Seventy pairs of human EC tissues and corresponding adjacent normal tissues were obtained from patients treated in People’s Hospital of Rizhao. All patients signed the informed consent form for the use of samples, and this study was approved by the Ethical Committee on Scientific Research of People’s Hospital of Rizhao. All tissue samples were immediately frozen in liquid nitrogen after surgical removal and stored at −80°C.

Four human EC cell lines (HEC-1A, HEC-1B, RL-95-2, and JEC) and human normal endometrial stromal cells (HESCs) were obtained from the ATCC. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS, Gibco) at 37°C in a 5% CO₂ incubator. Two small interfering RNAs (siRNAs) targeting lncRNA DCST1-AS1, negative control siRNA (si-NC), miR-665 mimic, miR-873-5p mimic, control mimic (miR-NC), miR-665 inhibitor, miR-873-5p inhibitor, control inhibitor, pcDNA-HOXB5, pcDNA-CADM1, and empty vector (pcDNA3.1) were synthesized by GenePharma Co., Ltd. (Shanghai, China). Briefly, EC cells were seeded into six-well plates at 70% cell confluence. Cell transfection was conducted using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the protocol of the manufacturer. After transfection, EC cells were cultured for the indicated time and subjected to the following experiments.

Total RNA was extracted from frozen tissues or cell lines using TRIzol reagent (Invitrogen), and then was reverse transcribed into complementary (cDNA) using PrimerScript RT reagent (TaKaRa, Dalian, China). The quantitative real-time PCR (qRT-PCR) assay was conducted using SYBR Green Mix (TaKaRa) on a Bio-Rad system. GAPDH was used as the internal control. The primer sequences for qRT-PCR are listed in Table 1.

Cell viability experiment was conducted using a Cell Counting Kit-8 (CCK-8) assay (Dojindo Laboratories, Kumamoto, Japan). Briefly, cells were transfected and seeded into 96-well plates. Twenty-four, 48, and 72 h later, the CCK-8 solution was added to each well and incubated for 4 h at 37°C. The absorbance of each well was measured at 450 nm wavelength using a microplate reader (PerkinElmer, Shanghai, China). All assays were performed in triplicate.

EC cells were seeded on six-well plates and maintained in the culture media at 37°C for 2 weeks. Then, cells were fixed with 4% paraformaldehyde for 10 min and stained using 0.1% crystal violet solution for 30 min at room temperature. The number of colonies was counted using ImageJ software.

The transfected cells were seeded in the top chamber in 200 µl of serum-free DMEM medium (Gibco) with Matrigel (BD Biosciences, San Jose, CA, USA), and a complete medium (750 µL) containing 10% FBS was added to the lower chamber. After 24 or 48 h, the migrated or invaded cells were fixed and stained with crystal for 15 min at room temperature. Five random fields per well were observed, and the cells were counted under the microscope.

The apoptosis of EC cells was examined using an Annexin V-FITC/propidium iodide (PI) staining assay (BD Biosciences). After washing with cold PBS, the cells were resuspended in binding buffer followed by staining with Annexin V-FITC/PI at room temperature for 15 min in the dark. Apoptotic cells were evaluated by a fluorescence-activated cell sorting (FACS) flow cytometer (BD Biosciences). All experiments were performed in triplicate.

The wild-type (WT) fragments of DCST1-AS1 3′-untranslated region (UTR), HOXB5 3′-UTR, and CADM1 3′-UTR were synthesized and cloned into pGL3 vector (Promega Corporation, Madison, WI, USA).

Mutations (MUT) of the miRNA-binding sites in the DCST1-AS1, HOXB5 3′-UTR, and CADM1 3′-UTR were generated using a QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). Then, miR-665 mimic, miR-873-5p mimic, or control mimic was co-transfected with the indicated reporter plasmids into EC cells using Lipofectamine 2000 (Invitrogen). Luciferase assays were performed using the Dual-Luciferase Reporter Assay System (Promega) according to the instructions of the manufacturer.

Animal experiments were approved by the Institutional Animal Care and Use Committee of the People’s Hospital of Rizhao. Four- to 6-week-old female nude mice were purchased from the Shanghai Laboratory Animal Center of the Chinese Academy of Sciences (Shanghai, China) and were randomly divided into two groups for further study. We established stable EC lines using shRNA against lncDCST1-AS1 or control shRNA. EC cells with stable knockdown of lncDCST1-AS1 or control cells were injected subcutaneously into the flanks of nude mice. Tumor volume was calculated as length (mm) × width² (mm²) × 0.5. After 40 days, tumors were harvested and weighed.

The expression level of lncRNA DCST1-AS1, miR-665, miR-873-5p, HOXB5, and CADM1 was examined using the Encyclopedia of RNA Interactomes (ENCORI) database (http://starbase.sysu.edu.cn/). The online website tool DIANA-TarBase (https://carolina.imis.athena-innovation.gr/diana_tools/web/index.php) was used to predict the interaction between microRNAs, DCST1-AS1, and mRNAs.

All data were presented as means ± standard deviation. Student’s t-test was used to compare the difference between the two groups. One-way analysis of variance (one-way ANOVA) was used to analyze the differences among multigroups. All of the statistical calculations were performed using GraphPad Prism 5 software (San Diego, CA, USA). P <0.05 was considered statistically significant: *P < 0.05, **P < 0.01, and ***P < 0.001.

We first compared the expression of lncRNA DCST1-AS1 in normal endometrial tissues (n = 35) and EC tissues (n = 548) using the online database ENCORI. As shown in Figure 1A, lncRNA DCST1-AS1 was significantly upregulated in EC tissues compared with normal endometrium tissues. Our qRT-PCR analysis of lncRNA DCST1-AS1 expression from 70 pairs of EC tissues and adjacent normal tissues showed that lncRNA DCST1-AS1 was overexpressed in EC patients (Figure 1B). Consistently, the expression of lncRNA DCST1-AS1 was significantly higher in EC cell lines than that in human normal endometrial stromal HESC cells (Figure 1C). Furthermore, 70 EC patients were classified into lncRNA DCST1-AS1 low and lncRNA DCST1-AS1 high groups. The results revealed that the expression of DCST1-AS1 was not related to the age and gender of the patients, but higher expression of DCST1-AS1 was significantly associated with larger tumor size, advanced TNM stage, and lymph node metastasis (Table 2).

Survival analysis using the Kaplan–Meier (KM)-plotter database suggested that those patients with high DCST1-AS1 expression had worse overall survival (OS) and progression-free survival (PFS) than those with low DCST1-AS1expression (Figure 1D). All these results suggested that DCST1-AS1 is overexpressed in EC and correlated with poor clinical outcomes.

To investigate the role of lncRNA DCST1-AS1 in EC, we silenced the expression of lncRNA DCST1-AS1 in EC cell lines using siRNAs and examined the effect of DCST1-AS1 knockdown on EC cell proliferation, migration, and invasion. First, we confirmed the transfection efficiency by qRT-PCR analysis (Figure 2A). Then, CCK-8, colony formation, migration, and invasion assays suggested significant inhibition of proliferation, growth, migration, and invasion of EC cells transfected with DCST1-AS1 siRNA compared with those transfected with control siRNA (Figures 2B–E). Moreover, silencing of lncRNA DCST1-AS1 enhanced cell apoptosis (Figure 2F). Together, these results suggested that lncRNA DCST1-AS1 functions as an oncogenic lncRNA in EC.

To explore how DCST1-AS1 exerts its function, we predicted miRNAs that can bind with DCST1-AS1 using the online database. The results showed that DCST1-AS1 contains the putative binding sites for miR-665 and miR-873-5p (Figure 3A). Using luciferase reporter assays, we showed that overexpression of miR-665 and miR-873-5p reduced the luciferase activity of the wild-type DCST1-AS1 reporter gene, but not the mutant DCST1-AS1 reporter gene (Figures 3B, C).

Furthermore, we found that silencing of lncRNA DCST1-AS1 increased the expression of miR-665 and miR-873-5p (Figures 3D, E). We tested the expression of miR-665 and miR-873-5p in EC tissues and normal tissues using the ENCORI database. We demonstrated that the expression of both miR-665 and miR-873-5p was significantly downregulated in EC samples (Figures 3F, G). The qRT-PCR assays showed that the expression of DCST1-AS1 was negatively correlated with miR-665 and miR-873-5p expression (Figure 3H). These results indicated that DCST1-AS1 directly sponges miR-665 and miR-873-5p in EC cells.

To find out genes sharing the regulatory role of DCST1-AS1 and miR-665, we predicted the target genes of miR-665 using the DIANA database. Among the predicted genes, HOXB5 could be potentially affected by miR-665 (Figure 4A). Our luciferase reporter assay demonstrated that overexpression of miR-665 suppressed the luciferase activity of the wild-type HOXB5 3′-UTR luciferase reporter, while it failed to inhibit the luciferase activity of the mutant HOXB5 3′-UTR (Figure 4A). Western blot analysis suggested that overexpression of miR-665 inhibited the mRNA and protein expression of HOXB5 (Figure 4B), suggesting that HOXB5 is the direct target gene of miR-665.

Since DCTS1-AS1 acts as a sponge of miR-665, we tried to examine whether DCTS1-AS1 could regulate HOXB5 expression via miR-665. We transfected EC cells with DCTS1-AS1 siRNA (or control siRNA), along with miR-665 inhibitor (or control inhibitor), and investigated the mRNA and protein levels of HOXB5 using qRT-PCR and Western blot assays. As shown in Figures 4C, D, silencing of DCST1-AS1 significantly decreased HOXB5 mRNA and protein expression, whereas inhibition of miR-665 rescued the expression of HOXB5. Our meta-analysis and qRT-PCR assays demonstrated that HOXB5 was dramatically upregulated in EC tissues compared with normal tissues (Figure 4E). We also detected a negative correlation between miR-665 and HOXB5 expression and a positive correlation between DCST1-AS1 and HOXB5 expression in EC tissues (Figure 4F). These findings suggested that DCST1-AS1 sponges miR-665 to increase HOXB5 expression in EC cells.

Similarly, using an online database, we found that CADM1 is a potential target gene of miR-873-5p (Figure 5A). The results from luciferase assays suggested a direct interaction between miR-873-5p and CADM1 3′-UTR (Figure 5A). Furthermore, we found that overexpression of miR-873-5p significantly decreased the mRNA and protein levels of CADM1 (Figure 5B). Our rescue experiments supported that knockdown of DCST1-AS1 downregulated the expression of CADM1 in EC cells, while inhibition of miR-873-5p had the opposite effects (Figures 5C, D). Meanwhile, CADM1 was found to be significantly upregulated in EC tissues (Figures 5E, F). The mRNA level of CADM1 was inversely correlated with miR-873-5p expression, but positively correlated with DCST1-AS1 expression in EC tissues (Figure 5F). These results suggested that DCST1-AS1 induces CADM1 expression in EC cells through sponging miR-873-5p.

HOXB5, a member of the HOX gene cluster, is a key regulator of developmental processes (24), and it is implicated in multiple human cancers including breast cancer, head and neck cancer, and bladder cancer (25–27). Since we have shown that DCST1-AS1 promotes HOXB5 expression by acting as a sponge for miR-665, we asked whether DCST1-AS1 facilitates EC progression by regulating the miR-665/HOXB5 signaling. To this end, we constructed a HOXB5 overexpression plasmid (Figure 6A).

Using CCK-8, colony formation, migration, and invasion assays, we confirmed that silencing of DCST1-AS1 suppressed EC cell proliferation, growth, migration, and invasion, whereas either miR-665 knockdown or overexpression of HOXB5 could reverse the effects of DCST1-AS1 silencing (Figures 6B–E). In addition, depletion of DCST1-AS1 increased EC cell apoptosis; however, cell apoptosis was reduced by the transfection with miR-665 inhibitor or HOXB5 overexpression plasmid (Figure 6F). These data suggested that DCST1-AS1 promotes EC cell proliferation, migration, and invasion and inhibited cell apoptosis, at least in part, by regulating the miR-665/HOXB5 pathway.

To explore whether the miR-873-5p/CADM1 axis mediates the function of DCST1-AS1 in EC cells, we constructed a CADM1 overexpression plasmid (Figure 7A). Our rescue experiments revealed that suppression of DCST1-AS1 significantly suppressed EC cell proliferation, migration, and invasion and induced cell apoptosis (Figures 7B–F). In contrast, either knockdown of miR-873-5p or CADM1 overexpression rescued the malignant features of EC cells and reduced cell apoptosis (Figures 7B–F). These data indicated that DCST1-AS1 promotes the aggressive phenotype in EC cells, at least in part, by regulating the miR-873-5p/CADM1 pathway.

To validate the effects of lncRNA DCST1-AS1 on tumor growth in vivo, we injected EC cells transfected with DCST1-AS1 shRNA or control shRNA into nude mice. The results from the tumor xenografts in nude mice showed that the volume and weight of subcutaneous tumors were significantly suppressed by DCST1-AS1 knockdown (Figures 8A–C). Our qRT-PCR analysis of the tumors suggested that knockdown of DCST1-AS1 significantly increased the expression of miR-665 and miR-873-5p, but decreased the expression of HOXB5 and CADM1 (Figures 8D–F). Taken together, our results suggested that DCST1-AS1 promotes EC growth in vivo.