18 β-glycyrrhetinic acid (98% of purity, determined by HPLC) was purchased from Aladdin Bio-Chem Technology Corp. (Shanghai, China) and dissolved in dimethyl sulfoxide (DMSO). Antibodies used in western blotting against MMP1, MMP2, MMP3 and MMP9 were bought from Proteintech Group Inc. (Chicago, IL, USA). Antibodies against Poly(ADP-ribose) polymerase (PARP), phospho-PI3K p85/p55, phospho-AKT, AKT, phospho-STAT3, STAT3, phospho-p38, p38, phospho-Erk1/2, Erk1/2, phospho-SAPK/JNK, SAPK/JNK, phospho-NF-κB p65, NF-κB p65 and β-actin were obtained from Cell Signaling Technology (Danvers, MA, USA).

The human CRC cell lines LoVo, SW480 and SW620 and a normal human colon mucosal epithelial cell line (NCM460) were obtained from the American Type Culture Collection and cultured in RPMI-1640 medium complemented with 10% heat-inactivated fetal bovine serum (FBS), penicillin/streptomycin (100 U/ml) and L-glutamine (2 nM). Cells were cultured at 37°C with 5% CO₂ in a humid incubator. All the cell culture reagents were purchased from Invitrogen (Shanghai, China).

Cells were plated in 96-well plate at a density of 1×10³ cells/well and cultured in complete medium supplemented with varied concentrations of GRA (0, 12.5, 25 and 50 µM). After culturing for 24, 48, 72 and 96 h, cell growth was analyzed by CCK-8 assay. A total of 10 µl of CCK-8 solution reagent (Beyotime Institute of Biotechnology, Haimen, China) was added to each well and cultured for another 1 h. Absorbance at 450 nm was measured on an ultra-microplate reader (EMax; Molecular Devices, Sunnyvale, CA, USA). The growth inhibition rate was calculated with the following formula: (Ab. control − Ab. treated)/Ab. control × 100. The IC₅₀ was taken as the concentration at which it caused 50% inhibition of cell proliferation (50% reduction in the absorbance value in the treated cells, in respect to control). IC₅₀ was calculated by curve fitting statistical analysis by GraphPad Prism (Graphpad Software, Inc., San Diego, CA, USA).

To determine apoptosis, the cells were stained with the Annexin V apoptosis kit according to the manufacturer's instructions (Lianke Biotech, Co., Ltd., Hangzhou, China). In brief, 1×10⁵ cells in 6-well plate were treated with or without GRA for 48 h before cells were trypsinized, washed with ice-cold PBS, resuspended in 500 µl binding buffer and incubated with 5 µl Annexin V-FITC and 10 µl propidium iodide (PI) working solution at room temperature for 5 min in the dark. Stained cells were then analyzed by flow cytometry (BD FACScan; BD Biosciences, San Jose, CA, USA).

Cells in 6-well plate at a density of 2×10⁵ were treated with GRA at different concentrations for 24 h and total RNA was extracted using Ultra-Pure RNA kit (Cwbiotech, Beijing, China). After DNase treatment, cDNA was synthesized by the HiFiScript cDNA Synthesis kit (Cwbiotech). Reverse transcriptase PCR was performed on 2720 Thermal Cycler (Applied Biosystems, Foster City, CA, USA). Quantitative PCR was then performed with 1 µl of cDNA using the iTaq Universal SYBR-Green Supermix (Bio-Rad Laboratories) on CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories). The specific primers for survivin and β-actin were as follows: survivin: 5′-GATTTGAATCGCGGGACCCGTTG-3′ (forward) and 5′-TCAAGACAAAACAGGAGCACAGT-3′ (reverse); β-actin: 5′-GTATCCTGACCCTGAAGTACC-3′ (forward) and 5′-TGAAGGTCTCAAACATGATCT-3′ (reverse). qPCR results were analyzed with Bio-Rad CFX Manager (Bio-Rad Laboratories). A relative quantification delta CT strategy was used to analyze the qPCR data. The calculated threshold cycle was normalized to the value of internal β-actin amplified from the same cDNA and the fold-change was calculated as referenced to control.

Transwell migration and invasion assays were carried out using cell culture chambers with 8-µm pore filter (Corning Costar Corp., Cambridge, MA, USA). For invasion assay, the upper surfaces of the membranes were coated with 50 µl Matrigel (BD Biosciences) 6 h before cells were seeded. Cells/well (2×10⁵) in 200 µl serum-free RPMI-1640 were added into the upper chamber of 24-well plates supplied with or without 50 µM GRA. Culture media containing 20% FBS were added in the lower chamber. After 48 h culturing, the cells in the upper chamber were removed by wiping the upper side of the membrane with a cotton swab and cells in the lower surface of the membrane were fixed with ice-cold methanol for 5 min and stained with crystal violet in 20% ethanol overnight. Images were acquired by light microscopy and the cells invaded to the underside of the membrane were quantitated by cell counting in five predetermined fields. To determine whether the invasion inhibition was caused by the decreasing of cells number, the number of the LoVo cells after invasion was recorded in both sides of the chamber.

The methods used for the migration assay were almost the same as the invasion assay described above, except for no Matrigel coating in the upper surfaces of the membranes.

Cells were seeded in 24-well plate at an initial density of 1.5×10⁵/well and allowed to grow to 90% confluency in complete medium. Cells were wounded by a plastic tip and then washed three times with phosphate-buffered saline (PBS) to remove the cell debris and incubated for 24 h in 5% FBS medium containing different concentrations of GRA. Images of the wound morphology were acquired by light microscopy.

Four- to five-week-old female BALB/c nude mice, weighing between 16–18 g were purchased from Shanghai Laboratory Animal Company (SLAC; Shanghai, China). Mice were maintained in the animal facility at the Zhejiang University, China. The animal experiments were approved by Animal Care and Use Committee of Zhejiang University. LoVo cells (5×10⁶) resuspended in 100 µl PBS were injected subcutaneously in the right flanks of the nude mice. When tumors reached an average volume of 100 mm³, the mice were randomly divided into two groups. Mice in GRA-treated group were intraperitoneally injected with GRA at 40 mg/kg once every 2 days in 1% DMSO in PBS. The mice in control group received the same volume of PBS. Tumor size and body weight were measured by periodic measurements with calipers every 5 days until day 45 after tumor cell implantation at which time some mice began to die. Tumor volume was calculated using the following formulae: V = (length × width²/2). The survival was recorded and the mice were sacrificed at 55 days after tumor cell implantation.

Cells plated in 6-well plate at a density of 2×10⁵ were treated with GRA at different concentrations for 24 h or cells were treated with the indicated concentrations of GRA for 2, 4, 6 and 12 h and total cellular proteins were extracted for western blotting. Cells were lysed in 1X RIPA buffer (Cell Signaling Technology, Danvers, MA, USA) containing 1 mM phenylmethylsulfonyl fluoride (PMSF; Cell Signaling Technology) and protease inhibitor cocktail. Protein concentration was determined by BCA assay (Cwbiotech). Total proteins (20–50 µg) were separated by SDS-PAGE and blotted onto to PVDF membranes. The blots were probed by antibodies against different proteins including MMP1, MMP2, MMP3, MMP9, PARP, phospho-PI3K p85/p55, phospho-AKT, AKT, phospho-STAT3, STAT3, phospho-p38, p38, phospho-Erk1/2, Erk1/2, phospho-SAPK/JNK, SAPK/JNK, phospho-NF-κB p65, NF-κB p65 and β-actin. Then the membranes were probed with horseradish-peroxidase-conjugated goat anti-rabbit secondary antibody (Lianke) followed by the detection of signal with FluorChem E System (ProteinSimple, Santa Clara, CA, USA). Western blot bands were quantified by ImageJ and normalized to their respective total protein and compared with Controls (no GRA treatment).

All assays were performed in triplicate and experiments were repeated three times. Data are presented as mean ± SEM. Significance of differences between the two independent groups was determined by the Student's t-test. Survival curves were drawn by the Kaplan-Meier method and analyzed using the log-rank test. Statistical calculations were performed using Graph Pad Prism. A P<0.05 was considered statistically significant.

To evaluate the general inhibitory effect of GRA on colorectal cancer cell growth, we first examined the effect of GRA on three colorectal cancer cell lines LoVo, SW620 and SW480 and the normal colon mucosal epithelial cell line NCM460. Cells were treated with GRA at varying concentrations ranging from 12.5 to 50 µM, as we had previously determined that the IC₅₀ value was >50 µM in all three colorectal cancer lines and the normal colon mucosal epithelial cell line. As shown in Fig. 2A, comparing with the NCM460, the proliferation level of LoVo, SW620 and SW480 was significantly inhibited. IC₅₀ value of GRA for NCM460, LoVo, SW620 and SW480 cells was calculated. As detailed in Table I, the IC₅₀ value of GRA for NCM460 cells was 866, 144.8 and 141.3 µM at 24, 48 and 72 h, respectively, which is much higher than that for SW620 (360.9, 82.28 and 60.19 µM), SW480 (153.8, 88.23 and 34.64 µM) or LoVo (99.48, 71.23 and 43.23 µM) cells. The inhibitory effect was both dose- and time-dependent. Furthermore, to evaluate whether GRA can inhibit tumorigenicity of colorectal cancer in vivo, nude mice were implanted subcutaneously with LoVo cells for 10 days followed by intraperitoneal injection of 40 mg/kg GRA or vehicle (DMSO) once every 2 days for 1 month. As seen in Fig. 2B, the tumor volume was significantly larger in the control group than that in GRA-treated group. Consistently, the body weight of the GRA-treated group decreased more slowly than the DMSO-treated group. Survival curve till 55 days after tumor cells implanted is shown in Fig. 2C. The survival periods of LoVo xenograft mice in GRA-treated group were significantly longer than that of DMSO-treated mice (P<0.05; Kaplan-Meier). These results show that GRA can effectively inhibit colorectal cancer development both in vitro and in vivo.

To investigate whether GRA inhibits cells proliferation by regulating apoptosis, Annexin V-FITC staining was performed in LoVo, SW620 and SW480 cells after treatment with GRA at 25 and 50 µM for 48 h. As shown in Fig. 3A, flow cytometric analysis showed a significant increase in the apoptotic population of the cells treated with GRA compared with that of the cells without treatment. Anti-apoptotic protein survivin is expressed abundantly in human colorectal cancer and its expression indicates poor prognosis and a pro-metastatic phenotype (19). Inhibition of survivin expression is associated with the growth suppression and apoptosis induction by antitumor agents (20), thus, we next investigated whether the pro-apoptotic effect of GRA was mediated by regulation of survivin expression. We detected the survivin expression in LoVo, SW620 and SW480 cells after treatment with GRA by RT-qPCR. As shown in Fig. 3B, GRA induced a dose-dependent reduction of survivin expression, of which the most significant reduction was at 50 µM, suggesting a potential role for survivin in GRA-induced colorectal cancer cells death. Poly-ADP-ribose polymerase (PARP), one essential substrate cleaved by both caspase-3 and -7, is an abundant DNA-binding enzyme. The presence of cleaved PARP-1 is one of the most potent indicators for the detection of apoptosis in many cell types (21,22). Thus, we also tested it and found that the active form of PARP was slightly induced upon GRA treatment as shown in Fig. 3C. These data indicated that GRA suppresses colorectal cancer development partly through promoting the apoptosis process.

To determine whether GRA was involved in the regulation of colorectal cancer metastasis, we performed Transwell assay on LoVo, SW620 and SW480 cells. Cells plated in the upper chamber in serum-free medium were treated with 50 µM GRA for 48 h. The invasive and migratory abilities were evaluated based on the numbers of cells passed through the membrane. As shown in Fig. 4A, in control group, the number of invaded LoVo, SW620, SW480 cells was 75.5±5.838, 106.5±12.98 and 122.3±11.74, but in GRA treated cells, the numbers decreased to 49.50±3.663, 41.75±4.385 and 47.25±3.966, respectively (P<0.05). The migration assay (Fig. 4B) showed similar results. With the number of migrated LoVo, SW620, SW480 cell being 222.8±19.44, 54.25±3.425 and 56.50±2.723, the treatment of GRA decreased the numbers to 110±4.435, 33.5±2.598 and 35.25±3.250, respectively (P<0.05). Taken together, the results indicated that GRA inhibits invasive and migratory ability of colorectal cancer cells. In addition, we also calculated the number of the LoVo cells after invasion in both sides of the chamber. As shown in Fig. 4C, GRA treatment significantly decreased LoVo cell invasion while had no significant effect on the cell numbers in the upper side of the chamber, indicating that the invasion and migration inhibitory effect of GRA was not caused by decrease of the cell number. Consistent with the invasion and migration results, GRA treatment induced a slower closing of scratch wounds in LoVo cells, indicating that GRA efficiently inhibited motility activity of LoVo cells (Fig. 4D). We next determined the changes of the expression of several matrix metalloproteinases (MMPs), which are strongly related with tumor metastasis and overexpressed in human colorectal cancers (23). MMP1, MMP2, MMP3 and MMP9 were significantly decreased after GRA treatment in LoVo, SW620 and SW480 cells (Fig. 4E). These data together clearly show that GRA could impede the invasion and migration of colorectal cancer cells.

Previous studies have shown multiple signaling pathways such as PI3K/AKT, p-STAT3, MAPKs and NF-κB are activated in colorectal cancers, and activation of these signaling pathways affect the proliferation, invasion and migration of tumor cells. Thus, we investigated whether GRA inhibited proliferation, invasion, and migration of colorectal cancer cells by altering the activation of these signaling pathways. LoVo, SW620 and SW480 cells were treated with GRA for 24 h before proteins were extracted for western blot analysis. As shown in Fig. 5, GRA dose-dependently reduced the protein expression of p-PI3K, p-AKT, p-STAT3, p-JNK, p-p38 and p-NF-κB whereas had no significant effect on the expression of p-Erk. Time course studies were also performed at 2, 4, 6 and 12 h after GRA treatment. As shown in Fig. 6, the phosphorylation levels of PI3K p85/p55 and STAT3 decreased as early as 2 h after GRA treatment, suggesting they are the main signaling pathways regulated by GRA. The phosphorylation levels of p-AKT, p-JNK, p-p38 and NF-κB p65 were also decreased, but in later time-points, suggesting these effects may be secondary to the drug treatment.