Circular RNA hsa_circ_0005909 modulates osteosarcoma progression via the miR-936/HMGB1 axis

54 patients with OS underwent surgical treatment at Henan Provincial People’s Hospital, and their tumors tissues and neighboring non-tumor tissues were used in this study. The inclusion criteria were as follows: preoperatively confirmed as osteosarcoma by puncture or biopsy; no immunotherapy, chemotherapy, radiotherapy, or chemotherapy before surgery; patients agreed to participate in this study and clinical data were complete. The exclusion criteria were as follows: suffering from other malignant tumors; suffering from other types of serious diseases; patients received treatment for other clinical diseases within 3 months before admission; pregnant or breastfeeding. All patients informed consents before undergoing surgical resection. The association between circ_0005909 expression and clinicopathological parameters of OS patients were presented in Table 1. This research was ratified by the Ethics Committee of Henan Provincial People’s Hospital.

Normal osteoblast hFOB1.19 and OS cell lines (HOS, U2OS, and SAOS-2) were purchased from the American Type Culture Collection (Manassas, VA, USA). All cells were kept in an incubator with 5% CO₂ at 37 °C and cultured in Dulbecco’s modified Eagle’s medium (DMEM, Life Technologies, Grand Island, NY, USA) complemented with fetal bovine serum (FBS, 10%, Sigma, Louis, Missouri, MO, USA), streptomycin (100 μg/mL, Sigma), and penicillin (100 U/mL, Sigma).

Short hairpin RNA targeting circ_0005909 (sh-circ#1 and sh-circ#2) and negative control (sh-NC) were bought from RiboBio (Guangzhou, China). MiRNA mimics and inhibitors targeting miR-936 (miR-936 and anti-miR-936) and their negative controls (miR-NC and anti-NC) were procured from RiboBio. The sequence of HMGB1 was cloned into the pcDNA3.1 vector (vector) (Invitrogen, Carlsbad, CA, USA) to construct the overexpression vectors for HMGB1. When the cells were cultured to approximately 80% confluence, OS cells (HOS and U2OS) were transiently transfected with the assigned oligonucleotides or vectors using Lipofectamine 3000 reagent (Invitrogen) and cultured for 48 h.

The nuclear and cytoplasmic RNA of OS cells were isolated with the PARIS KIT50 RXNS (Life Technologies). Total RNA was extracted through the TRIzol reagent (Invitrogen). For RNase R digestion, total RNA of OS cells was treated with RNase R (3 U/μg, Epicentre Technologies, Madison, WI, USA) at 37 °C for 15 min. Total RNA (5 μg) was reversely transcribed into complementary DNA using the PrimeScript RT reagent Kit (Takara, Dalian, China) or miRNA First-Strand Synthesis Kit (Takara). QRT-PCR was conducted through the SYBR Premix Ex Taq (Takara). The expression of circ_0005909, XPR1, miR-936, and HMGB1 was figured by the 2^−ΔΔCt method. The primers used in the research were displayed as below: Glyceraldehyde-3-phosphate dehydrogenase (GAPDH): (Forward: 5′-GACTCCACTCACGGCAAATTCA-3′ and Reverse: 5′-TCGCTCCTGGAAGATGGTGAT-3′); circ_0005909: (Forward: 5′-GTATCCACTTGCCCTTTA-3′ and Reverse: 5′-TTACTCCAGCCTGTCTC-3′); XPR1: (Forward: 5′-TCCACCTACGGAGGACAATC-3′ and Reverse: 5′-GGAGAAGTGCAGGCAAGAAC-3′); HMGB1 (Forward: 5′-GGGATGGCAAAGTTTTTCCCTTTA-3′ and Reverse: 5′-CACTAACCCTGCTGTTCGCT-3′); U6 small nuclear RNA (snRNA) (Forward: 5′-GCTCGCTTCGGCAGCACA-3′ and Reverse: 5′-GAGGTATTCGCACCAGAGGA-3′) and miR-936: (Forward: 5′-AACGAGACGACGACAGAC-3′ and Reverse: 5′-ACAGTAGAGGGAGGAATCGCAG-3′). GAPDH or U6 snRNA was served as an internal control.

The assigned vectors or oligonucleotides were transfected into OS cells for 48 h. After washing with phosphate buffer solution (PBS), the cells were incubated with DMEM mixed the CCK-8 reagent (10 μL, Dojindo, Tokyo, Japan) for 2 h. The absorbance at 450 nm was measured through the Microplate Absorbance Reader (Bio-Rad Labs., Richmond, CA, USA).

The transfected HOS and U2OS cells (1.5 × 10² cells/well) were seeded in 6-well plates and maintained at 37 °C in an incubator with 5% CO₂ for 2 weeks, and the medium was replaced every 4 days. After fixation with formaldehyde (10%, Sigma) for 20 min, cell colonies in each well were stained with crystal violet (0.1%, KeyGen, Jiangsu, China) for 5 min. The colonies (> 50 cells) were counted and photographed under a light microscope (Olympus, Tokyo, Japan).

After transfection for 48 h, the HOS and U2OS cells were collected. In the invasion assay, the transfected HOS and U2OS cells (1 × 10⁵) were seeded on the upper chamber of the transwell chamber (8 μm, Corning Costar, Corning, NY, USA) with matrigel matrix (BD Biosciences, San Jose, CA, USA). And the lower chamber of the transwell chamber was supplemented with the culture medium with FBS (10%). The cells on the upper membrane were removed after culturing for 24 h. Subsequently, the cells on the lower membrane were fixed with paraformaldehyde (4%, Beyotime, Shanghai, China) and stained with crystal violet (0.1%, KeyGen). The same procedures were used for the migration assay, but the upper chamber of the transwell chamber was not covered with the matrigel matrix. The migrated and invaded cells were figured with a light microscope (Olympus).

Western blot analysis was executed as previously described [24]. OS tissues and cells were lysed in lysis buffer (Beyotime). After the subjection of the sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), the total protein was transferred onto the polyvinylidene difluoride (PVDF) membranes (Invitrogen). The membranes were probed with anti-N-cadherin (N-cad) (ab18203, 1:1000, Abcam, Cambridge, MA, USA), anti-E-cadherin (E-cad) (ab40772, 1:10,000, Abcam), anti-HMGB1 (ab79823, 1:500, Abcam), and anti-GAPDH (ab128915, 1:5000, Abcam), and then incubated with goat anti-rabbit (ab97051, 1:10,000, Abcam) immunoglobulin G (IgG). GAPDH was regarded as a loading control. Protein bands were visualized by the ImmunoStar LD (Wako Pure Chemical, Osaka, Japan).

The animal experiment was ratified by the Ethics Committee of Henan Provincial People’s Hospital. 8 BALB/c nude mice (athymic, 4–5 weeks old) were purchased from Charles River Laboratories (Beijing, China) and were fed under specific-pathogen-free conductions. HOS cells (1 × 10⁶) with sh-NC or stable lentivirus-mediated sh-circ_0005909 (sh-circ#1, RiboBio) were subcutaneously injected into the dorsal side of the nude mice (4 mice/group). The tumor volume was measured once a week using the equation: Volume = (length × width²)/2. The mice were euthanized on day 35 after injection for subsequent analysis.

The binding sites of miR-936 in circ_0005909 were predicted by the Circinteractome database. The sequences of circ_0005909 and mutant circ_0005909 (within the binding sites to miR-936) was inserted into the pGL3-control vector (Promega, Madison, WI, USA) for the construction of the wild type (WT) and mutant (MUT) luciferase reporter vectors for hsa_circ_0000517, respectively. The binding sites of HMGB1 in miR-936 were predicted by the Targetscan database. The same method was employed to construct the luciferase reporter vectors for HMGB1. HOS and U2OS cells were cotransfected luciferase reporter vectors and miR-NC or miR-936. The dual-luciferase reporter assay kit (Promega) was employed to assess the luciferase intensities of the luciferase reporter vectors in OS cells.

In brief, the Biotinylated (Bio)-miR-NC and Bio-miR-936 from Sigma were transfected into OS cells, respectively. The lysates of OS were incubated with Dynabeads M-280 Streptavidin (Invitrogen) at 4 °C for 3 h to pull down the Bio-coupled RNA complexes. Following this, the beads were washed through the lysis buffer (Beyotime), and the complex was purified by TRIzol (Invitrogen). The enrichment of circ_0005909 in the complex was examined through qRT-PCR.

The experiments of the research in vitro were repeated at the last 3 times. GraphPad Prism 7.0 (GraphPad, La Jolla, CA, USA) and SPSS 18.0 software (SPSS, Chicago, IL, USA) were utilized for statistical analysis. The  association between circ_0005909 expression and clinicopathological parameters of patients with OS was analyzed with the Chi square test. The survival curve was assessed via the Kaplan–Meier method and the log-rank test. The differences between the OS tissues and neighboring non-tumor tissues were analyzed by using paired Student’s t test. The differences between the two groups were determined by unpaired Student’s t test, and the differences among more groups were analyzed with one-way analysis of variance (ANOVA) followed by Tukey post hoc test. The correlation was evaluated via Pearson’s correlation analysis. Differences were deemed significant if P < 0.05. Data were exhibited as mean ± standard deviation.

At first, we examined the expression pattern of circ_0005909 in 54 paired OS tissues and neighboring non-tumor tissues through qRT-PCR. We observed that circ_0005909 expression was overtly increased in OS tissues with respect to that in the neighboring non-tumor tissues (Fig. 1a). Consistently, a distinct augmentation of circ_0005909 was observed in OS cell lines (HOS, U2OS, and SAOS-2) compared to the hFOB1.19 cells (Fig. 1b). Moreover, circ_0005909 expression was associated with clinical staging and distant metastasis (Table 1). Furthermore, patients with OS were divided into two groups based on the median value: high circ_0005909 expression group and low circ_0005909 expression group, and the overall survival of OS patients with the high expression of circ_0005909 were lower than the OS patients with low circ_0005909 expression (Fig. 1c). Also, compared to the liner gene XPR1, circ_0005909 was resistant to RNase R (Fig. 1d, e). Besides, we detected the relative expression of circ_0005909 in the cytoplasm and nucleus of HOS and U2OS cells. QRT-PCR displayed that circ_0005909 was primary located in the cytoplasm of HOS and U2OS cells (Fig. 1f). These findings revealed that circ_0005909 was abundant and stable in OS tissues and cells, and high circ_0005909 expression might be associated with OS advancement.

Considering that circ_0005909 was upregulated in OS tissues and cells, we further surveyed the role of circ_0005909 in OS through loss-function-experiments. In contrast to the control sh-NC, circ_0005909 expression was drastically reduced in HOS and U2OS cells transfected with sh-circ#1 and sh-circ#2, while the expression of XPR1 mRNA did not change (Fig. 2a). Also, the expression of circ_0005909 was lower in the sh-circ#1 group with respect to the sh-circ#2 group, so we employed the sh-circ#1 to explore the impacts of circ_0005909 silencing on viability, colony formation, migration, and invasion of OS cells. CCK-8 assay revealed that decreased circ_0005909 expression  impeded cell viability in HOS and U2OS cells (Fig. 2b). Cell colony formation assay exhibited that the colony formation ability of HOS and U2OS cells was inhibited by circ_0005909 downregulation (Fig. 2c). Transwell assay displayed that circ_0005909 reduction evidently curbed cell migration and invasion capacities in HOS and U2OS cells (Fig. 2d, e). Moreover, the EMT associated proteins, E-cad and N-cad, were detected via western blot analysis in circ_0005909-silenced HOS and U2OS cells. The results exhibited that E-cad protein expression was upregulated and N-cad protein expression was downregulated in circ_0005909-silenced HOS and U2OS cells (Fig. 2f). Subsequently, we explored the role of circ_0005909 in vivo through xenograft models. The results presented that tumor volume and weight of mice in the sh-circ#1 group were remarkably curbed compared to the control group (Fig. 2g, h). Collectively, these results indicated that circ_0005909 knockdown repressed OS cell growth in vitro and in vivo.

To investigate the molecular mechanism of circ_0005909 in OS, we predicted the target for circ_0005909 through the Circinteractome database. We discovered that 8 miRNAs (miR-338-3p, miR-411, miR-498, miR-515-5p, miR-561, miR-605, miR-766, and miR-936) had possible binding sites for circ_0005909. Silenced circ_0005909 expression could elevate the expression of miR-338-3p and miR-936 in HOS and U2OS cells, while there was no distinct difference in the other miRNAs (Additional file 1: Fig. S1). It was reported that circ_0005909 acted as a sponge for miR-338-3p in OS cells [10], so we chose miR-936 for subsequent researches. The potential binding sites between circ_0005909 and miR-936 were shown in Fig. 3a. After miR-936 mimics transfection, miR-936 expression was overtly increased in HOS and U2OS cells compared to the control group (Additional file 2: Fig S2A). Moreover, dual-luciferase reporter assay exhibited that miR-936 overexpression obviously suppressed the luciferase intensity of the circ_0005909-WT reporter vectors in HOS and U2OS cells, while the luciferase activity of the circ_0005909-MUT reporter vectors did not change (Fig. 3b, c). Also, RNA pull-down assay manifested that circ_0005909 was enriched in HOS and U2OS cells with Bio-miR-936 compared to the control group (Fig. 3d). Moreover, miR-936 expression was elevated in circ_0005909-inhibited HOS and U2OS cells (Fig. 3e). Furthermore, miR-936 expression was effectively reduced in HOS, U2OS, and SAOS-2 cells compared with the hFOB1.19 cells (Fig. 3f). The same expression trend of miR-936 was observed in OS tissues (Fig. 3g). Additionally, circ_0005909 and miR-936 had a negative correlation in OS tissues (Fig. 3h). We also explored the influence of miR-936 overexpression on viability, colony formation, migration, and invasion of OS cells. The results exhibited that miR-936 upregulation decreased cell viability, colony formation, migration, and invasion of U2OS and SAOS-2 cells (Additional file 2: Fig S2B–S2E). Together, these data indicated that circ_0005909 acted as a sponging for miR-936 in OS cells.

Given that miR-936 served as a target for circ_0005909 in OS cells, we explored whether circ_0005909 played its role via miR-936. In comparison to the anti-NC, the expression of miR-936 was curbed in HOS and U2OS cells transfected with anti-miR-936 (Fig. 4a). Moreover, the repression of viability and colony formation of HOS and U2OS cells mediated by circ_0005909 silencing was abrogated by miR-936 inhibitors (Fig. 4b, c). Also, the repressive impacts of circ_0005909 inhibition on migration and invasion of HOS and U2OS cells were overturned by miR-936 downregulation (Fig. 4d, e). Besides, silenced miR-936 expression reversed the effects of circ_0005909 inhibition on the levels of N-cad and E-cad proteins in HOS and U2OS cells (Fig. 4f, g). Taken together, these results demonstrated that circ_0005909 mediated cell viability, colony formation, migration, invasion, and EMT in OS cells via miR-936.

Subsequently, we further surveyed the downstream target for miR-936 via Targetscan database. Online bioinformatics prediction (TargetScan) exhibited that 8 genes (E2F3, GSK3B, MAPK8, TWIST1, HMGB1, wmt5a, ATM, and IGF1) might be downstream targets of miR-936. Furthermore, miR-936 mimics could reduce the level of HMGB1 mRNA in HOS and U2OS cells, so that HMGB1 was selected for subsequent studies (Additional file 3: Fig. S3). As exhibited in Fig. 5a, the 3′UTR of HMGB1 had latent binding sites for miR-936. Dual-luciferase reporter assay suggested that the luciferase activity of the luciferase reporter vectors with HMGB1 3′UTR-WT was reduced in HOS and U2OS cells transfected with miR-936, while there was no overt difference in the luciferase reporter vectors containing HMGB1 3′UTR-MUT (Fig. 5b, c). Elevated miR-936 expression reduced HMGB1 mRNA and protein levels in HOS and U2OS cells (Fig. 5d, e). Also, HMGB1 mRNA and protein levels were decreased in circ_0005909-inhibited HOS and U2OS cells, while this influence was restored by miR-936 downregulation (Fig. 5f, g). Therefore, these findings disclosed that HMGB1 acted as a target for miR-936 in OS cells.

Knowing that circ_0005909 regulated HMGB1 expression via miR-936 in OS cells, we further surveyed whether circ_0005909 modulated cell viability, colony formation, migration, invasion, and EMT in OS cells via HMGB1. Compared to the control vectors, HMGB1 protein level was evidently increased in HOS and U2OS cells after HMGB1 transfection (Fig. 6a). Furthermore, HMGB1 elevation recovered the inhibitory impacts of circ_0005909 silencing on viability and colony formation of HOS and U2OS cells (Fig. 6b, c). Moreover, forced HMGB1 expression overturned the suppressive influence on migration and invasion of HOS and U2OS cells caused by circ_0005909 inhibition (Fig. 6d, e). Additionally, both the upregulation of E-cad and the downregulation of N-cad in circ_0005909-blocked HOS and U2OS cells were reversed by HMGB1 overexpression (Fig. 6f, g). Therefore, these results suggested that circ_0005909 modulated cell malignant behaviors in OS cells via HMGB1.