A total of 86 human glioma specimens and 20 normal brain tissues were collected from the People’s Hospital of Gaozhou. Diagnoses were confirmed by pathologist. None of the patients received preoperative immunotherapy, chemotherapy or radiotherapy. These clinical samples were immediately fixed with formalin and embedded with paraffin after surgical resection. The written informed consents were obtained from every patient included in this study. The clinical stage was based on World Health Orgnization (WHO) grade (19). The demographic and clinical information of patients are shown in Table I. The protocols involving clinical tissues in the present study were approved by the Ethics Review Committee of the People’s Hospital of Gaozhou.

The normal human astrocyte (NHA) cell line was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). The human glioma cell lines (U87, T98, A172 and U251) were purchased from the Chinese Academy of Sciences (Shanghai, China). All cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; HyClone Laboratories, Logan, UT, USA) along with fetal bovine serum (10%) (FBS; HyClone Laboratories), penicillin (100 U/ml) and streptomycin (100 μg/ml). Cell cultures were kept in an incubator containing 5% CO₂ and humidified atmosphere at 37°C.

Before IHC staining, glioma specimens and normal brain tissues were fixed with formalin and embedded with paraffin. Then, the embedded tissues were cut into 4 μm thick sections. IHC staining followed standard protocol to evaluate the expression level of BCL3 (sc-185; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and STAT3 (#9139; Cell Signaling Technology, Boston, MA, USA) in human tissues. The percentage of positive tumor cells was graded as per the following criteria: 0, <10%; 1, 10–30%; 2, 31–50%; and 3, >50%.

BCL3 shRNA (sc-29789-SH) and corresponding non-target (NT) shRNA (sc-108060) were obtained from Santa Cruz Biotechnology. shRNAs transfection was performed using plasmid transfection reagent (sc-108061; Santa Cruz Biotechnology). Retroviral vector pMMP-BCL3 and pMMP-STAT3 were generated by inserting the cDNA into pMMP. Retrovirus packaging and transduction were previously described (20). A STAT3 specific siRNA and scrambled siRNA was obtained from OriGene Technologies, Inc. (Shanghai, China). The siRNAs were transduced into glioma cells with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol.

Glioma cells that were treated with corresponding vectors were seeded into 96-well plates (1.5×10³ cells/well); 24, 48, 72 and 96 h after transfection, the cell proliferation assay was performed by addition of 10 μl Cell Counting kit-8 (CCK-8) solution (Beyotime Institute of Biotechnology, Shanghai, China) to each well, followed by incubation at 37°C for 2 h. Absorbance was measured at a wavelength of 490 nm using a microplate reader (FlexStation III ROMV2.1.28; Molecular Devices, Sunnyvale, CA, USA). For 5-bromodeoxyuridine (BrdU) assays, 48 h after transfection, glioma cells grown on coverslips were incubated with BrdU at room temperature for 60 min and subsequently were incubated with the antibody of BrdU (Sigma-Aldrich, St. Louis, MO, USA) following the manufacturer’s protocol.

Forty-eight hours after transfection, glioma cells were collected with trypsin and fixed overnight at 4°C using 80% ethanol. Glioma cells were then incubated with propidium iodide (PI; Sigma-Aldrich) for 20 min. Then, flow cytometry assays for glioma cells were performed with a FACSCalibur (BD Biosciences, Bedford, MA, USA).

Apoptosis of cells was evaluated with the Annexin-V-FLUOS staining kit (Roche Diagnostics, Indianapolis, IN, USA) following standard protocol provided by the manufacturer. Then flow cytometry assay was performed with a FACSCalibur (BD Biosciences).

Total proteins were collected with RIPA lysis buffer, and 40 μg protein were subjected to 4–20% SDS gel electrophoresis and were then transferred to PVDF membranes. Then, 5% milk blocked membranes were incubated with BCL3 (Santa Cruz Biotechnology), STAT3 (Cell Signaling Technology), p-STAT3 (Cell Signaling Technology), BCL2 (Cell Signaling Technology), MCL-1 (Cell Signaling Technology) or cyclin D1 (Cell Signaling Technology) antibodies, respectively, and subsequently incubated with matched secondary antibodies (Cell Signaling Technology). Then, signals for each protein expression were detected with the Bio-Rad Gel imaging system. GAPDH (Cell Signaling Technology) was used as a loading control.

The subcutaneous tumor formation experiments were performed on nude mice. U251 cells (5×10⁶) with BCL3 knockdown or cells of control group were suspended in PBS and were implanted into the back of mice via subcutaneous injection. Tumor volume was measured with calipers every 3 days. The xenograft tumor tissues were subjected to immunoblotting for Ki-67 (Cell Signaling Technology) and caspase-3 (Cell Signaling Technology). The protocols of in vivo experiments were approved by the Institutional Animal Care and Use Committee of the People’s Hospital of Gaozhou.

All data were collected and showed as mean ± standard error (SE). GraphPad Prism 5 software (GraphPad Software, Inc., San Diego, CA, USA) was used in the present study to perform statistical analysis. P<0.05 was considered to be statistically different.

IHC was initially performed to detect BCL3 expression in glioma specimens and normal brain tissues. IHC sections were scored according to the percentage of positive cells (Fig. 1). The level of BCL3 was significantly upregulated in glioma tissues as compared with normal brain tissues (P<0.05; Fig. 2A). Furthermore, our data revealed that BCL3 expression in four glioma cell lines (A172, U87, U251 and T98 cells) was evidently increased compared to an NHA cell line (P<0.05, respectively; Fig. 2B). BCL3 staining was considered as positive expression in 46 cases of glioma according to IHC scores (1–3). As shown in Table I, positive expression of BCL3 in glioma patients was positively correlated with large tumor size (P=0.012), high Karnofsky performance status (KPS) score (P=0.020) and advanced WHO grade (P=0.029). Then, we examined the prediction value of BCL3 for the prognosis of glioma patients. Compared with those with negative expression of BCL3, patients that showed positive expression of BCL3 had significant reduced 5-year overall survival (P<0.001; Fig. 2C). The above indicate that BCL3 can potentially act as biomarker and prognostic indicator for glioma.

Next, we disclosed the exact roles of BCL3 in glioma cells. Since increased proliferation, aberrant cell cycle progression and reduced apoptosis are well-recognized hallmarks of human cancer (21), we examined whether BCL3 regulated these biological processes of glioma cells. BCL3 knockdown was confirmed by immunoblotting after BCL3 shRNA transfection (P<0.05; Fig. 3A). CCK-8 and BrdU assays indicated that BCL3 knockdown significantly reduced the proliferative ability of U251 cells (P<0.05, respectively; Fig. 3B and C). Moreover, BCL3 knockdown led to a prominent cell cycle arrest at G1-phase (P<0.05; Fig. 3D). The portion of apoptotic cells were evidently increased after BCL3 silencing (P<0.05; Fig. 3E). On the contrary, infection of BCL3 retroviruses led to a significantly increased expression of BCL3 in U87 cells (P<0.05; Fig. 4A). Functionally, BCL3 overexpression facilitated cell proliferation (P<0.05, respectively; Fig. 4B and C) and cell cycle progression (P<0.05; Fig. 4D), while suppressed apoptosis in U87 cells (P<0.05; Fig. 4E). Furthermore, we used a subcutaneous tumor formation model to investigate whether BCL3 could affect the growth of glioma cells in a mouse xenograft model. As shown in Fig. 5A, BCL3 knockdown significantly reduced the tumor growth of U251 cells in mice (P<0.05). Furthermore, immunoblotting of Ki-67 and caspase-3 indicated BCL3 knockdown inhibited proliferation and induced apoptosis of glioma cells in vivo (P<0.05, respectively; Fig. 5B). Thus, BCL3 functions as an oncogene by promoting proliferation and the process of cell cycle, and inhibiting apoptosis in glioma.

A previous study demonstrated that STAT3, an important oncogene in human cancer, was a bona fide target of BCL3 in cervical cancer cell line (17). U251 cells that were transduced with NT shRNA and BCL3 shRNA, respectively, were subjected to immunoblotting for BCL3, STAT3 and p-STAT3 expression. As expected, BCL3 knockdown significantly reduced STAT3 and p-STAT3 expression in vitro (Fig. 6). Since BCL2, MCL-1 and cyclin D1 are downstream molecules involved in STAT3 signaling pathway (22). Notably, BCL3 knockdown coordinately reduced the levels of BCL2, MCL-1 and cyclin D1 in U251 cells (Fig. 6). In contrast, BCL3 overexpression resulted in an obvious increase of STAT3, p-STAT3, BCL2, MCL-1 and cyclin D1 protein expression (Fig. 6) in U87 cells. Furthermore, IHC was performed to detect the expression of STAT3 in glioma specimens. Interestingly, Spearman’s correlation analysis indicated that BCL3 was positively correlated with STAT3 expression in glioma specimens (r=0.673, P<0.001; Fig. 7).

To determine whether STAT3 silencing could attenuate the oncogene function of BCL3 in glioma, a series of rescue experiments were performed. STAT3 siRNA were employed for STAT3 knockdown in BCL3 overexpressing U87 cells (P<0.05; Fig. 8A). Notably, STAT3 knockdown abolished the oncogenic effects of BCL3 overexpression on U87 cells with decreased cell proliferation (P<0.05, respectively; Fig. 8B and C), cell cycle arrest (P<0.05; Fig. 8D) and increased apoptosis (P<0.05; Fig. 8E). Our data indicate that STAT3 mediates the functions of BCL3 in glioma cells.