Knockdown of circ-ABCB10 promotes sensitivity of lung cancer cells to cisplatin via miR-556-3p/AK4 axis

Human bronchial epithelial cell (HBE) and human NSCLC cells (H-1299, H-125, NCI-H292, A549) were purchased from Chinese Academy of Sciences (Beijing, China). The RPMI-1640 medium (Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS; Invitrogen) and 1% penicillin/streptomycin (Sigma-Aldrich, Milan, Italy) was applied for culturing cells, and cells were cultured in an incubator with 5% CO₂ at 37 °C. To study cellular sensitivity to drugs, 2 μg/ml of 5-fluorouracil (5-Fu), 1 μM of cisplatin, 10 μM of Sorafenib and 1 μM of Sunitinib were all utilized for treating A549 or NCI-H292 cells, all from Sigma-Aldrich. 0.1% DMSO (Sigma-Aldrich) was added to culture medium as a solvent-only negative control group.

A549 and NCI-H292 cells were transfected with specific shRNAs against circ-ABCB10 (sh-circ-ABCB10#1#2), AK4 (sh-AK4#1#2), negative control (sh-NC), pcDNA3.1/AK4 or the empty pcDNA3.1 vector (GenePharma, Shanghai, China), separately. The miR-556-3p mimics and NC mimics were gained from GenePharma. Each plasmid was transfected into cells by Lipofectamine 2000 (Invitrogen).

TRIzol reagent (Takara, Otsu, Japan) was utilized for extracting total RNA from cultured A549 or NCI-H292 cells, and RNA concentration was measured with bicinchoninic acid (BCA) Kit (Invitrogen). Total RNA was reverse-transcribed into cDNA utilizing Reverse Transcription Kit (Applied Biosystems, Foster City, USA) for RT-qPCR. SYBR Green real-time PCR Kit (Takara, Tokyo, Japan) was then used on Bio-Rad CFX96 (Bio-Rad, Hercules, CA). Fold changes of gene expression were calculated by 2^−ΔΔCt method, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH)/U6 as endogenous control.

2 mg/ml of actinomycin D (Sigma-Aldrich) was cultivated with A549 or NCI-H292 cells for 0, 4, 8, 12 h to block mRNA transcription, with DMSO as control. For treating cells with Rnase R, 2 μg of total RNA was cultivated for 30 min with or without 3 U/μg of Rnase R (Epicentre Technologies, Madison, WI, USA) at 37 °C, while cells without Rnase R treatment were seen as control group. RT-qPCR was employed for determining the expression levels of circABCB10 and ABCB10 mRNA.

After amplifying ABCB10 mRNA and circ-ABCB10, the cDNA and genomic DNA (gDNA) were obtained as templates, their PCR products were assessed by the agarose gels with TE buffer (Thermo Scientific, Waltham, MA, USA). GAPDH was used as a negative control.

The A549 and NCI-H292 cells (1 × 10⁴) were inoculated into 96-well plates, and then incubated with 10 mg/ml of CCK-8 (Sigma-Aldrich) for another 4 h. The absorbance at 450 nm was measured with a plate reader (Bio-Tek Instruments, Hopkinton, MA, USA).

1 × 10³ cells were cultured in 6-well plates and the culture medium was changed every 3 days. After 2 weeks, cells were fixed in ethanol (Sigma-Aldrich), and were then dyed using crystal violet (Sigma-Aldrich). Finally, the visible colonies containing ≥50 cells were then counted manually under a microscope (Olympus, Tokyo, Japan).

Using TUNEL Apoptosis Kit (Invitrogen), the apoptosis of A549 and NCI-H292 cells were determined. The above cells were stained by utilizing DAPI (Koritai Biotechnology, Beijing, China). Then, the TUNEL positive cells were captured under a fluorescence microscopy (Olympus).

The migration abilities of 1 × 10³ A549 or NCI-H292 cells were examined using transwell chambers (Millipore, Billerica, MA, USA). Serum-free medium containing transfected cells was added to top compartment, while medium with 10% FBS was plated into bottom chamber. 48 h later, methanol (Sigma-Aldrich) and crystal violet (Solarbio, Beijing, China) were applied for fixing and dying cells, severally. 5 chosen fields at random under a microscope were used for counting migratory cells.

Total protein from cells were extracted by RIPA lysis buffer with protease inhibitors (Beyotime, Shanghai, China) and quantified through BCA Kit (Invitrogen). Then, total protein was isolated by SDS-PAGE and was moved to PVDF membranes (Millipore). After being blocked with skimmed milk, membranes were cultivated overnight with specific primary antibodies for AK4 (ab232888, Cambridge, USA) and GAPDH (ab8245). Later, secondary antibodies were added for cultivating for 1 h. The amount of protein was detected via chemiluminescence detection system.

The fractions of cytoplasmic and nuclear were isolated via NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Invitrogen) and gathered by RNeasy Midi Kit (Qiagen, Hilden, Germany) to determine the cellular localization of circ-ABCB10. The RT-qPCR was used for exploring the expression levels of circ-ABCB10, U6 (nuclear control) and GAPDH (cytoplasmic control).

Fluorescence-conjugated circ-ABCB10 probes were produced by Bersinbio Company (Guangzhou, China). NSCLC cells were immobilized in 4% paraformaldehyde (Sigma-Aldrich). After being cleaned using PBS and dehydrated in ethanol, the air-dried cells were denatured for further cultivation with FISH probes by hybridization reaction buffer. Later, cells were washed with × 2 saline-sodium citrate (SSC). Finally, DAPI was used for dying cells and results were visualized through a confocal laser microscopy (Olympus).

RIP assay was completed using an Imprint RNA immunoprecipitation kit (Sigma-Aldrich). A549 and NCI-H292 cells were reaped and lysed in RIP lysis buffer. Cell lysate was then incubated with magnetic beads which were conjugated with anti-Ago2 antibody or with a negative control normal anti-IgG antibody. The relative enrichment of circ-ABCB10, miR-556-3p and AK4 were determined by RT-qPCR analysis.

The wild-type (WT) reporter vector psiCHECK-circ-ABCB10-WT (Promega, Madison, WI, USA) were constructed based on the bioinformatics-predicted binding sites of miR-556-3p, while binding sequences were mutated to form psiCHECK-circ-ABCB10-Mut vector. The two vectors were co-transfected with miR-556-3p mimics or NC mimics into A549 and NCI-H292 cells. The wild-type and mutant interacting sequences of miR-556-3p in AK4 3′-UTR fragments were obtained to generate psiCHECK-AK4-WT/Mut vectors, then co-transfected into cells with indicated transfection plasmids. The luciferase activity was examined at 48 h post-transfection via Dual-Luciferase Reporter Assay System (Promega, MA, USA).

Previous studies have revealed the cancer-promoting role of circ-ABCB10 in clear cell renal cell carcinoma and breast cancer [14, 15]. Nevertheless, the critical role of it in lung cancer and its association with cisplatin resistance of lung cancer are kept to be explored. To begin with, we conducted RT-qPCR analysis and observed an abnormally high expression of circ-ABCB10 in lung cancer cell lines (H125, H1299, NCI-H292 and A549) compared with that in normal human bronchus epithelium cells (HBE) (Fig. 1a). Then, data from nucleic acid gel electrophoresis analysis revealed that divergent primers could produce the circular isoform of ABCB10 in complementary DNA (cDNA) but not genomic DNA (gDNA), whereas convergent primers were able to amplify the linear isoform of ABCB10 from both cDNA and gDNA in A549 and NCI-H292 cells (Fig. 1b). Under the treatment of ActD, the mRNA synthesis inhibitor, RT-qPCR results presented that ABCB10 mRNA degraded whereas circ-ABCB10 was stable within 12 h in A549 and NCI-H292 cells (Fig. 1c). Additionally, after treatment with Rnase R in A549 and NCI-H292 cells, ABCB10 expression was observably lowered whereas no obvious change of circ-ABCB10 expression could be observed (Fig. 1d). Subsequently, we aimed to explore whether dysregulation of circ-ABCB10 was implicated in the sensitivity of lung cancer cells to cisplatin. As illustrated in Fig. 1e, in comparison with that in control cells, the expression of circ-ABCB10 was significantly declined in cisplatin-treated cells rather than other drugs including 5-FU, Sorafenib, and Sunitinib, indicating that circ-ABCB10 was related to cellular response to cisplatin in lung cancer. To uncover the effect of circ-ABCB10 on cellular sensitivity to cisplatin and lung cancer progression in vitro, we knocked down circ-ABCB10 in NCI-H292 and A549 cells. RT-qPCR depicted a better satisfactory efficiency of circ-ABCB10 knockdown in A549 and NCI-H292 cells by transfection with sh-circ-ABCB10#1 (Fig. 1f). Next, CCK-8 and colony formation assays depicted that compared with that sh-circ-ABCB10#1 strikingly weakened the proliferation ability of lung cancer cells compared with sh-NC group. Cisplatin treatment abrogated lung cancer cell proliferation, and such effect was further enhanced by sh-circ-ABCB10#1 in A549 and NCI-H292 cells were treated with cisplatin (Fig. 1g-h). Moreover, TUNEL assay testified that cell apoptosis capability in sh-circ-ABCB10#1 group was significantly increased when compared with sh-NC group. The treatment of cisplatin increased apoptosis ratio, such increase was further enhanced in sh-circ-ABCB10#1 + cisplatin group (Fig. 1i). Furthermore, transwell assay demonstrated that the migration ability was weakened in lung cancer cells transfected with sh-circ-ABCB10#1 compared with sh-NC control. Cisplatin treatment also decreased migrated cells, and sh-circ-ABCB10#1 could strengthen the effect of cisplatin (Fig. 1j). In sum, circ-ABCB10 is expressed at high levels and featured with circular structure in lung cancer; circ-ABCB10 deletion represses lung cancer progression whereas sensitizes lung cancer cells to cisplatin.

To investigate the underlying mechanism of circ-ABCB10 in lung cancer, we carried out subcellular fractionation and FISH assays to investigate the localization of circ-ABCB10. Results showed that circ-ABCB10 was mainly distributed in cytoplasm of A549 and NCI-H292 cells (Fig. 2a). Thus, we suspected that circ-ABCB10 played vital role in lung cancer via sponging specific miRNA. Subsequently, under the particular screening condition (CLIP Data: strict stringency > = 3; Degradome Data: low stringency > = 1) in starBase (http://​starbase.​sysu.​edu.​cn/​), miRNAs (miR-3163 and miR-556-3p) that was predicted to bind with circ-ABCB10 were listed in Fig. 2b. Then, we further find out the miRNA related to circ-ABCB10 in lung cancer. Data from RT-qPCR analysis showed that miR-556-3p, rather than miR-3163, was upregulated by sh-circ-ABCB10#1 in A549 and H292 cells compared with sh-NC control (Fig. 2c). Thus, we focused on investigating miR-556-3p. We delineated via RT-qPCR that miR-556-3p expression was significantly decreased in lung cancer cells versus the normal HBE cells (Fig. 2d). In addition, a binding site between circ-ABCB10 and miR-556-3p was predicted via starBase (Fig. 2e). After transfection with miR-556-3p mimics, miR-556-3p expression was conspicuously elevated in A549 and NCI-H292 cells (Fig. 2f). Luciferase reporter assay displayed that forced expression of miR-556-3p led to an obvious decrease in the luciferase activity of psiCHECK-circ-ABCB10-WT whereas caused no distinct alteration on the luciferase activity of psiCHECK-circ-ABCB10-Mut (Fig. 2g). It was revealed through RIP analysis that circ-ABCB10 and miR-556-3p were notably enriched in anti-Ago2 group (Fig. 2h). To further verify the effect that the circ-ABCB10 regulated lung cancer progression and sensitivity to cisplatin via interacting with miR-556-3p, we applied rescue assays in transfected cells treated with or without cisplatin. RT-qPCR confirmed the significant inhibition of miR-556-3p expression by miR-556-3p inhibitor (Fig. 2i). As displayed in Fig. 2j-m, miR-556-3p inhibitor reversed the effect of circ-ABCB10 knockdown on inhibiting cell proliferation, inducing apoptosis, and retarding migration. Also, miR-556-3p counteracted the sensitizing effect of sh-circ-ABCB10#1 on cellular response to cisplatin (Fig. 2 j-m). Taken together, circ-ABCB10 directly interacts with miR-556-3p to regulate proliferation, migration, apoptosis, and sensitivity to cisplatin in lung cancer cells.

In subsequence, we attempted to screen out the target mRNA of miR-556-3p. Through searching PITA, miRmap, microT and RNA22 databases, AK4 was predicted to be the target gene of miR-556-3p (Fig. 3a). Then RT-qPCR and western blot revealed that silencing circ-ABCB10 or overexpressing miR-556-3p could decrease the expression of AK4 in A549 and NCI-H292 cells (Fig. 3b-c). Additionally, a prominently elevated expression of AK4 in lung cells was observed via RT-qPCR and western blot (Fig. 3d). Later on, RIP assay depicted that circ-ABCB10, miR-556-3p and AK4 were all aggregated in the immunoprecipitates of anti-Ago2, which revealed the coexistence of them in RNA-induced silencing complex (RISC) (Fig. 3e). Thereafter, a binding site between miR-556-3p and AK4 predicted by starBase was presented in Fig. 3f. After transfection with pcDNA3.1/AK4, the expression of AK4 in A549 and NCI-H292 cells was remarkably increased (Fig. 3g). Afterwards, data from luciferase reporter assay uncovered that the reduced luciferase activity of psiCHECK-circ-ABCB10-WT resulted from overexpression of miR-556-3p could be recovered by overexpressing AK4. However, the luciferase activity of psiCHECK-circ-ABCB10-Mut exhibited no clear changes in different groups (Fig. 3h). Then, we studied the potential influence of AK4 on lung cancer cell progression and sensitivity to cisplatin. First, we adopted RT-qPCR analysis to detect the efficiency of AK4 knockdown. As shown in Fig. 3i, AK4 expression was decreased most significantly in sh-AK4#1-transfected cells. Subsequently, cell proliferation and apoptosis assays illuminated that compared with sh-NC group, sh-AK#1 reduced proliferation and induced apoptosis in A549 and NCI-H292 cells. Also, the anti-proliferation and pro-apoptosis effects of cisplatin in two lung cancer cell lines were further strengthened by sh-AK#1 (Fig. 3j-l). Additionally, transwell assay demonstrated that the migration ability of lung cancer cells in sh-AK4#1 group was markedly lower than that in sh-NC group, and the migration retarded by cisplatin was further decreased by sh-AK4#1 in lung cancer cells (Fig. 3m). Briefly, AK4 is a downstream target of miR-556-3p and its depletion inhibits lung cancer progression and sensitizes lung cancer cells to cisplatin.

Based on the above findings, we intended to validate the role of circ-ABCB10/miR-556-3p/AK axis in regulating lung cancer progression and cell sensitivity to cisplatin. According to the data from RT-qPCR and western blot, the expression of AK4 was cut down by silencing circ-ABCB10 and then AK4 expression was restored by overexpressing AK4 (Fig. 4a). Then CCK-8, colony formation and TUNEL assays validated that upregulation of AK4 reversed the effect of circ-ABCB10 depletion on cell decreasing proliferation and increasing apoptosis (Fig. 4b-d). Knockdown of circ-ABCB10#1 strengthened the anti-proliferation and pro-apoptosis effects of cisplatin on A549 cells, and such strengthening effect of circ-ABCB10#1 was counteracted by AK4 overexpression (Fig. 4b-d). Similarly, upregulating AK4 expression could offset the suppressive effect of circ-ABCB10 knockdown on the migration of A549 cells treated without cisplatin (Fig. 4e). The sh-circ-ABCB10#1 enhanced the suppressive effect of cisplatin on migration of A549 cells, and such enhancement was countervailed by AK4 overexpression (Fig. 4e). To be concluded, circ-ABCB10 weakened the cellular sensitivity to cisplatin via miR-556-3p/AK4 axis in lung cancer.