ADAMTS9-AS2 regulates PPP1R12B by adsorbing miR-196b-5p and affects cell cycle-related signaling pathways inhibiting the malignant process of esophageal cancer

ABSTRACT

This study explored the mechanism that ADAMTS9-AS2/miR-196b-5p/PPP1R12B/cell cycle pathway axis in inhibiting the malignant progression of esophageal cancer (EC), providing a new idea for targeted molecular therapy of EC. The expression data of EC tissue were acquired from TCGA database. The target lncRNA, downstream miRNA and its target gene were determined by bioinformatics analysis. ADAMTS9-AS2, miR-196b-5p and PPP1R12B levels in EC tissue and cells were assayed through qRT-PCR. Western blot was applied to assess protein level of PPP1R12B in cells and tissues, as well as protein expression of CDK1, cyclin A2, cyclin B1 and Plk1 in EC cells. Cell proliferation was assayed via CCK-8 assay. Cell cycle distribution was analyzed by flow cytometry. Cell migratory and invasive abilities were measured through scratch healing and transwell assays. Pearson correlation analysis was utilized to analyze relationship among ADAMTS9-AS2, miR-196b-5p and PPP1R12B. RIP was introduced to assess binding among the three. Dual-luciferase assay was utilized to verify targeted binding sites. The tumor formation in nude mice assay was utilized to detect tumorigenesis of EC cells in vivo. ADAMTS9-AS2 was significantly lowly expressed while miR-196b-5p was increased in EC tissue and cells. ADAMTS9-AS2 bound to miR-196b-5p and constrained its expression. Overexpressed ADAMTS9-AS2 inhibited EC cell malignant progression via downregulating miR-196b-5p, while overexpressed miR-196b-5p reversed this inhibitory effect. ADAMTS9-AS2 modulated PPP1R12B level by competitively inhibiting miR-196b-5p. PPP1R12B played a modulatory role in EC by inhibiting cell cycle pathway. Overexpressed ADAMTS9-AS2 regulated the tumor-forming ability of EC cells in vivo through miR-196b-5p/PPP1R12B/cell cycle signaling pathway axis. ADAMTS9-AS2 downregulated PPP1R12B by adsorbing miR-196b-5p, so as to regulate the cell cycle signaling pathway to inhibit EC malignant progression.

Introduction

The esophageal cancer (EC) is a malignancy in gastrointestinal tract [1], the 5 y survival rate of which is about 15%–25% [2]. Esophagectomy is a therapeutic option for early EC. When EC is in advanced stage, neoadjuvant chemotherapy and multi-mode management combining surgery and chemoradiotherapy are required [3]. Therefore, it is still very urgent to study the malignant progression of EC.

Long non-coding RNAs (lncRNAs) lack the ability to encode proteins [4]. It was reported that lncRNAs may affect the development of EC, and have potential for diagnosis and treatment of EC [5]. ADAMTS9-AS2 is a novel tumor suppressor [6]. It acts as an antisense transcript of ADAMTS9 in glioma, and the extracellular protease ADAMTS9 inhibits angiogenesis, thereby inhibiting EC and nasopharyngeal cancer formation [7,8]. ADAMTS9-AS2 deletion upregulates miR-130a-5p, thereby promoting breast cancer resistance to tamoxifen [9]. ADAMTS9-AS2 is related to promoting tongue squamous cell carcinoma progression via regulating miR-600/EZH2 axis [10]. ADAMTS9‐AS2 constrains biological processes of ovarian cancer through directly mediating miR-182-5p/FOXF2 axis, thereby becoming a basic biomarker for ovarian cancer treatment [11]. In addition, it can inhibit the progression of EC by mediating CDH3 promoter methylation [1]. However, the mechanism of ADAMTS9‐AS2 in EC is still worth exploring.

In this work, it was found that ADAMTS9-AS2 was underexpressed in EC tissue and cells. Then, mechanism of ADAMTS9-AS2 targeting PPP1R12B by adsorbing miR-196b-5p to regulate cell cycle signaling pathway, thereby repressing EC malignant progression was further studied. This study identified the role of ADAMTS9-AS2 in occurrence and development of EC, so as to disclose new targets for EC treatment.

1. Methods

1.1. Bioinformatics analysis

Expression data of mature miRNAs (Normal: 13, Tumor: 185) and mRNAs (Normal: 11, Tumor: 160) of EC were accessed from The Cancer Genome Atlas (TCGA). LncRNAs, miRNAs, and mRNAs in normal and tumor groups were analyzed differentially (|logFC|>1.5, padj<0.05) using “edgeR” package to obtain differential lncRNAs (DElncRNAs), miRNA (DEmiRNAs), and mRNA (DEmRNAs). Target lncRNA was confirmed by literature citation. LncBase database was used to predict downstream target miRNAs of the target lncRNA, and the target miRNA was obtained by taking the intersection of the downstream target gene miRNAs and the DEmiRNAs. Target mRNAs of the target miRNA were predicted by miRDB, mirDIP, miRWalk and starBase databases, and the target gene mRNA was determined by overlapping predicted target mRNAs and the DEmRNAs. GSEA was used for target gene mRNA pathway enrichment analysis.

1.2. Tissue samples

The 50 EC tissue and corresponding adjacent normal tissue samples from 50 patients with EC (41 males and 9 females, mean age 59.37 ± 6.4 y old) were collected from the First Affiliated Hospital of Zhengzhou University from January 2015 to December 2017. Patients meeting the following criteria could be included in this study: the surgically removed nodules were pathologically confirmed as EC; no other treatments were received previously; the nodules were completely excised and the surface of the adjacent normal tissue was verified by pathological examination; complete clinicopathologic and follow-up data. If the patient died from a non-esophageal disease or accident, his (her) tissue samples were excluded. The surgically removed cancer tissue was immediately cryopreserved in a liquid nitrogen tank. The study was approved by patients, complied with the Declaration of Helsinki and got the approval from the Ethics Committee of the First Affiliated Hospital of Zhengzhou University.

1.3. Cell culture

Human normal esophageal epithelial cell line HEEC was accessed from Shanghai Xinyu Biotechnology Co., Ltd. (No. XY-H253). Human EC cell lines EC109 (No. 3111C0001CCC000246) and NEC (No. 3111C0001CCC000259) were bought from Cell Resource Center, Institute of Basic Medicine, Chinese Academy of Medical Sciences. Human EC cell lines TE-1 (No. 3131C0001000700089) and TE-10 (No. 3131C0001000700090) were accessed from The Cell Resource Center, Shanghai Institute of Life Sciences, Chinese Academy of Sciences. HEEC cell line was cultivated in RPMI-1640 medium (Gibco, Carlsbad, CA) with 10% FBS. Cell lines EC109, NEC, TE-1 and TE-10 were cultivated in RPMI-1640 (w/o Hepes) medium with 10% FBS. All cells were maintained under routine conditions.

1.4. Cell transfection and grouping

ADAMTS9-AS2 and PPP1R12B were amplified by RT-qPCR. Then amplified products were cloned into lentivirus gene overexpression vectors pLV-EGFP-N (Cat. No. VL3211; Inovogen Tech Co Beijing, China) at ECOR1 and NOT1 sites by using the cold fusion kit (System Biosciences [SBI], Mountain View, CA). Stable transfected cell lines were obtained through introducing lentivirus particles with empty vector or overexpressed ADAMTS9-AS2 and PPP1R12B into cells. Gene knockout expression cell line was generated by inserting shRNA sequences into gene silencing vector pSIH1-H1-copGFP (lentivirus shRNA fluorescent expression vector). PPP1R12B shRNA and shRNA negative control (sh-NC) were synthesized by Shanghai Genepharma Co., LTD. MiR-196b-5p mimic and inhibitor were purchased from RiboBio, Guangzhou, China. EC109 cells (No. ZY-H010, ATCC, USA) and NEC cells (resource number: 3111C0001CCC000259) were placed in RPMI-1640 complete medium plus 10% FBS. Lentiviruses were collected and transfected into EC109 and NEC cells with associated vectors. EC cells were isolated by trypsin and the cells were separated into cell suspension (5 × 10⁴ cells/mL). The suspension was added in 6-well plates (2 mL/well) at 37°C overnight. Lentiviral particles (1 × 10⁸ TU/mL) were introduced into cells. A fluorescence microscopy was introduced to assess transfection efficiency 48 h after infection.

1.5. qRT-PCR

Total RNA from esophageal cancer tissues and cells was isolated with Trizol kit (Sigma-Aldrich). RNA was reversely transcribed into cDNA with PrimeScript RT Kit (Takara Biotechnology Ltd., Dalian, Liaoning, China). SYBR premixed Ex-Taq-II kit (Tli-RNaseH-Plus, Takara Biotechnology Co., LTD) was used for qPCR. Then, Thermal Cycler Dice Real-Time System amplification device (TP800; Takara Biotechnology Co., LTD.) was used for qRT-PCR. The qRT-PCR amplification procedure was performed by a two-step method: pre-denaturation at 95°C for 30 s; PCR reaction with 30 cycles of 95°C for 5 s and 60°C for 34 s, and three replicate wells were set for each sample. Primer sequences are listed in Table 1. Primers for template amplification (synthesized by Ribobio, Guangzhou, China) were designed on Primer Express 2.0 software. MiR-196b-5p took U6, while ADAMTS9-AS2 and PPP1R12B took GAPDH as the internal references. Formula 2^−ΔΔCt was used for calculating relative expression.

Table 1.

Primer sequences

Gene	Primer sequence
miR-196b-5p	F: 5’-AGAGATTAGCCGCAGCGTCC-3’
	R: 5’-TTCCTGGGCAGAGCCTGAAG-3’
U6	F: 5’-CTACTACGACGGGGATGTTGG-3’
	R: 5’-GAGTCATGCGGATTCGGTGAG-3’
ADAMTS9-AS2	F: 5’-CGTAGGGTTTAAATAGGTGATAACG-3’
	R:5’-AAATAATAAAAACTCGAACACCGAA-3’
PPP1R12B	F: 5’-TGTAGGGTTTAAATAGGTGATAATGA-3’
	R: 5’-AAATAATAAAAACTCAAACACCAAA-3’
CDK1	F: 5’-GGGGTCAGCTCGTTACTCAA-3’
	R: 5’-TGACATGGGATGCTAGGCTT-3’
GAPDH	F: 5’-CACCCACTCCTCCACCTTTG-3’
	R: 5’-CCACCACCCTGTTGCTGTAG-3’

1.6. Western blot

48 h after transfection, EC109 and NEC cells were harvested. Proteins were lysed and extracted in RIPA lysis buffer (CWBIO). Protein concentration was measured using the BCA protein assay kit (Thermo Fisher). Then, the same amount of proteins were added to 10% polyacrylamide gel. After separated by SDS-PAGE, the proteins were transferred to nitrocellulose membrane (Millipore, Bedford, MA, USA). Membrane was then sealed with 5% skim milk for 1 h at room temperature and cultured with primary antibodies overnight at 4°C. After rinsing, membrane was maintained with HRP-conjugated secondary antibody at 37°C for 1 h. Then, membrane was washed again and the protein bands were examined with ECL detection kit (Pierce Biotechnology, USA). Primary antibodies were rabbit anti-PPP1R12B (ab76785, 1:100), anti-CDK1 (ab131450, 1:500), anti-cyclin A2 (ab181591, 1:2000), anti-cyclin B1 (ab181593, 1:2000), anti-Plk1 (ab155095, 1:1000) and anti-GAPDH (ab181602, 1:10,000). Secondary antibody was goat anti-rabbit (ab181602, 1:10,000). All of the antibodies were purchased from Abcam (Cambridge, UK). Image J software (http://rsb.info.nih.gov/ij/, Bethesda, USA) was employed to quantify strength of protein bands.

1.7. Immunohistochemistry assay

EC or nude mice tumor tissue was fixed in 4% paraformaldehyde solution, paraffin-embedded, cut into 5-mm sections, and then routinely dewaxed. Following antigen repair, sections were blocked with normal goat serum sealant. Known positive sections of EC were taken as positive controls, and PBS (Phosphate Buffer Saline) was used instead of primary antibodies as negative control (NC). Histostain TMSP-9000 immunohistochemical staining kit (Zymed Laboratories, California) was used to stain sections. Primary antibodies rabbit anti-PPP1R12b (ab76785, 1:200) and anti-CDK1 (ab131450, 1:100) were added into the sections and cultivated overnight at 4°C. Following reheating and rinsing with PBS, sections were added with the secondary antibody goat anti-rabbit (ab6721, 1:1000) and incubated at 37°C for 30 min and were then maintained with a horseradish-labeled working solution. DAB was recommended to color sections for 5–10 min. Following counterstaining with hematoxylin for 1 min, sections were blocked with neutral gum, dried, and photoed. Positive expression of the cells was presented as brown.

1.8. CCK-8 assay

Cells transfected for 24 h were plated into 96-well plates (2 × 10³ cells/well) and cultured at 37°C. A 10 μL CCK-8 reagent solution (Beijing Solarbio Science & Technology, China) was supplemented to the cells, followed by cell incubation at 37°C for 1 h. Cell viability was assessed at 0, 24, 48, and 72 h, respectively. OD value was measured at 450 nm and proliferation curve was drawn. The assay was repeated three times.

1.9. Scratch healing assay

EC109 cells and NEC cells transfected for 24 h were planted in 6-well plates (5 × 10⁵ cells/well), and maintained at 37°C for 12 h. A scratch was created in the middle of each well with pipette tip, and cells were maintained for 24 h after washed. ImageJ software was used to image wound and measure wound closure.

1.10. Transwell invasion assay

Cell invasion was evaluated in serum-free DMEM in the Transwell chamber (Millipore, USA) coated with 1:6 diluted Matrigel. EC109 cell and NEC cell suspension was prepared in serum-free medium 24 h after transfection. The upper chamber was added with 500 μL cell suspension (3 × 10⁴ cells) and the lower chamber with 1 μL culture medium containing FBS. After incubation overnight, cells in the upper chamber were swabbed. After PBS wash, the invaded cells were fixed with 4% paraformaldehyde for 30 min and dyed with 0.1% crystal violet for 20 min. Following rinsing with PBS, three fields were selected randomly under the microscope and photoed (×100) for observation and counting.

1.11. Flow cytometry

Stable EC109 cells and NEC cells were harvested, rinsed with PBS, resuspended in 75% precooled ethanol and fixed overnight at 4°C. After ethanol was dropped off, cells were suspended with 500 μL PBS and added with 20 μL RNAse A (100 μg/mL) at 37°C for 30 min. Fixed cells were stained with 400 µL propidium iodide (50 μg/mL) at 4°C in the dark for 30 min. Cell cycle was analyzed by flow cytometry (BD Biosciences, USA).

1.12. Dual-luciferase reporter gene analysis

The mutant (MUT) or wild-type (WT) binding sites of miR-196b-5p on ADAMTS9-AS2 or PPP1R12B were inserted into pmirGLO vectors. Luciferase vector PRL-TK (Promega) and constructed luciferase vectors were transfected into EC cells of mimic NC or miR-196b-5p mimic group, respectively. 48 h later, luciferase activity was assessed on dual-luciferase reporter assay system (Promega), and it was normalized relative to renilla luciferase activity.

1.13. RNA Immunoprecipitation (RIP)

EC109 cells (1 × 10⁷ cells) and NEC cells (1 × 10⁷ cells) were lysed in the RIP buffer by using Magna- RIP-RNA binding protein immunoprecipitation kit (Merck Millipore, MA). To bind target proteins, cell extracts were cultured at 4°C in RIP buffer which contained magnetic beads and coated with Ago2 (ab186733, 1:50) or immunoglobulin G (IgG; ab205718, 1:100). Antibodies above were purchased from Abcam. After the buffer was centrifuged instantaneously, the centrifuge tube was placed on the magnetic separator and supernatant was discarded. Magnetic bead antibody complex was rinsed six times with 500 μL RIP washing buffer. In the last wash, 50 μL was isolated from a total of 500 μL buffer system and supernatant was discarded. After buffer solution was added, western blot was used to assess whether magnetic beads were coated with antibodies. Samples were maintained with protease K at 55°C for 30 min to isolate proteins. Trizol was then applied to extract immunoprecipitated RNA. qRT-PCR was used to assay enrichment of ADAMTS9-AS2 and miR-196b-5p.

1.14. Transplanted tumor in nude mice assay

Twelve female nude mice (4–5 wks old, 18–21 g in weight) with no specific pathogen free (SPF) grade were selected and purchased from Xipuer-Bikai, Shanghai. 24 h later, EC109 cells in each group were subcutaneously transplanted into mice. Prior to transplantation, cells were mixed with Matrigel in a 1:1 ratio. A 0.2 mL suspension (3 × 10⁶ cells) was inoculated into the right armpit of mice. Tumor size was assessed and recorded weekly. The tumor size was calculated as follows: V (mm³) = (L × D [2])/2, where L is the maximum long diameter of the tumor and D is the minimum short diameter of the tumor. Tumor growth curve was then plotted. After 5 wks, nude mice were killed by CO₂ method. Tumors were collected and weighed. Subsequently, tumors were removed and divided into two parts, one in 4% paraformaldehyde and the another in liquid nitrogen.

1.15. Analysis of statistics

Statistical analysis was done on GraphPad Prism 8 Software (GraphPad Software Inc., USA). The measured data were expressed as mean ± SD. Paired t-test was used to compare the paired design data that followed normal distribution and homogeneous variance between two groups. Unpaired t-test was utilized to compare the unpaired design data that followed normal distribution and homogeneous variance between two groups. One-way ANOVA was utilized to compare multiple data sets. Tukey method was used for validation. Pearson correlation analysis was utilized to assess express correlation. P < 0.05 indicates statistically significant difference, while p < 0.01 indicates extremely significant difference.

2. Results

2.1. ADAMTS9-AS2 is significantly lowly expressed in EC cells and can inhibit malignant progression of EC cells

A differential analysis was conducted based on EC-related expression data from TCGA database, where 908 significant differently expressed lncRNAs were screened out (Figure 1a). Among them, ADAMTS9-AS2 was significantly low-expressed in EC tissues (Figure 1b), yet has rarely been researched in EC. Therefore, to identify the molecular mechanisms and biofunctions of ADAMTS9-AS2 in EC, qRT-PCR assayed ADAMTS9-AS2 level in 126 EC tissue and adjacent normal tissue samples. It was found that compared with the normal samples, the ADAMTS9-AS2 expression in the EC tissues was noticeably downregulated (p < 0.05) (Figure 1c). qRT-PCR was again employed to assess ADAMTS9-AS2 level in human normal esophageal epithelial cell line HEEC and 4 human EC cell lines (EC109, NEC, TE-1, and TE-10). The results displayed that in relevant to HEEC cell line, the ADAMTS9-AS2 expressions in EC109, NEC, TE-1 and TE-10 cell lines were notably downregulated (p < 0.05), in which EC109 and NEC were selected for the following assays (Figure 1d).

Figure 1.

ADAMTS9-AS2 is significantly lowly expressed in EC cells and can inhibit malignant progression of EC cells.

A: Volcano map of DElncRNAs of normal and tumor groups; Red dots: upregulated genes; green dots: downregulated ones; B: The ADAMTS9-AS2 level in the EC data set, green box plot: normal samples; red box plot: tumor samples; C: ADAMTS9-AS2 level in EC tissue and anormal tissue assayed through qRT-PCR (n = 50, p < 0.0001); D: ADAMTS9-AS2 expression in HEEC, EC109, NEC, TE-1 and TE-10 detected via qRT-PCR (p = 0.0002, p = 0.0005, p = 0.0009, p = 0.0048); E: ADAMTS9-AS2 expression in cells in oe-NC group and oe-ADAMTS9-AS2 group assessed through qRT-PCR after ADAMTS9-AS2 was overexpressed (EC109 p = 0.0004; NEC p = 0.001); F: Cell viability of EC109 cells and NEC cells in each group assessed by CCK-8 after ADAMTS9-AS2 was overexpressed; G: Cell cycle distribution of EC109 cells and NEC cells measured via flow cytometry after ADAMTS9-AS2 was overexpressed (EC109 p < 0.0001; NEC p < 0.0001); H: Migratory capability of EC109 cells and NEC cells assessed via scratch healing assay after ADAMTS9-AS2 was overexpressed (40×) (EC109 p = 0.0005; NEC p = 0.0006); I: Invasive property of EC109 cells and NEC cells assayed through Transwell after ADAMTS9-AS2 was overexpressed (100×). * p < 0.05, ** p < 0.01. Each assay was repeated three times (EC109 p = 0.0005; NEC p = 0.0027).

Next, the effects of ADAMTS9-AS2 on the malignant phenotypes of EC cells were examined. The ADAMTS9-AS2 upregulated cell lines were constructed based on EC109 and NEC cell lines, and qRT-PCR assessed ADAMTS9-AS2 level in cells. ADAMTS9-AS2 level in the oe-ADAMTS9-AS2 group was prominently upregulated relevant to the oe-NC group (p < 0.05) (Figure 1e). CCK-8 assays were applied to examine cell activity of each group, where the results presented that in comparison to the control group, the cell activity in the oe-ADAMTS9-AS2 group was evidently reduced (p < 0.05) (figure 1f). Flow cytometry assay was employed to evaluate the cell cycle condition of each group with the results exhibiting that the number of cells in G0/G1 phase in the oe-ADAMTS9-AS2 group was relatively increased, while the number of cells in S phase was decreased compared to the control (p < 0.05) (Figure 1g). Also, the cell migratory capability was detected by scratch healing assay, in which observed that upregulating ADAMTS9-AS2 could significantly reduce cell migratory ability (P < 0.05) (Figure 1h). Transwell assay was carried out to investigate cell invasive ability. It was observed that, compared with oe-NC group, cell invasive trend in the oe-ADAMTS9-AS2 group was significantly reduced (p < 0.05) (Figure 1i). The above results indicated that ADAMTS9-AS2 was decreased in EC cells, and overexpressed ADAMTS9-AS2 had repressive effects on EC cell phenotypes.

2.2. ADAMTS9-AS2 binds to miR-196b-5p and suppresses its expression

Some researchers reported that ADAMTS9-AS2 can regulate gene expression as a ceRNA in EC [12,13]. To extensively understand the molecular mechanisms of ADAMTS9-AS2 in EC, its downstream target miRNAs were predicted bioinformatically. Mature miRNA expression data samples of EC were accessed from TCGA and 85 DEmiRNAs were screened by a differential analysis (Figure 2a). The lncBase database was introduced to predict downstream target miRNAs of ADAMTS9-AS2, followed by overlapping the upregulated DEmiRNAs and the predicted candidates obtaining eight final candidates (Figure 2b). Via Pearson correlation analysis for ADAMTS9-AS2 and the eight miRNAs, we discovered that miR-196b-5p was negatively associated with ADAMTS9-AS2 with the highest correlation coefficient (Figure 2c). The differential analysis of miRNA expression samples in EC tissue of TCGA database indicated that miR-196b-5p was notably increased in EC tissue (Figure 2d). MiR-196b-5p is overexpressed in various cancers [14–16]. Therefore, miR-196b-5p was used as a potential downstream target of ADAMTS9-AS2 and verified. Firstly, miR-196b-5p level in EC and normal tissues was assessed by qRT-PCR, which displayed that miR-196b-5p in EC tissue was noticeably upregulated compared with that in adjacent normal tissue (p < 0.05) (Figure 2e). Pearson correlation analysis assessed correlation between expression of ADAMTS9-AS2 and miR-196b-5p in clinical samples of EC. The results denoted that the negative association of ADAMTS9-AS2 level and miR-196b-5p level (p < 0.05) (figure 2f), which was consistent with the trend of correlation analysis of database samples. qRT-PCR assessed miR-196b-5p level in human normal esophageal epithelial cells HEEC and 4 human EC cell lines. MiR-196b-5p in EC109, NEC, TE-1 and TE-10 cell lines was remarkably upregulated compared with that in HEEC cells (p < 0.05) (Figure 2g). Targeted binding sites of ADAMTS9-AS2 and miR-196b-5p were predicted by lncBase database (Figure 2h), and binding of ADAMTS9-AS2 and miR-196b-5p was assayed via RIP (Figure 2i). In relevant to IgG group, ADAMTS9-AS2 and miR-196b-5p in Ago2 group were dramatically enriched (p < 0.05). Dual-luciferase assay validated the targeted sites between ADAMTS9-AS2 and miR-196b-5p. The results showed that in relevant to mimic NC group, luciferase activity of co-transfection group of miR-196b-5p mimic and ADAMTS9-AS2-WT was notably reduced (p < 0.05), while the luciferase activity of co-transfection group of miR-196b-5p mimic and ADAMTS9-AS2-MUT was not dramatically different (p > 0.05) (Figure 2j). qRT-PCR assessed miR-196b-5p level when ADAMTS9-AS2 was overexpressed. MiR-196b-5p expression in oe-ADAMTS9-AS2 group was relatively downregulated (p < 0.05) (Figure 2k). Thus, ADAMTS9-AS2 could bind to miR-196b-5p and suppress its expression.

Figure 2.

ADAMTS9-AS2 binds to miR-196b-5p and restrains its expression.

A: Volcano map of DEmiRNAs of normal and tumor groups in EC data set; B: Venn diagram of predicted downstream target miRNAs of ADAMTS9-AS2 and downregulated DEmiRNAs in EC data set; C: Pearson correlation analysis of ADAMTS9-AS2 and eight differential miRNAs with targeted binding sites; D: MiR-196b-5p level in the EC data set; Green box: normal samples; Red box: tumor samples; E: MiR-196b-5p level in EC tissue and adjacent normal tissue assessed through qRT-PCR (n = 50, p < 0.0001); F: Pearson correlation analysis of ADAMTS9-AS2 and miR-196b-5p expression in clinical samples of EC; G: MiR-196b-5p level in HEEC, EC109, NEC, TE-1 and TE-10 cells assessed by qRT-PCR (p = 0.0004, p = 0.0014, p = 0.024, p = 0.062); H: Binding sites between ADAMTS9-AS2 and miR-196b-5p predicted by lncBase database; I: Binding relationship between ADAMTS9-AS2 and miR-196b-5p detected by RIP (EC109 p = 0.0006, p = 0.0005; NEC p = 0.0002, p < 0.0001); J: Targeted binding sites between ADAMTS9-AS2 and miR-196b-5p verified by dual-luciferase assay (EC109 p = 0.0007; NEC p = 0.0006); K: MiR-196b-5p level when ADAMTS9-AS2 was overexpressed measured via qRT-PCR. * p < 0.05. Each assay was repeated three times (EC109 p = 0.0027; NEC p < 0.0001).

2.3. ADAMTS9-AS2 regulates EC cell progression via hampering miR-196b-5p

In order to identify whether ADAMTS9-AS2 could exert an effect on EC cells by modulating miR-196b-5p, overexpression of ADAMTS9-AS2 was combined with overexpression of miR-196b-5p for experiments. ADAMTS9-AS2 and miR-196b-5p expression in cells was assayed through qRT-PCR. With oe-NC+mimic NC group as control, ADAMTS9-AS2 level in oe-ADAMTS9-AS2+ mimic NC group was noticeably upregulated, and miR-196b-5p level was markedly downregulated (p < 0.05). In relevant to NC mimic group, ADAMTS9-AS2 level in the oe-ADAMTS9-AS2+ miR-196b-5p mimic group did not change significantly (p > 0.05), and miR-196b-5p level was notably upregulated (p < 0.05) (Figure 3a). CCK-8 assessed cell viability of each group in combined experiment (Figure 3b), and flow cytometry measured cell cycle distribution (Figure 3c). The results illustrated that ADAMTS9-AS2 overexpression notably reduced cell viability, increased G0/G1 phase cell number and decreased S phase cell number (p < 0.05), while upregulated expression of miR-196b-5p markedly decreased inhibitory effect of ADAMTS9-AS2 on the proliferative ability (p < 0.05). Afterward, cell migratory property was detected via scratch healing assay (Figure 3d), and the cell invasive capacity was assessed via Transwell assay (Figure 3e). The results manifested that ADAMTS9-AS2 overexpression significantly reduced the number of migrated cells and invasive cells, while overexpression of ADAMTS9-AS2 and miR-196b-5p significantly restored the migratory and invasive capabilities of EC109 cells and NEC cells (p < 0.05). The above findings illustrated that ADAMTS9-AS2 overexpression could constrain EC cell progression via hindering miR-196b-5p expression.

Figure 3.

ADAMTS9-AS2 regulates EC cell progression via downregulating miR-196b-5p level.

A: ADAMTS9-AS2 and miR-196b-5p levels in EC cells in oe-NC+mimic NC, oe-ADAMTS9-AS2+ mimic NC and oe-ADAMTS9-AS2+ miR-196b-5p mimic groups assayed via qRT-PCR (EC109 p = 0.0005; p = 0.0008, p = 0.0015; NEC p = 0.0004, p = 0.0262; p = 0.0014, p = 0.0002); B: Cell viability of EC109 cells and NEC cells at 0, 24, 48 and 72 h measured through CCK-8 (EC109 p < 0.0001, p < 0.0001; NEC p < 0.0001, p < 0.0001); C: Cell cycle distribution of EC109 cells and NEC cells measured via flow cytometry (EC109 p = 0.0233, p = 0.148; p = 0.0089, p = 0.0081; NEC p = 0.0018, p = 0.0035; p = 0.0010, p = 0.0004); D: Migratory property of EC109 cells and NEC cells assessed via scratch healing assay (40×) (EC109 p = 0.0005, p = 0.0003; NEC p = 0.0036, p = 0.0016); E: Invasive property of EC109 cells and NEC cells assessed via Transwell assay (100×). * p < 0.05. Each assay was repeated three times (EC109 p = 0.0015, p = 0.0017; NEC p = 0.0019, p = 0.0010).

2.4. MiR-196b-5p targets and downregulates PPP1R12B expression

Next, the mechanism of ADAMTS9-AS2 targeting miR-196b-5p was further studied. Target mRNAs of miR-196b-5p were identified using miRDB, mirDIP, miRWalk and starBase databases. The predicted results were overlapped with 1,105 downregulated DEmRNAs in TCGA data samples (Figure 4a) to finally obtain 8 DEmRNAs that had binding sites with miR-196b-5p (Figure 4b). After Pearson correlation analysis of ADAMTS9-AS2 and these eight mRNAs, it was found that PPP1R12B was positively correlated with ADAMTS9-AS2 and the correlation coefficient was the highest (Figure 4c). The differential analysis of mRNA expression samples in EC tissue in TCGA database showed that PPP1R12B expression was remarkably low in EC tissue (Figure 4d).

Figure 4.

MiR-196b-5p targets and downregulates PPP1R12B.

A: Volcano map of the DEmRNAs between normal and tumor groups, red dots: upregulated genes, green dots: downregulated ones; B: Venn diagram of predicted downstream target mRNAs of miR-196b-5p and upregulated DEmRNAs; C: Pearson correlation analysis of ADAMTS9-AS2 and eight DEmRNAs that had target-binding sites with miR-196b-5p; D: PPP1R12B level in the EC data set; E: PPP1R12B expression in EC tissue and adjacent normal tissue assessed through qRT-PCR and western blot (n = 50, p < 0.0001); F: Correlation of expression between miR-196b-5p and PPP1R12B, ADAMTS9-AS2 and PPP1R12B in clinical samples of EC analyzed by Pearson correlation; G: The expression of PPP1R12B in HEEC, EC109, NEC, TE-1 and TE-10 cells detected by qRT-PCR (p = 0.0007, p = 0.0020, p = 0.0408, p = 0.0217); H: StarBase website predicted that miR-196b-5p bound to PPP1R12B; I: Binding relationship between miR-196b-5p and PPP1R12B detected by RIP (EC109 p = 0.0005, p = 0.0009; NEC p = 0.0002, p < 0.0001); J: Targeted relationship between miR-196b-5p and PPP1R12B validated by dual-luciferase assay (EC109 p = 0.0011; NEC p = 0.0004); K: PPP1R12B mRNA level after overexpression of miR-196b-5p measured through by qRT-PCR (EC109 p = 0.0015; NEC p = 0.0016); L: PPP1R12B protein level with forced miR-196b-5p level assessed via western blot. * p < 0.05. Each assay was repeated three times.

Similarly, the expression of PPP1R12B in clinical EC tissue samples was detected through qRT-PCR and western blot. PPP1R12B was lowly expressed in EC tissue (p < 0.05) (Figure 4e). Pearson correlation analysis results denoted a negative association between miR-196b-5p and PPP1R12B expression, and a positive association between ADAMTS9-AS2 and PPP1R12B expression in EC tissue (p < 0.05) (figure 4f). qRT-PCR result disclosed that with HEEC cells as control, PPP1R12B level in EC109, NEC, TE-1 and TE-10 cell lines was markedly reduced (p < 0.05) (Figure 4g). According to website prediction (starBase), miR-196b-5p was observed to bind 3’-untranslated region (UTR) of PPP1R12B (Figure 4h). RIP was introduced to verify binding between miR-196b-5p and PPP1R12B. Compared with IgG group, miR-196b-5p and PPP1R12B in the Ago2 group were remarkably enriched (p < 0.05) (Figure 4i). Dual-luciferase assay displayed that with NC mimic group as control, fluorescence activity of miR-196b-5p mimic and PPP1R12B-WT co-transfection group was prominently reduced (p < 0.05), and the fluorescence activity of miR-196b-5p mimic and PPP1R12B-MUT co-transfection group was without significant difference (p > 0.05) (Figure 4j). qRT-PCR and western blot measured PPP1R12B levels after miR-196b-5p overexpression (Figure 4k-l). Results showed that mRNA and protein levels of PPP1R12B in miR-196b-5p mimic group were dramatically downregulated in relevant to those in mimic NC group (p < 0.05).

2.5. PPP1R12B plays a modulatory role in EC by mediating cell cycle signaling pathway

GSEA pathway enrichment analysis result denoted that PPP1R12B was notably enriched in cell cycle signaling pathway (Figure 5a), and the dysfunction of cell cycle signaling pathway was confirmed to be correlated with tumor progression [17,18]. Firstly, the regulation effect of PPP1R12B on cell cycle signaling pathway was assessed. PPP1R12B level in EC109 cells was upregulated or downregulated, and the expression of PPP1R12B and CDK1 in EC cells was assessed via qRT-PCR, denoting that mRNA level of PPP1R12B in the oe-PPP1R12B group was significantly upregulated, while CDK1 expression was dramatically downregulated when compared with the oe-NC group (p < 0.05). As compared with the sh-NC group, PPP1R12B expression in the sh-PPP1R12B group was notably downregulated, while CDK1 expression was upregulated (p < 0.05) (Figure 5b). Then, the protein levels of PPP1R12B, CDK1, cyclin A2, cyclin B1 and Plk1 were assessed via western blot. Overexpression of PPP1R12B remarkably downregulated CDK1, cyclin A2, cyclin B1 and Plk1 protein levels in EC cells (p < 0.05), while silencing PPP1R12B significantly upregulated their protein expression levels (p < 0.05) (Figure 5c).

Figure 5.

PPP1R12B plays a regulatory role in EC by mediating cell cycle signaling pathway.

A: GSEA pathway enrichment analysis results of PPP1R12B; B: PPP1R12B and CDK1 levels in EC cells in oe-NC, oe-PPP1R12B, sh-NC and sh-PPP1R12B groups detected by qRT-PCR (EC109 p = 0.0005, p = 0.0005; p = 0.0005, p = 0.0053; NEC p = 0.0005, p = 0.0008; p = 0.0014, p = 0.0004); C: PPP1R12B, CDK1, cyclin A2, cyclin B1 and Plk1 protein levels assayed through western blot. * p < 0.05. Each assay was repeated three times.

2.6. ADAMTS9-AS2 competitively inhibits miR-196b-5p to regulate PPP1R12B/cell cycle pathway to inhibit the progression of EC

Next, transfection groups of EC109 cells including oe-NC+mimic NC, oe-ADAMTS9-AS2+ mimic NC and oe-ADAMTS9-AS2+ miR-196b-5p mimic were constructed, and expression of ADAMTS9-AS2 and miR-196b-5p in transfection groups was assessed by qRT-PCR. After ADAMTS9-AS2 was overexpressed, its expression was significantly upregulated. Overexpression of ADAMTS9-AS2 remarkably downregulated miR-196b-5p level, while after ADAMTS9-AS2 and miR-196b-5p were overexpressed together, miR-196b-5p level was restored (Figure 6a). Then, qRT-PCR and western blot assays assessed PPP1R12B expression in the combined experiment (Figure 6b-c). With oe-NC+mimic NC group as control, mRNA and protein levels of PPP1R12B in oe-ADAMTS9-AS2+ mimic NC group were significantly upregulated (p < 0.05). Compared with oe-ADAMTS9-AS2+ mimic NC group, PPP1R12B level was dramatically downregulated in oe-ADAMTS9-AS2+ miR-196b-5p mimic group (p < 0.05). Thus, ADAMTS9-AS2 hampered miR-196b-5p level by adsorbing it to modulate its downstream target PPP1R12B level.

Figure 6.

ADAMTS9-AS2 hinders EC progression via adsorbing miR-196b-5p to regulate PPP1R12B/cell cycle pathway.

A: The expression levels of ADAMTS9-AS2 and miR-196b-5p in EC cells after overexpression of ADAMTS9-AS2 and miR-196b-5p (EC109 p = 0.0014; p = 0.0004, p = 0.0006; NEC p < 0.0001, p = 0.0011; p = 0.0028, p = 0.0052); B: PPP1R12B level in EC cells with overexpression of ADAMTS9-AS2 and miR-196b-5p (EC109 p = 0.0025, p = 0.0013; NEC p = 0.0001, p = 0.0003); C: PPP1R12B level in combined experiment assayed by western blot (EC109 p = 0.0006, p = 0.0006; NEC p = 0.0076, p = 0.0009); D: CDK1 expression after ADAMTS9-AS2 overexpression combined with miR-196b-5p overexpression detected through qRT-PCR; E: CDK1, cyclin A2, cyclin B1 and Plk1 protein levels in EC cells measured through western blot. * p < 0.05. Each assay was repeated three times.

Then, effect of ADAMTS9-AS2 and miR-196b-5p upstream regulation on the cell cycle pathway in EC cells was detected. CDK1 level was assayed through qRT-PCR after ADAMTS9-AS2 and miR-196b-5p were overexpressed together. With oe-NC+mimic NC group as control, expression of CDK1 in oe-ADAMTS9-AS2+ mimic NC group was dramatically downregulated (p < 0.05). Compared with oe-ADAMTS9-AS2+ mimic NC group, CDK1 level in oe-ADAMTS9-AS2+ miR-196b-5p mimic group was significantly upregulated (p < 0.05) (Figure 6d). Western blot was utilized to detect CDK1, cyclin A2, cyclin B1 and Plk1 protein levels in EC cells (Figure 6e). Overexpression of ADAMTS9-AS2 noticeably downregulated CDK1, cyclin A2, cyclin B1 and Plk1 protein levels (p < 0.05), while overexpression of miR-196b-5p dramatically decreased inhibitory effect of ADAMTS9-AS2 on CDK1, cyclin A2, cyclin B1 and Plk1 protein levels (p < 0.05).

2.7. Overexpressed ADAMTS9-AS2 affects tumorigenicity of EC cells in vivo

Furthermore, the effect of ADAMTS9-AS2 on EC in vivo was assessed. After ADAMTS9-AS2 level was forced in nude mice, size and weight of tumors were measured. It could be noted that compared with oe-NC group, tumor growth in the oe-ADAMTS9-AS2 group was prominently slowed down and weight was significantly reduced (p < 0.05) (Figure 7a-c). In addition, ADAMTS9-AS2, miR-196b-5p, PPP1R12B and CDK1 levels in the tumor tissue of nude mice were assayed by qRT-PCR (Figure 7d). The results exhibited that in relevant to the control, levels of ADAMTS9-AS2 and PPP1R12B in oe-ADAMTS9-AS2 group were notably upregulated, while miR-196b-5p and CDK1 levels were prominently decreased (p < 0.05). PPP1R12B and CDK1 protein levels in the tumor tissue were then detected through western blot and immunohistochemistry assay. In relevant to the control, PPP1R12B protein level in the oe-ADAMTS9-AS2 group was notably upregulated, and CDK1 protein level was dramatically downregulated (p < 0.05) (Figure 7e-f). In summary, overexpression of ADAMTS9-AS2 could regulate the tumor-forming ability of EC cells in vivo through miR-196b-5p/PPP1R12B/cell cycle pathway axis.

Figure 7.

Overexpressed ADAMTS9-AS2 affects the tumorigenicity of EC cells in vivo.

A: Solid tumor picture of nude mice in oe-NC group and oe-ADAMTS9-AS2 group after ADAMTS9-AS2 was overexpressed; B: Growth volume curve of transplanted EC xenograft tumor in nude mice with forced ADAMTS9-AS2 expression (p < 0.0001); C: Weight of EC xenograft tumor of nude mice after ADAMTS9-AS2 was overexpressed (p = 0.0019); D: ADAMTS9-AS2, miR-196b-5p, PPP1R12B and CDK1 levels in tumor tissue of nude mice assessed through qRT-PCR (p = 0.0005, p = 0.0013, p = 0.0010, p = 0.0007); E: PPP1R12B and CDK1 levels in tumor tissue of nude mice detected via western blot; F: PPP1R12B and CDK1 protein levels in tumor tissue of nude mice analyzed by immunohistochemical assay (400×). * p < 0.05. Each assay was repeated three times.

3. Discussion

EC is a kind of very common malignant tumor with a poor prognosis [19]. EC pathogenesis is observed to be associated with many lncRNAs [20]. In this study, the purpose was to investigate the impact of ADAMTS9-AS2 on miR-196b-5p level through sponging it, so as to regulate the downstream PPP1R12B gene, and further mediated the cell cycle signaling pathway and repressed EC malignant progression.

Firstly, the data of EC tissue samples in TCGA database were analyzed, and ADAMTS9-AS2 was found to be lowly expressed in EC tissue. As mentioned in the introduction, ADAMTS9-AS2 is underexpressed in varying cancers and exerts multiple tumor repressive effects. Therefore, in this paper, the ADAMTS9-AS2 expression was detected in clinical tissue samples and cancer cell lines of EC, and low expression of ADAMTS9-AS2 was confirmed. Then, cellular assays were conducted to detect biological behavior of malignant progression of EC. It was found that the overexpression of ADAMTS9-AS2 had a significant inhibitory effect on malignant progression of EC. In addition, it was reported that ADAMTS9-AS2 is decreased in EC tissue, and downregulated ADAMTS9-AS2 can facilitate the development of cancer cells, while upregulated ADAMTS9-AS2 can hamper cancer cell development [1], which is congruous with research trend in this paper. However, the specific mechanism of ADAMTS9-AS2 in EC is still unclear.

It was reported that ADAMTS9-AS2 can regulate gene expression as a ceRNA in EC [12,13]. Therefore, the miRNA expression data of EC tissue samples were utilized to analyze DEmiRNAs, and the miRNAs that could be targeted by ADAMTS9-AS2 were predicted through the bioinformatics website. MiR-196b-5p was highly expressed in EC tissue and could bind to ADAMTS9-AS2. Warnecke-eberz U et al. [21] manifested that miR-196b-5p was overexpressed in EC. Moreover, miR-196b-5p was found to have a pro-tumor effect in non-small cell lung cancer [22], breast cancer [23], gastric cancer [24] and other cancers. However, there is no relevant study on its mechanism of action in EC. We displayed that ADAMTS9-AS2 inhibited malignant progression of EC cells via regulating miR-196b-5p level. In addition, downstream target PPP1R12B of miR-196b-5p was also found from DEmRNAs in EC tissue samples from TCGA database, and miR-196b-5p could inhibit the PPP1R12B level via binding to 3’-UTR region of PPP1R12B. PPP1R12B has a marked inhibitory impact on malignant phenotypes of colorectal cancer cells [25], and functions on key genes that control tumor cell migration in breast cancer cells [26,27]. However, no studies about it have been reported on EC. Through combined experiment, ADAMTS9-AS2 targeted and modulated PPP1R12B level by competitively inhibiting miR-196b-5p expression. In addition, GSEA manifested that PPP1R12B was remarkably enriched in cell cycle signaling pathway, and the dysfunction of the cell cycle signaling pathway has been confirmed to be correlated with tumor progression [18,27]. PPP1R12B could play a regulatory role in EC cells by mediating cell cycle signaling pathway.

In conclusion, this study revealed that ADAMTS9-AS2 could regulate PPP1R12B/cell cycle pathway by adsorbing miR-196b-5p, and exerted an inhibitory role in EC malignant progression. But whether overexpressed ADAMTS9-AS2 in vivo can play a negative role as a cytotoxic agent, and the induction factor for ADAMTS9-AS2 low expression in EC tissue remains to be further clarified. Altogether, the findings of this study may offer a novel target for EC treatment.

Funding Statement

This research was supported by Zhejiang Provincial Natural Science Foundation of China under Grant No. 2020KY160. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethics approval and consent to participate

This study was conducted in accordance with the ARRIVE guidelines and was approved by the Ethical Review Committee of Laboratory Animal Welfare, Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University (SRRSH20220328).

Consent for publication

All authors consent to submit the manuscript for publication.

Availability of data and materials

The data used to support the findings of this study are included within the article.

Authors’ contributions

All authors contributed to data analysis, drafting and revising the article, gave final approval of the version to be published, and agreed to be accountable for all aspects of the work.