation, apoptosis and the cell cycle in BC
The correlation analysis based on TCGA demonstrated that the expression of EZH2 was positively correlated with the expression of MKI67 (R = 0.64, P = 0, Figure 4F) and CDK1 (R = 0.51, P = 0, Figure 4G) and negatively correlated with the expression of CASP9 (R = −0.15, P = 1.5 × 10 −6 , Figure 4H) in BC.

| EZH2 inhibitor and knockdown of EZH2 can inhibit the migration, invasion and proliferation of TNBC cells, promote apoptosis and block the cell cycle
To

| Curcumin inhibits the migration, invasion and proliferation of TNBC cells, promotes apoptosis and blocks the cell cycle
To confirm the effect of curcumin on TNBC cells, we also performed migration cell apoptosis, cell cycle, proliferation, invasion and   Figure 6G).  Figure 7D). IHC showed that the expression of EZH2 was down-regulated and the expression of DLC1 was up-regulated in TNBC xenograft tumour tissues treated with curcumin ( Figure 7E). can inhibit the migration and proliferation of HCC cells 38