The SHG139s GSC line was developed and provided by the Neurosurgery and Brain and Nerve Research Laboratory, The First Affiliated Hospital of Soochow University, (Suzhou, China). The SHG139s cell line was maintained in stem-cell permissive medium [Dulbecco's modified Eagle's medium (DMEM)-F12 containing 20 ng/ml epidermal growth factor, basic fibroblast growth factor (bFGF; R&D Systems, Inc., Minneapolis, MN, USA), nitrogen gas (dilution, 1:50) and B27 (dilution, 1:50; Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA)].

The cells were dissociated using 0.25% trypsin (Beyotime Institute of Biotechnology, Haimen, China) and resuspended in phosphate-buffered saline (PBS). All reagents and supplies for MACS separation were purchased from Miltenyi Biotec GmbH (Bergisch-Gladbach, Germany). Selection of CD133⁻ SHG139s cells was performed, according to the manufacturer's instructions, using CD133/1 Micro Beads. Subsequently, the CD133⁻ cells were separated from the A2B5⁺ cells and A2B5⁻ cells, according to the manufacturer's instructions, using A2B5 Micro Beads.

The CD133⁻/A2B5⁺and CD133⁻/A2B5⁻ cells were transferred onto Matrigel-coated invasion chambers (24-well insert, 8-µm pores; BD Biosciences, Franklin Lakes, NJ, USA), containing serum-free DMEM. DMEM containing 10% fetal bovine serum was added to the lower chamber as a chemoattractant. Following an incubation period at 37.5°C for 48–72 h, non-invading cells were removed from the inner part of the insert using a cotton swab. The cells on the lower membrane surface were fixed in 4% formaldehyde (Beyotime Institute of Biotechnology) and stained with 0.1% crystal violet (Beyotime Institute of Biotechnology). The number of invading cells were manually counted in five randomly-selected fields under a microscope (CKX41SF inverted microscope; Olympus Corporation, Tokyo, Japan) and images were captured.

The primary antibodies used in the present study were polyclonal rabbit anti-human tissue inhibitor of metalloproteinase 3 (TIMP3; cat. no. BA0577), polyclonal rabbit anti-human E-cadherin (cat. no. PB0583), polyclonal rabbit anti-human matrix metalloproteinase (MMP) 2 (cat. no. BA0569) and anti-MMP9 (cat. no. BM0573) all purchased from Wuhan Boster Bioengineering Co., Ltd. (Wuhan, China). Total protein from the cells was directly extracted in lysis buffer (Beyotime Institute of Biotechnology) and the concentration of total protein was quantified using an ultraviolet spectrophotometer (Multiskan Mk3; Thermo Fisher Scientific, Waltham, MA, USA). Protein samples (100 or 50 µg) were separated using 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Beyotime Institute of Biotechnology). The proteins were then transferred onto nitrocellulose membranes (Beyotime Institute of Biotechnology) and non-specific binding was blocked by incubating the membranes in 5% non-fat milk. The membranes were incubated at 37°C with primary antibodies overnight at 4°C. Following incubation, the membranes were washed with PBS with Tween-20 and incubated at 37°C for 2 h with horseradish peroxidase (HRP)-conjugated monoclonal goat anti-rabbit secondary antibodies (cat no. Beyotime Institute of Biotechnology; cat. no. A0208), followed by detection and visualization using electrochemiluminescence western blotting detection reagents (Pierce Biotechnology, Inc.; Thermo Fisher Scientific, Inc.). Quantification of protein expression was performed by measuring the gray-scale value of bands using ImageJ software (version 2.1.4.7; National Institutes of Health, Bethesda, MD, USA).

A total of 16 nude mice were separated into two groups (each group comprised of four female and four male mice). The mice were aged ~4–6 weeks and weighed 24–28g. The mice were raised in specific-pathogen free conditions, and the temperature was maintained between 26 and 28°C. The mice were exposed to ~10 h light per day, and given food and water following high-temperature sterilization. The mice were raised separately, but under the same conditions and were fed standard chow. To investigate the effects of A2B5 expression on tumor growth in vivo, cells (1×10⁶) from the two groups (CD133⁻/A2B5⁺ and CD133⁻/A2B5⁻) were injected into the left axilla of nude mice (n=8). For the injection, the cells were suspended in DMEM-F12 to the same concentration (1×10⁷ cells/ml) and 100 µl cell suspension was injected into each of the mice on the same day. On day 17 post-implantation, caliper measurements were performed to assess tumor growth.

Mice received an intraperitoneal injection of 45 mg/kg pentobarbital sodium (Shanghai Westang Bio-tech Co., Ltd., Shanghai, China) for anesthetization. The tumors (weight, 1–2.5 g) were excised using tweezers and scissors, and the mice were sacrificed by cervical dislocation. The formalin-fixed paraffin-embedded (Qilin Environmental Technology Firm, Guangzhou, China) SHG139s tumors were cut into 6-µm sections using a microtome (RM2016; Leica Biosystems, Nussloch, Germany). Antigen retrieval was performed in 10 mM sodium citrate buffer (pH 6; Beyotime Institute of Biotechnology) for 16 min at 96–98°C. The slides were then incubated with primary antibodies against MMP9, MMP2, E-cadherin, TIMP3 and Ki-67 (cat. no. BA2888; Wuhan Boster Bioengineering Co., Ltd.), and with antibodies against vascular endothelial growth factor (VEGF; cat. no. ab46154), VEGF receptor 2 (VEGFR2; cat. no. ab2349) and CD34 (cat. no. ab81289; Abcam, Tokyo, Japan). The sections were subsequently incubated with a cell and tissue HRP-DAB staining system (R&D Systems, Inc., Minneapolis, MN, USA), according to the manufacturer's instructions. Immunostaining was performed using tumor controls that were positive (tumors formed by A2B5⁺ cells) and negative (tumors formed by A2B5⁻ cells), and was evaluated by a pathologist in a blinded-manner. For ICC staining, the cells were seeded onto coverslips (~5×10⁵ cells/ml) and fixed with 4% paraformaldehyde (Sigma-Aldrich), treated with 3% hydrogen peroxide for 10 min and incubated with the antibodies, described above, overnight at 4°C. Fluorescein isothiocyanate- or tetramethyl-rhodamine isothiocyanate-labeled goat anti-rabbit secondary antibodies (1:200; cat. no. ab97178; Abcam) were then added and incubated for 2 h at 37°C. The 4′,6-diamidino-2-phenyl-indole reagent (Wuhan Boster Bioengineering Co., Ltd.) was used to stain the cell nuclei, and the cells were visualized using fluorescence microscopy (BX40F4; Olympus Corporation).

The cells, which were isolated using MACS were seeded (100 cells/well) into 96-well plates in the presence of the stem cell-permissive medium (0.2 ml). The cultures were maintained by replacing half of the medium every 3 days. A subsphere-forming assay (also termed passaging) was repeated every 2 weeks. Following each passage, the number and the size of the spheres were assessed, on days 7 and 14, under a CKX41SF inverted microscope.

Cells were collected in an exponential growth phase and then fixed with ethanol. Subsequently, RNase A treatment (Beyotime Institute of Biotechnology) and propidium iodide staining were performed. The cells were detected using flow cytometry with a FACSCalibur (BD Biosciences). The number of cells at the G₀/G₁, S and G₂/M phases were quantified using Modfit software (BD Biosciences), excluding the calculation of cell debris and fixation artifacts. Cell proliferation was quantified using a Cell Counting Kit-8 (CCK-8; Beyotime Institute of Biotechnology). Briefly, 100 µl cells (suspended in stem-cell permissive medium) from the two groups (CD133⁻/A2B5⁺and CD133⁻/A2B5⁺ cells) were seeded onto a 96-well plate at a concentration of 2,000/cells per well and incubated at 37°C. At daily intervals (1, 2 and 3 days), the optical density was measured at 450 nm using a microtiter plate reader (Thermo Multiskan MK3; Thermo Fisher Scientific, Inc.) with the cell survival rate expressed as the absorbance. The results represent the average of six replicates under the same conditions.

Statistical analyses were performed using SPSS software, version 13.0 (SPSS, Inc., Chicago, IL, USA). Statistical significance was determined using two-tailed Student's t-test and data are expressed as the mean ± standard error. P<0.05 was considered to indicate a statistically significant difference.

The SHG139s cells were dissociated using 0.25% trypsin, following which the cells were subjected to MACS. The results revealed two groups of cells: CD133⁻/A2B5⁺ and CD133⁻/A2B5⁻ cells. The purity of the isolated cells was determined using ICC, and the purity of the two groups was >90% (Fig. 1A–C).

The ICC staining also demonstrated that the majority of cells in the two subpopulations expressed Nestin, sex determining region Y-box 2 (Sox2) and Nanog stem cell markers (Fig. 1D and E). To assess the differentiation potential of the cells, the two populations of cells were cultured in serum-containing medium. The cells presented with adherent growth. The expression levels of glial fibrillary acidic protein (GFAP), β-III tubulin and galactosylceramidase (Galc) were assessed using ICC. The adherent cells in the two groups exhibited expression of the three differentiation markers (Fig. 1F and G).

An in vitro sphere formation assay was used to examine whether the expression of A2B5 was involved in cell renewal upon serial passaging. It was found that the high expression level of A2B5 not only affected the size of the spheres, but also led to the reduction in the numbers of spheres in subsequent generations (Fig. 2A and B). To investigate whether the expression of A2B5 affected the proliferation of cells in vitro, a CCK-8 assay (Fig. 2C) and flow cytometric analysis of the cell cycle were performed. The results showed that the proliferation abilities of the A2B5⁻-derived cells were more marked than those of the A2B5⁺-derived cells. Furthermore, to probe the effects of high expression levels of A2B5 on cancer cell growth in vivo, a mouse model of human glioma was used. The A2B5⁻-derived cells and A2B5⁺-derived cells were injected into the left axilla of nude mice. The results showed that the two A2B5 subtypes had similar tumorigenicity in nude mice in vivo. However, that the expression levels of Ki-67 in the tumors formed from the A2B5⁻-derived cells were higher, compared with those formed by the A2B5⁺-derived cells (Fig. 2D). In addition, the growth of the tumors from the A2B5⁺-derived cells was significantly inhibited, compared with that of the tumors formed by the A2B5⁻-derived cells (Fig. 2E and F; P<0.05).

The results of flow cytometry indicate that A2B5 expression increased the percentage of G₀/G₁ phase cells, and decreased the percentage of S and G₂/M phase cells (P<0.05; Fig. 3).

To assess the effects of high expression levels of A2B5 on the invasiveness of glioma cells, a Transwell invasion system was used. The number of invasive cells from the A2B5⁻-derived cells was significantly reduced, compared with the A2B5⁺-derived cells (Fig. 4A and B). To further examine the molecular associations between the high expression levels of A2B5 and invasiveness in human glioma, the relative expression levels of MMP9, MMP2, E-cadherin and TIMP3 were analyzed using western blot analysis. The results revealed no difference in the expression levels of MMP2 and MMP9, whereas E-cadherin and TIMP3 were expressed at high levels in the A2B5⁻-derived cells (Fig. 4C and D). The subcutaneous tumors were removed and sectioned, and the sections were stained with antibodies against MMP2, MMP9, E-cadherin and TIMP3. Subsequent IHC analysis showed that the tumors formed from A2B5⁻-derived cells had higher expression levels of E-cadherin and TIMP3; however, no difference was observed between the expression levels of MMP2 and MMP9 (Fig. 4E). Thus, the data indicated that higher expression levels of A2B5 led to enhancement of glioma cell invasion in the tumor xenografts.

To assess the angiogenesis of the two groups of cells, the expression levels of CD34, VEGF and VEGFR2 in the differentiated cells of each group were assayed using IHC (13–15). No significant difference was observed between the percentages of CD34-positive cells between the two groups (Fig. 5A). However, the percentage of VEGF-positive and VEGFR2-positive cells in the differentiated cells from the A2B5⁻-derived cells were higher than those in the A2B5⁺-derived cells (Fig. 5B). In addition, these three indicators were assessed in vivo using IHC. A small number of CD34⁺ cells were involved in the formation of tumors in the two groups. Tumors formed by A2B5⁻-derived cells exhibited higher expression levels of VEGF and VEGFR2 (Fig. 5C).