The 293T and human pancreatic cancer PANC-1 and BXPC-3 cell lines were purchased from the American Type Culture Collection and were incubated at 37°C in a humidified incubator containing 5% CO₂. The cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) (both from Thermo Fisher Scientific, Inc.), 100 U/ml penicillin and 100 μg/ml streptomycin (both Jiangsu KeyGEN BioTECH Corp., Ltd.).

A total of 20 female nude mice (BALB/cJNju-Foxn1nu/Nju; age, 5 weeks; weight, 16±2.5 g) were purchased from Nanjing Biomedical Research Institute (http://www.nbri-nju.com/en-us/about-animal-facility). All the animals were raised under pathogen-free conditions at 20-26°C, 40-70% relative humidity, 15 times/h ventilation and a 12 h light/12 h dark cycle. The animals were provided with water and food freely. All animal experiments were approved by the Animal Care and Use Committee of Sun Yat-Sen University (Guangzhou, China).

Synthetic miR-623 mimics (5′-ACCCAACAGCCCCUGCAAGGGAU-3′), inhibitors (5′-CAGUACUUUUGUGUAGUACAA-3′), miR- 623 mutant (5′-ATAACTTGAAGGGGCTGTTGGGT-3′), mimics negative control (NC; 5′-UUCUCCGAACGUGUCACGUTTACGUGACACGUUCGGAGAATT-3′) and inhibitor NC (5′-CAGUACUUUUGUGUAGUACAA-3′) miRNAs were purchased from Shanghai GenePharma Co., Ltd. Recombinant human IL-8 (rIL-8) was purchased from PeproTech, Inc. Rabbit-anti-human MMP1 (cat. no. 54376), ERK (cat. no. 4695), phosphorylated (p-)ERK (cat. no. 4370), Twist-related protein 1 (TWIST1; cat. no. 46702), zinc finger protein SNAI1 (cat. no. 3879) and E-cadherin (cat. no. 14472) polyclonal primary antibodies, mouse GAPDH (cat. no. 51332) monoclonal primary antibody and horseradish peroxidase-conjugated secondary antibodies (anti-rat cat. no. 7077; anti-rabbit cat. no. 7074) were purchased from Cell Signaling Technology, Inc. Phalloidin was purchased from Sigma-Aldrich; Merck KGaA, and DAPI was from Wuhan Boster Biological Technology, Ltd.

The microRNA.org (http://www.microrna.org/), PITA (http://genie.weizmann.ac.il/pubs/mir07/mir07_prediction.html), TargetScan (Release 6.2, http://www.targetscan.org) and miRTar.human (http://mirtar.mbc.nctu.edu.tw/human/) databases were screened for potential miRNAs binding to the 3′-UTR of the MMP1 transcript (acc. no. NM_002421). The minimum free energy (MFE) of hybridization was calculated using the RNAhybrid program (http://bibiserv2.cebitec.uni-bielefeld.de/rnahybrid).

The pGL3-Control vector (Promega Corporation) and the pGL3-CU vector were constructed according to the protocol outlined in previous studies (20,21). The sequence of the MMP1 3′-UTR was amplified via PCR using primers listed in Table I. The PCR products were digested with USER enzyme (New England BioLabs, Inc.) and ligated into the pGL3-CU vector, resulting in a pGL3-MMP1 vector, according to the manufacturer's protocols. The authenticity of pGL3-MMP1 vector was confirmed.

To verify the precise target of the miRNAs, the Dual-Luciferase Reporter Assay system (Promega Corporation) was used. Briefly, 293T cells were seeded in 96-well plates at a density of 1×10⁵ cells per well and cultured in complete medium without antibiotics for 24 h. The luciferase reporter plasmids pGL3-NC (Promega Corporation) and pGL3-MMP1 (100 ng/well) were then co-transfected with a mimic NC, hsa-miR-623 mimic, inhibitor NC, hsa-miR-623 inhibitor or hsa-miR-623 mutant (50 nM) using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). Firefly and Renilla luciferase activities were measured at 24 h post-transfection using the Dual-Luciferase Reporter Assay kit. Firefly luciferase was normalized to Renilla luciferase activity. Three independent experiments were performed.

An RNA EMSA was performed using a LightShift Chemiluminescent RNA EMSA kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The binding buffer mixture was prepared with 1X REMSA binding buffer supplemented with 5% glycerol, 200 mM KCl and 100 mM MgCl₂. Next, 200 nmol IRDye1800-labeled hsa-miR-623 and/or Cy5.5-labeled hsa-miR-623 TAR (targeting sequences in the MMP1 3-'UTR) oligonucleotides (Integrated DNA Technologies, Inc.) were added to the binding buffer mixture. Cytoplasmic protein from PANC-1 cells was extracted using NE-PER Nuclear and Cytoplasmic Extraction reagents (Thermo Fisher Scientific, Inc.). The oligonucleotide sequences are listed in Table I.

To confirm the binding specificity between hsa-miR-623 and its target, 50-fold molar excess of cold-hsa-miR-623 or cold-NC probe was added to the reaction system 5 min prior to the addition of the dye-labeled probes. A total of 2 μg PANC-1 cell cytoplasmic protein and 1 μg tRNA were added to the hsa-miR-623/target mixtures to examine the RNA-protein interactions. Following a 20-min incubation at room temperature, the reaction mixture was loaded onto a 10% gel for PAGE and separated by electrophoresis (100 V; 150 min) at 4°C. Finally, the bands were detected using an Odyssey CLx Infrared Imaging system (LI-COR Biosciences).

PANC-1 and BXPC-3 cells were seeded at a density of 6×10⁵ cells per well in 6-well plates and cultured for 12 h to allow cells to fully attach. The medium was then replaced with serum-free medium, and rIL-8 was added to a final concentration of 100 ng/ml (22,23), followed by incubation for 6, 12, 24, 36 and 48 h. The cells were then harvested for further analyses.

Synthetic miR-623 mimics, inhibitors and NC miRNAs were transfected into PANC-1 and BXPC-3 cells using Lipofectamine 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.). PANC-1 and BXPC-3 cells were seeded at a density of 6×10⁵ cells per well in 6-well plates, 3 μl transfection reagent was added into each well, and the final concentration of microRNA (mimic NC, mimic, inhibitor NC or inhibitor) was 20 nmol/l in a 2-ml final volume. At 6 h following incubation, the medium was changed to fresh complete medium, and the cells were harvested 48 h post-transfection. The transfection efficiency was determined by reverse transcription-quantitative PCR (RT-qPCR).

For stable transfections, 3 μg miR-623-GFP or NC-GFP vector were packaged into lentiviral virus vectors, according to the manufacturer's protocol (Vigene Biosciences, Inc.). The PANC-1 cells were then transfected with the lentiviruses and selected with 2 μg/ml puromycin (Sigma-Aldrich; Merck KGaA). The transfection efficiency was evaluated by RT-qPCR.

Total RNA was extracted from PANC-1 and BXPC-3 cells using the miRNeasy Mini kit (Qiagen GmbH). The RNA concentration and purity were determined according to the A260/280 values, using a NanoDrop spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc.). A High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Thermo Fisher Scientific, Inc.) or miScript^® II RT kit (Qiagen GmbH) was used for first-strand cDNA synthesis. The samples were reverse transcribed at 37°C for 15 min and heated to 85°C for 5 sec to inactivate the reverse transcriptase.

The qPCR was performed on an ABI Prism 7900 Sequence Detection system (Applied Biosystems; Thermo Fisher Scientific, Inc.) using a QuantiFast SYBR1 Green RT-PCR kit (Qiagen GmbH). The reactions were preheated for 2 min at 50°C and denatured for 2 min at 90°C, followed by 40 cycles of denaturation at 95°C for 15 sec and annealing/extension at 60°C for 15 sec. The qPCR results, recorded as threshold cycle numbers (Ct), were calculated by the 2^−ΔΔCq method with normalization against GAPDH mRNA as an internal control (24). The relative expression levels of MMP1 and hsa-miR-623 were normalized to the housekeeping gene GAPDH and U6 small nuclear RNA, respectively. The primers used in the RT-qPCR were obtained from Integrated DNA Technologies, Inc. and their sequences are listed in Table I.

Protein extracts from PANC-1 and BXPC-3 cells were prepared using RIPA lysis buffer with protease inhibitors and quantified using a bicinchoninic acid protein assay kit (Pierce; Thermo Fisher Scientific, Inc.). A total of 20 μg protein from each experimental group was separated on a polyacrylamide gel containing 10% sodium dodecyl sulfate and subsequently transferred to a 0.45-μm polyvinylidene difluoride membrane (EMD Millipore). The membrane was then blocked with 5% non-fat milk at room temperature for 1 h. Following an incubation with the primary polyclonal primary antibodies MMP1 (1:1,000 dilution), TWIST1 (1:500 dilution), SNAI1 (1:1,000 dilution), E-cadherin (1:400 dilution), ERK (1:1,000 dilution), p-ERK (1:2,000 dilution) or GAPDH (1:2,000 dilution) overnight at 4°C and incubation with the secondary antibodies (1:2,000 dilution) for 1 h at room temperature, protein bands were detected using enhanced chemiluminescence reagents (Thermo Fisher Scientific, Inc.) and visualized after exposure using the Tanon-5200 Multi-imaging system (Tanon Science and Technology Co., Ltd.). ImageJ version 1.8.0 software (National Institutes of Health) was used to analyze the images.

Following transfection with synthetic miR-623 mimics, inhibitors and NC miRNAs, and co-culturing with IL-8, PANC-1 and BXPC-3 cells were fixed with 4% formaldehyde for 20 min at room temperature. The cells were then stained with 5 μg/ml phalloidin for 30 min, counterstained with 1 μg/ml DAPI for 10 min at room temperature, and finally visualized using a fluorescence scanning microscope at ×400 magnification.

PANC-1 and BXPC-3 cells were transfected with synthetic miR-623 mimics, inhibitors and NC miRNAs, and a wound-healing assay was performed 48 h after transfection. When the cells reached 100% confluence, a wound was created by manually scraping the cell monolayer with a 200-μl pipette tip. At 0 and 24 h after incubation, images were captured using a light microscope (Zeiss GmbH) at ×40 magnification, and analyzed quantitatively by ImageJ version 1.8.0 software.

An invasion assay was performed using 8-μm pore, 24-well Transwell chambers coated with Matrigel (Corning Life Sciences). PANC-1 and BXPC-3 cells were transfected with synthetic miR-623 mimics, inhibitors and NC miRNAs. At 48 h after transfection, the cells were harvested, 5×10⁴ cells in 200 μl serum-free medium were reseeded into the top chamber, and the bottom chamber was filled with 750 μl DMEM containing 10% FBS. Cells that migrated to the underside of the membrane were stained with 0.1% crystal violet (Phygene Life Sciences Co., Ltd.) for 30 min at room temperature, followed by imaging and counting under a light microscope (Zeiss GmbH) at ×40 magnification. All experiments were performed in triplicate and repeated ≥3 times.

In vivo efficacy studies

To investigate the effect of hsa-miR-623 on the metastasis of pancreatic cancer cells, 1×10⁶ PANC-1 cells stably expressing miR-623 or NC were injected into the tail veins of mice, which were divided into 4 groups (n=5): PANC-1^Lenti-GFP, PANC-1^{Lenti-miR-623-GFP}, PANC-1^{Lenti-GFP+IL-8} and PANC-1^{Lenti-miR623-GFP+IL-8}. The rIL-8 was administered via intraperitoneal injection at 100 ng/animal once a week for 5 weeks. The weight, activity and appearance of the mice were monitored every 3 days. If the body weight of the animal dropped markedly in a short time, the subcutaneous fat disappeared, the animals exhibited spinal curvature, and the activity and reaction ability decreased, euthanasia was carried out in advance, otherwise euthanasia was carried out at 5 weeks. The mice were anaesthetized by intraperitoneal injection of 0.5% pentobarbital sodium salt (CAS, 57-33-0; Beijing Huayehuanyu Chemical Co., Ltd.) at a dosage of 100 mg/kg, and then placed in transparent glass containers. 100% carbon dioxide was introduced into the bottles with the filling rate of 10-30% volume per min. The death of the mice was confirmed when no spontaneous breathing was detected for 2-3 min and they exhibited no blinking reflex. Subsequently, the livers and lungs of the animals were dissected and imaged under a fluorescence scanning microscope. All dissected organs were fixed overnight with 4% formalin at room temperature and then embedded in paraffin. The livers and lungs were sliced to 5-μm thickness and stained with hematoxylin for 15 min and eosin for 1 min at room temperature. Immunochemical assays were performed according to the protocols of the immunohistochemical staining kit manufacturer (cat. no. SA1022; Wuhan Boster Biological Technology, Ltd.), using the aforementioned primary antibodies incubated overnight at 4°C. The numbers of liver and lung metastases were independently calculated and assessed by two pathologists.

The clinicopathological characteristics, mRNA expression levels of MMP1 and IL-8, and the hsa-miR-623 levels in pancreatic ductal adenocarcinoma (PDAC) tissues were obtained from TCGA database (https://cancergenome.nih.gov/), which contained 172 patients with PDAC. As there are only 4 normal samples included in TCGA database, the present study also searched other GEO (https://www.ncbi.nlm.nih.gov/geo/) public datasets for further data, including datasets GES28735 (45 paired PDAC and adjacent non-tumor tissues), GSE32678 and GSE32676 (data for miRNA and mRNA expression levels in 25 PDAC tissues) (25,26).

All statistical analyses were performed using SPSS version 17.0 software (SPSS, Inc.). The results are presented as the mean ± standard deviation. Pearson's correlation analysis was used to analyze the correlation between MMP1 mRNA levels and hsa-miR-623 levels, and between MMP1 mRNA levels and IL-8 mRNA levels in the original data from human PDAC tissues published in the GEO or TCGA databases. One-way analysis of variance was used to compare data from multiple groups, and the Bonferroni method was used to determine the significant differences between the groups. P<0.05 was considered to indicate a statistically significant difference.

The screening results from all four databases suggested that several miRNAs may target the 3′-UTR of the MMP1 transcript. The correlation between the expression of miRNA and the expression of MMP1 in human PDAC tissues was analyzed, and the results are presented in Table II. According to our previous studies, the value of MFE ≤-20 kcal/mol was considered to be essential for the miRNA/mRNA hybridization (20,21,27-29). As presented in Table II, the MFE of hybridization between hsa-miR-623 and its cognate target at the 3′-UTR of MMP1 was -27 kcal/mol. Based on the database analysis, compared with adjacent non-tumor tissues, the expression of MMP1 in PDAC tumor tissues was increased, whereas the expression of hsa-miR-623 was decreased, and a negative correlation was observed between their expression (r=−0.402, P=0.046; Fig. 1A). Based on these results, hsa-miR-623 was identified as a candidate miRNA for further investigation.

RNA EMSA was performed to explore whether direct binding occurs between hsa-miR-623 and its target within the MMP1 3′-UTR. As illustrated in Fig. 1B, complexes formed between hsa-miR-623 and their cognate MMP1 mRNA target sequences in vitro (lanes 3-5, the duplex is indicated by an arrowhead). Competition assays indicated that the complex of hsa-miR-623 and its cognate targets was sequence-specific, as the complex formation was inhibited by a 50-fold molar excess of cold hsa-miR-623 (lane 5); however, was not influenced by a 50-fold molar excess of unlabeled NC oligonucleotide (lane 4). PANC-1 cytoplasmic extracts were subsequently used to observe the formation of miRNA/mRNA/protein complexes (lanes 6-8, complexes A and B). Taken together, the competition assays revealed that complex A was inhibited by a 50-fold molar excess of cold hsa-miR-623 (lane 8), suggesting that hsa-miR-623 directly and specifically interacts with its target within the MMP1 3′-UTR.

Luciferase reporter assays were performed to explore whether hsa-miR-623 could regulate the 3′-UTR of MMP1. The mutant hsa-miR-623 was designed with 3 cytosines replaced by adenines; consequently, the free energy of hybridization between the mutated miRNA and its target mRNA changed from -27 to -16.8 kcal/mol (Fig. 1C). Following dual transfection, the luciferase activity produced by the reporter gene plasmid was significantly downregulated by the wild-type hsa-miR-623 mimic (0.64±0.11; P<0.05) and increased by the hsa-miR-623 inhibitor (1.52±0.15; P<0.05). Conversely, the mutant hsa-miR-623 mimic failed to influence the luciferase activity in the 293T cells (Fig. 1D). These results indicate that hsa-miR-623 modulates its target in the 3′-UTR of MMP1 in a direct and specific manner. Following synthetic miRNA transfection, mimics increased and inhibitors decreased hsa-miR-623 expression levels in the PANC-1 and BXPC-3 cells (Fig. 2A and D). Consistent with the results from the EMSA and luciferase assay, the hsa-miR-623 mimic suppressed endogenous MMP1 expression at the mRNA and protein levels, whereas the hsa-miR-623 inhibitor led to an increase in MMP1 expression (Fig. 2B, C, E and F). Furthermore, the hsa-miR-623 mimic resulted in a decrease in the migration and invasion ability of PANC-1 and BXPC-3 cells, whereas the hsa-miR-623 inhibitor displayed the opposite effects (all P<0.05; Fig. 3A and B).

In rheumatoid arthritis, MMP1 can be induced by TNF-α and IL-1β (30,31), however, whether MMP1 can be induced by inflammatory factors in PDAC cells remains unclear. First, the correlation was investigated between the expression of MMP1 mRNA and 3 commonly studied inflammatory factors, IL-1β, IL-6 and IL-8. The results based on TCGA database (Fig. 4A-C) revealed that the correlation between IL-8 and MMP1 (r=0.445) was higher than that with IL-1β (r=0.257), and both were statistically significant. However, only IL-8 exhibited significant correlation with MMP1 expression in the GEO dataset GSE28735 (Fig. 4D, E and F). Based on the results of the analysis from the two databases, IL-8 was selected as an inducer of MMP1. As presented in Fig. 5A, at 24, 36 and 48 h of co-culturing with rIL-8, the expression of MMP1 was significantly upregulated (all P<0.05). When the hsa-miR-623 mimic was transfected into PANC-1 and BXPC-3 cells prior to treatment with rIL-8, the induction of MMP1 was attenuated, whereas the hsa-miR-623 inhibitor transfection further increased the MMP1 expression (Fig. 5B). In addition, the expressions of TWIST1 and SNAI1 were increased following IL-8 treatment for 6, 12 and 24 h, and E-cadherin was decreased at 12 and 24 h (Fig. 5C). As the ERK signaling pathway is commonly involved in mechanisms underlying IL-8 function, and exerts pro-tumor activities, western blot assays were subsequently performed to investigate whether hsa-miR-623 could suppress the phosphorylation of ERK. As hypothesized, the transfection of hsa-miR-623 mimic not only decreased ERK phosphorylation, but also decreased IL-8-induced TWIST1 and SNAI1 expression, and reversed the downregulation of E-cadherin. By contrast, transfection of the hsa-miR-623 inhibitor exhibited the opposite results (Fig. 5D). As indicated by the phalloidin staining (Fig. 5E), the morphology of the miR-623 mimic-transfected PANC-1 and BXPC-3 cells displayed relatively tight cell-cell adhesion, whereas miR-623 inhibitor-transfected cells exhibited many new spike-like filopodia at their edges, indicating a spindle-like and fibroblastic phenotype. These results suggest that upregulation of miR-623 results in inhibition of the epithelial-mesenchymal transition (EMT). Furthermore, the hsa-miR-623 mimic attenuated IL-8-induced migration and invasion of PANC-1 and BXPC-3 cells, whereas miR-623 inhibition resulted in more aggressive features (Fig. 6).

A xenograft nude mouse model was established to further investigate whether hsa-miR-623 could suppress distant metastasis of PDAC in vivo. The transfection efficiency evaluated by qPCR demonstrated that Lenti-miR-623-GFP significantly increased the expression of miR-623 compared with NC transfection (Fig. 7A). The transfected PANC-1 cells were injected into the tail vein of nude mice to mimic the spread of cancer cells throughout the blood. All animals survived until the end time point of the experiment. As presented in Fig. 7B and C, heightened metastasis was observed in mice injected with rIL-8; however, when hsa-miR-623 was overexpressed in PANC-1 cells, metastasis was markedly inhibited in mice with or without IL-8 treatment. Compared with that in the control mice, the mean fluorescence density of lungs and livers in the hsa-miR-623 overexpression groups was markedly lower (P<0.05), indicating decreased metastasis of cancer cells. In addition, the number of lung and liver metastasis nodules was also significantly decreased in the hsa-miR-623-overexpression group compared with those in the control group (Fig. 7C and D). Furthermore, the expression of MMP1, TWIST1 and SNAI1 protein in liver metastatic nodules in the hsa-miR-623 overexpression group was also significantly decreased (Fig. 7E). These results were consistent with the in vitro results, indicating that hsa-miR-623 attenuates IL-8-induced PDAC metastasis in vitro and in vivo.