Long non-coding RNA CTBP1-AS2 enhances cervical cancer progression via up-regulation of ZNF217 through sponging miR-3163

This study was executed between 2014 and 2019, with the ethical approval from the Ethics Committee of Harbin Medical University Cancer Hospital. Patients without Human papillomavirus (HPV) infection were excluded from this study. All 72 participants had signed the written informed consent. Highly sensitive polymerase chain reaction (PCR) techniques were used to detect the HPV. The number of patients infected with HPV-18, HPV-11, HPV-45 and HPV-68 were separately 25, 19, 15, 13. The 72 CC samples and adjacent normal samples from CC patients were collected and instantly maintained in the liquid nitrogen at − 80 °C.

Human cervical cancer cell lines, including SiHa (HPV positive), HeLa (HPV positive), MS751 (HPV positive) and C33A (HPV negative) as well as the normal cervical epithelial cells (H8) were purchased from Shanghai Institute of Cell Biology (Shanghai, China), cultured routinely in RPMI-1640 medium (Invitrogen, Carlsbad, CA, USA) at 37 °C with 5% CO₂. 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics were applied to supplement the culture medium.

Total RNA was extracted from HeLa and SiHa cells using 1 mL of TRIzol (Invitrogen) and reversely transcribed into cDNA using PrimeScript RT reagent Kit (Takara, Kyoto, Japan) or miRNA reverse transcription PCR kit (Ribobio; Guangzhou, China). The relative gene expression level was measured by SYBR Green PCR Master Mix (Invitrogen) or SYBR^® PrimeScript^® miRNA RT-PCR Kit (Takara), and Step-One Plus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), and quantified by the comparative 2^−ΔΔCt method. GAPDH mRNA or U6 snRNA served as the endogenous control. The sequences of PCR primers were provided in Additional file 1: Table S1.

When the cell density was about 70%, cell transfection was performed in 24-well plates with CO₂ at 37 °C for 48 h utilizing Lipofectamine 2000 (Invitrogen). The duplicate short hairpin RNAs for CTBP1-AS2, ZNF217 (sh-CTBP1-AS2#1/2, sh-ZNF217#1/2), plasmid pcDNA3.1/ZNF217 were designed by Genepharm (Shanghai, China), as well as their relative negative control RNA (sh-Ctrl, pcDNA3.1). MiR-3163 mimics and NC mimics, miR-3163 inhibitor and NC inhibitor were also produced by Genepharm. Relevant sequences were provided in Additional file 1: Table S1.

HeLa or SiHa cells were planted into 96-well plates at 3 × 10³ per well. After incubated with different time points, cell viability was evaluated by adding 10 μl of CCK-8 solution (Beyotime, Shanghai, China) for 2 h following suppliers guide. The proliferation activity (OD value) was detected at 450 nm by microplate reader (Bio-Rad, Hercules, CA, USA).

EdU assay kit from Ribobio (Guangzhou, China) was added into cell culture medium in 96-well plates for 3 h. Then, 5 × 10⁴ cells were subjected to 4% paraformaldehyde fixation, 0.5% Troxin X-100 incubation and 1 × Apollo^® 488 fluorescent staining. Cell nucleus was subjected to DAPI staining in the dark, and observation using microscope (Thermo Fisher Scientific).

TUNEL staining was used to detect cell apoptosis following the guidelines of in situ Cell Death Detection Kit (Roche Diagnostics GmbH, Penzberg, Germany). Transfected cells (1 × 10⁵) were washed in PBS and stained by TUNEL kit. After treatment with DAPI solution, positively stained cells were all counted using EVOS FL microscope (Thermo Fisher Scientific).

The transwell chamber (Corning Incorporated, Corning, NY, USA) coated with Matrigel (BD Biosciences, Franklin Lakes, NJ) at high concentration or not was employed for cell invasion or migration assay. HeLa and SiHa cells (1 × 10⁵) were added into the upper chamber with serum-free medium. Conditioned culture medium was put into the lower chamber. The invaded or migrated cells were treated with 4% paraformaldehyde fixation and crystal violet solution after 48 h, followed by counting under the microscope (Thermo Fisher Scientific).

Protein samples of 5 × 10⁵ cells were prepared in RIPA lysis buffer (Beyotime) on ice and quantified. 50 μg of samples were subjected to 10% SDS PAGE separation, then transferred to PVDF membranes (Millipore, Billerica, MA, USA). Following sealing with 5% skimmed milk for 1 h, membranes were incubated with primary antibodies including anti-Bax (ab32503), anti-Bcl-2 (ab196495), anti-caspase 3 (ab13847), anti-cleaved caspase-3 (ab2302), anti-ZNF217 (ab136678) and anti-GAPDH (ab128915), together with corresponding HRP-tagged secondary antibodies (all from Abcam, Cambridge, MA, USA). GAPDH served as internal control. Samples were analyzed by enhanced chemiluminescence reagent (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

The separation of nucleus and cytoplasm was run in HeLa and SiHa cells (1 × 10⁷) with PARIS Kit (Invitrogen) on the basis of protocol. After centrifugation, cells were treated with cell fractionation buffer to isolate cell cytoplasm. Cell nucleus was acquired via adding cell disruption buffer. GAPDH and U6 acted as the fractionation indicators for cell cytoplasm and cell nucleus, respectively. Quantification of CTBP1-AS2, GAPDH and U6 in different cellular fractions was made by qRT-PCR.

The RNA FISH probe for CTBP1-AS2 was bought from RiboBio and utilized as suppliers requested. Cells were cultivated with FISH probe in hybridization buffer. Cell nuclei were then subjected to Hoechst counterstaining, finally imaged by laser scanning confocal microscope from ZEISS (Jena, Germany).

The wild type (WT) dual-luciferase reporter gene vectors pmirGLO-CTBP1-AS2 WT and pmirGLO-ZNF217 WT were obtained using the predicted miR-3163 binding sites to CTBP1-AS2 sequence or 3′ un-translated region (3′UTR) of ZNF217. The mutant (MUT) vectors pmirGLO-CTBP1-AS2 MUT and pmirGLO-ZNF217 MUT were established with point mutations of miR-3163 binding sites. Relevant sequences were provided in Additional file 1: Table S1. All vectors were co-transfected into cells with miR-3163 mimics or NC mimics for 48 h. The pmirGLO dual-luciferase reporter vectors were bought from Promega (Madison, WI). Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) was applied for detecting vector activity.

RNA pull down assay was undertaken using Pierce Magnetic RNA–Protein Pull-Down Kit (Thermo Fisher Scientific, Waltham, MA, USA). The wild type or mutant CTBP1-AS2 sequences containing the putative miR-3163 binding sites were labeled with the Biotin. Cell lysates of 1 × 10⁶ cells were mixed with Biotin labeled CTBP1-AS2 for 1 h, then with streptavidin beads for 30 min. The enrichment of miR-3163 was analyzed by qRT-PCR.

Using EZMagna RIP Kit (Millipore), RIP assay was conducted in HeLa and SiHa cells (1 × 10⁷). Lysates from RIP lysis buffer were subjected RIP buffer incubation with anti-Ago2 or anti-IgG antibodies-coated beads (Millipore) for 4 h. At length, the precipitated RNAs were isolated and purified, result was analyzed by RT-qPCR.

BALB/c female nude mice from Shanghai SIPPR-BK Laboratory Animal (Shanghai, China) were used for in vivo experiment and maintained under SPF-condition lab. The animal-related protocol was approved by the Animal Research Ethics Committee of Harbin Medical University Cancer Hospital. SiHa cells stably transfected with sh-Ctrl and sh-CTBP1-AS2#1 were injected into nude mice at a density of 5 × 10⁶. The tumor volumes were recorded every 4 days and calculated in accordance with a formula (length × width² × 0.5). Twenty-eight days later, tumors were excised from killed mice and weighed for further analysis.

At first, we performed qRT-PCR to explore the expression of CTBP1-AS2 in CC cell lines (C33A, MS751, HeLa and SiHa). CTBP1-AS2 was found abnormally up-regulated in CC cell lines than in normal cervical epithelial cell line (H8) (Fig. 1a). The highest level of CTBP1-AS2 was detected in SiHa and HeLa cells (HPV-positive). Then, we performed loss-of-function experiments in Hela and SiHa cell lines after guaranteeing the transfection efficiency of sh-CTBP1-AS2#1/2 plasmid by qRT-PCR (Fig. 1b). Then, CCK8 and EdU assays were performed to determine the effect of CTBP1-AS2 on cell proliferation. We observed that knockdown of CTBP1-AS2 significantly suppressed cell proliferation (Fig. 1c, d). TUNEL assay was performed to investigate cell apoptosis. We found significantly enhanced apoptosis ability after silencing CTBP1-AS2 (Fig. 1e). Results of transwell assays indicated that knockdown of CTBP1-AS2 notably reduced the number of migrated and invaded cells (Fig. 1f, g). As for cell apoptosis, we found that knockdown of CTBP1-AS2 significantly increased the expression of Bax and cleaved caspase-3, while decreased the expression of Bcl-2 was observed through western blot assay (Fig. 1h). To further prove the oncogenic role of CTBP1-AS2 in CC, we also conducted gain-of function assays in normal H8 cells. After overexpression of CTBP1-AS2 in H8 cells (Additional file 2: Figure S1A), cell proliferation was strengthened, whereas apoptosis rate was decreased (Additional file 2: Figure S1B–D). In addition, the migration and invasion were both stimulated by the overexpression of CTBP1-AS2 (Additional file 2: Figure S1E–F). Taken together, CTBP1-AS2 exerts positive effects on CC cell growth, migration and invasion in vitro.

It has been widely reported that lncRNAs exert different functions in nucleus and cytoplasm at a transcription or post-transcription level. In order to determine the potential role of CTBP1-AS2 in CC cells, we detected its subcellular location. Subcellular fractionation manifested that CTBP1-AS2 was principally located in the cytoplasm of CC cells (Fig. 2a). The result of FISH assay further confirmed this conclusion, demonstrating the post-transcriptionally regulatory role of CTBP1-AS2 (Fig. 2b). We detected combinable miRNAs for CTBP1-AS2 via Starbase database (http://​starbase.​sysu.​edu.​cn/​) and restricted binding conditions (medium stringency >=2 in Clip data, with or without data in degradome data, class >=7 mer-m8 and Ago ExpNum >=5). Three miRNAs (miR-3150b-3p, miR-3163 and miR-4784) were identified. qRT-PCR was performed to measure miRNA expression after silencing CTBP1-AS2. The expression of miR-3163 exhibited the most significant up-regulation after knockdown of CTBP1-AS2 compared with negative control group (Fig. 2c). Besides, miR-3163 has been reported to be a tumor suppressor in Retinoblastoma Cancer Stem Cells (RCSCs) [26]. Therefore, we selected miR-3163 as candidate miRNA for this study. In the present study, we noticed a significant down-regulation of miR-3163 in CC cell lines, which is contrary to the expression profile of CTBP1-AS2 (Fig. 2d). We obtained the putative miR-3163 binding site in the sequence of CTBP1-AS2 by utilizing Starbase (Fig. 2e). Next, dual luciferase reporter assays were performed in HeLa and SiHa cells to determine the physical interaction between CTBP1-AS2 and miR-3163. We observed that transfection of miR-3163 mimics evidently attenuated the luciferase activity of CTBP1-AS2-WT, but not CTBP1-AS2-MUT (Fig. 2f). RNA pull down assay manifested that miR-3163 was significantly enriched in biotinylated CTBP1-AS2-WT compared to negative control groups, while no products were observed in biotinylated CTBP1-AS2-MUT (Fig. 2g), which validated the combination between CTBP1-AS2 and miR-3163. From above results, we verified that miR-3163 can bind with CTBP1-AS2.

To determine the effects of miR-3163 on CTBP1-AS2-mediated biological functions in CC cells, we knocked down the expression of miR-3163 (Fig. 2h) and performed rescue experiments in CC cells. EdU assay revealed that miR-3163 inhibition could abrogate the anti-proliferation effect of sh-CTBP1-AS2 (Fig. 2i). Results of TUNEL assay demonstrated that miR-3163 inhibition could reverse the pro-apoptosis effect of sh-CTBP1-AS2 (Fig. 2j). Data of transwell assays manifested that miR-3163 inhibition could reverse the inhibitory effects of silencing CTBP1-AS2 on migration and invasion (Fig. 2k, l). Based on these results, we identified miR-3163 as a tumor suppressor in CC.

Numerous miRNAs have been discovered to be dysregulated in various cancers and exert their functions through modulating their downstream target genes. We found four target genes (MORF4L1, PFN1, ZNF217 and SERBP1) of miR-3163 using Starbase database (high stringency >=3 in Clip data, with or without data in degradome data, microT program, and Ago ExpNum >=35). qRT-PCR was performed to detect the effects of overexpressing miR-3163 on expression of 4 mRNAs. We observed that the expression of ZNF217 was notably down-regulated after overexpressing miR-3163 while the other three mRNAs were not influenced (Fig. 3a). Besides, ZNF217 was discovered to be an oncogene in several cancers. For example, ZNF217 expression is predominantly increased in prostate cancer (PCa) and promotes PCa growth [27]. ZNF217 has been found to be an indicator of bone metastasis in breast cancer [28]. In the present study, we found that ZNF217 was also significantly highly expressed in CC cell lines (Fig. 3b), which was contrary to the expression status of miR-3163 in CC cell lines. We obtained putative miR-3163 binding site in 3′ UTR sequence of ZNF217 from starBase (Fig. 3c). Luciferase reporter assay performed in HEK-293T manifested that miR-3163 mimics could weaken the luciferase activity of ZNF217-WT (Fig. 3d). RIP was performed to verify the interaction among CTBP1-AS2, miR-3163 and ZNF217. The results revealed the significant enrichment of miR-3163, ZNF217 and CTBP1-AS2 in anti-Ago2 group compared with IgG control group (Fig. 3e). To determine the role of ZNF217 in CC cells, we conducted loss-of-function experiments in HeLa and SiHa cells. qRT-PCR analysis determined the transfection efficiency of sh-ZNF217 firstly (Fig. 3f). According to the results of EdU assay, knockdown of ZNF217 inhibited cell proliferation (Fig. 3g). TUNEL results showed that silencing ZNF217 promoted CC cell apoptosis ability (Fig. 3h). Silencing ZNF217 also curbed CC cell migration and invasion, as shown on Fig. 3i and j. Western blot was performed to measure the expression level of apoptosis-related proteins. Consist with the result of TUNEL assay, ZNF217 knockdown promoted the apoptosis of CC cells (Fig. 3k). Collectively, these functional assays revealed that ZNF217 acts as an oncogene in CC progression through facilitating CC cell proliferation, invasion, migration and inhibiting CC cell apoptosis.

To further investigate the CTBP1-AS2/miR-3163/ZNF217 axis in CC, we performed rescue assays. Firstly, we performed qRT-PCR analysis to identify the transfection efficiency of pcDNA3.1/ZNF217 (Fig. 4a). qRT-PCR and western blot analysis were conducted to study the expression of ZNF217 after knockdown of CTBP1-AS2. The results showed that the mRNA and protein levels of ZNF217 were inhibited significantly after knockdown of CTBP1-AS2, but increased again when co-transfected with pcDNA3.1/ZNF217 (Fig. 4b, c). Subsequently, a series of rescue experiments was performed to study the influence of up-regulated ZNF217 on CTBP1-AS2-induced cellular function. CCK8 showed that cell vitality was suppressed after silencing CTBP1-AS2, but enhanced again after ZNF217 overexpression (Fig. 4d). Cell cycle distribution was also analyzed by flow cytometry analysis. Cell cycle was arrested at G0/G1 phase by the silencing of CTBP1-AS2, whereas this tendency was reversed by the upregulation of ZNF217 (Additional file 3: Figure S2A). The proliferation marker PCNA and cell cycle related proteins (CDK1 and Cyclin D1) were also detected in two CC cells transected with sh-Ctrl, sh-CTBP1-AS2#1 or co-transfected with sh-CTBP1-AS2#1 and pcDNA3.1/ZNF217. As expected, all the levels of above proteins were reduced by the downregulation of CTBP1-AS2 but was recovered by the overexpression of ZNF217 (Additional file 3: Figure S2B). Results of TUNEL assay manifested that the apoptotic cells increased significantly after knockdown of CTBP1-AS2, but decreased again after transfecting pcDNA3.1/ZNF217 (Fig. 4e). Results of transwell revealed that ZNF217 overexpression abrogated the anti-migration and anti-invasion effects of CTBP1-AS2 knockdown in vitro (Fig. 4f, g). In addition, western blot assay further validated that ZNF217 overexpression could reverse the pro-apoptosis effects of sh-CTBP1-AS2 (Fig. 4h). These results indicated that ZNF217 overexpression could rescue the oncogenic function of CTBP1-AS2.

PI3K/AKT pathway can involve in cell proliferation and apoptosis. Moreover, Src/FAK pathway is known as a regulator for cell migration and invasion. Here, we detected whether CTBP1-AS2 regulated CC cell growth and migration through these two signaling pathway. Through western blot analyses, the phosphorylated levels of PI3K, AKT, mTOR, Src and FAK were all reduced by the inhibition of CTBP1-AS2, while were enhanced again by the enhancement of ZNF217 expression (Additional file 3: Figure S2C-D).

Xenografts model was used to further explore the tumor suppressing effects of sh-CTBP1-AS2 in vivo. We observed that the tumor growth speed was slower on nude mice after subcutaneously injection with sh-CTBP1-AS2#1 than control group (Fig. 5a). Besides, we found that the tumor volume and weight were apparently lessened in sh-CTBP1-AS2#1 group compared with control group (Fig. 5b–d). Moreover, as exhibited in Fig. 5e, positivity of Ki-67 was reduced and that of cleaved caspase-3 was enhanced by silenced CTBP1-AS2. TUNEL assay further determined the positive effect of CTBP1-AS2 silencing on apoptosis (Fig. 5f). Additionally, the levels of CTBP1-AS2 and ZNF217 were lower in tumor tissues removed from mice in sh-CTBP1-AS2#1 group than that in sh-Ctrl group (Fig. 5g). The level of miR-3163 was relatively higher in sh-CTBP1-AS2#2 group. These in vivo experiments demonstrated that knockdown of CTBP1-AS2 inhibited CC tumor growth in vivo. Taken together, CTBP1-AS2 exerted its oncogenic properties in CC via up-regulating the expression of ZNF217 through sponging miR-3163 (Fig. 6).

In addition to in vitro and in vivo experiments, the clinical significance of CTBP1-AS2/miR-3163/ZNF217 axis was also analyzed. As presented in Additional file 4: Figure S3A, CTBP1-AS2 was expressed at a high level in CC tissues. Then, we divided CTBP1-AS2 into high or low expression of CTBP1-AS2 based on the medium expression value. The Kaplan–Meier survival analysis was conducted and the results depicted that high expression of CTBP1-AS2 led to unfavorable prognosis of CC patients (Additional file 4: Figure S3B). Besides, miR-3163 was significantly down-regulated in CC tissues (Additional file 4: Figure S3C) and up-regulation of miR-3163 predicted favorable prognosis of CC patients (Additional file 4: Figure S3D). Our present study also verified the overexpression of ZNF217 in CC tissues (Additional file 4: Figure S3E) and validated that up-regulated ZNF217 predicted poor prognosis of CC patients (Additional file 4: Figure S3F).