Glioma tissues and samples of corresponding adjacent normal brain tissue (36 paired samples) were obtained from patients diagnosed with glioma and admitted to Soochow University. The tissues were stored in liquid nitrogen immediately after excision. Written informed consent was obtained from all the glioma patients who participated in this study.

Human normal astrocytes were purchased from Jennio Biological Technology (Guangzhou, China). Human glioma cell lines U87, LN229, U251 and T98G were purchased from Procell (Wuhan, China). All the cell lines were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (FBS, Gibco, USA). The cell lines were maintained in a humidified incubator (37°C and 5% CO₂). Small interfering RNA (siRNA), miR-545-3p mimic, miR-545-3p inhibitor, overexpression plasmid (full-length sequence of circ_0001367 was cloned into pGL3-Basic Vector) and the corresponding negative controls used in this study were all synthesized by GenePharma (Shanghai, China). Once the cells reached approximately 80% confluence, they were transfected with synthesized oligonucleotides using Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer’s protocol.

Total RNA was isolated from tissues and cells by using TRIzol (Invitrogen, USA). RNA was reverse transcribed into complementary DNA with the First Strand cDNA Synthesis Kit (MBI, Canada). SYBR Green Master Mix II (Takara, Japan) and the ABI 7900 system (Applied Biosystems, USA) were used to conduct qRT-PCR according to the manufacturers’ protocols. GAPDH and U6 were used as controls. The expression level of targets was determined by 2^−ΔΔCt method. The following primers used in this study. Circ_0001367, 5′-TGG GTC TAT CGT GCC GTT GA-3′ (forward) and 5′-GGA CAT CAT TTC ATT CCC AAG TA-3′ (reverse); miR-545-3p, 5′-TGG CTC AGT TCA GCA GGA AC-3′ (forward) and 5′-TGG TGT CGT GGA GTCG-3′ (reverse); LUZP1, 5′-ATG GCC GAA TTT ACA AGC TAC-3′ (forward) and 5′-TCA GTT CTC CTC AGC ACA GG-3′ (reverse); GAPDH, 5′-CAT CAC TGC CAC CC AG-3′ (forward) and 5′-ATG CCAG TGA GCT TC CC-3′ (reverse); U6, 5′-CTC GCT TCG GCA GCA CA-3′ (forward) and 5′-AAC GCT TCA CGA ATT TGC GT-3′ (reverse).

Total protein was extracted from cells with the Total Protein Extraction Kit (KeyGEN BioTECH, Shanghai, China). Protein concentration was determined with the BCA Protein Assay Kit (KeyGEN BioTECH, Shanghai, China). Then, the proteins were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes. After being blocked with blocking buffer for 10 min, the membranes were incubated with primary antibodies against LUZP1 (Proteintech, USA) and GAPDH (Proteintech, USA) overnight at 4°C. The membranes were then incubated with horseradish peroxidase-conjugated secondary antibody at room temperature for 1 h. Image Quant LAS (GE, USA) was used for signal detection.

The wild-type (WT) and mutant (MUT) sequences of circ_0001367 and miR-545-3p were synthesized and inserted into pmirGLO vectors. Next, LN229 and T98G cells were transfected with miR-545-3p mimic or NC mimic along with WT or MUT reporter plasmids by using Lipofectamine 3000 (Invitrogen, USA). After 48 h, the luciferase activity was determined by the Dual-Luciferase Reporter Assay (Promega, USA).

TUNEL staining was conducted by using the TUNEL Assay Kit (Beyotime, Shanghai, China). Cells were fixed with anhydrous ethanol and washed three times with PBS. Next, the cells were successively treated with Triton-X-100, terminal deoxynucleotide transferase (TdT), streptavidin HRP solution and DAB substrate according to the manufacturer’s protocol. The results were evaluated via microscope (Nikon, Japan).

After being isolated from LN229 and T98G cells, total RNA was incubated with RNase R or the mock treatment for 30 min (37°C). Then, the expression of circ_0001367 and GAPDH was determined by qRT-PCR (24).

Actinomycin D (2 mg/ml, APExBIO, USA) was added to culture medium to culture LN229 and T98G cells for 24 h. Next, the cells were harvested, and their expression of circ_0001367 and GAPDH was determined by qRT-PCR (24).

MTT assays were performed with an MTT kit (Beyotime, Shanghai, China). Cells were incubated with MTT solution (20 µl, 5 mg/ml) for 0, 24, 48 or 72 h. Then, dimethyl sulfoxide (DMSO, 4% 200 µl, Beyotime, Shanghai, China) was added to the cells. The absorbance at 490 nm was then determined with a microplate reader (Bio-Tek, Germany).

EdU assays were performed by using the EdU Cell Proliferation Kit with Alexa Fluor 488 (Beyotime, Shanghai, China). Cells were incubated with culture medium and EdU labelled for 4 h. After being fixed with anhydrous ethanol and treated with Triton-X-100, the cells were incubated with reaction buffer for 30 min and stained with DAPI for 5 min in darkness. The number of EdU-positive cells was measured with a fluorescence microscope (Olympus, Japan).

Cells were suspended in FBS-free culture medium (200 μl) and added to the upper chamber of Transwell inserts (Millipore, Germany). Medium (500 μl) containing 10% FBS was added to the lower chamber. After incubation in a humidified incubator (37°C and 5% CO₂) for 24 h, the non-migrated cells in the upper chamber were wiped with a wet cotton swab. The remaining cells were fixed in anhydrous ethanol and stained with 0.5% crystal violet. After being washed with PBS, the stained cells were imaged and counted under a microscope (Nikon, Japan). The invasion assay was conducted following the same steps described above for the migration assay except that the chambers were precoated with Matrigel (BD Biosciences, USA).

Nude mice aged 4-5 weeks were obtained from Shanghai SLAC Laboratory Animal Company (Shanghai, China). Stable cell lines LN229 overexpressing circ_0001367 or corresponding negative control were administered into the dorsal surface of the mice shoulder. Three weeks later, the tumors formed in the mice were excised. Tumor volume and weight were determined (volume = 0.5 ×length×width²). In addition, the expression of circ_0001367, miR-545-3p and LUZP1 was determined by qRT-PCR, western blot or immunohistochemistry staining.

The results of this study are presented as the mean ± standard deviation. SPSS software 19.0 (SPSS, USA) was used for statistical analysis. Student’s t-test was used for comparisons between groups. P <0.05 was considered statistically significant.

qRT-PCR was used to measure circ_0001367 in the clinical samples. As shown in Figure 1A , circ_0001367 was significantly downregulated in glioma tissues compared with adjacent normal brain tissues (ANTs). To determine whether circ_0001367 expression was associated with clinical features, Kaplan-Meier analysis was performed. The results indicated that patients in the low circ_0001367-expression group had a shorter survival time than those in the high-expression group ( Figure 1B ). In addition, we detected circ_0001367 expression in glioma cell lines and found that circ_0001367 was obviously downregulated in glioma cell lines compared with NAs ( Figure 1C ). RNase R digestion demonstrated that circ_0001367 was resistant to RNase R ( Figure 1D ). In addition, actinomycin D treatment showed that the circular structure of circ_0001367 was stable ( Figure 1E ). These findings suggested that circ_0001367 is involved in gliomagenesis.

To explore the role of circ_0001367 in gliomagenesis, circ_0001367 was upregulated or knocked down in LN229 and T98G cells by transfecting the cells with overexpression plasmid or small interfering RNA (siRNA). The transfection efficiency was verified by qRT-PCR ( Figures S1A, B ). MTT assays and EdU assays demonstrated that circ_0001367 overexpression obviously suppressed the proliferation of glioma cells. In contrast, the proliferation capacity of glioma cells was markedly promoted in circ_0001367-knockdown groups ( Figures 2A–F ). Furthermore, Transwell assays were performed to assess the effects of circ_0001367 on the migration and invasion of glioma cells. We found that circ_0001367 overexpression significantly inhibited the migration and invasion ability of glioma cells, whereas circ_0001367 knockdown markedly enhanced the migration and invasion of glioma cells ( Figures 2G, H ). Taken together, these results suggest that circ_0001367 inhibits the proliferation, migration and invasion of glioma cells in vitro.

The data obtained from starBase showed the presence of binding sites between circ_0001367 and miR-545-3p ( Figure 3A ). The luciferase reporter assay indicated that miR-545-3p was able to obviously reduce the luciferase activity of circ_0001367-WT cells; however, it had no pronounced effect in the circ_0001367-MUT group ( Figure 3B ). Furthermore, RIP assay verified that circ_0001367 and miR-545-3p were enriched in the anti-Ago2 group ( Figures 3C, D ). Next, we detected the expression of miR-545-3p in glioma tissues and cell lines. The results showed that miR-545-3p was upregulated in glioma tissues and cell lines compared with ANTs and NAs, respectively ( Figures 3E, F ). We also detected the level of miR-545-3p in circ_0001367-overexpression LN229 and T98G cells. The results demonstrated that miR-545-3p was downregulated when circ_0001367 was overexpressed ( Figure 3G ). These findings suggest that circ_0001367 acts as a sponge of miR-545-3p.

To investigate whether circ_0001367 exerts its function in LN229 and T98G cells by sponging miR-545-3p, function assays were conducted. The transfection efficiency of circ_0001367-overexpression plasmid and miR-545-3p mimics was verified by qRT-PCR ( Figure S1C ). The MTT assays and EdU assays demonstrated that circ_0001367 overexpression obviously decreased the proliferation of glioma cells and that this inhibitory effect could be reversed by miR-545-3p mimic ( Figures 4A–D ). Similarly, the Transwell assays showed that miR-545-3p mimic could reverse the inhibition of migration and invasion caused by circ_0001367 overexpression ( Figures 4E, F ). These results suggest that circ_0001367 suppresses the proliferation, migration and invasion of glioma cells by absorbing miR-545-3p.

The data from starBase indicated that the 3’-UTR of LUZP1 contains sites that can bind to miR-545-3p ( Figure 5A ). The luciferase reporter assay indicated that miR-545-3p could reduce the luciferase activity of LUZP1-WT cells much more than that of LUZP1-MUT cells ( Figure 5B ). By searching the online database GEPIA, we found that LUZP1 was downregulated in glioma tissues ( Figure 5C ). Next, we detected LUZP1 expression in clinical samples by qRT-PCR. The results demonstrated that LUZP1 was downregulated in glioma tissues compared with ANTs ( Figure 5D ). Consistent with this finding, LUZP1 was downregulated in LN229 and T98G cells compared with NAs ( Figures 5E, F ). Furthermore, we detected LUZP1 expression in the cell models we constructed previously. qRT-PCR and western blot indicated that hsa_circ_0001367 overexpression could promote expression of LUZP1, meanwhile, miR-545-3p mimic could suppress expression of LUZP1 ( Figures 5G–J ). These findings indicate that LUZP1 is the direct target of miR-545-3p.

To investigate whether LUZP1 is the functional target of miR-545-3p in LN229 and T98G cells, we constructed four cell models. The efficiency of miR-545-3p inhibitor and si-LUZP1 transfection was verified by qRT-PCR ( Figure S1D ). The MTT and EdU assays suggested that LUZP1 deletion could block the suppressive effect of miR-545-3p inhibitor on glioma proliferation ( Figures 6A–D ). Consistent with this finding, the Transwell assays showed that the migration and invasion ability of LN229 and T98G cells was impaired by miR-545-3p inhibitor, whereas LUZP1 deletion reversed this inhibitory effect ( Figures 6E, F ). In addition, function assays indicated that LUZP1 deletion could block the suppressive effect of circ_0001367 overexpression on glioma proliferation, migration and invasion ( Figures 7A–F ). These results suggested that LUZP1 downregulation restored the effect of miR-545-3p knockdown and circ_0001367 overexpression on glioma cells.

To evaluate the effect of circ_0001367 on glioma growth in vivo, LN229 cells transfected with circ_0001367-overexpression plasmid or corresponding negative control were subcutaneously injected into mice. Three weeks later, tumors had formed. Compared with those in the negative control group, the tumors formed in the circ_0001367-overexpression group were lower in weight and smaller in volume ( Figures 8A–D ). qRT-PCR showed that circ_0001367 was upregulated whereas miR-545-3p was downregulated in the tumors formed in the circ_0001367-overexpression group ( Figure 8E ). Furthermore, qRT-PCR and western blot revealed that LUZP1 was upregulated in the circ_0001367-overexpression group relative to the control group ( Figures 8E, F ). Furthermore, Ki-67 staining indicated that circ_0001367 overexpression decreased the proportion of Ki-67-positive cells ( Figure 8G ). TUNEL staining showed that more apoptosis occurred in the circ_0001367-overexpression group than in the control group ( Figure 8H ).