Human CRC HT-29, SW480 and HCT116 cell lines were obtained from Stem Cell Bank of Chinese Academy of Sciences (Shanghai, China) and were cultured in complete RPMI-1640 medium (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) with 10% fetal bovine serum (Thermo Fisher Scientific, Inc., Waltham, MA, USA), penicillin (100 U/ml) and streptomycin (100 µg/ml; both Sigma-Aldrich; Merck KGaA). Cells were maintained at 37°C with 5% CO₂ in a humidified cell culture incubator (Sanyo, Tokyo, Janpan). For all experiments, five biological groups were designed according to the concentrations of GA (0, 0.33, 1, 3.3 and 10 µM), which was purchased from Sigma-Aldrich; Merck KGaA. For each indicated time points (24, 48, and 72 h), cells were sampled for the following assay.

An MTT assay kit (Beyotime Institute of Biotechnology, Shanghai, China) was used to measure cell growth for indicated time points following the protocol of manufacturer. Briefly, HT-29 cells at the logarithmic growth phase were seeded into 96-well plate at 5×10³/well. 200 µl GA with dose escalation (0, 0.33, 1, 3.3 and 10 µM) was added into each well in triplicate when cells were totally adhered at 24 h. Cells were cultured at 37°C (5% CO₂) and sampled at 24, 48 and 72 h. MTT (5 mg/ml, Sigma-Aldrich; Merck KGaA) solution was added (20 µl/well) at each sampling time point and cells were further incubated for additional 4 h in a cell incubator. At the end of incubation, the supernatants were removed with a pipette. Before reading under the microplate reader, 150 µl DMSO (Sigma-Aldrich; Merck KGaA) was added to each well. The proliferation rate was calculated with the absorbance value (OD) at a wavelength of 480 nm. The cell viability percentage was measured with the following formula: [(drug treated group/control group) ×100]. Each assay was performed triplicate and the results are presented as the mean ± SD.

After incubation with GA (0, 1 or 3.3 µM) for 24, 48 or 72 h, CRC cells were collected and total RNA was isolated from these cells using a PicoPure™ RNA Isolation kit (Arcturus, Sunnyvale, CA, USA) according to the manufacturer's instructions. Subsequently, 1 µg isolated RNA was transcribed into cDNA using SuperScript III RNase H Reverse Transcriptase (Thermo Fisher Scientific, Inc.). Following the reverse transcription reactions, amplification was performed using a Vii™ 7 system (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instructions. Transcript quantities were compared by relative Ct numbers, and the miR-21 expression level was normalized to an endogenous reference gene, GAPDH, and then the relative miR-21 mRNA levels to control sample was calculated using the 2^−ΔΔCt method. Both primers for miR-21and GAPDH were purchased from Thermo Fisher Scientific, Inc.

Apoptotic cells were analyzed according to the previously described method (34). Briefly, HT-29 cells were exposed to 3.3 µM GA, GA combined with miR-21 mimics or vehicle for 72 h. Subsequently, these three groups of HT-29 cells were harvested and resuspended in PBS. Apoptotic cells were identified with dual-staining of Annexin V-FITC and propidium iodide (PI; Thermo Fisher Scientific, Inc.). For each group, the experiments were performed in triplicate.

For the wound healing assay, the procedure was performed according to the previous reports by Ni et al (35). HT-29 cells were seeded into 6-well plates (6×10⁴ cells/well) and treated with GA, GA combined with miR-21 mimics or a vehicle. When the cells grew to 90% confluence, vertical scratches were induced down through the monolayer cell surfaces using 1,000 µl pipette tips. The media and cell debris were carefully aspirated before further culture with serum-free RPMI-1640 medium at 37°C. At 24 h, images of each sample were captured under a microscope at magnification, ×100 (Leica DM500; Leica Buffalo Grove, IL, USA). The images were used to analyze the distance of one side of the wound to the other side using a scale bar.

MiR-21 mimics, scramble, wild-type and mutant 3′UTR PTEN vectors were provided by Ambion (Thermo Fisher Scientific, Inc.). Vectors carried the PTEN sequence, which contained the predicted miR-21 binding sites with wild-type or mutant 3′UTR. 5×10³ HT-29 cells were plated into 24-well plates, which were then transfected with PTEN wild-type or mutant vector for 4 h using Lipofectamine^® 2000 transfection reagent (Thermo Fisher Scientific, Inc.). Subsequently, the cells were treated with 100 nM miR-21 mimics or scramble miRNA. 48 h later, the cells were lysed using 0.2% trypsin at 37°C and a dual luciferase assay kit was used to detect the luciferase activities (Promega Corporation, Madison, WI, USA).

Following treatment with GA, GA combined with miR-21 mimics or vehicle for 72 h, HT-29 cells were collected and lysed in RIPA buffer (Sigma-Aldrich; Merck KGaA). The total protein concentration was quantified using a BCA Protein Assay kit (Beyotime Institute of Biotechnology), and 20 µg protein from each group was separated with 10% Tris-SDS gel. After the electrophoresis, the gel was electro-transferred onto polyvinylidene fluoride membranes (Thermo Fisher Scientific, Inc.). For the subsequent blocking, washing and antibody incubation at room temperature, an iBind kit (Thermo Fisher Scientific, Inc.) was used according the manufacturers' instructions. The following primary antibodies were used: anti-PTEN (1:250), anti-PI3K (1:300), anti-p-Erk (1:300), anti-matrix metalloproteinase (MMP2; 1:300), anti-MMP9 (1:300; all Cell Signaling Technology, Inc., Danvers, MA, USA) and anti-GAPDH (1:1,000; Santa Cruz Biotechnology, Inc., CA, USA). Horseradish peroxidase-conjugated secondary antibodies (1:3,000; anti-rabbit or anti-goat IgG; Santa Cruz Biotechnology, Inc.) were used. The luminescent signal was detected by adding super sensitive regent (Thermo Fisher Scientific, Inc.) and quantified using the Image Lab™ system from Bio-Rad Laboratories.

The online protocol published by Justus et al (36) was used in order to validate the effect of GA on HT-29 invasion. Briefly, 100 µl HT-29 cells (1×10⁴ cells/well) mixed with 0.33 µM GA, GA combined with miR-21 mimics or vehicle were added to the upper chamber of Transwell chambers (Corning Life Science, Tewksbury, MA, USA) and 100 µl RPMI-1640 medium with 30% FBS was added to the lower chamber. Following incubation for 72 h at 37°C, cells on the upper surface of the microporous membrane were carefully removed with cotton swabs, whereas cells on the lower surface of the membrane were fixed and stained with crystal violet. The cells in five selected views were counted under a microscope (Leica DM500; Leica) at magnification, ×100. The average sum of the cells was used to calculate the invasion rate.

All experiments were performed in triplicate and all data are presented as the mean ± SD. One-way ANOVA followed by Dunnett's test was used to determine the difference between groups with P<0.05 being considered to indicate a statistically significant difference (*P<0.05, **P<0.01).

The viability of HT-29 cells was detected at indicated time points with dose escalation GA (Fig. 1A) from 0.33 to 10 µM using the MTT assay. As shown in Fig. 1B, GA inhibited the growth of HT-29 cells in a time- and dose-dependent manner. GA at all doses significantly inhibited HT-29 cell proliferation compared with vehicle at 72 h (P<0.01). miR-21 was mostly found to be overexpressed in CRC. In order to explore the effect of GA on miR-21 expression, the cells treated with 1 or 3.3 µM GA for 24, 48 and 72 h were subject to qPCR. The results indicated that GA downregulated miR-21 RNA expression at both doses, even at 24 h (Fig. 1C; P<0.05). Moreover, we used the other two CRC cell lines, SW480 and HCT116, to investigate the involvement of miR-21 on the effects of GA. The results were consistent with the data obtained using HT-29 cells (Fig. 1D-F). All these results indicated that GA exerted anti-proliferation effects on HT-29 cancer cells through downregulation of miR-21 expression.

To further investigate how GA inhibited HT-29 cancer cell growth and whether miR-21 was the effector of GA, apoptotic cells were detected with FACS by double staining with Annexin V and PI. In addition to the vehicle and GA (3.3 µM) treatment groups, the GA with miR-21 mimics group was added to determine whether miR-21mimics could interfere with the antitumor effect of GA. The results in Fig. 2A and B demonstrated that the majority of HT-29 cells were subject to apoptosis following GA treatment compared with the control group (P<0.01). Moreover, when miR-21 mimics were added to the GA 3.3 µM group, the apoptotic rate of the HT-29 cells was reduced to a lower level, which was similar to that of the control group and was significantly less than that of the GA 3.3 µM group (Fig. 2B; P<0.01). These results indicated that GA inhibited HT-29 cancer cell growth by inducing cell apoptosis via downregulation of miR-21 expression.

In order to validate the effect of GA on HT-29 cell migration, a wound healing assay was performed. The results in Fig. 3A indicated that cell migration was significantly inhibited by 0.33 µM GA compared with control group. Additionally, in the miR-21 mimics treated group, the wound was similar to that in the control group, which further confirmed that the anti-migration effect of GA was exerted through miR-21. Quantification of the migration activity revealed significant differences among the control, GA and GA with miR-21 groups (Fig. 3B, P<0.01, P<0.05). Thus, this result suggested that GA inhibited HT-29 cell migration via blocking miR-21 activity.

In the present study, miR-21 gene expression was inhibited by GA (Fig. 1B) and, as Zhang et al (15) had previously reported that miR-21 was an onco-miRNA that downregulated the tumor suppressor gene, PTEN in gastric cancer, we needed to determine whether GA could affect PTEN expression in CRC. From the luciferase assay, we found a statistically significant decrease in firefly luciferase activity in the cells transfected with miR-21 mimics and wild-type 3′UTR PTEN vectors (Fig. 4A). In addition, GA-induced cell growth inhibition was reversed by miR-21 mimics (Fig. 4B). The western blot results further confirmed that GA significantly increased the PTEN protein level and miR-21 mimics suppressed the enhancement induced by GA (Fig. 4C and D). Moreover, PTEN is a suppressor of the PI3K/Akt pathway and therefore, the enhanced expression of PTEN induced by GA would have definitely blocked the PI3K/Akt signaling pathway. The western blot results shown in Fig. 4C, E and F indicated that GA significantly decreased PI3K and p-Akt protein expression (P<0.01). All these findings demonstrated that GA showed antitumor activity through targeting miR-21 and, in turn, blocking of the PI3K/Akt signaling pathway.

A Transwell assay was used in order to further investigate the biological function of GA on HT-29 cell invasion. The invasion of HT-29 cells treated with 0.33 µM GA was markedly less than that in the control group under the microscope view (Fig. 5A). The invaded cell number was quantified and is shown in Fig. 5B, which reveals a significantly smaller number of invading cells compared with the vehicle group. Moreover, in the GA and miR-21 mimics treated group, the cell number was comparable with that in the control group (Fig. 5B). As MMP2 and MMP-9 are closely associated with cell migration and invasion, the expression of these factors was also detected. The results revealed that MMP-2 and MMP-9 expression was significantly decreased following treatment with GA (Fig. 5C-E). These results indicated that GA inhibited the activity of MMP2 and MMP-9 and thus showed its anti-invasion activity in HT-29 cancer cells.