Human colon cancer cell lines SW480, SW620, HT29, and the human embryonic kidney HEK293 cell line were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and grown in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) (both from Hyclone, GE Healthcare Life Sciences, Logan, UT, USA), 100 U/ml penicillin G and 100 mg/ml streptomycin (both from Beyotime Institute of Biotechnology, Shanghai, China). The human colonic epithelial cell line NCM460 was purchased from INCELL Corporation LLC (San Antonio, TX, USA) and grown in M3:10 medium (INCELL Corporation LLC). The cells were cultured at 37°C in a 5% CO₂ incubator.

Full-length Notch1 and Jagged1 cDNA lacking the 3′-untranslated region (UTR) were purchased from GeneCopoeia (Rockville, MD, USA) and subcloned into the eukaryotic expression vector pIRES (Clontech Laboratories, Inc., Mountain View, CA, USA). The empty pIRES vector was used as a negative control. Rabbit monoclonal anti-Notch1 (cat no. EP1238Y) and anti-Jagged1 (cat no. EPR4290) were purchased from Epitomics (Abcam, Cambridge, UK), mouse monoclonal anti-GAPDH (cat no. MAB374) was purchased from EMD Millipore (Billerica, MA, USA), and mouse monoclonal anti-vimentin (cat no. sc-6260), anti-fibronectin (cat no. C6F10) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (cat no. sc-2004) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Synthetic miR-34a mimic and negative control (NC) were purchased from GenePharma Co., Ltd. (Shanghai, China). The miR-34a LNA 5′-digoxigenin (DIG)-labeled (cat no. 612537-340) and scrambled LNA 5′-DIG-labeled probes (cat no. 699004-340) were purchased from Exiqon A/S (Vedbaek, Denmark). The probe sequences are as follows: miR-34a, AGG GCA GTA TAC TTG CTG AT; and scramble, GTG TAA CAC GTC TAT ACG CCC A. The probes were detected using an enhanced sensitive ISH detection kit (cat no. MK1030; Wuhan Boster Biological Technology, Ltd., Wuhan, China).

A tissue microarray (TMA) containing 103 samples of human CRC tissue specimens with their matched distant normal mucosa tissues (>10 cm away from the primary tumor) was used in the present study. The high-density CRC tissue microarray from Chinese specimens collected in March 2006 from the Third Xiangya Hospital of Central South University (Changsha, China) was constructed by the study group. Written informed consent was obtained from patients. The present study was approved by the Ethical Review Committees of the Third Xiangya Hospital of Central South University, Hunan Key Laboratory of Nonresolving Inflammation and Cancer and the Central South University (Changsha, China). Histology of all slides was reviewed by two expert pathologists. The cancer stage was made according to the TNM classification criteria (25). The TMAs (4 µm) were dewaxed and treated with 3% H₂O₂ and proteinase K to inactivate endogenous peroxidases. The slides were treated with pepsin in 3% citric acid. The slides were hybridized at 56°C overnight with 50 nM miR-34a LNA or scrambled LNA. Following washing 3 times with 2X saline-sodium citrate buffer, sections were blocked in blocking buffer (cat no. MK1030; Wuhan Boster Biological Technology, Ltd.) at 37°C for 30 min and then incubated with 1X anti-DIG Fab fragment antibody (cat no. MK1030; Wuhan Boster Biological Technology, Ltd.) at room temperature for 2 h for the hybridization reaction. The immunoreaction was detected using 3-diaminobenzidine tetrahydrochloride (cat no. 0031; Fuzhou Maixin Biotech Co., Ltd., Fuzhou, China). The sections were counterstained with 5 g/l hematoxylin at room temperature for 3 min. Following dehydration and mounting, the sections were observed under an inverted microscope (Nikon Corporation, Tokyo, Japan) with magnifications of ×100 or ×200. The staining intensity of miR-34a was scored as: 0–1, negative; 1–2, weak; 2–3, medium; and 3–4, strong. The percentages of miR-34a cells in 3 representative high-power fields of individual samples were determined. The miR-34a expression was calculated in terms of the product of the intensity scores and the percentage of positive cells. Individual samples were evaluated by at least 2 blinded pathologists. Scores ≥2 were used to define specimens with a high expression, and a score of <2 indicated low expression.

Potential target genes of miR-34a were predicted and analyzed using 4 web-based bioinformatics tools: TargetScan, miRWalk, microRNA.org and RNA22. The number of false positive predictions was markedly decreased by selecting the putative target genes that were predicted by ≥2 programs.

The SW480 cells were seeded into 6-well plates. The cells were transfected by nucleofection with 100 nM miR-34a mimic or mimic NC using HiPerFect Transfection reagent (Qiagen, Inc., Valencia, CA, USA), according to the manufacturer's protocol. At 48 h post-transfection, the reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was performed to verify the transfection efficiency.

Total RNA from cells and tissues was extracted using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA). The Hairpin-it™ MicroRNA Quantitation PCR kit was purchased from GenePharma Co., Ltd. The RT-qPCR was performed according to the manufacturer's protocol. At least three biological replicates were performed for each case. The qPCR analysis was performed using the iQ5 Real-Time PCR system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The miRNA PCR quantification used the 2^−ΔΔCq method (26) against U6 for normalization.

The expression of mRNA was evaluated using the SYBR Premix Ex Taq II (Tli RNaseH Plus) Real-Time qPCR kit (Takara Biotechnology Co., Ltd., Dalian, China). The mRNA PCR quantification used the 2^−ΔΔCq method (26) against GAPDH for normalization. The RT-qPCR primers, including E-cadherin, vimentin, fibronectin, zinc finger protein SNAI2 (Slug), zinc finger protein SNAI1 (Snail), GAPDH and zinc finger E-box-binding homeobox 1 (ZEB1), are summarized in Table I.

Human Notch1 (accession number NM_017617) and Jagged1 (NM_000214) 3′-untranslated regions (UTRs) are predicted to contain two putative binding sites of miR-34a. Notch1 and Jagged1 wild-type (WT) 3′-UTRs and mutant 3′-UTRs with 6 base mutations at putative seed regions were synthesized and cloned into the pIRES vector with the restriction sites for HindIII and SpeI at the two ends of the oligonucleotides. The WT and mutant oligonucleotide sequences used are summarized in Table II. In total, 9 luciferase reporters (LR) were constructed and referred to as LR-blank (no insertion), LR-34a/Notch1w1 (where w1 indicates WT site 1), LR-34a/Notch1w2 (where w2 indicates WT site 2), LR-34a/Notch1m1 (where m1 indicates site 1 mutant), LR-34a/Notch1 m2 (where m2 indicates site 2 mutant), LR-34a/Jagged1w1, LR-34a/Jagged1w2, LR-34a/Jagged1m1 and LR-34a/Jagged1m2. HEK293 cells were cultured in RPMI-1640 medium supplemented with 10% FBS in 24-well plates at 37°C overnight, and then co-transfected with 100 ng of one of the reporter plasmids and 10 pmol miR-34a mimic or mimic NC using MegaTran 1.0 Transfection reagent (OriGene Technologies, Inc., Rockville, MD, USA). At 24 h post-transfection, firefly and Renilla luciferase activities were consecutively measured using the commercial dual-luciferase reporter system (Promega Corporation, Madison, WI, USA). The Renilla luciferase signal was normalized to the firefly luciferase signal for each individual analysis.

Cells were lysed in radioimmunoprecipitation assay buffer containing 1% protease inhibitor cocktail (Pierce; Thermo Fisher Scientific, Inc.). The protein concentrations in the lysates were measured using a bicinchoninic acid protein assay kit (Pierce; Thermo Fisher Scientific, Inc.). Total protein (50 µg/lane) was separated on a Bio-Rad Bis-Tris Gel system (Bio-Rad Laboratories, Inc.) and electrotransferred onto polyvinylidene fluoride membranes. Subsequent to blocking with 5% bovine serum albumin (cat no. SW3015; Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) at room temperature for 1 h, the membranes were serially incubated with anti-Notch1 (1:500 dilution), anti-Jagged1, anti-vimentin (1:200 dilution), anti-fibronectin (1:500 dilution) and anti-GAPDH (1:1,000) primary antibodies at 4°C overnight, followed by incubation with the HRP-conjugated goat anti-rabbit IgG secondary antibody (1:1,000 dilution) at 37°C for 2 h. Chemiluminescence using Luminata Forte Western horseradish peroxidase substrate (Merck KGaA, Darmstadt, Germany) was visualized using a ChemiDoc XRS+ system (Bio-Rad Laboratories, Inc.). The densities of the bands were quantified using Image Lab 2.0 software (Bio-Rad Laboratories, Inc.).

SW480 cells (1×10⁶ cells/well) were cultured overnight at 37°C. SW480 cells were transfected with miR-34a (or mimic NC) followed by Notch1 or Jagged1 transfection. Following the cells in the 6-well dish attaining 90% confluence, a wound was created using a 10 µl pipette tip. Following washing 3 times in PBS to remove cellular debris, cells were cultured in RPMI-1640 medium with 2% serum. Cell migration at the wound sites was documented using an inverted microscope (Nikon Corporation) with a magnification of ×200 at the indicated time-points (0 and 48 h). Cells in 3 randomly selected areas were counted.

Corning Costar Transwell 24-well plates (8-µm pores; Corning Incorporated, Corning, NY, USA) coated with BD Matrigel matrix (BD Biosciences, Franklin Lakes, NJ, USA) were maintained for 1 h at 37°C, followed by the addition of 1×10⁵ transfected cells suspended in 200 µl medium with 1% serum into the top of each well insert. Normal growth medium was added to the bottom wells. The cells were allowed to migrate for 24 h at 37°C. The migrated cells were fixed with 10% methanol for 15 min. The invading cells on the lower surface of the membrane were stained with 0.5% crystal violet for 5 min at room temperature. The stained cells were counted under a microscope (Nikon Corporation). To minimize the bias, ≤5 randomly selected fields at ×100 magnification were counted, and the average number was calculated.

Statistical analysis was performed using the SPSS for Windows (version 17.0; SPSS, Inc., Chicago, IL, USA) and GraphPad Prism (version 5; GraphPad Software, Inc., La Jolla, CA, USA). χ² tests were used for categorical variables, including categorized miR-34a expression levels, clinical stage, histological differentiation, sex, age and metastasis of the colorectal tumors. Independent unpaired t-tests and one-way analysis of variance with Bonferroni post hoc testing were performed for the analysis of continuous variables in categories. Survival analysis was determined using the Kaplan-Meier estimator method. P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR was used to detect the expression of miR-34a in colon cancer cell lines SW480, SW620 and HT29, and normal colonic epithelial cell line NCM460. Compared with the normal colonic epithelial cells, miR-34a was identified to be downregulated ~2.5-fold in SW480 cells, 3-fold in HT29 cells and 6-fold in SW620 cells (Fig. 1A). In addition, compared with SW480, the relatively high metastatic SW620 and HT29 cells exhibited significantly decreased miR-34a expression, particularly in SW620 cells (P=0.014).

To verify the altered expression of miR-34a in human colorectal carcinogenesis, ISH was first conducted to evaluate miR-34a expression levels in 103 pairs of CRC tissues and their matched adjacent normal tissues using a TMA (representative images are presented in Fig. 1B). It was identified that miR-34a was apparently downregulated in the colorectal tumors compared with the normal tissues (P=0.000; Table III). Next, the potential clinicopathological implications of altered miR-34a expression were assessed. Clinical samples were divided into low- and high-expression groups based on the miR-34a expression scores >2 or <2. Of the 103 normal colorectal samples, 88 (85%) exhibited increased expression of miR-34a (Table III). Thus, miR-34a is underexpressed in CRC tissue compared with the normal colorectal mucosa (representative images are presented in Fig. 1C). In the 103 individuals with CRC, the miR-34a level was inversely associated with distant metastasis, and positively associated with differentiation and age (P=0.045, 0.010 and 0.020, respectively; Table III). However, no association between miR-34a expression and sex and TNM stage was observed in patients with CRC. To additionally assess the significance of miR-34a in terms of clinical prognosis, a Kaplan-Meier estimator survival analysis was conducted using overall survival (OS). The results demonstrated that patients with a diminished miR-34a expression exhibited shorter OS times compared with patients expressing high miR-34a levels (P=0.029; Fig. 1D). These data demonstrated overt downregulation of miR-34a in CRC, suggesting that the expression levels of miR-34a was associated with the metastatic potential of CRC.

The biological function of miR-34a was explored using several bioinformatics algorithms in order to identify potential miR-34a target genes. Among the candidate target genes, Notch1 and its ligand Jagged1 were identified. As presented in Fig. 2A, 2 miR-34a-binding sites were identified in the 3′-UTRs of Notch1 and Jagged1 mRNAs, with a perfect base pairing between the seed sequence of mature miR-34a and the miRNA-recognition element in the 3′-UTRs. A number of cellular functions and biochemical processes associated with tumorigenesis are modulated by Notch signaling, including EMT, proliferation, apoptosis, adhesion and angiogenesis (27). A previous study has identified Notch1 overexpression in various solid tumors, including CRC (28). Previous studies have also suggested that miR-34a suppressed tumor invasion via the downregulation of Notch1 and Jagged1 in cervical carcinoma and choriocarcinoma cells (15). Therefore, an initial screen of the effects of miR-34a on Notch1 and Jagged1 expression was performed. Initially, miR-34a inhibition of the expression of Notch1 and Jagged1 was determined by transfecting miR-34a mimic into the SW480 cells. It was identified that the miR-34a mimic significantly decreased the mRNA and protein levels of Notch1 and Jagged1 compared with the mimic NC (Fig. 2B).

Whether Notch1 and Jagged1 were direct targets of miR-34a was investigated using a luciferase reporter assay to test the binding of miR-34a to the 3′-UTR of Notch1 and Jagged1. The Notch1 and Jagged1 3′-UTRs containing the binding sites for miR-34a were subcloned into a luciferase reporter vector. The addition of an miR-34a mimic significantly suppressed the luciferase activity of the Notch1 and Jagged1 3′-UTRs following co-transfection of the luciferase vector (WT, mutant or blank control) with the miR-34a mimic into SW480 cells (Fig. 2C). Thus, these results confirm that miR-34a directly recognizes the 3′-UTRs of the Notch1 and Jagged1 mRNAs, leading to mRNA degradation and translational inhibition.

Considering the aforementioned data, the role of miR-34a in the human colon cancer SW480 cell line was next investigated. The effects of miR-34a were assessed using wound-healing assays and Matrigel invasion assays following transfection of the cells with miR-34a mimic or mimic NC. It was identified that overexpressed miR-34a in SW480 cells significantly attenuated cell migration and invasion. The re-expression of Notch1 or Jagged1 (lacking an endogenous 3′-UTR) prevented this inhibition, suggesting that miR-34a specifically targets Notch1 or Jagged1 to suppress CRC metastasis (Fig. 3A and B). Previous evidence suggests that the activation of Notch1 and Jagged1 induces EMT and promoted the invasion and dissemination of cancer cells (29–31). Therefore, whether miR-34a affected EMT was investigated. A panel of EMT markers, including epithelial cadherin, vimentin, fibronectin, Slug, Snail and ZEB1, were detected in SW480 cells following transfection with miR-34a mimic or mimic NC (Fig. 3C). Significant attenuation of the mRNA expression levels of two key mesenchymal markers, vimentin and fibronectin, was observed (Fig. 3C and D). In addition, the re-expression of Notch1 or Jagged1 recovered the protein expression levels of vimentin and fibronectin in miR-34a-overexpressing SW480 cells, suggesting that miR-34a suppresses CRC metastasis by targeting Notch1/Jagged1, and the downstream molecules vimentin and fibronectin.