A total of EC tissues and paired adjacent normal esophagus tissues from 66 patients with EC diagnosed by histopathological examination and undergoing resection between 2009 and 2015 were collected. Clinical characteristics of these patients were collected retrospectively for all patients. Informed consent was obtained from all the patients.

EC9706 and KYSE cell lines were obtained from the Chinese Academy of Sciences Cell Bank. Cells were cultured in RPMI-1640 (HyClone Laboratories, Inc., Logan, UT, USA) medium supplemented with 10% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA), 100 U/ml penicillin and 100 U/ml streptomycin. These cell lines were cultured in humidified 5% CO₂ at 37°C.

Total RNA was isolated using TRIZol reagent (Invitrogen, Carlsbad, CA, USA). First strand cDNA was generated using the Reverse EasyScript One-Step gDNA Removal and cDNA Synthesis SuperMix (Transgene Biotek, Ltd., Hyderabad, India). qRT-PCR was performed by using SYBR Green Master Mix (Roche) reagent in ABI 7500 real-time PCR instrument. ACTB was employed as an endogenous control. The relative levels of gene expression were calculated by the 2^−ΔΔCt method. Primers used for RT-RCR are as follows: UCA1-F, 5′-GAG GATTCCCAGCCATATGAAG-3′ and UCA1-R, 5′-CGGCAGTTGGTGTGCTATAA-3′; Sox4-F, 5′-CCAACTCCTTAGTGCCGATT-3′ and Sox4-R, 5′-TCTCTAGCACTTAGTCCCTCTC-3′; ACTB-F, 5′-CACCAGGCACCCAGTTTAAT-3′ and ACTB-R, 5′-AGTCTCTGCTCTCTCTTCCTATG-3′.

Cell proliferation was determined by the Cell Counting kit-8 (CCK-8) assay. Cells (2×10³) were seeded in 96-well plates. Each well was incubated with WST-8 solution (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) for 3 h, and the absorbance was measured at 450 nm using a spectrophotometer. All experiments were performed in triplicate.

Western blotting was performed with the standard protocol. Primary antibodies used were rabbit anti-Sox4 (Abcam) and mouse anti-β-actin (Proteintech, Chicago IL, USA).

To construct UCA1 knockdown cells, cells were infected with lentivirus particles expressing shRNA against UCA1 and selected by puromycin (1 μg/ml) for one week. The target sequence is as follow: GCCATATGAAGACACCCTA. To construct stable UCA1 overexpressed cells, cells were transfected with pcDNA3.1 containing full-length UCA1 and selected with G418 (400 μg/ml) for one week.

siRNAs against Sox4 were purchased from Guangzhou RiboBio Co., Ltd. (Guangzhou, China). The target sequence against Sox4 is as follows: GGACTAAGTGCTAGAGACT.

The Sox4 3′ untranslated region (3′UTR) was cloned to luciferase reporter plasmid, pmirGLO. The pmirGLO-Sox4 was cotransfected with miR-NC (negative control) or miR-204 mimics into cells using TurboFect (Thermo Fisher Scientific). The relative luciferase activity was normalized to Renilla luciferase activity 48 h after transfection. The luciferase activity was measured using a Dual-luciferase reporter gene assay system (Promega, Madison, WI, USA).

Anti-MS2 RIP and anti-AGO2 RIP was performed as previously describe by using the EZ-Magna RNA immunoprecipitation kit (Millipore) (9).

All statistical analyses were performed by SPSS 20.0 software. For comparisons, Student's t-test and Pearson's Chi-square test were performed as appropriate. A value of P<0.05 was considered to indicate a statistically significant result.

To determine the role of UCA1 in EC, the UCA1 expressions were detected in EC tissues and their paired adjacent normal tissues from 66 EC patients by using qRT-PCR. We found that UCA1 expression was significantly increased in EC tissues compared with the paired normal tissues (Fig. 1A).

To gain insight to the clinical significance of UCA1 expression, we analyses the correlation of UCA1 expression with clinicopathological features (Table I). These 66 patients were classified into two groups using the median expression level of UCAL1 in EC tissues as the cut-off value. Notably, higher UCA1 expression in EC was significantly associated with poor differentiation and higher TNM stage.

Next, we determined the association of UCA1 and prognosis by performing Kaplan-Meier analysis. The results showed that patients with higher UCA1 expression had a significantly poorer prognosis compared to patients with lower UCA1 expression (Fig. 1B).

To determine the function of UCA1 in EC cells, we constructed stable cells with knockdown or overexpression of UCA1 (Fig. 2A). We then detected the effect of UCA1 on proliferation by CCK-8 assay. The results showed that both UCA1 overexpressed EC9706 and KYSE cells was significantly accelerated compared with that of the control cells (Fig. 2B), while knockdown of UCA1 effectively inhibits cell proliferation (Fig. 2C).

It has been reported that lncRNAs can function as microRNA (miRNA) sponge to regulate the activity of miRNAs (22). A recent study reported that UCA1 could sponge endogenous miR-204 and inhibit its activity in colorectal cancer (16). However, the direct interaction of UCA1 and miR-204 in EC remains unclear. To confirm the direct binding between UCA1 and miR-204, we performed RIP assay to pull down miRNAs interacted with UCA1. The results showed that UCA1 was significantly enriched for miR-204 compared to the empty vector (MS2) and UCA1 with mutations in miR-204 targeting sites (mut-UCA1-MS2) (Fig. 3A and B). In addition, we constructed luciferase reporter plasmids containing the wild-type UCA1 or mutant UCA1 (mut-UCA1) with mutated miR-204 binding sites. We found that miR-204 overexpression decreased the luciferase activities of the wild-type UCA1 reporter vector but not empty vector or mutant UCA1 reporter vector (Fig. 3C). Finally, we performed anti-AGO2 RIP to detect whether UCA1 was regulated by miR-204 in an AGO2-dependent manner (Fig. 3D). Endogenous UCA1 pull-down by AGO2 was significantly enriched in miR-204 overexpressed cells, suggesting that miR-204 is a bona fide UCA1-targeting miRNA.

Recent studies have reported that Sox4 is a target gene of miR-204 (23,24). We suspected that UCA1 may regulate Sox4 expression through functioning as a ceRNA. To confirm this hypothesis, we detected the mRNA and protein level of Sox4 in UCA1 overexpressed and UCA1 knockdown cells. Surprisingly, we found that overexpression of UCA1 increased Sox4 expression, while silence of UCA1 suppressed Sox4 expression (Fig. 4A and B). For rescue experiment, we transiently transfected wild-type UCA1 or mutant UCA1 overexpressed cells with miR-204. Overexpression of wild-type UCA1, but not the mutant UCA1, increased Sox4 expression (Fig. 4C). Overexpression of miR-204 abolished the increase of Sox4 induced by UCA1. In addition, we suppressed miR-204 expression in UCA1 knockdown cells and found that inhibition of miR-204 rescued the decrease of Sox4 induced by UCA1 knockdown (Fig. 4D).

To further confirm whether the UCA1 influenced Sox4 expression depends on the regulation of Sox4 3′UTR, we constructed luciferase reporters containing Sox4 3′UTR (pmirGLO-Sox4). The luciferase plasmid (pmirGLO-Sox4) was transfected into the wild-type or mutant UCA1 overexpressed cells with or without miR-204. We found that overexpression of wild-type UCA1, but not the mutant UCA1, increased the luciferase activity of Sox4 3′UTR, while miR-204 abolished this effect (Fig. 4E). Similarly, the silence of UCA1 decreased the luciferase activity of Sox4 3′UTR, which were rescued by inhibition of miR-204 (Fig. 4F).

We next detected the expression levels of Sox4 in 66 EC and paired normal tissues. As shown in Fig. 4G, UCA1 expression was significantly correlated with Sox4 mRNA expression. All these results suggest an important role of UCA1 in regulating Sox4 expression by competitively binding miR-204.

Finally, we detected whether UCA1 promotes cell proliferation through regulation with Sox4 expression. We transfected siRNA against Sox4 in UCA1 overexpressed cells (Fig. 5A). We found that silence of Sox4 abolished the cell proliferation induced by UCA1 overexpression (Fig. 5B). Consistently, overexpression of Sox4 rescued the inhibition of cell proliferation mediated by UCA1 knockdown (Fig. 5C and D). These results suggest that UCA1 promotes cell proliferation through regulation of Sox4.