The human ESCC cell line, Eca109, was obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences, and was cultured in RPMI-1640 with 10% fetal bovine serum (FBS). A total of 64 pairs of human ESCC samples were obtained from patients who underwent surgery at the Tangdu Hospital (Xi’an, China) between 2011 and 2012. The corresponding normal esophageal mucosa samples were obtained from the farthest margin of the resected specimen. All the samples were fixed in 10% formalin and embedded in paraffin. Following H&E staining, the tumor tissue sections with maximum tumor content were carefully selected. This study was approved by the Research Ethics Review Committee at Tangdu Hospital. All patients signed informed consent approving the use of their tissues for research purposes.

Total RNA was extracted and purified from 8 pairs of ESCC tumors and corresponding normal esophageal mucosa samples using a RecoverAll Total Nucleic Acid Isolation kit (AM1975; Ambion, Austin, TX, USA) following the manufacturer’s instructions. Prior to microarray analysis, the quality of the RNA was assessed using the Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). The miRNA was distinguished from the remaining RNA using the miRNA Complete Labeling and Hyb kit (Agilent Technologies). Subsequently, each miRNA microarray slide (Agilent Version 16.0) was hybridized with 100 ng labeled RNA and scanned using an Agilent Microarray Scanner (Agilent Technologies). The data were normalized and analyzed using the quantile algorithm and Gene Spring Software 11.0 (Agilent Technologies). Hierarchical cluster analysis was performed using Cluster 3.0 software. The fold change was calculated by comparing the expression level of miRNAs in the ESCC tumor pool to that of the normal tissue pool.

Total RNA was extracted from the paraffin-embedded samples using a miRNeasy FFPE kit (Qiagen, Valencia, CA, USA). A total of 8 sections of 5 μm in thickness were freshly cut and placed in a 1.5 ml microcentrifuge tube. Subsequently, 1 ml xylene was added for deparaffinization. The samples were then washed in ethanol washes. Proteins were degraded by proteinase K digestion and genomic DNA was removed with DNase I step. The concentrated RNA was purified using RNeasy MinElute Spin Columns and eluted in 20 μl RNase-free water. The quality and quantity were measured using a NanoDrop spectrophotometer (Thermo Scientific, Waltham, MA, USA).

RNA was reverse transcribed using a miScript Reverse-Transcription kit (Qiagen) for first strand complementary DNA synthesis. The reaction conditions involved the incubation of the samples at 37°C for 60 min followed by 95°C for 5 min. Quantitative PCR (qPCR) was then performed using primers for miR-1 (miScript Primer Assay: MS00008358) and RNU6B (miScript Primer Assay: MS00033740) and the miScript SYBR-Green PCR kit (both from Qiagen) with the ABI 7500 Fast Real-time PCR System (Applied Biosystems, Carlsbad, CA, USA). Briefly, the samples were incubated at 95°C for 15 min, and qPCR was repeated for 40 cycles with incubation at 94° for 15 sec followed by 55°C for 35 sec and then at 70°C for 30 sec. These reactions were run in triplicate.

The RNA mimic for miR-1 and the mimic negative control were purchased from GenePharma Co., Ltd. (Shanghai, China). The Eca109 cells were transfected with 100 nM mimic or the mimic negative control using Lipofectamine 2000 in accordance with the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). The expression level of miR-1 was assayed by RT-qPCR. For foci formation assay, 200 transfected cells were seeded in 60-mm Petri dishes. Surviving colonies (50 cells per colony) were counted using Giemsa staining following 14 days in culture. For MTT assay, 1,000 Eca109 cells were plated and transfected with the mimic or the mimic negative control in a 96-well culture dish. The cells were then incubated for 24, 48, 72, 96 and 120 h following transfection. The assay was performed by the addition of MTT solution (Sigma-Aldrich, St. Louis, MO, USA) for 4 h and DMSO for 15 min. The absorbance (OD) of each well was then measured at a wavelength of 490 nm. The experiment was performed in triplicate.

The Eca109 cells were seeded in a 6-well culture plate and transfected with either the miR-1 mimic or the negative control after the cells reached subconfluence. Wounds were created using a 200 μl pipette tip. Wound healing was observed by taking photos under a microscope at 0, 24 and 48 h after scratching. Cell invasion assays were performed using chambers consisting of a Transwell pre-coated Matrigel membrane filter (8 μm pore size; Millipore Corp., Billerica, MA, USA). The transfected cells were harvested and transferred to the upper chamber filled with serum-free medium. In the lower chamber, 10% FBS-containing medium was added. After 24 h, the invaded cells were stained and 3 randomly selected fields were counted. Duplicate wells for each condition were examined, and each experiment was repeated 3 times.

The expression level was normalized against endogenous small nuclear RNA gene U6 for the miRNAs examined. The relative quantification of miRNA expression was calculated using the 2^−ΔΔCt method. Statistical analysis was carried out using SPSS 13.0 software for Windows. A paired t-test was used to analyze the differences between the tumor and normal tissue. ANOVA was used to evaluate the association between miRNA expression and the clinicopathological characteristics. Differences were considered statistically significant when the P-value was <0.05.

Given the limitations of any single prediction program, TargetScan, PicTar and miRanda were used to analyze the potential target genes of the deregulated miRNAs. To determine which genes were most likely targeted by the given miRNA, the predicted results from the different databases were integrated. GO analysis was applied to analyze the main function of the differentially expressed genes according to the GO database. DAVID 6.7 was used to annotate the molecular functions of the putative miRNA-targeted genes.

To investigate the role of miRNAs in ESCC, the Agilent Human miRNA microarray (version 16.0) containing 1,205 human miRNAs was performed to compare the miRNA profiles between the ESCC tumor samples and the corresponding normal esophageal mucosa samples pooled from 8 primary cases of ESCC. A paired t-test was used to assess the differentially expressed miRNAs between the cancer and normal esophageal tissue samples and differences at P<0.05 were considered statistically significant. With the 2-fold cut-off point, 27 downregulated and 16 upregulated miRNAs were detected in the ESCC tumor tissue samples compared with their normal counterparts (Fig. 1A). Hierarchical cluster analysis of these 43 miRNAs among the 8 pairs of test samples was performed. The analysis revealed that the cancer tissue samples were grouped separately from the normal esophageal tissue samples and this created 2 major cluster branches (Fig. 1B). However, with the 4-fold cut-off point, 15 downregulated and 9 upregulated miRNAs were detected in the ESCC tumor tissue samples compared with their normal counterparts (Table I).

To confirm the results from microarray analysis, 5 downregulated miRNAs (miR-375, miR-1, miR-133b, miR-30a^* and miR-101) and 4 upregulated miRNAs (miR-196a, miR-424, miR-96 and miR-130b) were selected for validation by RT-qPCR in the same 8 pairs of test samples. All the miRNAs examined showed a pattern of up- or downregulation similar to that obtained by microarray analysis (Fig. 1C).

RT-qPCR was performed to confirm the expression levels of miR-1 in 64 pairs of primary ESCC samples. The fold change was calculated by comparing the relative expression level of miR-1 in the cancer tissue samples and corresponding normal esophageal tissue samples. The small nuclear RNA gene U6 served as a reference gene. The results revealed that miR-1 was significantly downregulated in the primary ESCC samples compared with the corresponding normal esophageal tissue samples. Among all the samples, 90.6% (58/64) of the ESCC tissue samples showed a 2-fold lower expression of miR-1 (Fig. 2).

In addition, in order to determine the clinical significance of the downregulation of miR-1 in primary ESCC, the correlation of miR-1 downregulation with clinicopathological characteristics from 64 patients was investigated. The results revealed that the downregulation of miR-1 significantly correlated with tumor invasion and an advanced clinical stage (P=0.022 and P=0.028, respectively; Table II), while no significant correlation was observed with respect to other clinicopathological characteristics. These results suggest that the downregulation of miR-1 plays an important role in the development of ESCC.

To explore the potential tumor-suppressive role of miR-1 in ESCC, the synthetic miR-1 mimic was transfected into the ESCC Eca109 cells. The expression of mature miR-1 was validated by RT-qPCR (Fig. 3A). MTT assay was used to evaluate cell viability. The assay revealed that the ectopic expression of miR-1 inhibited the cell growth rate (Fig. 3B). To further determine the effects of miR-1 on the growth of ESCC cells, a colony formation assay was performed. The results revealed that the colony-forming ability of the ESCC cells transfected with miR-1 declined by approximately 40% compared with the cells transfected with the negative control (Fig. 3C).

In addition to cell growth inhibition, the effects of miR-1 on tumor migration and invasion were also investigated in this study. Wound healing assays revealed that the cells transfected with the miR-1 mimic migrated very small distances and were unable to achieve wound closure within 48 h (Fig. 4A). To validate that the suppression of mobility was not caused by the inhibitory effect on tumor cell growth, a Matrigel invasion assay was also performed. The results revealed that the number of invading cells was significantly decreased in the miR-1-transfected cells (Fig. 4B).

In order to investigate the potential target and possible mechanisms underlying the regulatory effects of miR-1 in ESCC, we searched for potential targets using TargetScan, PicTar and miRanda. A total of 584, 535 and 3,017 target genes of miR-1 were predicted using the 3 databases stated above, respectively. Among these, 181 genes were common to all 3 databases. The predicted target genes of miR-1 were then classified by GO enrichment analysis. We observed that the significantly enriched target genes of miR-1 were associated with the following: i) transcription regulator activity; ii) sequence-specific DNA binding; iii) the endomembrane system; iv) regulation of transcription; v) vasculature development (Fig. 5).