Knockdown of circRAD18 Mitigates Breast Cancer Progression through the Regulation of miR-613/HK2 Axis.

BACKGROUND: Breast cancer (BC) remains the most prevalent malignancy and the leading cause of cancer death. Circular RNAs (circRNAs) have been discovered to serve as crucial regulators in BC. In the current work, we aimed to study the impact of circRAD18 (hsa_circ_0002453) on BC progression and mechanism governing it.
MATERIALS AND METHODS: The expression levels of circRAD18, miR-613 and hexokinase 2 (HK2) mRNA were determined by quantitative real-time polymerase chain reaction (qRT-PCR). CircRAD18 identification was performed using RNase R digestion and actinomycin D assay. Cell viability, colony formation, apoptosis, migration, invasion and glycolysis were measured by Cell Counting Kit-8 assay, colony formation assay, flow cytometry, transwell analysis and extracellular acidification rate detection assay, respectively. Western blot was used to assess the levels of E-Cadherin, Vimentin, N-Cadherin and HK2 protein. The targeted interplay between miR-613 and circRAD18 or HK2 was detected by dual-luciferase reporter assay. Xenograft model assay was performed to observe the role of circRAD18 in vivo.
RESULTS: CircRAD18 was highly expressed in BC tissues and cells. CircRAD18 depletion hindered BC cell malignant behaviors, as evidenced by the inhibition in cell viability, colony formation, migration, invasion, epithelial to mesenchymal transition and glycolysis, as well as the promotion in cell apoptosis. CircRAD18 directly interacted with miR-613, and miR-613 mediated the repressive effect of circRAD18 knockdown on BC cell malignant behaviors. Moreover, HK2 was a direct target of miR-613, and circRAD18 positively regulated HK2 expression via sponging miR-613. Additionally, circRAD18 knockdown repressed tumor growth in vivo by miR-613.
CONCLUSION: Our current work suggested that circRAD18 silencing suppressed BC cell malignant behaviors in vitro and tumor growth in vivo at least partly via the regulation of the miR-613/HK2 axis, highlighting that circRAD18 might be a promising therapeutic target for BC treatment.

Introduction

Breast cancer (BC) remains the most prevalent malignancy and the leading cause of cancer-related death worldwide.1 Despite significant advances in treatment, effective agents against BC, especially high-risk BC, are still limited.2 Therefore, it is very urgent to identify novel therapeutic biomarkers and targets for BC management. One promising strategy is based on circular RNAs (circRNAs).

CircRNAs, a new class of non-coding RNA molecules, exist as a closed loop without 5ʹ caps and 3ʹ polyadenylated tails and thus make them more stable.3 Recent studies have discovered that circRNAs serve as crucial regulators in a wide variety of human cancers, including BC.4,5 For instance, Zhang and colleagues uncovered that circRNA hsa_circ_0052112 enhanced BC cell migration and invasion through sequestering microRNA (miRNA)-125a-5p by acting as a sponge.6 Gao et al underscored that hsa_circRNA_0006528 acted as a miR-7-5p sponge to drive BC progression via regulating the mitogen-activated protein kinase (MAPK) signaling pathway.7 As for circRAD18 (hsa_circ_0002453), it derives from the RAD18 gene, which has been reported to contribute to the carcinogenesis of multiple human cancers.8–10 A recent document highlighted that circRAD18 (hsa_circ_0002453) was highly expressed in triple-negative breast cancer, and it facilitated the development of this disease by sponging miR-208a/3164 and modulating insulin like growth factor 1 (IGF1) and fibroblast growth factor 2 (FGF2) expression.11 In the current project, we aimed to study the impact of circRAD18 on BC progression and mechanism governing it.

MiRNAs are a type of small non-coding RNA molecules that are reported to negatively modulate expression of many oncogenes or tumor suppressive genes.12 Previous studies had identified that miR-613 acted as a tumor suppressor in BC via the regulation of cell proliferation, invasion and migration.13,14 Recently, the competing endogenous RNAs (ceRNAs) hypothesis suggests that circRNAs could function as efficient miRNA sponges to implicate in tumorigenesis.4 The computational method using starBase v.3 software predicted two putative target sites between miR-613 and circRAD18 or hexokinase 2 (HK2), eliciting a potential ceRNA network of the circRAD18/miR-613/HK2 axis in BC.

In the current research, we firstly validated a significant up-regulation of circRAD18 in BC. Consequently, we examined the influence of circRAD18 on BC cell malignant behaviors and whether the miR-613/HK2 axis was involved in the underlying mechanism.

Materials and Methods

Clinical Samples and Cells

Clinical samples used in the current study were obtained from 45 BC patients who received surgery at China-Japan Union Hospital of Jilin University from April 2017 to June 2018, with written informed consent. The corresponding healthy samples were collected at a > 5 cm distance from the tumor. All samples were histologically examined and evaluated by experienced pathologists. The Institutional Ethical Committee of China-Japan Union Hospital of Jilin University approved all aspects.

Human BC cell lines (MCF-7, MDA-MB-231, MDA-MB-468, BT474, and BT549, all from the American Type Culture Collection, ATCC, Manassas, VA, USA) were maintained in RPMI-1640 medium (Gibco, Life Technologies, Breda, The Netherlands) with 10% (v/v) fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Gibco). Human normal breast epithelial cell line MCF-10A (ATCC) was grown in Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) plus 5% horse serum, 20 ng/mL epidermal growth factor (EGF), 10 µg/mL insulin, 100 ng/mL cholera toxin and 500 ng/mL hydrocortisone (all from Gibco). All cells were incubated at 37°C in a humidified incubator containing 5% CO2.

RNA Preparation, RNase R Digestion and Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)

These experiments were performed essentially as previously described.15 TRIzol reagent (Invitrogen, Life Technologies) was used for RNA extraction, and RNase R assay was carried out with RNase R (TaKaRa, Dalian, China) at 37°C for 60 min. For qRT-PCR, cDNA was generated using M-MLV First Strand Kit (Life Technologies) for circRAD18 and mRNA and miScript II RT Kit (Qiagen, Surrey, UK) for miR-613. The SYBR Green PCR Kit (Life Technologies) was used for qRT-PCR with following primers (5ʹ-3ʹ): circRAD18 forward: CAGCTCATTAAAAGGCACCA and circRAD18 reverse: CACACAGCAAGTTGGACACTG, RAD18 mRNA forward: TTGAATTTTGCACGGAATCA and RAD18 mRNA reverse: ATTAACCTGCTCCCCTGCTT, HK2 mRNA forward: TCAATATTAGAGTCTCAACCCCCA and HK2 mRNA reverse: GAAGGCGCTTGTGGAGAAGG, β-actin forward: GCACCACACCTTCTACAATG and β-actin reverse: TGCTTGCTGATCCACATCTG, miR-613 forward: GTGAGTGCGTTTCCAAGTGT and miR-613 reverse: TGAGTGGCAAAGAAGGAACAT, U6 forward: CTCGCTTCGGCAGCACA and U6 reverse: AACGCTTCACGAATTTGCGT. The levels of circRAD18, mRNA and miR-613 were normalized against β-actin or U6.

Actinomycin D Assay

Cells (1.0 × 106) cultured in 6-well plates were exposed to actinomycin D (Sigma-Aldrich, Saint-Quentin Fallavier, France) at a final concentration of 2 mg/L. At the indicated time points (0, 6, 12 and 24 h), the levels of circRAD18 and RAD18 mRNA were determined using qRT-PCR.

Subcellular Fractionation

The commercial PARISTM Kit (Invitrogen) was used to isolate nuclear and cytoplasmic RNA from BC cells, referring to the manufacturer’s protocols. Then, qRT-PCR was used for circRAD18 measurement, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and U6 as the internal control.

Cell Transfection

Small interfering RNA (siRNA) against circRAD18 (5ʹ-AAUCAGACUGCUCUCUCUGUA-3ʹ) and nontarget siRNA (si-NC, 5ʹ-UUCUCCGAACGUGUCACGUTT-3ʹ), pcDNA3.1-based circRAD18 overexpression vector (circRAD18) and negative control vector (Vector), mature miR-613 mimic (5ʹ-CCGUUUCUUCCUUGUAAGGA-3ʹ) and a scrambled oligonucleotide sequence (miR-NC mimic, 5ʹ-UUCUCCGAACGUGUCACGUTT-3ʹ), the inhibitor of miR-613 (anti-miR-613, 5ʹ-UCCUUACAAGGAAGAAACGG-3ʹ) and negative inhibitor control (anti-NC, 5ʹ-UCUACUCUUUCUAGGAGGUUGUGA-3ʹ) were obtained from GenePharma (Shanghai, China). MCF-7 and BT549 cells were transfected with 50 nM of the indicated oligonucleotide or 100 ng of the vector using the commercially available lipid-based Lipofectamine 3000 reagent (Invitrogen).

Measurement of Cell Viability, Colony Formation and Apoptosis

Viability and apoptosis of MCF-7 and BT549 cells after various transfections were evaluated using the Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan) and Annexin V-FITC/PI Apoptosis Assay Kit (Invitrogen), respectively, referring to the recommendation of producers. Cell viability was proportional to the absorbance, which was detected at 450 nm using an iMark microplate reader (Bio-Rad, Munich, Germany). The FACScan flow cytometer (BD Biosciences, Le Pont de Claix, France) was used for the determination of cell apoptosis with Cell Quest software. Cell colony formation was evaluated using a standard colony formation assay as previously described.16

Transwell Migration and Invasion Assay

Migration and invasion of transfected MCF-7 and BT549 cells were observed using a 24-transwell plate (8 µm pore inserts, Millipore, Watford, UK) and a Matrigel-coated Boyden chamber (Millipore), respectively, as previously described.17 Briefly, cells (1.0 × 104) in serum-free media were seeded into the upper compartment of transwell chambers, and medium plus 10% FBS was added into the lower compartment. 24 h later, the penetrated cells through the inserts were stained with 0.2% crystal violet and counted at 100× magnification.

Western Blot

The analysis of Western blot was carried out as previously described.18 Protein extracts were separated on an 8–10% SDS polyacrylamide gel electrophoresis, followed by the transferring to a nitrocellulose membrane (Millipore). The primary antibodies used in the experiments were as follows: anti-E-Cadherin (ab1416), anti-Vimentin (ab8978), anti-N-Cadherin (ab76057), anti-HK2 (ab227198) and anti-GAPDH (ab181602) (all from Abcam, Cambridge, UK).

Detection of Extracellular Acidification Rate (ECAR) and Oxygen Consumption Rate (OCR)

The Seahorse Extracellular Flux analyzer (XFe 96, Seahorse Bioscience, Billerica, MA, USA) was used to measure the ECAR and OCR of transfected MCF-7 and BT549 cells with the Seahorse XF Glycolysis Stress Test Kit, referring to the suggestion of manufacturers. To be brief, for ECAR, transfected cells were sequentially incubated with glucose, the oxidative phosphorylation inhibitor oligomycin and the glycolytic inhibitor 2-DG. For OCR, the cells were sequentially incubated with oligomycin, p-trifluoromethoxy carbonyl cyanide phenylhydrazone (FCCP) and the mitochondrial complex I inhibitor rotenone plus the mitochondrial complex III inhibitor antimycin A (Rote/AA). The Seahorse XF-96 Wave software was used for data analyses, and ECAR was expressed in mPH/min.

Bioinformatics and Dual-Luciferase Reporter Assay

Analyses for the potentially targeted miRNAs of circRAD18 and miR-613 molecular targets were implemented using online software starBase v.3 at . pmirGLO vector-based wild-type reporter plasmids (circRAD18-WT and HK2-3ʹ-untranslated region (UTR)-WT) harboring the complementary sequence for miR-613 and their mutants (circRAD18-MUT and HK2-3ʹ-UTR-MUT) in the seed region were synthesized by HanBio (Shanghai, China). These reporter constructs were introduced into MCF-7 and BT549 cells, respectively, together with miR-NC mimic or miR-613 mimic. 48 h post-transfection, the Dual-luciferase Reporter Assay System Kit (Promega, Paisley, UK) was employed to measure the luciferase activity.

Lentiviral Vector Transduction

Lentiviral vectors expressing short hairpin RNA (shRNA) targeting circRAD18 (sh-circRAD18) or nontarget shRNA (sh-NC) were obtained from HanBio. MCF-7 and BT549 cells were infected with sh-circRAD18 or sh-NC with various multiplicities of infection in media plus 8 µg/mL polybrene (Solarbio, Beijing, China). 24 h later, the cells with positive transduction were selected using puromycin (10 µg/mL, Solarbio).

Xenograft Model Assay

In the animal experiments, 5-week-old female BALB/c nude mice (n = 72, Henan Research Center of Laboratory Animal, Zhengzhou, China) were used following the approval of the Institution Ethics Committee of China-Japan Union Hospital of Jilin University. All processes were performed according to the Council of Agriculture Guidebook for the Care and Use of Laboratory Animals. These mice were randomly divided into 12 groups (n = 6): MCF-7+sh-NC+PBS, MCF-7+sh-circRAD18+PBS, MCF-7+anti-NC, MCF-7+anti-miR-613, MCF-7+sh-circRAD18+anti-NC, MCF-7+sh-circRAD18+anti-miR-613, BT549+sh-NC+PBS, BT549+sh-circRAD18+PBS, BT549+anti-NC, BT549+anti-miR-613, BT549+sh-circRAD18+anti-NC or BT549+sh-circRAD18+anti-miR-613. Cells (5.0 × 106) with sh-circRAD18 or sh-NC transduction were subcutaneously inoculated into the nude mice. One week later, intratumor injection of PBS, anti-NC or anti-miR-613 was carried out, and tumor volume measurement was began and implemented every one week. In the end, all mice were sacrificed, and tumor tissues were removed.

Statistical Analysis

All data were analyzed using a two-sided Student’s t-test or one-way ANOVA followed by Tukey’s post hoc test and were presented as mean ± standard deviation (SD). Correlations between circRAD18, miR-613 and HK2 mRNA expression were determined by the Spearman test in BC tissues. Statistical significance was defined at p-values of less than 0.05.

Results

CircRAD18 Level was Increased in BC Tissues and Cells

For a preliminary observation of circRAD18 in BC progression, we firstly estimated its expression in BC tissues and cell lines. As shown by qRT-PCR, circRAD18 was highly expressed in BC tissues compared with matched normal tissues (Figure 1A). Notably, circRAD18 level was positively correlated with clinical stage of these BC patients (Figure 1B). Moreover, circRAD18 expression was higher in BC cells than that of control (Figure 1C). To validate that circRAD18 was indeed a circular transcript, we performed RNase R and actinomycin D assays. As shown in Figure 1D, circRAD18 could not digested by RNase R compared with the negative control. The data of actinomycin D assays revealed that circRAD18 had an exceeded half-life with higher stability than the linear transcript (Figure 1E). Additionally, the analysis of subcellular localization demonstrated that circRAD18 was prominently present in the cytoplasm fraction of BC cells (Figure 1F).

Figure 1

Characterization of circRAD18 and it was up-regulated in BC tissues and cells. CircRAD18 expression by qRT-PCR in 45 pairs of BC tissues and matched healthy breast tissues (A), clinical stage (I, II, III and IV) BC tissue samples (B), MCF-10A, MCF-7, MDA-MB-231, MDA-MB-468 and BT549 cells (C). (D) The levels of circRAD18 and linear RAD18 mRNA in BC cells by qRT-PCR after digestion with or without RNase R. (E) The levels of circRAD18 and linear RAD18 mRNA in BC cells after treatment with actinomycin D for the indicated time point. (F) CircRAD18 level in nuclear and cytoplasmic fractions of BC cells by qRT-PCR. *P < 0.05, **P < 0.01.

Knockdown of circRAD18 Hampered the Malignant Behaviors of BC Cells

To explore the influence of circRAD18 on BC progression, loss-of-function experiments were implemented by using siRNA against circRAD18 (si-circRAD18). In comparison to the negative group, si-circRAD18 transfection led to a significant down-regulation (about 68% in MCF-7 cells, 59% in BT549 cells and 70% in BT474 cells) of circRAD18 expression (Figure 2A and ). But, RAD18 mRNA expression was not affected by si-circRAD18 (Figure 2A and ). CCK-8 and colony formation assays revealed that cell viability and colony formation were remarkably mitigated by circRAD18 silencing compared with the negative control (Figure 2B and C, and ). Flow cytometry analysis showed that cell apoptosis was markedly elevated by circRAD18 depletion in the three BC cells (Figure 2D, ). Moreover, cell migration and invasion abilities were strikingly declined when circRAD18 knockdown (Figure 2E and F, ). The data of Western blot revealed that circRAD18 silencing resulted in increased E-Cadherin expression and decreased Vimentin and N-Cadherin levels (Figure 2G and ), indicating the inhibitory effect of circRAD18 knockdown on cell epithelial to mesenchymal transition (EMT). Additionally, circRAD18 depletion triggered a remarkable reduction of ECAR (Figure 2H and ) and a striking augment of OCR (Figure 2I and ), which reflects overall glycolytic flux, suggesting the repression of circRAD18 depletion on BC cell glycolysis.

Figure 2

CircRAD18 silencing retarded the malignant behaviors of BC cells. MCF-7 and BT549 cells were transfected with si-circRAD18 or si-NC. (A) CircRAD18 expression and RAD18 mRNA level by qRT-PCR after 48 h transfection. (B) Cell viability by CCK-8 assay 48 h after transfection. (C) Cell colony formation using a standard colony formation assay after 48 h transfection. (D) Cell apoptosis by flow cytometry after 48 h transfection. (E and F) Cell migration and invasion by transwell assay after 24 h transfection. (G) The levels of E-Cadherin, Vimentin and N-Cadherin by Western blot 48 h after transfection. (H and I) Measurement of ECAR and OCR in transfected cells. *P < 0.05.

CircRAD18 Directly Interacted with miR-613

To further understand the role of circRAD18 in BC progression, we used online software starBase v.3 to help identify the miRNAs that potentially bind to circRAD18. The predicted data revealed a putative binding site for miR-613 in circRAD18 (Figure 3A). To affirm this, dual-luciferase reporter assays were performed. Cotransfection of wild-type reporter plasmid and miR-613 mimic caused a significant down-regulation in luciferase activity (Figure 3B). However, upon transfection of the mutant-type reporter, the reduction of miR-613 mimic in luciferase activity was dramatically abolished (Figure 3B). Moreover, the data of qRT-PCR showed that miR-613 level was decreased in BC tissues and cells (Figure 3C–E). Importantly, miR-613 expression was prominently elevated when circRAD18 deficiency in the two cells (Figure 3F).

Figure 3

CircRAD18 directly interacted with miR-613 in BC cells. (A) Schematic of the complementary site for miR-613 in circRAD18 and the mutated target site. (B) Relative luciferase activity in MCF-7 and BT549 cells cotransfected with circRAD18-WT or circRAD18-MUT and miR-NC mimic or miR-613 mimic. MiR-613 expression by qRT-PCR in 45 pairs of BC tissues and matched healthy breast tissues (C), clinical stage BC (I, II, III and IV) tissue samples (D), MCF-10A, MCF-7, and BT549 cells (E), MCF-7 and BT549 cells transfected with si-NC or si-circRAD18 (F). *P < 0.05.

The Repressive Impact of circRAD18 Knockdown on BC Cell Malignant Behaviors was Mediated by miR-613

Then, we validated whether circRAD18 exerting regulatory effect on BC progression was mediated by miR-613. Transient introduction of anti-miR-613, but not the anti-NC negative control, evidently reduced miR-613 expression in both MCF-7 and BT549 cells (Figure 4A). Subsequent function experiments revealed that si-circRAD18-mediated anti-viability (Figure 4B), anti-colony formation (Figure 4C), pro-apoptosis (Figure 4D), anti-migration (Figure 4E), anti-invasion (Figure 4F) and anti-EMT (Figure 4G and H) effects were significantly abrogated by anti-miR-613 cotransfection. Moreover, the decreased influence of circRAD18 on cell glycolysis was prominently reversed by the cotransfection of anti-miR-613 (Figure 4I and J).

Figure 4

MiR-613 mediated the repressive impact of circRAD18 deficiency on BC cell malignant behaviors. (A) The expression of miR-613 by qRT-PCR in MCF-7 and BT549 cells transfected with anti-NC or anti-miR-613. MCF-7 and BT549 cells were transfected with si-NC, si-circRAD18, si-circRAD18+anti-NC, or si-circRAD18+anti-miR-613, followed by the determination of cell viability by CCK-8 assay (B), colony formation ability by a colony formation assay (C), cell apoptosis by flow cytometry (D), cell migration (E) and invasion (F) by transwell assay, E-Cadherin, Vimentin and N-Cadherin levels by Western blot (G and H), cell glycolysis by ECAR and OCR measurement (I and J). *P < 0.05.

HK2 was Directly Targeted and Suppressed by miR-613

MiRNAs play important roles in biological processes by the post-transcriptional repression of their target mRNAs. Hence, we carried out a detailed analysis for the targets of miR-613. Using starBase v.3 computational method, a potential target sequence for miR-613 was demonstrated within the 3ʹ-UTR of HK2 (Figure 5A). When we performed dual-luciferase reporter assays, the luciferase activity of the wild-type reporter was remarkably down-regulated by miR-613 overexpression, and this effect was completely abolished with the mutation of the miR-613 putative sequence (Figure 5B). Additionally, the results of qRT-PCR and Western blot showed that HK2 expression was strongly increased in BC tissues and cells (Figure 5C–F). To verify whether the miR-613 putative sequence was functional, we transfected with anti-miR-613 or miR-613 mimic into MCF-7 and BT549 cells. As demonstrated by qRT-PCR, miR-613 expression was significantly increased by miR-613 mimic introduction compared with the negative control (Figure 5G). Moreover, in contrast to their counterparts, HK2 expression was strikingly elevated when miR-613 depletion and reduced upon miR-613 up-regulation (Figure 5H). Additionally, qRT-PCR data showed that HK2 mRNA expression was positively correlated with circRAD18 level and inversely correlated with miR-613 level in BC tissues (Figure 5I).

Figure 5

HK2 was a direct target of miR-613. (A) Schematic of the miR-613-binding sequence in HK2 3ʹ-UTR and the mutant in the target sequence. (B) Relative luciferase activity in MCF-7 and BT549 cells transfected with HK2-3ʹ-UTR-WT or HK2-3ʹ-UTR-MUT and miR-NC mimic or miR-613 mimic. The expression of HK2 in 45 pairs of BC tissues and matched healthy breast tissues (C), clinical stage (I, II, III and IV) BC tissue samples (D), 3 pairs of BC tissues and matched healthy breast tissues (E), MCF-10A, MCF-7, and BT549 cells (F). (G) MiR-613 level by qRT-PCR in MCF-7 and BT549 cells transfected with miR-NC mimic or miR-613 mimic. (H) HK2 protein level by Western blot in MCF-7 and BT549 cells transfected with anti-NC, anti-miR-613, miR-NC mimic, or miR-613 mimic. (I) The correlations between circRAD18, miR-613 and HK2 mRNA expression were determined in BC tissues using the Spearman test. *P < 0.05.

CircRAD18 Modulated HK2 Expression through Acting as a miR-613 Sponge

Next, we investigated whether, if so, how circRAD18 affected HK2 expression in BC cells. As expected, in comparison to their counterparts, HK2 expression at both mRNA and protein levels were prominently up-regulated by the transfection of circRAD18 overexpression, while they were strongly down-regulated upon circRAD18 knockdown (Figure 6A–D), indicating the positive regulation of circRAD18 on HK2 expression. Whereas, these effects were significantly abrogated by the restoration of miR-613 expression in the two BC cells (Figure 6A–D).

Figure 6

CircRAD18 acted as a sponge of miR-613 to regulate HK2 expression. (A and B) HK2 mRNA expression by qRT-PCR and HK2 protein level by Western blot in MCF-7 cells transfected with Vector, circRAD18, circRAD18+miR-NC mimic, or circRAD18+miR-613 mimic. circRAD18: circRAD18 overexpression vector. (C and D) HK2 mRNA expression by qRT-PCR and HK2 protein level by Western blot in BT549 cells transfected with si-NC, si-circRAD18, si-circRAD18+anti-NC, or si-circRAD18+anti-miR-613. *P < 0.05.

CircRAD18 Knockdown Suppressed Tumor Growth by miR-613 in vivo

Given our data that circRAD18 silencing mitigated BC cell malignant behaviors, we further explored its influence on tumor growth using the xenograft model of MCF-7 and BT549 cells. These results revealed that the transduction of sh-circRAD18 prominently declined tumor growth, while this effect was dramatically reversed by anti-miR-613 introduction (Figure 7A–D). Moreover, the levels of circRAD18 and HK2 were strongly decreased, and miR-613 expression was remarkably increased in tumor tissues derived from sh-circRAD18-transducing BC cells (Figure 7E–H). However, these expression alterations were significantly abolished by the introduction of anti-miR-613 (Figure 7E–H). Besides, the in vivo data showed that anti-miR-613 alone led to a striking enhancement in tumor growth compared with the negative control ().

Figure 7

CircRAD18 depletion inhibited tumor growth via regulating miR-613 in vivo. The nude mice were subcutaneously inoculated with sh-circRAD18-transducing MCF-7 or BT549 cells (5.0 × 106). We then performed intratumor injection of PBS, anti-NC or anti-miR-613 one week after implantation. In the end, all mice were sacrificed, and tumor tissues were removed. (A and B) After 7 days implantation, tumor volume measurement was began and implemented every one week. (C and D) Tumor weight was determined. (E and F) The levels of circRAD18, miR-613 and HK2 were assessed by qRT-PCR in xenograft tissues. (G and H) HK2 protein level was detected by Western blot in excised tumor tissues. *P < 0.05.

Discussion

As a new star of non-coding RNAs, circRNAs are being found to implicate in tumorigenesis and have potential as therapeutic biomarkers and targets in further tumor medicine.19 Emerging evidence has highlighted that some circRNAs, such as hsa_circ_000911 and hsa_circ_0007294, play pivotal roles in BC carcinogenesis.20,21 Moreover, Tang et al reported that hsa_circ_0001982 was highly expressed in BC, and its silencing weakened BC cell proliferation and invasion and enhanced apoptosis through sponging miR-143.22 Liang and colleagues uncovered that circ-ABCB10 promoted BC progression through the regulation of BC cell proliferation and apoptosis via sponging miR-1271.23 In the present work, we focused on the effect and mechanism of circRAD18 on BC cell malignant behaviors in vitro and in vivo.

Firstly, we determined the expression pattern of circRAD18 in BC tissues and cells, and our data revealed a striking up-regulation of circRAD18 in BC as compared with normal control, in agreement with a recent study.11 Due to the closed loop structure, circRNAs are very stable and resistant to RNase R digestion.24 Actinomycin D is capable of the repression of RNA synthesis in various types of tumor cells.25 As a result, we validated that circRAD18 was indeed a circular transcript using RNase R and actinomycin D assays. In the current study, the expression of RAD18 mRNA was not affected by circRAD18 transfection in the two BC cells. Hence, our data highlighted that circRAD18 silencing weakened BC cell malignant behaviors, as evidenced by the inhibition in cell viability, colony formation, migration, invasion, EMT and glycolysis, as well as the promotion in cell apoptosis. Moreover, circRAD18 depletion hindered tumor growth in vivo. In short, circRAD18 knockdown exerted a tumor suppressive effect in BC. Besides, circRAD18 was demonstrated to be prominently present in the cytoplasm fraction of BC cells, hinting the possibility of the interplay of circRAD18 and miRNAs in BC.

Then, online software starBase v.3 was used to help predict the miRNAs that potentially bind to circRAD18. Among these predicted candidates, miR-613 was particularly interesting in our current project considering its tumor suppressive role in a wide series of human cancers, such as cervical cancer, bladder cancer and non-small cell lung cancer.26–28 Moreover, Wu et al manifested that miR-613 hampered BC cell proliferation and invasion by targeting vascular endothelial growth factor A (VEGFA).14 Song et al underscored that miR-613 suppressed the progression of triple-negative breast cancer by the repression of WW domain-binding protein 2 (WBP2) via inactivating YAP-mediated gene expression.29 In the present work, we were first to demonstrate that circRAD18 directly interacted with miR-613 in BC cells. Although the introduction of anti-miR-613 led to the increase in tumor growth in vivo, the promotional extent was different with circRAD18+anti-miR-613 group. Therefore, our data substantiated that miR-613 mediated the repressive effect of circRAD18 knockdown on BC cell malignant behaviors in vitro and tumor growth in vivo.

CircRNAs are widely thought to protect against the repression of gene expression by acting as molecular sponges of miRNAs. Herein, we predicted and confirmed that HK2 was a direct target of miR-613 in BC cells. HK2, the major glycolytic isozyme, is required for tumor initiation and development.30,31 Targeting HK2 has been considered as a promising therapeutic strategy for cancer intervention.32 Moreover, high HK2 expression was reported to be associated with metastasis to the brain in BC.33 HK2 overexpression induced by miR-155 enhanced glycolysis in relation to inflammation in a type of tumor cells, including BC.34 In the present research, we also highlighted that circRAD18 positively regulated HK2 expression via functioning as a miR-613 sponge. Furthermore, in vivo assays demonstrated that HK2 was involved in the regulation of the circRAD18/miR-613 axis on tumor growth. Therefore, more researches about the regulatory relationship between HK2 and the circRAD18/miR-613 axis in BC progression will be carried out in further work.

In conclusion, the current study suggested that circRAD18 knockdown repressed BC cell malignant behaviors in vitro and tumor growth in vivo possibly by the regulation of the miR-613/HK2 axis. We highlighted that the circRAD18/miR-613/HK2 axis might be a novel mechanism of BC carcinogenesis.