The present study was reviewed and approved by the ethical review board of Xinxiang Medical University (Xinxiang, China). All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. All procedures performed involving human participants were in accordance with the ethical standards of the institutional and national research committee, and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was performed following the provision of written informed consent from patients. A total of 72 fresh glioma tissues and their adjacent non-glioma tissues were collected during surgery between December 2014 and December, 2015. The levels of CISD2 were evaluated in these paired tissues. In order to determine the association between CISD2 and the prognosis of patients with glioma, 120 paraffin-fixed glioma specimens were collected from patients between January 2008 and January 2010, and these patients were followed up 40 months later.

The U87 glioma cell line was purchased from the Chinese Academy of Sciences (Shanghai, China) and was cultured in DMEM (Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% FBS (Thermo Fisher Scientific, Inc.). Please note that the U87 cell line is known to be cross-contaminated with another cell line, which is most likely to be a glioblastoma cell line (23). The cells were maintained in a humidified 37°C incubator containing 5% CO₂. For autophagy inhibition, the cells were treated with 1 mM 3-MA (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) for 24 h at 37°C.

BALB/c nude male mice (6 weeks old) were purchased from the Laboratory Animal Center of Xinxiang Medical University. All mice were housed in a strictly pathogen-free conditions at room temperature with free access to food and water and 12 h/12 h light/dark cycle. The protocols for the experiments involving mice complied with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Bethesda, MD, USA) and approved by the Animal Ethics Committee of the Xinxiang Medical University.

Following being transfected with the small interfering RNA to knock down CISD (si-CISD) or a scramble construct, the U87 cells (5×10⁷) were harvested and subcutaneously inoculated into the right groin of the nude mice (n=3 per group). Following growth for 40 days, the formed tumors were carefully excised. The weight of the formed tumor was measured and the volume was calculated using the following formula: Length × width² × π/6.

TRIzol reagent (Thermo Fisher Scientific, Inc.) was used to extract RNA from samples. Equal quantities of RNA were then reverse transcribed into cDNA with SuperScript^® IV Reverse Transcriptase (Thermo Fisher Scientific, Inc.). A total of 2 µg CISD2 cDNA was amplified under the following thermal conditions: 95°C for 10 min; followed by 40 cycles of 95°C for 15 sec and 60°a for 1 min; and 4°C holding; the total duration was 1 h 48 min. GAPDH was used as an internal control. The final results were calculated using the 2^−ΔΔCq (24) method and presented as fold changes. The following primers were used: CISD2, forward 5′-GCAAGGTAGCCAAGAAGTGC-3′ and reverse 5′-CCCAGTCCCTGAAAGCATTA-3′; GAPDH, forward 5′-GCGAGATCGCACTCATCATCT-3′ and reverse 5′-TCAGTGGTGGACCTGACC-3.

Briefly, the tissues were lysed with radioimmunoprecipitation assay lysis buffer containing inhibitor cocktail for protease and phosphatase (Thermo Fisher Scientific, Inc.). The concentration of protein sample was determined using the bicinchoninic acid method. Following boiling for 15 min, 15-µg protein samples were separated on a 10% SDS gel, blocked with 5% nonfat milk and transferred onto a PVDF membrane (EMD Millipore, Billerica, MA, USA). The membrane was incubated with primary antibody overnight at 4°C. Following washing with TBST five times, the membrane was incubated with secondary antibody for 2 h at 27°C. The membrane was washed with TBST five times, following which the membrane was visualized with Pierce^™ Enhanced Chemiluminescence Plus Western Blotting Substrate (Thermo Fisher Scientific, Inc.). GAPDH was used as an internal control. The protein band intensity was measured using Image J software (version 1.49; National Institutes of Health, Bethesda, MD, USA).

The primary antibodies used in the present study were all purchased from Sigma-Aldrich (Merck KGaA) and were as follows: Anti-CISD2 (1:1,000; cat no. AV44552), anti-LC3-II (1:800; cat. no. ABC432), anti-beclin-1 (1:1,000; cat. no. SAB1306484), anti-autophagy related 7 (Atg7) (1:1,000; cat. no. MABN1124), anti-p62 (1:1,000; cat. no. MABC32) and anti-GAPDH (1:1,000; cat. no. G9545). The secondary antibodies used were as follows: Horseradish peroxidase (HRP)-labeled goat anti-rabbit immunoglobulin (Ig)G (1:3,000; cat. no. A0208; Beyotime Institute of Biotechnology, Haimen, China) and HRP-labeled goat anti-mouse IgG (1:3,000; cat. no. A0216; Beyotime Institute of Biotechnology).

Immunohistochemical staining was performed on the collected clinical glioma tissues sections to detect the levels of CISD2. Briefly, formalin and paraffin were used to fix and embed the tissues, respectively. The samples were heated to retrieve the antigen. The sections were incubated with anti-CISD2 primary antibody (1:500; cat. no. AV44552; Sigma-Aldrich; Merck KGaA) at 4°C overnight. Following three washes with PBS, the sections with incubated with HRP-labeled goat anti-rabbit IgG (1:1,000; cat. no. A0516; Beyotime Institute of Biotechnology) for 2 h. Following three washes with PBS, the sections were visualized with DBA solution, followed by counterstaining of nuclei with hematoxylin. Images were captured using a microscope (Eclipse Ci-E; Nikon, Tokyo, Japan).

A Pierce Co-IP kit (Thermo Fisher Scientific, Inc.) was used to examine the binding activity between CISD2 and beclin-1 in the indicated groups. The general procedure was performed according to the manufacturer's protocol, as previously described (25). The protein extracts were precipitated using anti-CISD2 (Sigma-Aldrich; Merck KGaA), and the precipitated protein was evaluated using western blot analysis with anti-beclin-1 (Sigma-Aldrich; Merck KGaA).

Immunofluorescence was also used to determine the levels of CISD2 in the glioma tissue sections. Briefly, following fixation with 4% paraformaldehyde, the sections were incubated with anti-CISD2 primary antibody overnight at 4°C. Following three washes with PBS, the sections were incubated with Cy3-labeled goat anti-rabbit IgG (1:1,000; cat. no. A0516; Beyotime Institute of Biotechnology) for 2 h at 27°C. Images were captured using a laser confocal microscope (A1; Nikon Corporation, Tokyo, Japan).

Briefly, when the cells reached a confluence of 70%, the cells were transfected with negative scramble siRNA, a CISD2-overexpression plasmid or an si-CISD2 and/or si-beclin-1 plasmid using Lipofectamine^® 2000 transfection reagent (Thermo Fisher Scientific, Inc.). After 6 h, the DMEM was replaced with normal medium. All plasmids were purchased from GenePharma Co., Ltd. (Shanghai, China).

The U87 glioma cells were transfected with the scramble or si-CISD2 plasmid and seeded into a 96-well plate at a confluence of 30% (3×10⁴ cells/well). At the indicated time points, the number of cells in each well was determined using a Scepter Handheld Automated Cell Counter (EMD Millipore).

A TUNEL assay was used to measure apoptosis. Briefly, cells grown on a cover slip were fixed with 4% paraformaldehyde. Following washing with PBS, the cells were incubated with 0.3% H₂O₂ to block endogenous peroxidase activity. The cells were then incubated with TUNEL reaction solution (Sigma-Aldrich; Merck KGaA) for 1 h at 37°C. Images were captured under a laser confocal microscope (A1; Nikon).

The U87 glioma cells were transfected with scramble or si-CISD2 plasmid and seeded into a 6-well plate at a confluence of 30% (3×10⁴ cells/well). When cell confluence reached 95%, the cells were starved for 12 h. A 100-ml pipette tip was then used to scratch a straight line in the cell layer. Following incubation for another 24 h, the cells were fixed and images were captured under a microscope (Eclipse Ci-E; Nikon Corporation). The length of the wound was measured using Image J software.

CISD2 wild-type 3-untranslated region (UTR) and mutated 3′-UTR constructs were sub-cloned into the pGL3 Luciferase Promote Vector (Sangon Biotech Co., Ltd., Shanghai, China) with XbaI and NotI restriction sites. Using Lipofectamine^® 2000 transfection reagent (Thermo Fisher Scientific, Inc.), the pGL3 vector containing the CISD2 wild-type 3′-UTR or mutated form was co-transfected with or without miR-449a mimic (GenePharma Co., Ltd.) into the U87 cells. At 48 h post-transfection, the luciferase activity was measured using a Luciferase Reporter Assay kit (Sangon Biotech Co., Ltd.).

Data are expressed as the mean ± standard deviation of at least three independent experiments. Comparisons between two groups were analyzed using Student's t-test (two-tailed). Comparisons among groups were analyzed using one-way analysis of variance followed by Student-Newman-Keuls test. All analyses were performed with SPSS 19.0 software (IBM SPSS, Armonk, NY, USA). P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR analysis was performed to measure the mRNA levels of CISD2 in 72 fresh glioma tissue samples and corresponding non-glioma tissue samples. The results showed that the mRNA level of CISD2 was significantly increased in the glioma tissues, compared with that in the non-glioma tissues (Fig. 1A). In addition, elevated expression of CISD2 in glioma tissues at the protein level was confirmed using western blot analysis (Fig. 1B), immunohistochemistry and immunofluorescence (Fig. 1C). Taken together, the above data demonstrated that the mRNA and protein levels of CISD2 were markedly elevated in glioma tissues, compared with non-glioma tissues. As CISD2 was significantly increased in the glioma tissues, it was hypothesized that CISD2 may be a predictor of the survival rates of patients with glioma. A total of 120 paraffin-fixed glioma specimens were collected between January 2008 and January 2010 and the corresponding survival rates of the patients were recorded. The results showed that the level of CISD2 was negatively correlated with the survival rates of the patients (Fig. 1D). The associations between the level of CISD2 and clinicopathological characteristics were also analyzed in these patients (Table I). A high level of CISD2 was associated with advanced clinical stage (P<0.05), relapse (P<0.05), vascular invasion (P<0.05) and increased tumor size (P<0.05), but not differentiation (P>0.05).

To examine the role of CISD2 in glioma cells, the level of CISD2 was upregulated using a CISD2 expression construct or knocked down using siRNA in U87 glioma cells. These effects were validated using western blot analysis (Fig. 2A). Compared with the vector control cells, the upregulation of CISD2 significantly increased the proliferation rate of the U87 cells (Fig. 2B). The overexpression of CISD2 also significantly increased the mean number of colonies in the colony formation assay (Fig. 2C). However, the knock down of CISD2 significantly reduced the proliferation rate and number of colonies of glioma cells, as indicated in the proliferation and colony formation assays, compared with the control or CISD2 upregulation groups (Fig. 2B and C). In addition, the TUNEL assay revealed that the inhibition of CISD2 markedly increased the apoptosis of U87 cells, compared with the control or CISD2 overexpression groups (Fig. 2D). The wound-healing assay showed that, compared with the control group, the upregulation of CISD2 enhanced the invasive ability of the U87 cells, whereas si-CISD2 significantly inhibited the invasion of U87 cells (Fig. 2E). Taken together, these results suggested that CISD2 promoted the proliferation and survival of glioma cells.

The present study also established a xenograft model in BALB/C nude mice using U87 cells to determine the in vivo effects of silencing CISD2. The results showed the silencing of CISD2 led to a significantly slower growth rate, compared with that in the control group (Fig. 3A). In addition, the CISD2-deficit U87 cells formed smaller tumors, compared with those in the control vector-transfected cells (Fig. 3B). The average volume and weight of tumors were significantly lower in the CISD2-deficit group, compared with those in the control group (Fig. 3C and D). Western blot analysis confirmed that CISD2 was effectively knocked down in the formed tumors of the si-CISD2 group, compared with that in the control group (Fig. 3E). Taken together, these data demonstrated that silencing of CISD2 significantly reduced the ability of glioma cells to form tumors in vivo.

In order to examine the involvement of autophagy in the tumor-inhibitory effects of si-CISD2 in glioma cells, CISD2 was manipulated and the levels of autophagy activity markers were evaluated using western blot analysis. The resulting data showed that, compared with the control group, silencing of CISD2 led to significant increases in the levels of LC3-II, beclin-1 and Atg7, and a decrease in selective autophagy target p62; this was reversed by the overexpression of CISD2 (Fig. 4A-E). The above data demonstrated that the silencing of CISD2 activated autophagy whereas the overexpression of CISD2 inhibited autophagy in glioma cells.

In order to confirm whether CISD2 silencing-induced glioma cell death was dependent on autophagy, the specific autophagy inhibitor, 3-MA, was used. Compared with the control group, si-CISD2 led to significant increases in cell apoptosis (Fig. 5A) and necrosis (Fig. 5B), however, these effects were eliminated by 3-MA. Consistently, the colony formation assay showed that si-CISD2 markedly reduced the mean number of colonies, compared with that in the control group, which was reversed by 3-MA (Fig. 5C). Finally, the wound-healing assay revealed that si-CISD significantly attenuated the invasive ability of U87 cells, which was also abrogated by 3-MA (Fig. 5D). Taken together, these data suggested that inhibiting autophagy significantly alleviated the CISD2 silencing-induced suppression of glioma cell proliferation and survival.

Previous studies have shown that beclin-1 is vital in the activation of autophagy (26). Therefore, the present study hypothesized that the enhanced autophagy by CISD2 silencing is dependent on beclin-1. The Co-IP assay showed that endogenous CISD2 was able to bind with beclin-1 (Fig. 6A). In addition, U87 cells with si-CISD2 and si-beclin-1 vectors were co-transfected, and the activity of autophagy activity was evaluated using western blot analysis. The results of the western blot analysis confirmed that CISD2 and beclin-1 were effectively knocked down (Fig. 6B-D). The data revealed that the downregulation of beclin-1 significantly eliminated the si-CISD2-induced increase in LC3-II and Atg7, and decrease in p62 in the U87 cells (Fig. 6E-G). Taken together, these data demonstrated that the CISD2 silencing-induced activation of autophagy was mediated by beclin-1.

Using TargetScan, the present study found that CISD2 was one of the potential targets of miR-449a. The predicted binding site of miR-449a with the CISD2 3′-UTR is shown in Fig. 7A. To confirm the interaction between miR-449a and CISD2, the CISD2 complementary sites, with or without mutations, were cloned into the 3′-UTR of the firefly luciferase gene and co-transfected with miR-449a mimics into U87 cells. The results revealed that the presence of miR-449a led to a significant reduction in the relative luciferase activity of the wild-type construct of the CISD2 3′-UTR in U87 cells. However, the mutant construct of the CISD2 3′-UTR eliminated the suppressive effect of miR-449a in the U87 cells (Fig. 7B). The above results demonstrated that CISD2 is a direct target of miR-449a.