HT-29, SW480, SW620 and LoVo human CRC cell lines were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The normal human mucosal epithelial cell line NCM460 was purchased from the National Institute of Cells (Shanghai, China). All cells were maintained in RPMI-1640 medium (HyClone; GE Healthcare, Chicago, IL, USA) supplemented with 10% fetal bovine serum (FBS; Gibco™, Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 1% penicillin-streptomycin in a humidified incubator containing 5% CO₂ at 37°C. Luteolin was purchased from Sigma-Aldrich/Merck (Shanghai, China) and dissolved in dimethyl sulfoxide (DMSO).

We collected a total of 50 cases of CRC and matched normal tissues from patients who underwent surgery at Hangzhou Hospital of Traditional Chinese Medicine from May 2017 to May 2018. All postoperative pathology was determined to be colorectal adenocarcinoma. The following patient characteristics were recorded: age, sex, histologic differentiation, TNM stage and serum levels of carcinoembryonic antigen (CEA) before surgery and all the data are listed in Table I. Approval of the present study was obtained from the Ethics Committee of Hangzhou Hospital of Traditional Chinese Medicine and informed consent was acquired from all the patients.

Cell proliferation ability was detected by Cell Counting Kit-8 (CCK-8) assay kit (Beyotime Institute of Biotechnology, Shanghai, China). In brief, CRC cells were plated in a 96-well plate at the initial density of 1,000 cells/well and cultured in medium without or with luteolin at concentrations of 0, 10, 50 and 100 µM for 24, 48 and 72 h. CCK-8 solutions were added into the cell cultures 2 h before harvesting. The optical density at 450 nm was measured using an ultra-microplate reader (EMax; Molecular Devices, Sunnyvale, CA, USA).

Thirty-nine BALB/c nude mice (17–20 g) were utilized to evaluate the effects of luteolin on tumor growth and tumor metastasis in vivo. Five- to six-week-old female BABL/c nude mice were purchased from Shanghai Slac Animal (Shanghai, China) and maintained in the animal facility at the Zhejiang University (Hangzhou, China). Mice were provided with water and food ad libitum and kept under standard conditions (temperature 24±2°C, humidity, 50–70 %, 12-h light/dark cycle). The study received ethical approval from the Animal Care and Use Committee of Zhejiang University, and experiments and animal care were performed according to the approved protocols.

For the tumor growth assay, HT-29 cells (1×10⁶) were subcutaneously injected into the right flanks of the nude mice. When the volumes of xenograft tumors reached an average of 100 mm³, the mice were randomly divided into two groups: the control group and the luteolin group with 10 mice in each group (20 mice in total). Mice received luteolin were intragastrically administered with luteolin at 100 mg/kg every 2 days for 30 days. The control group mice received the same volume of phosphate-buffered saline (PBS). Tumor measurements were performed with a caliper by measuring the largest diameter and its perpendicular length every 3 days. According to the rules of our ethics committee, humane endpoints set for the tumor experiments were as follows: the tumor weight was >10% of the original body weight or the average tumor diameter was >20 mm in mice. Thus, when the tumor almost reached 20 mm, the experiment was terminated. The mice were sacrificed by cervical dislocation, and tumors were harvested and weighed on day 30. Tumor volume was calculated using the following formula: V = ½ × (length × width²). To note, we repeated this experiment for the same setting for three times (8 mice for the first experiment, and (6 mice for the other two experiments).

For liver metastasis assay, mice (n=19) were anesthetized with 75 mg/kg pentobarbital by intraperitoneal injection. Mice were fixed and then a small left abdominal flank incision (1 cm) was made and the spleen was exteriorized for the intrasplenic injection. HT-29 cells (5×10⁶) suspended in 100 µl ice-cold PBS were injected into the spleen with a 27-gauge needle. A sterile cotton was held over the site of injection for 30 sec to prevent tumor cell leakage and bleeding. The injected spleen was returned to the abdomen and the wound was sutured with 5-0 black silk. Mice receiving luteolin were intragastrically administered with luteolin at 100 mg/kg every 2 days for 30 days. The control mice received the equal volume of PBS. After 3 weeks, the mice were sacrificed by cervical dislocation, spleen and liver were removed and observed. The liver metastatic tumors on the surface were calculated and H&E staining was performed. H&E staining was performed at Servicebio Co., Ltd. (Wuhan, China). To note, we also repeated this experiment three times (6 mice for the first two experiments and 7 mice for the third trial).

HT-29 cells were seeded in 24-well plate until reaching 80% confluency in the complete medium. The monolayer was wounded using a sterile 200-µl micropipette tip. The cells were washed with PBS three times to rinse off the detached cells. Cells were incubated for 48 h in 2% FBS medium containing the indicated concentrations of luteolin. Images of the wound morphology were acquired under a light microscopy at ×100 magnification.

Cell migration and invasion potential were assessed using 8-µm Transwell chambers (Corning Costar, Corning, NY, USA). Briefly, 8×10⁴ CRC cells suspended in 200 µl serum-free medium treated with luteolin or not were seeded in the upper chambers with the lower chamber filled with 600 µl of medium supplemented with 20% FBS. For the invasion assay, the upper surfaces of the membranes were coated with 50 µl Matrigel (BD Biosciences, Franklin Lakes, NJ, USA) 6 h before cells were seeded. Cells were allowed to migrate for 36 h before being fixed with ice-cold methanol and stained with 0.1% crystal violet for 30 min. Cells that passed through the membrane were photographed and counted with an Olympus CKX41 light microscope (Olympus, Shanghai, China; magnification, ×100).

Cells or CRC tissues for protein extraction were lysed with RIPA buffer (Beyotime Institute of Biotechnology). Protein concentration was determined by BCA assay (Cwbiotech, Beijing, China). Total proteins (30 µg) were separated by 10% SDS-PAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were incubated with primary antibodies against human MMP-16 (cat. no. E-AB-32127), MMP-2 (cat. no. 10373-2-AP), MMP-3 (cat. no. 17873-1-AP), MMP-9 (cat. no. 10375-2-AP), PTN (cat. no. 27117-1-AP) (all diluted 1:1,000) and GAPDH (cat. no. 60004-1-Ig) (dilution 1:5,000) at 4°C overnight. Subsequently, the membranes were incubated with the peroxidase (HRP)-conjugated goat anti-rabbit (dilution 1:5,000; cat. no. SA00001-2) or goat anti-mouse (dilution 1:5,000; cat. no. SA00001-1) secondary antibody at room temperature for 2 h. MMP-16 was purchased from Elabscience (Shanghai, China). All the other antibodies were purchased from Proteintech Group Inc. (Chicago, IL, USA). The protein bands were detected with FluorChem E System (ProteinSimple, Santa Clara, CA, USA).

Total RNA was extracted from the cells or the surgically resected CRC and adjacent non-tumor tissues using TRIzol reagent (Thermo Fisher Scientific, Inc.). Then, 1 µg of total RNA was reverse transcribed using the HiFiScript cDNA Synthesis kit or miRNA cDNA Synthesis kit (Cwbiotech) according to the manufacturer's instructions. The PTN expression levels were determined with UltraSYBR Mixture (Cwbiotech) and the miR-384 levels were determined with miRNA qPCR Assay Kit (Cwbiotech) on CFX96 Touch Real-Time PCR Detection system (Bio-Rad Laboratories, Hercules, CA, USA). The miR-384 and U6 primers were synthesized by Shanghai GenePharma Co., Ltd. (Shanghai, China), and the PTN and β-actin primers were synthesized by Thermo Fisher Scientific, Inc. The primers were as follows: miR-384 forward, 5′-TGTTAAATCAGGAATTTTAA-3′ and reverse, 5′-TGTTACAGGCATTATGAA-3′; U6 forward, 5′-CTCGTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′; PTN forward, 5′-ACCAGTGAGTCATCCGTCCA-3′ and reverse, 5′-TGCAAATTTTCGACGCTGCT-3′; β-actin forward, 5′-GTATCCTGACCCTGAAGTACC-3′ and reverse, 5′-TGAAGGTCTCAAACATGATCT-3′. The thermocycling conditions were as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C. Relative fold-change in expression of miR-384 and PTN was normalized to U6 and β-actin expression, respectively. The relative expression was analyzed using the 2^−∆∆Cq (27) method.

miR-384 mimic, miR-384 inhibitor and its relative control miRNAs were synthesized by Shanghai GenePharma Co., Ltd. The sequences are as follows (28): miR-384 mimic: 5′-AUUCCUAGAAAUUGUUCAUA-3′, miR-384 inhibitor: 5′-UAUGAACAAUUUCUAGGAAU-3′, miR-384 mimic NC: 5′-UUCUCCGAACGUGUCACGU-3′, miR-384 inhibitor NC: 5′-CAGUACUUUUGUGUAGUACAA-3′. Full length clone DNA of human pleiotrophin was obtained from Sino Biological Inc. (Shanghai, China). Transfections were performed using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions.

Target genes for miR-384 were predicted using miRDB (http://www.mirdb.org) and TargetScan v.5.1 (http://www.targetscan.org/). From the TargetScan, we found that PTN ranked the top 2 genes that was the targets of miR-384. We then confirmed it in the miRDR database and found that PTN ranked 86. Thus, we chose PTN as one candidate of miR-384 target genes.

Wild-type (WT) or the mutated type (MUT) of PTN-3′-UTR containing miR-384 binding sites was synthesized by Shanghai GenePharma Co., Ltd., and cloned downstream of the luciferase reporter gene. HT-29 cells were seeded into 24-well plates (5×10⁴ cells/well) one day before and then co-transfected with the WT PTN-3′-UTR or the MUT PTN-3′-UTR vectors (500 ng) and the miR-384 mimics or negative controls with Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific, Inc.). After 48 h of transfection, cells were harvested, and the luciferase activity was detected with the Dual-Luciferase Reporter Assay system (Promega Corporation, Madison, WI, USA).

Oncomine database (http://www.oncomine.org) was selected to search the PTN expression levels between the CRC and normal groups.

Data are expressed as means ± SEM. Statistical significance was determined by Student's t-test or one-way analysis of variance (ANOVA) followed by Dunnett's test. The relationship between the PTN expression and the CRC clinical features was analyzed with the Chi-square test. P-value <0.05 was considered as indicative of statistical significance.

In order to explore the functions of luteolin in CRC, we first detected HT-29 and SW480 cells proliferation after luteolin treatment by CCK-8 assay. The results showed that luteolin had no effects on CRC cells proliferation (Fig. 1A). We then examined cells migration and invasion by Transwell and wound healing assays. As shown in Fig. 1B, we observed that luteolin-treated HT-29 and SW620 cells displayed significantly reduced cells migration and invasion. Consistently, luteolin treatment induced a slower closing of scratch wounds in the HT-29 cells (Fig. 1C), indicating that luteolin inhibited cells migratory ability. We also detected the expressions of several matrix metalloproteinases (MMPs), which are strongly related with tumor migration and invasion and overexpressed in human CRCs. As shown in Fig. 1D, luteolin obviously downregulated MMP-2, MMP-3, MM-9 and MMP-16 expressions.

We then performed the in vivo assay to observe the tumor growth and tumor metastasis by injecting HT-29 cells into the nude mice subcutaneously and into the spleen, respectively. In accordance with the in vitro results, intake of luteolin did not impact the tumor growth as evidenced by no differences in tumor volume and tumor weight (Fig. 2A). However, luteolin significantly inhibited HT-29 cells metastasis from the spleen to the liver (Fig. 2B). The number of metastatic nodules in the liver was significantly decreased. Taken together, all these data suggest that luteolin plays a significant role in suppressing the migration and invasion of CRC cells both in vitro and in vivo while has no effects on cells proliferation.

To identify the specific miRNA involved in the luteolin-induced inhibition of CRC cells migration and invasion, the expression level of miR-384 was determined in CRC cells treated with luteolin. It was found that the expression of miR-384 was significantly elevated in the CRC cell lines following treatment with luteolin in a dose-dependent manner compared with no-luteolin stimulation (Fig. 3A). Then, we constructed the miR-384 inhibitor and transfected into the CRC cells before luteolin treatment. As shown in Fig. 3B, the miR-384 inhibitor partially reversed the inhibition of cells migration and invasion induced by luteolin. Simultaneously, the expression levels of MMP-2, MMP-3, MMP-9 and MMP-16 which were decreased by luteolin treatment showed an increase compared to the miR-384 inhibitor NC group (Fig. 3C). All these results imply that miR-384 is involved in the anticancer effects of luteolin. We then detected the miR-384 expression level in CRC tissues. Downregulation of miR-384 expression was noted in the CRC tissues compared with that observed in the non-tumor tissues (P<0.001; Fig. 3D). Moreover, with the progression of the tumor, the expression was decreased significantly. The expression of miR-384 was lowest in patients with TNM stage IV.

miRNAs exert their function by regulating the expression of their target genes. Thus, we took advantage of the prediction algorithms (miRDB and TargetScan v.5.1) to identify the potential targets of miR-384. Among the potential target genes, pleiotrophin (PTN) caught our interest. Thus, we detected the PTN expression after luteolin stimulation. The results showed that PTN expression was downregulated in a dose-dependent manner (Fig. 4A). Overexpression of PTN was able to significantly reverse the luteolin-mediated inhibition of cell migration and invasion as well as MMP expressions (Fig. 4B-D). This finding suggests that luteolin inhibited CRC cells migration and invasion by downregulating PTN expression. We next detected the relationship between PTN and miR-384. The expression of PTN was decreased when cells transfected with miR-384 mimic (Fig. 4E). Then miR-384 mimic or miR-384 mimic NC and PTN were co-transfected into CRC cells and the luciferase activity was detected. Compared with the control, a decrease in relative luciferase activity was observed when the WT PTN 3′-UTR was co-transfected with miR-384. However, the miR-384 mimic did not affect luciferase activity in the mutant construct (Fig. 4F). These data indicate that miR-384 directly modulates PTN expression by binding to its 3′-UTR. We also detected PTN expression after cells transfected with the miR-384 inhibitor followed by treatment with luteolin. Results showed that miR-384-inhibitor rescued the PTN downregulation induced by luteolin (Fig. 4G), confirming that luteolin-induced suppression of PTN expression was mediated by miR-384.

Data from Oncomine database were analyzed to examine the differential expression levels of PTN DNA between CRC and normal tissues. As shown in Fig. 5A, PTN was upregulated in tumor tissues (P<0.001). We also collected 50 cases of human CRC tissues and the paired adjacent normal tissues. The relative expression level of PTN was measured by qPCR (Fig. 5B) and western blotting (Fig. 5C). Results showed that PTN was highly expressed in CRC tissues. Furthermore, as the tumor progressed, the expression level of PTN increased as well. We also measured the PTN expression level in four CRC cell lines (SW480, SW620, LoVo and HT-29) and one normal colon epithelial cell (NCM460). The expression level of PTN was markedly higher in CRC cell lines than that noted in the NCM460 cells (Fig. 5D). Then, we analyzed the expression levels of miR-384 and PTN in tumor tissues, and we found that the Pearson's correlation coefficient was −0.58, demonstrating a negative correlation between miR-384 and PTN (Fig. 5E).

We next divided the patients into two groups according to the PTN expression level (25 patients in each group). The relationship between PTN expression and the clinical characteristics including age, sex, tumor differentiation, TNM stage and serum CEA level was determined. The results are shown in Table I. The differential expression of PTN was significantly associated with tumor size, N stage and M stage of CRC patients (all P<0.05).