Luteolin suppresses gastric cancer progression by reversing epithelial-mesenchymal transition via suppression of the Notch signaling pathway

Hs-746T and MKN28 GC cells were cultured with 0, 10, 20 and 30 μM luteolin. CCK-8 assay was performed every 24 h, and results showed that proliferation of GC cells were effectively inhibited by luteolin in a dose- and time-dependent manner (Fig. 1a, b). Moreover, luteolin treatment also significantly reduced the number of colonies compared with the control group (for Hs-746T, P = 0.0097; for MKN28, P = 0.0014; Fig. 1c, d).

The percentage of early and late apoptosis was increased upon treatment with 10 and 30 μM luteolin compared with the control group (Fig. 2a, b). The proportion of apoptotic Hs-746T (0 vs. 10 μM, P = 0.0047, 0 vs. 30 μM, P = 0.0009, Fig. 2c) and MKN28 (0 vs. 10 μM, P = 0.0014, 0 vs. 30 μM, P = 0.0010, Fig. 2d) cells increased in a dose-dependent manner. Since PI3K/Akt signaling is implicated in cell apoptosis in a majority of tumors, we examined the phosphorylated Akt levels in GC cells after treatment with luteolin. The results showed that phosphorylated Akt (Ser-473) was decreased by luteolin treatment (Fig. 2e).

NCI-N87 GC cells showed a mesenchymal phenotype, as evidenced by F-actin staining, in the absence of luteolin treatment (Fig. 3a). However, when NCI-N87 cells were treated with luteolin, the cytoskeleton shrank and cell size decreased (Fig. 3b). These findings indicate that luteolin can suppress the motility of GC cells. Transwell assays showed that invasion and migration of GC cells was significantly inhibited by luteolin treatment (P < 0.01, Fig. 3c–f).

The remodeling of the cytoskeleton upon luteolin treatment indicated that luteolin may regulate this process by inhibition of EMT in GC cells. We observed that the epithelial biomarker E-cadherin was increased and the mesenchymal biomarkers N-cadherin, vimentin, and Snail were reduced in a dose-dependent manner upon luteolin treatment (Fig. 4b). Luteolin treatment also caused a decrease in β-catenin levels (Fig. 4c). We also found that Notch1, cyclin-D1, and Hes-1 were downregulated due to luteolin treatment (Fig. 4d–f), suggesting that luteolin prevented GC progression by suppressing Notch signaling.

To investigate the suppressing effects of luteolin on GC progression whether through regulating Notch1 or not, Notch1 was downregulated or overexpressed in GC cells. Notch1 knockdown in Hs-746T and MKN28 cells decreased the expression of its target genes Hes-1, Hey-1, and cyclin-D1 (Fig. 5a). Moreover, proliferation and migration were inhibited in Notch1-silenced GC cells compared with the control cells (Fig. 5b, c). In addition, Notch1 knockdown promoted cell apoptosis and reversed EMT in GC cells (Fig. 5d, e). However, overexpression of Notch1 recovered EMT in Hs-746T following luteolin treatment as well as elevated AKT phosphorylation (Fig. 5f). The inhibiting effect on cell migration by luteolin treatment was also partially reversed by overexpression of Notch1 (Fig. 5g). These observations confirm that luteolin treatment suppressed GC progression by inhibiting Notch signaling. Furthermore, NICD directly bound with β-catenin to form a complex, while the interaction between NICD and β-catenin was abrogated subsequent to luteolin treatment in vitro and in vivo (Fig. 5h). The interaction between NICD and β-catenin may contribute to promote cell proliferation, cell migration, and inhibit cell apoptosis in GC by regulating downstream target genes (Fig. 5i), which is blocked by luteolin treatment.

To test the effects of luteolin on tumor growth in vivo, MKN28 cells were injected subcutaneously into nude mice. After the tumors were formed, nude mice were injected 6 times intraperitoneally with PBS or luteolin (10 mg/kg). We found that the tumor volume (P < 0.01) and tumor weight (P < 0.05) in luteolin-treated mice were less than that in the control group (Fig. 6a–c). Furthermore, β-catenin, Notch1 and Ki-67 expression were decreased in tumors from luteolin-treated mice (Fig. 6d); while in contrast, TUNEL staining was elevated in tumors from mice treated with luteolin (Fig. 6e). Analysis of data available online on KMplot indicated that higher expression of Notch1 correlated with a poor overall survival (OS) (P = 0.00022, Fig. 6f) and a poor time to first progression (FP) (P = 0.00062, Fig. 6g). These results suggest that luteolin can suppress GC progression by inhibiting Notch1 expression.

Based on our observations, we propose a model for the roles of luteolin, Notch1, and β-catenin in GC (Fig. 7). Ligand binding activates Notch, causing translocation of NICD to the nucleus, which forms a complex with activated β-catenin. This results in the regulation of target genes and induces cell proliferation and metastasis in GC. The PI3K/Akt signaling enhances the formation of this complex. Luteolin can block Notch signaling, and exerts an anti-tumorigenic effect in GC by inhibiting cell proliferation and migration and increasing cell apoptosis. Thus, luteolin may be an effective drug for the treatment of cancers.

The human GC cell lines NCI-N87 and MKN28 were maintained in our lab, and Hs-746T was purchased from American Type Culture Collection. Cells were cultured at 37 °C in 5% CO₂ and saturation humidity in RPMI-1640 medium with 10% fetal bovine serum containing penicillin and streptomycin. Luteolin was purchased from Aladdin Industrial Corporation (Shanghai). The molecular weight is 286.24 g/mol and the purity is greater than 98% (HPLC).

Cell proliferation was monitored by Cell Counting Kit-8 (CCK-8). In brief, GC cells were suspended in medium with or without luteolin treatment and then plated in 96-well plate at the concentration 2000 cells/well. Cell proliferation was measured every 24 h for 5 days after adding CCK-8 reagent 2 h at the absorbance 450 nm using Epoch Microplate Spectrophotometer (Bio Tek). For colony formation, Hs-746T and MKN28 cells were plated in 6-well plates at the concentration 1000 cells/well. The experimental groups were treated with luteolin for 24 h groups and then replaced for fresh medium. After 10–14 days, the plates were stained with 1% crystal violet.

Hs-746T and MKN28 cells were plated in 6-well plates treated with or without luteolin for 24 h. Then cells were collected and examined by apoptosis detection kit (BD Pharmingen). Briefly, 3 μl annexin V-FITC and 5 μl propidium iodide were added into cells successively for 15 min. And then GC cells were monitored by flow cytometry. Right upper quadrant represents percentage of late apoptosis. Right lower quadrant represents percentage of early apoptosis.

A number of 1 × 10⁵ GC cells were suspended in serum-free medium with or without Matrigel (BD Bioscience, CA, USA) in upper chambers (Corning Costar, NY, USA) and luteolin or PBS were added into 24-well plates. Next, GC cells were fixed by 10% formalin and stained by 0.5% crystal violet after 24 h. Finally, GC cells that passed through membrane were photographed and counted.

Cells were plated into 8-well glass (Merck Millipore) for overnight and then fixed, permeated and blocked according to protocols. We next stained the cells with β-catenin antibody (1:100; Cell Signaling Technology, CST), followed by incubation with fluorescent secondary antibody for 1 h at room temperature. The nuclei were stained with DAPI. And to visualize the cytoskeleton of GC cells, rhodamine phalloidin (1:20; CST) was used. Slides were analyzed and imaged on a fluorescence microscope.

Notch1 shRNA, negative control, pCMV-Notch1 (H3176 pLenti-CMV-MCS-HA-3Flag-P2A-EGFPT2A-Puro), and control vectors were purchased from Oobio Corporation (Shanghai, China) and the vectors carry puromycin-resistance function. siRNA sequences of Notch1 or negative control (NC) were as follows: Notch1, GCAACAGCTCCTTCCACTT; NC, TTCTCCGAACGTGTCACGT. Lip2000 (Invitrogen, Carlsbad, USA) was used to transfect vectors into GC cells and then transfected cells were selected by treatment with puromycin. The effects of shRNA and pCMV-Notch1 were confirmed at protein level using western blot assay.

The method was consist with the previous [46]. In brief, proteins of cells were separated by SDS-PAGE and then were transferred into PVDF membranes. Primary antibody (1:1000 dilutions) AKT, p-AKT, NICD, β-catenin, E-cadherin, N-cadherin, Snail and vimentin were purchased from cell signaling technology (CST, USA). Notch-1, Hes-1, Hey-1, Cyclin-D1 (1:1000 dilutions) and GAPDH (1:10000 dilutions) were purchased from Proteintech. After incubation with primary antibody, secondary antibody followed. Finally, the results were visualized by Tanon system.

Cells were lysed at 4 °C using RIPA and the followed procures were based on Co-IP kit (Pierce, Rockford, USA) according to manufacturer’s instructions. In brief, a total of 300 μg proteins were incubated overnight with specific primary antibodies β-catenin and NICD at 4 °C. Immune complexes were precipitated with protein A/G Sepharose beads and next were examined by western blot.

RNA was extracted from GC cells treated with luteolin or without, and then RNA was reversed to cDNA. Primers for Notch-1, forward GCTTGTGGTAGCAAGGAAGC (20b), reverse CCACATTCAAGTGGCTGATG (20b); Hes-1, forward ACACGACACCGGATAAACCAA (21b), reverse CGAGTGCGCACCTCGGTA (18b); Cyclin-D1, forward GGGTGGGTTGGAAATGAACT (20b), reverse CTTCCTCTCCAAAATGCCAG (20b). RT-PCR was performed using SYBR-green according to manufacturer’s instructions.

BALB/c male nude mice (Institute of Zoology, China Academy of Sciences) were used to evaluate the role of luteolin in tumor growth in vivo. Nude mice received humane care and the study protocols were carried out according to a protocol approved by the Institutional Animal Care and Use Committee (IACUC) at Shanghai Jiao Tong University, Shanghai, China. Tumor nodules were measured every week, and was calculated using formula: tumorous volume = (width² × length)/2. Mice were killed at 4 weeks after injection and then tumors were weighed and fixed for immunohistochemistry staining (IHC).

For IHC, sections staining was performed according to the DAKO protocol, using primary antibody (1:200 dilutions) Notch1, β-catenin and Ki-67. The tunel assay was performed using In situ cell death detection kit (Roche). Blue represents nucleus, green represents death cells.

Differences between experimental groups were assessed by the Student’s t test or one-way ANOVA. Student’s t test was used to examine the statistical differences between the two groups. The significant inhibited effect on cell growth by luteolin was observed at 4th and 5th day after luteolin treatment. The results of 4th and 5th day were compared to that in their control groups using the Student’s t test in Fig. 1a, as well as in Fig. 1b. Survival was analyzed with the Kaplan–Meyer method comparing survival curves by log-rank test. Data are shown as mean ± SD. A two-tailed value of P < 0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS 19.0 software (SPSS Inc).