Gene expression profiles of GSE13937 were downloaded from the GEO repository. These data were based on the GPL8835 platform. The data of 44 patients with ESCC (including 44 samples of normal adjacent esophageal tissues and 44 samples of tumor tissues) were chosen to perform further analysis and create the heat maps with Morpheus online tool (https://software.broadinstitute.org/morpheus/). Then, the top 20 of differentially expressed (DE) miRNAs were screened. This study was published by Mathé et al (28). Total RNA was extracted and detected by miRNA microarray chips. Version 3. R (BioConductor) (http://www.bioconductor.org/install/) was used for background correction and normalization of the data.

All samples were collected at the Panyu Central Hospital and the Third Affiliated Hospital of Southern Medical University from February 27, 2010 to May 2, 2017. One patient was female and 19 patients were male, with a mean age of 62.4±5.8 years, and diagnosed with clinicopathological characteristics of ESCC. These samples were used only for the detection of the gene expression of miRNA and 10 patient samples among the 20 were utilized for target gene detection. All samples were stored at −80°C after collection.

This study was approved by the Ethics Committee of Panyu Central Hospital (Guangzhou, China). All patients agreed to participate provided informed consent for this study.

The cell line Eca-109 (human esophageal squamous carcinoma cell) with metastatic ability and 293T cell were obtained from the State Laboratory of Oncology in South China, Sun Yat-Sen University Cancer Center (Guangzhou, China). The cell line KYSE510 (well differentiated squamous cell carcinoma cell line) was obtained from the Central Laboratory, Nanfang Hospital, Southern Medical University (Guangzhou, China). The Eca-109 and KYSE510 cells were cultured in RPMI-1640 media, supplemented with 10% fetal calf serum (FCS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and were maintained under 5% CO₂ at 37°C. The 293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco; Thermo Fisher Scientific, Inc.), supplemented with 10% FCS and 1% GlutaMAX and 1% penicillin with streptomycin.

Ten BALB/C nude mice aged 4–6 weeks old with a male to female ratio of 1:1 weighing 20–25 g were used in the present study. These mice were obtained from Chase Reward Ltd. (Guangzhou, China) and kept under the specific pathogen-free conditions (temperature, 20–26°C; 12-h light-dark cycle; aseptic food and water). The animal experiments were approved by the Ethics Committee of Panyu Central Hospital (no. K20170002).

miR-21 mimics (overexpression), miR-21 inhibitor (downregulation) and negative control (NC including mimic NC, inhibitor NC) were purchased from Guangzhou RiboBio Co., Ltd. (Guangzhou, China). Transfections were performed on Eca-109 cells with Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) following the manufacturer's protocol. Briefly, 5×10⁵ cells were cultured in 6-well plates for 24 h prior to transfection. The miR-21 inhibitor, mimics and their NC nucleotides were added to the cell with transfection reagent to a final concentration of 50 and 100 nm/l, respectively. The cells were treated for 48 h and harvested for analysis.

miRWalk (miR Walk1.0) (http://zmf.umm.uniheidelberg.de/apps/zmf/mirwalk/micrornapredictedtarget.html) was utilized to predict the target gene and RASA1 was predicted as the target gene of miR-21. Thus, the vector for RASA1 was constructed. The 3′ untranslated regions (3′UTR) of RASA1 mRNA were amplified using the PCR instrument (GeneAmp PCR System 2400; Applied Biosystems; Thermo Fisher Scientific, Inc.), which was bound to the mature miR-21 (MIMAT0000076). To prevent formation of a termination signal, ‘taagcta’ was selected as the region in a mutational expression vector template and was amplified (amplification primers are presented in Table I). Then, the aforementioned template was transfected with psi-CHECK-2 luciferase plasmid and used for dual-luciferase reporter assay.

The 3′UTR fragments of the RASA1 gene were amplified and then transfected and cloned into the psi-CHECK-2 luciferase miRNA expression reporter vector. The 293T cells (2×10⁴) were cultured in 24-well plates and transfected with 50 nM miR-21 mimics, 100 nM inhibitor or NC, 0.5 µg of psi-CHECK-2 luciferase reporter vector that contained the wild-type or mutant 3′UTR of RASA1 or empty plasmid. Transfections were performed using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Forty-eight hours after transfection, the luciferase activity was assessed by the Dual-Luciferase Reporter Assay System (GloMax; Promega Corporation, Madison, WI, USA). The experiments were repeated in 3 independent experiments in duplicate.

Eca-109 cells (100) were cultured in 6-well plates until colonies could be observed. Then the cells were fixed by 4% of triformol and stained with Giemsa for 15 min at room temperature (Nanjing Jiancheng Taihao Biotechnology Co., Ltd., Nanjing, China). Finally, the clones were counted under a light microscope (magnification, ×100). The clone which contained >10 cells was regarded as one colony formation, and the colony rate was calculated using the following formula: colony formation rate = (colony formation numbers/100) × 100%.

Eca-109 cells (8.0×10⁵) were cultured in 6-well plates for 24 h and confluence (>90%) without any vacant space. Subsequently the cells were scratched using 10-µl tips with a wound midline of the culture well, and then replaced in 1% FBS DMEM. Finally, the difference in width of the wounds was measured at 0, 24 and 48 h, respectively to evaluate the migration of cells and the healing rate was calculated.

Cell migration activities were detected using Transwell migration. After the transfected Eca-109 or KYSE510 cells were cultured for 24 h, the cells (1×10⁵) were resuspended in serum-free media and cultured in the inserts (8-µm diameter pore size; Corning Inc., Corning, NY, USA), which were placed in the 6-wells plates with 10% FBS serum media. The cells migrated from the upper surface to the lower surface of the membrane after 48-h incubation. Subsequently, the migrated cells were fixed with 100% methanol and stained with 1% toluidine. Finally, these stained cells were counted in 5 random optical fields under a light microscope (magnification, ×200). The cell invasion assay was similar to the cell migration assay except the Transwell membrane was pre-coated with 24 mg/ml Matrigel (Corning Inc.).

Transfected Eca-109 cells and tissues were processed in TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and total RNA was extracted according to the manufacturer's instructions. First-strand cDNA was synthesized with 1 µg total RNA per sample using the PrimeScript™ RT Synthesis system (Takara Bio, Inc., Shiga, Japan). Subsequently, the cDNA sample was amplified using QuantiTect reagents (GeneCopoeia, Inc., Rockville, MD, USA) in a final volume of 20 µl under the LightCycler 480II detector (Roche Diagnostics, Basel, Switzerland). The amplifications were performed as follows: predestination for 2 min at 50°C, denaturation for 30 sec at 95°C, followed by 40 cycles of 95°C for 5 sec, and 60°C for 34 sec. The melting curve analysis was performed to detect the specificity of amplification products. The experiments were performed in triplicate. U6 and 18S rRNA genes were used as the miRNA-21 and RASA1 endogenous reference controls, respectively. β-actin was used for the rest of the genes as an endogenous reference control. The relative gene expression level was calculated using the 2^−ΔΔCq method (29). Specific sense primers are presented in Table I.

The Eca-109 (2×10⁶) cells were subcutaneously injected in the posterior right gluteal area. When the tumor volume reached ~0.1 mm³, the mice were randomly assigned into 2 groups. One group received an injection with antagomir NC (10 nmol, 50 µl) and the other antagomir (10 nmol, 50 µl) every 3 days with 5 mice in each group. The tumor volume was measured before and every 3 days after treatment. The volume was calculated with the following formula: V = 1/2 (a × b²), in which a was the long diameter and b was the short diameter of the tumor. Twenty-one days after treatment, the mice were euthanized with CO₂ and the tumor tissues were dissected.

The data was calculated and analyzed with Excel and SPSS 13.0 packages (SPSS, Inc., Chicago, IL, USA). Data was expressed as the mean ± standard deviation (SD) of separate experiments. Statistical analysis was presented by the repeated measure, independent-samples t-test and one-way analysis of variance (ANOVA) followed by Dunnett's post hoc test. P<0.05 was regarded as a significant difference.

A total of 627 DE miRNAs were identified. The top 20 DE miRNAs (top 20 upregulated and 20 downregulated miRNAs) was analyzed with the Morpheus online tool to form a heat map, which was screened by SNR value. The heat map revealed that miR-21 was one of the principal genes and was selected for further research (Fig. 1).

The cell colonies assay revealed that the cell colony formation rate was significantly decreased in the miR-21 inhibitor group compared with the inhibitor NC group, while the rate was significantly increased in the miR-21 mimic group compared with the mimic NC group (P<0.05) (Fig. 2). The results of the scratch assay revealed that, the cell healing rate was significantly inhibited in the miR-21 inhibitor group compared with the inhibitor NC group after the cells were transfected for 48 h. Conversely, the healing rate was significantly enhanced in the miR-21 mimic group and the siRAS1 group compared with the mimic NC and siNC group, respectively (all P<0.05) (Fig. 2).

The number of migrated and invasive cells was significantly decreased in the miR-21 inhibitor group compared with the inhibitor NC group (all P<0.05). Conversely, the cells were significantly increased in the miR-21 mimic group compared with the mimic NC group (all P<0.05). Notably, we observed that the number of migrated and invasive Eca-109 cells was also significantly increased in the siRASA1 group compared with the siNC group (all P<0.05). Similar results were also obtained with the KYSE510 cells (Figs. 3 and 4).

The sequence of RASA1 3′UTR which was combined to miR-21 (mature sequence: UAGCUUAUCAGACUGAUGUUGA) was amplified and detected at 1043 bp using electrophoresis. The vector with mutated or wild-type RASA1 was constructed and used for the luciferase reporter assay (Fig. 5A-C).

The result of the dual-luciferase reporter assay revealed that the luciferase activity in the miR-21 group transfected with the RASA1 vector was significantly inhibited compared with the NC and the blank group (P<0.05) (Fig. 5D), while the activity in the miR-21 group transfected with the RASA1 mutated vector was not significantly altered compared with the NC and the blank group (P>0.05) (Fig. 5E).

It predicted that RASA1 may be the target gene of miR-21 by miRWalk. Their relative expression by PCR revealed that miR-21 was significantly increased and RASA1 was decreased in the tumor tissues compared to the normal adjacent tissues in patients with ESCC (Fig. 1B). The gene level of RASA1 was significantly increased when the Eca-109 cells were transfected with the miR-21 inhibitor compared with the inhibitor NC (P<0.05). Conversely, the expression was significantly reduced in the miR-21 mimic group when compared with the mimic NC group (P<0.05). The expression of miR-21 was significantly increased in the group of cells which were transfected with siRASA1 when compared with the siNC group (P<0.05) (Fig. 5F). The relative gene expression of Snail and vimentin was significantly increased by upregulated miR-21 or downregulated RASA1 (Fig. 6).

Nude mice bearing tumors were established to evaluate the effects of miR-21 inhibitor (antagomir) on esophageal carcinoma growth in vivo. Compared to the NC group, application of antagomir significantly inhibited the growth of the tumor 6 days after treatment (P<0.05). The mice volume in the 2 groups was not significantly different before treatment (P>0.05) (Fig. 7). No diarrhea or body weight loss was observed in any of the animals.