circ_0084927 promotes cervical carcinogenesis by sponging miR-1179 that suppresses CDK2, a cell cycle-related gene

CC tissue samples were collected from Yantai Affiliated Hospital of Binzhou Medical College, China, and the study protocols were approved by the Ethics Committee of Yantai Affiliated Hospital of Binzhou Medical College. CC cell lines (HeLa, CaSki, SW756 and C-33A) and normal cervical epithelial cell lines (HcerEpic) were purchased from the Beijing Beina Chuanglian Biotechnology Research Institute (BNBIO.com). HeLa, C-33A, SW756, and CaSki cells were cultured in 5% CO₂ at 37 °C in RPMI-1640 medium (E600028; Sangon, Shanghai, China) with 10% fetal bovine serum (16140071; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) in addition to 100 μg/ml streptomycin. HcerEpic cells were cultured in 5% CO₂ at 37 °C in MEM medium (E600024; Sangon, Shanghai, China) with 10% fetal bovine serum and 100 μg/ml streptomycin. The characteristics of the patients are shown in Table 1, and the representative histopathological examination results are illustrated in Additional file 1: Figure S1.

Tissue sections were deparaffinized twice using xylene treatment (10 min each time), and they were re-hydrated by decreasing the alcohol concentration. After washing the tissue sections in distilled water for 1 h, they were stained by hematoxylin solution for 8 min and by eosin for 3 min. After that, the tissue sections were dipped in 0.2% saturated lithium carbonate solution for 30 s. The eosin solution was then used to stain the tissue sections for 1 min after washing the sections in running tap water. Finally, the H&E staining images were photographed with the Nikon TE2000-U inverted microscope (Japan).

The small interfering RNAs of circ_0084927 (si-circ_0084927) and CDK2 (si-CDK2), as well as the negative control siRNA (si-NC), were synthesized by GenePharma (Shanghai, China). Some items were purchased from RiboBio Co., Ltd. (Guangzhou, China), such as miR-1179 control, miR-1179 negative control, miR-1179 mimic (for luciferase reporter gene assay) and miR-1179 inhibitor. HeLa and C-33A cells were transfected with si-NC, miR-1179 inhibitor, si-circ_0084927, si-CDK2, miR-1179 inhibitor plus si-circ_0084927 or miR-1179 inhibitor plus si-CDK2 via Lipofectamine ™ 2000 (11668019; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and via lipofectamine transfection method for 20 min. After the cells were incubated for 2 days at 37 °C, they were analyzed by qRT-PCR.

Before the nuclear and cytoplasmic RNA isolation, nuclear and cytoplasmic fractions were separated using the PARIS Kit (AM1921; Thermo Fisher Scientific, Waltham, Mass., USA). The isolated RNA products in nuclei and cytoplasm were analyzed by qRT-PCR. Then, the expression of circ_0084927 and ESRP1 mRNA was detected in the nuclei and cytoplasm. GAPDH and U2 were subsequently employed as a reference control for cytoplasmic expression and nuclear expression, respectively.

The trizol reagents (15596026; Thermo Fisher Scientific, Inc., Waltham, MA, USA) were first used, according to the instruction manual, to isolate and detect total RNA from the tissue samples and cell lines. The obtained RNA was then reverse-transcribed into cDNA. Then, miR-1179 was reverse-transcribed using the protocol of mirVana™ qRT-PCR miRNA Detection Kit (AM1558; Invitrogen™; Thermo Fisher Scientific, Inc., Waltham, MA, USA). The reverse-transcription of CDK2 mRNA and circ_0084927 was conducted with SuperScript III First-Strand Synthesis SuperMix for qRT-PCR (11752050; Thermo Fisher Scientific, Inc., Waltham, MA, USA). StepOnePlus Real-Time PCR System (4376600; Thermo Fisher Scientific, Inc., Waltham, MA, USA) was later used to perform qRT-PCR. The qPCR products were then validated using the agarose gel electrophoresis method. Next, the data were analyzed with the 2^−ΔΔCt method. GAPDH was then utilized as the internal control of circ_0084927 and CDK2 mRNA, while U6 was used as the internal control of miR-1179. The primer sequences are displayed in Table 2.

Oligonucleotides comprising the circ_0084927 mutant (the sequence containing the miR-1179 binding site was mutated to GAUACGA) or the CDK2 mRNA 3′UTR mutant (the sequence containing the miR-1179 binding site was mutated to GAUACGA) were synthesized by GenePharma (Shanghai, China). Inserted into the dual-luciferase miRNA target expression vector (pGL4) were wild-type circ_0084927, circ_0084927 mutant, CDK2 3′UTR mutant, and wild-type CDK2 3′UTR. This insertion was performed to construct luciferase reporter plasmid. HeLa and C-33A cells were also co-transfected with luciferase porter plasmid and miR-1179 mimic. After 48 h of incubation, the culture medium was removed to collect the cells. The collected cells were then lysed to obtain cell lysates. The luciferase activity was measured by Pierce Renilla-Firefly Luciferase Dual Assay Kit (16185; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the protocol.

The Hela and C-33A cells transfected with miR-1179 mimic were cultured to an appropriate density. The cultured cells were digested using trypsin and were collected after trypsin treatment. The cells were lysed using RIP lysis buffer. The cell lysates were then incubated with RIP buffer containing magnetic beads coupled with anti-Argonaute2 (MA5-23515; Thermo Fisher Scientific, Inc., Waltham, MA, USA) or IgG for 1 h, with IgG serving as a negative control. The mixture was subsequently incubated with Proteinase K. After that, the immunoprecipitated RNA was isolated and analyzed using qRT-PCR.

RNA pull-down assay was conducted to further validate the regulatory binding relationship between miR-1179 and CDK2 mRNA. The biotinylated double-stranded RNA of miR-1179 (Bio-miR-1179) and biotinylated negative control RNA (Bio-NC) were designed by GenePharma (Shanghai, China). Whereas the sense sequence of bio-miR-1179 was 5′-AAGCAUUCUUUCAUUGGUUGG-biotin-3′, the antisense sequence of bio-miR-1179 was 5′-CCAACCAAUGAAAGAAUGCUU-3′. 1 × 10⁵ Hela and C-33A cells were cultured in 6-well plates for 1 day, resuspended in 1 ml lysis buffer, and incubated in ice for 20 min. The lysate was centrifuged at 12,000×g for 15 min before the supernatant was collected. The mixture of bio-miR-1179 or bio-NC and streptavidin-coated magnetic beads (Invitrogen, USA) was added to the supernatant and incubated at 4 °C for 2 h. The pulled-down CDK2 mRNA in the bio-miR-1179 or bio-NC group was detected by qRT-PCR.

After the transfected cells underwent trypsinization, 100 μl of the transfected cell suspension was seeded into a 96-well plate (2 × 10³ cells/well). The plate was then placed in a 37 °C incubator for some hours (24 h, 48 h, and 72 h). Then, 10 μl of CCK-8 solution was added to each well, based on the manual guidelines of CCK-8. After the cells were incubated with CCK-8 for 2 h, the absorbance was measured at 450 nm.

BrdU cell proliferation assay kit was used to detect cell-proliferation ability. Anti-BrdU antibodies were used to detect 5-bromo 2′-deoxyuridine (BrdU), which was incorporated into the cell DNA during cell proliferation. Then, a trypsin-treated suspension containing 10⁴ cells was added to each well of a 24-well plate. The culture medium was changed every 6 h. After the cells were cultured for 24 h, 10 μm BrdU (E607203; Sangon, Shanghai, China) was added, and the culturing was continued for 4 h to allow the proliferating cells to incorporate BrdU into their DNA. The cultured cells were then fixed using the fixing solution and were permeabilized with 0.5% Triton (R) X-100 for 10 min. Mouse anti-IgG and anti-BrdU antibodies (diluted at 1:50) were then incubated with the cells overnight at 4 °C. Subsequently, cells were washed with PBST and incubated in the dark with HRP-conjugated secondary antibodies (A24494; Thermo Fisher Scientific, Inc., Waltham, MA, USA) at 1:500 in PBS at room temperature. In the end, the absorbance at 450 nm was proportional to the amount of BrdU incorporated into the cell, which directly reflected cell proliferation.

Before the cell suspension (10,000 cells/well) was seeded in the 96-well plates (4414133; Thermo Fisher Scientific, Inc., Waltham, MA, USA) that were previously coated with 10 μg/ml type I collagen (C7661, Sigma-Aldrich, USA), the cells were deprived of serum for at least 8 h. After 30 or 60 min adherence at 37 °C in a 5% CO₂ atmosphere, the wells were washed with PBS for at least three times to remove the non-adherent cells. The remaining cells were then treated with MTT for two more hours at 37 °C. Finally, the MTT-treated cells were treated with 100 µl DMSO. The absorbance recorded using a microplate reader (Benchmark, Bio-Rad, USA) was 570 nm.

Caspase-3 is an active cell-apoptosis protease and an early indicator of the onset of apoptosis. Colorimetric detection at 405 nm of p-nitroaniline (pNA), after the cleavage from the peptide substrate DEVD-pNA, may reflect the cell apoptosis level. The transfected Hela and C-33A cells (1x10⁵) in different groups were briefly harvested and lysed in 50 ml of ice-cold cell lysis buffer. Cell lysates were centrifuged at 10,000g for 10 min to obtain the supernatant. Then, 50 μl of 2× Reaction Buffer/DTT Mix and 5 μl of 1 mM DEVD-pNA (substrate for caspase-3) from caspase 3 colorimetric assay kit (630217, Takara Biomedical Technology (Beijing) Co., Ltd., China) were added to the cell lysates. The absorbance was determined by measuring OD405 of the released pNA using a microplate reader (Benchmark, Bio-Rad, USA).

Quantitative analysis was performed after all the protein was extracted with RIPA lysis buffer (C500005, Sangon; Shanghai, China) from Hela and C-33A cells in different groups. An equal amount of protein was separated by 10% SDS-PAGE. The gel was immersed in a transfer buffer to achieve equilibrium before transferring it to a polyvinylidene fluoride membrane. Primary antibodies were diluted at a ratio of 1:1000. The membrane was later incubated for 2 h with diluted primary antibodies against CDK2 (D220395; Rabbit-Human; Sangon; Shanghai, China) and β-actin (SAB5500001; Rabbit-Human; Sigma-Aldrich, China). Following that, the hybrid membrane was blocked with 5% skimmed milk and incubated at 4 °C overnight. Next, the membrane was incubated for 2 h with diluted secondary antibodies (A32731; Goat-Rabbit; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Finally, a hypersensitive ECL chemiluminescence kit (C510043; Sangon; Shanghai, China) was used to detect proteins according to the reagent instructions. The intensity of the protein bands was read using ImageJ software.

The transfected HeLa and C-33A cells were re-suspended once in pre-chilled 1xPBS and were subsequently diluted to 1 × 10⁵ cells/ml in 1× Annexin binding buffer. In every assay, 100 µl of cell suspension (10,000 cells) was used. The transfected HeLa and C-33A cells were then re-suspended and treated with pure ethanol for 30 min. After that, cells were incubated with RNase for 30 min not only to remove RNA but also to eliminate the influence of the binding between PI and RNA. Cells were subsequently stained with the red-fluorescent stain, PI (V13242; Thermo Fisher Scientific, Inc., Waltham, MA, USA), in a dark room at room temperature to allow PI to bind to the DNA of the cells. The stained cells were finally put into a flow cytometer before the proportion of cells in each phase of the cell cycle was obtained from the linked BD FACSuite software.

We analyzed the GSE102686 data series using the GEO2R algorithm and identified 21 differentially expressed genes (DEGs). The top five most significantly up-regulated circRNAs included circ_0084927, circ_0106385, circ_0099591, circ_0081723, and circ_0084912 (Fig. 1a). The expression of these five circRNAs was evaluated in the obtained tissue samples. The results of qRT-PCR showed that excluding circ_0106385, circ_0084927, circ_0099591, circ_0081723, and circ_0084912 were significantly up-regulated in CC tissues than were in paired healthy cervical tissues (Fig. 1b–f). We then selected circ_0084927, which had the -highest expression level, as the research object. Further analysis of circ_0084927 expression revealed that compared to the normal cervical epithelial cell line (HcerEpic), circ_0084927 was significantly up-regulated in CC cell lines, including HeLa, CaSki, SW756, and C-33A. In particular, Hela and C-33A cell lines showed more than twofold circ_0084927 expression of the HcerEpic cell line (Fig. 1g). For this reason, they were selected for follow-up studies. It was observed that circ_0084927 was a closed circular RNA generated from and contained exons 7, 8 and 9 of its host gene, ESRP1 (Fig. 1h). To further characterize circ_0084927, we performed RNase R degradation experiments on HeLa and C-33A cells. Our analysis showed that RNase R greatly reduced linear_0084927 expression. Yet it had little effect on circ_0084927 expression in both cell lines (Fig. 1i). In addition, subcellular fractionation location analysis of HeLa and C-33A cells suggested that circ_0084927 and linear_0084927 were mainly located in the cytoplasm (Fig. 1j).

Metascape.org was first employed to analyze the enriched terms of 904 DEGs (selection criteria: adj. P < 0.01 and log|FC| ≥ 2) from GSE63514 data series. As shown in the bar graph (Additional file 2: Figure S2A), the cell cycle was the most significantly enriched terms by the algorithm of Metascape.org. In particular, critical genes involved in the cell cycle-process of CC can be seen through an MCODE profile (Additional file 2: Figure S2B).

After conducting a GSEA analysis of all genes in GSE63514, we found that cell cycle-related processes were all up-regulated in CC, such as the cell cycle, regulation of cell cycle phase transition, and cell cycle checkpoint process (Additional files 3, 4, 5: Figures S3, S4, S5). By intersecting the 904 DEGs of the GSE63514 data series, genes involved in the 3 cell cycle-related processes were identified by GSEA analysis. A total of 14 genes was identified (Fig. 2a). The 14 genes then went through STRING v11 (https://​string-db.​org/​) protein–protein interaction network analysis. As shown in the network, CDK2 was significantly related to other proteins (Fig. 2b). Even though CDK2 had been studied thoroughly in cervical cancer, studies that focused on its networking with circRNAs were limited. Because of this, CDK2 was chosen as the gene of interest in this study.

By querying GEPIA expression data, we also found that CDK2 was significantly up-regulated in cervical squamous cancer (CESC) tissue samples (Fig. 2c). We then intersected downstream miRNAs of circ_0084927 predicted by circular RNA interactome (https://​circinteractome.​nia.​nih.​gov/​), and the upstream miRNAs of CDK2 mRNA predicted by TargetScan Human 7.2 (http://​www.​targetscan.​org/​vert_​72/​). Finally, miR-1179 was identified (Fig. 2d). The relative expression of miR-1179 in our collected tissue samples was measured, and findings revealed that miR-1179 was significantly downregulated in CC tissues rather than in healthy control tissues (Fig. 2e).

In this study, circ_0084927 and miR-1179 were predicted to bind with each other via pairing in the GAAUGCU-CUUACGA manner (Fig. 3a). To verify the interaction of circ_0084927 and miR-1179, we mutated the GAAUGCU sequence of circ_0084927 that bound miR-1179 to GAUACGA. Luciferin-containing circ_0084927 mutant or circ_0084927 wild-type plasmids and miR-1179 mimic were co-transfected into HeLa and C-33A cells. It was found that the introduction of miR-1179 mimic into the cells attenuated the luciferase expression of circ_0084927 wild-type plasmids but that it had no effect on circ_0084927 mutant (Fig. 3b). RIP results also demonstrated that adding miR-1179 mimic precipitated circ_0084927 (Fig. 3c). However, miR-1179 expression displayed a negative correlation with circ_0084927 in CC tissues (Fig. 3d). Later, miR-1179 was discovered to be significantly downregulated in CC cell lines. Hela and C-33A cell lines showed approximately ½ lower miR-1179 levels than HcerEpic cell lines (Fig. 3e).

To investigate what role circ_0084927 plays through sponging miR-1179 in CC, we transfected circ_0084927 siRNA (si-circ_0084927), miR-1179 inhibitor or si-circ_0084927 plus miR-1179 inhibitor into HeLa and C-33A cells. We found that si-circ_0084927 decreased and increased circ_0084927 level and miR-1179 level, respectively. Meanwhile, miR-1179 inhibitor reduced miR-1179 but had no effect on circ_0084927 expression compared to the control group. In addition, si-circ_0084927 seemed to compromise the effects of the miR-1179 inhibitor on miR-1197 expression, while miR-1179 inhibitor did not in turn affect the circ_0084927 expression (Fig. 4a). The results above indicated that HeLa and C-33A cells were successfully transfected. The results of subsequent CCK-8 and BrdU incorporation assays on the successfully transfected HeLa and C-33A cells indicated that the transfection of si-circ_0084927 declined cell proliferation, while the transfection of miR-1179 inhibitor stimulated cell growth. The proliferation of the HeLa and C-33A cells co-transfected with si-circ_0084927 and miR-1179 inhibitor nonetheless had no significant change compared with the control group. This result implied that their effects could be antagonized (Fig. 4b, c).

Furthermore, flow cytometry was employed to assess cell cycle progression. Results confirmed that si-circ_0084927 significantly increased the proportion of HeLa and C-33A cells in S phase (by approximately 25% in Hela cell line and 50% in C-33A cell line) but decreased the proportion in G2/M phase (by approximately 40% in Hela cell line and 40% in C-33A cell line). In contrast, miR-1179 inhibitor reduced the proportion of cells in S phase by approximately 30% in both cell lines, which were recovered by si-circ_0084927 (Fig. 4d). Cell adhesion experiments also showed that compared to the control group, transfecting with si-circ_0084927 reduced the cell’s adhesion ability by a third; however, transfecting with miR-1179 inhibitor improved the cell’s adhesion ability by approximately 25% in both cell lines. In short, miR-1179 inhibitor transfection could restore the decrease in cell adhesion ability caused by si-circ_0084927 (Fig. 4e).

Caspase 3 is usually activated during apoptosis, irrespective of the specific death-initiating stimulus [42, 43]. This activation could reflect the cell apoptosis level, however. The results of the caspase 3 activation assay revealed that the apoptosis of HeLa and C-33A cells was significantly facilitated by circ_0084927 silence (increased by over fivefold), and repressed by miR-1179 inhibition (decreased by up to 50%). Also noted was that miR-1179 inhibition comprised the increased apoptosis caused by si-circ_0084927 (Fig. 4f).

It’s predicted that miR-1179 paired with the 205–211 position of the 3′UTR of the CDK2 mRNA (Fig. 5a). The detection of the luciferase intensities demonstrated that introducing miR-1179 mimics decreased the fluorescence intensity of cells transfected with wild-type CDK2 3′UTR plasmids even though it did not affect the cells transfected with CDK2 3′UTR mutant (Fig. 5b). RNA pull-down experiment results also showed that miR-1179 interacted with CDK2 mRNA (Fig. 5c). After analyzing the CDK2 mRNA level in the obtained tissue samples, it was found that CDK2 was up-regulated in CC tissues (Fig. 5d). Findings also indicated that miR-1179 was negatively correlated with CDK2 expression (Fig. 5e).

To identify the effect of CDK2 expression in CC, we transfected CDK2 siRNA (si-CDK2) or miR-1179 inhibitor into HeLa and C-33A cells. Also, miR-1179 inhibitor and si-CDK2 were co-transfected into HeLa and C-33A cells to neutralize each other’s effects on CDK2 expression. qRT-PCR analysis (results in Fig. 6a) of transfected cells showed that compared with the control group, si-CDK2 transfection reduced CDK2 mRNA, while the transfection of miR-1179 inhibitor increased it. The reduction of CDK2 mRNA caused by si-CDK2 was restored by miR-1179 inhibitor, thus indicating that the cells were transfected successfully. The efficiency of si-CDK2 was approximately 70%. Western blotting analysis of CDK2 also demonstrated successful transfection at the protein level (Fig. 6b).

We explored the cellular phenotypes of the HeLa and C-33A cells transfected successfully. The results of the CCK-8 assay indicated that silencing CDK2 attenuated HeLa and C-33A cell proliferation, while miR-1179 down-regulation offset the inhibitory effect caused by CDK2 silence at 48 h and 72 h (Fig. 6c). The BrdU experiment also showed the same cell proliferation results: CDK2 silencing inhibited cell proliferation (Fig. 6d). While silencing CDK2 significantly augmented the proportion of S-phase cells compared to the control group, it diminished the proportion at the G1 phase. Besides, the growth in the proportion of S-phase cells caused by CDK2 silencing was attenuated by the down-regulation of miR-1179 (Fig. 7a). In short, CDK2 silencing significantly reduced HeLa and C-33A cell adhesion, and this reduction was restored by simultaneous down-regulation of miR-1179 (Fig. 7b). Finally, in caspase 3 activation experiments, CDK2 silencing increased HeLa and C-33A cell apoptosis, which was compromised by miR-1179 down-regulation (Fig. 7c).