Journal of Neuro-Oncology

, Volume 100, Issue 2, pp 165–176

IL-6 promotion of glioblastoma cell invasion and angiogenesis in U251 and T98G cell lines

Authors

  • Qinglin Liu
    • Department of Neurosurgery, Qi Lu HospitalShandong University
    • Department of Neurosurgery, Qi Lu HospitalShandong University
  • Ronghui Li
    • Department of Neurosurgery, Qi Lu HospitalShandong University
  • Jie shen
    • Department of Neurosurgery, Qi Lu HospitalShandong University
  • Qiaowei He
    • Department of Neurosurgery, Qi Lu HospitalShandong University
  • Lin Deng
    • Department of Neurosurgery, Qi Lu HospitalShandong University
  • Cai Zhang
    • Institute of Immunopharmacology and ImmunotherapySchool of Pharmaceutical Sciences, Shandong University
  • Jian Zhang
    • Institute of Immunopharmacology and ImmunotherapySchool of Pharmaceutical Sciences, Shandong University
Laboratory Investigation - Human/Animal Tissue

DOI: 10.1007/s11060-010-0158-0

Cite this article as:
Liu, Q., Li, G., Li, R. et al. J Neurooncol (2010) 100: 165. doi:10.1007/s11060-010-0158-0

Abstract

Interleukin-6 (IL-6) is a growth and survival factor in human glioblastoma cells and plays an important role in malignant progression. However, its role in glioblastoma invasion is still unknown. This study shows how IL-6 promotes cell invasion and migration in U251 and T98G glioblastoma cell lines. The underlying mechanism includes both protease-dependent and -independent manners. Stimulation with IL-6 increased MMP9 expression in the two cell lines but had no influence on MMP2 expression. Fascin-1 is a cell skeleton binding protein and plays a key role in cell migration and invasion. Its binding style directly influences cell morphology and tendency to become deformed. After IL-6 exposure, fascin-1 expression increased in an IL-6 dose-dependent manner. Immunofluorescence also revealed that the binding style of fascin-1 had changed after IL-6 exposure, resulting in a more invasive phenotype of the cells. Three most commonly emphasized invasion-associated signaling pathways, including JAK-STAT3, p42/44 MAPK, and PI3K/AKT, were verified to further illustrate its underlying mechanism. Only phosphorylation of STAT3 at ser 727 site paralleled the IL-6 stimulation, and JSI-124, a specific JAK-STAT3 pathway blocker, deterred the invasion and migration promotive effect of IL-6, indicating that the JAK/STAT3 pathway mediates signal transduction. Furthermore, IL-6 also acts in a paracrine fashion to promote vascular endothelial cell migration, thus facilitating tumor angiogenesis and invasion. These results suggest that IL-6 promotes glioblastoma cell invasion and angiogenesis and may be a potential anti-invasion target.

Keywords

IL-6GlioblastomaInvasionMigrationAngiogenesisFascin-1

Introduction

Gliomas account for more than 70% of all brain tumors. Of these, glioblastoma is the most frequent and malignant histological type of tumor (World Health Organization, WHO, grade IV) [1]. Glioblastoma remains a prevalent form of brain tumor despite advances in surgical and medical therapy. Two major aspects of glioma biology that contributes to its resistance are the formation of new blood vessels through the process of angiogenesis and the invasion of glioma cells through white matter tracts [2]. Therefore, in identifying new therapeutic strategies for glioblastoma, it is important to first comprehend their underlying mechanisms.

Interleukin-6 (IL-6), a cytokine mostly involved in the modulation of immuno- and inflammatory responses, was recently reported to be expressed in many malignant tumors, including prostate [3], breast [4], lung cancer [5], and glioblastoma [6]. In patients with glioblastoma, its expression level can indicate the prognosis of the patients [7], which could be partially explained by its proliferation promotive effect [8]. Recently, a variety of biological properties, such as proliferation [3, 5], apoptosis [3], and invasion [4] in different malignancies, were reportedly regulated by IL-6. However, whether IL-6 also contributes to the invasion and migration ability of glioblastoma has yet to be investigated.

Tumor metastasis is a multistep process wherein a cancer cell spreads from a primary tumor to distant secondary organs and tissues. However, controversies remain on the full understanding of protease-dependent and -independent processes in cell migration [9]. Protease-dependent manner refers to the ability of matrix metalloproteinases (MMPs) family members to lyse the extracellular matrix (ECM), while protease-independent manner refers to the ability of morphology deformity to squeeze gaps. Proteolysis and the remodeling of ECM make up one of several initiating events that allow cancer cells to invade the stroma [10]. MMPs are major hydrolytic enzymes targeting ECM during metastasis, especially MMP2 and MMP9. Fascin-1, a 55KD actin-bundling protein, is associated with pathological grade, prognosis, invasion, and migration in glioblastoma [1114]. To illustrate the mechanism inducing invasion of IL-6 in glioblastoma, MMPs and fascin-1 were chosen as downstream targets.

As a tumor mass grows larger, new vessels form to supply the tumor cell with oxygen and glucose to support its high metabolism. Angiogenesis seems to be critical in the development and progression of tumor cells. Glioblastoma is one of the abundant vasculature tumors, and the associated microvessel density correlates to the degree of malignancy, aggressiveness, clinical recurrence, and decreased survival of the patient [15]. Anti-angiogenesis strategies, such as the vascular endothelial growth factor (VEGF) monoclonal antibody, Bevacizumab, have been used in clinical trials, and have revealed some progress when combined with chemotherapy or radiotherapy. In this study, we propose another aspect of angiogenesis. We found that the supernatant of glioblastoma could dramatically increase the migration ability of endothelial cells and may promote the development of tumor vessels in vivo. In the process, IL-6 was detected as one of the mediators. Identifying the interaction between tumor cells and endothelial cells [16] is also proposed.

Several IL-6-related signal pathways have been identified as transmitting signals associated with cell invasion and migration in various tumors. Among them, STAT3, PI3K/Akt, and MAPK/ERK have been the most emphasized [3, 4, 1722]. In different cell lines, blockage of indicated pathways results in reduced invasion and migration potential. In this study, we investigate these three signal pathways to determine the pathways responsible for IL-6-induced invasion and migration in glioblastoma.

Materials and methods

Human samples

All human samples were collected from the archives of the Neurosurgery Institute of Shandong University Qilu Hospital (Jinan, China). All samples were resected for intracranial decompression. This study was approved by the Institutional Review Board of Shandong University. Written informed consent forms were also obtained from all participants.

Cell culture

The human glioblastoma cell line U251 was acquired from our laboratory, while the T98G cell line was provided by Professor JingDe Zhu, who is affiliated with a tumor research center in Shanghai, China. The two-cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco/BRL, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS) (TBD, Tianjin, China) at 37°C in a 5% CO2 atmosphere. Vascular endothelial cell line ECV-304 was conserved in our laboratory and cultured in RPMI 1640 (Gibco/BRL) containing 10% FBS (TBD) at 37°C in a 5% CO2 atmosphere.

Source of antibodies

The following antibodies were used: anti-Stat3 and anti-p-Stat3 (Tyr705, Ser727), anti-MAPK, anti-p-MAPK, anti-Akt, anti-p-Akt (Cell Signaling Technology, Beverly, MA), anti-β-actin, anti-MMP2, and anti-fascin-1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-MMP9 (Millipore, Daiichi Fine Chemical, USA). The dilution range was from 1:500 to 1:2,000 for western blot, depending on the different cell lines and antibodies, and 1:100 for fluorescence immuno-staining for fascin-1.

Invasion and migration assay

Matrigel invasion assay in vitro was measured in a Transwell culture chamber system. The filter membrane with 8-μm pores was coated with Matrigel (BD Biosciences, USA). Cells that were starved for 6 h were suspended in serum-free DMEM with various concentrations of IL-6 at a density of 1 × 105/ml. The 200 μl suspension was seeded into the upper chambers. Then, 600 μl FBS containing DMEM (10% for U251 cell and 20% for T98G cell) was added to the lower chamber. After incubation for 24 h, the upper chamber was taken out and the cells on the upper surface of the membrane were very lightly removed using a medical cotton bud. The cells on the other side were then fixed and stained with 0.1% crystal violet in methanol for several minutes. The chamber was viewed under light microscope and photographs were taken.

Transwell migration assay was employed to evaluate the protease-independent migration ability. No Matrigel coating was carried out on the 8 μm pore membrane.

The wound-healing assay (WHA) method was also used to evaluate cell migration ability. In brief, every 1 × 106 cells were seeded into a well of a six-well plate. After cells grew to about 70% confluence, the medium was changed into a serum-free medium for 6 h. The tip of a 10-μl pipette was used to create the wound line. Cells were washed three times with PBS and 2 ml DMEM medium containing 10% FBS with various concentrations of IL-6 were added to the wells. Photographs were taken at 0, 12, and 24 h at the same site, and the migration distance was measured.

Cell proliferation assay

For the cell proliferation assay, cells were seeded into a 96-well micro-plate at a density of 5 × 103 cells in 200 μl per well. Briefly, cells were cultured overnight and then starved for 6 h. Then, 200 μl DMEM containing 10% FBS and various concentrations of IL-6 were used. Cells were incubated at 37°C based on the indicated time, and 20 μl MTT (5 mg/ml) was added into each well. The medium was incubated for another 4 h in darkness. Consequently, the medium was expirated, but the formazan grain was conserved in the wells. A total of 200 μl DMSO was used to dissolve the formazan grain. The absorbance rate at 570 nm was read using an ELISA plate reader (Bio-Rad, Model 680).

RNA isolation and RT–PCR

Cells were starved for 6 h and treated with various concentrations of IL-6 for another 24 h. Then, the cells were harvested. Total mRNA was extracted using Trizol (Invitrogen). RT–PCR was conducted using an RT–PCR kit (Transgene, Beijing, China) following manufacturer instructions. The primers used were all synthesized by sangon (Shanghai, China). The sequences were as follows: β-actin: forward: 5′-ATC ATG TTT GAG ACC TTC AAC A-3′; reverse: 5′-CAT CTC TTG CTC GAA GTC CA-3′; MMP2: forward: 5′-GGC CCT GTC ACT CCT GAG AT-3′; reverse: 5′-GGC ATC CAG GTT ATC GGG GA-3′; gp130: forward: 5′-CTG TAT CAC AGA CTG GCA ACA AG-3′; reverse: 5′-GCA TTT GCT CTC TGC TAA GTT CC-3′; gp80: forward: 5′-GCT CCT CTG CAT TGC CAT TG-3′; reverse: 5′- GCA TCT GGT CGG TTG TGG CT-3; IL-6: forward: 5′-TCT CCA CAA GCG CCT TCG-3′; and reverse: 5′-CTC AGG GCT GAG ATG CCG-3′. Amplification was performed over 30 cycles: 94°C/60 s (denaturation), 58°C/60 s (annealing), and 72°C/60 s (extension). Agarose gel electrophoresis was performed and results were photographed and analyzed using an AlphaEaseFC software (Version 4.0.0; Alpha Innotech, USA).

Zymography

For this, 3 × 104 cells were seeded into the wells of a 24-well plate for 24 h. Then, the cells were starved for 6 h, and the culture medium was replaced with fresh serum-free DMEM containing various concentrations of IL-6. After 24 h incubation, the supernatant was collected and stored at −80°C. Next, 15 μl supernatant mixed with 5 μl 4 × loading buffer was added into each lane for SDS–PAGE with gels containing 0.1% gelatin (W/V) and 10% polyacrylamide (W/V). The gels were washed in solution A (2.5% Triton X-100, 50 mM Tris-HCl, 5 mM CaCl2, pH 7.6) for 40 min twice, and then solution B (50 mM Tris-HCl, 5 mM CaCl2, pH 7.6) for 20 min twice. Then the gels incubated in the substrate buffer (50 mM Tris-HCl, 5 mM CaCl2, 150 mM NaCl, 5 mM ZnCl2, pH 7.6) overnight at 37°C. After Coomassie Brilliant Blue staining and subsequent decoloration with acetic acid, the gel was photographed and analyzed using an AlphaEaseFC software (Version 4.0.0; Alpha Innotech).

Fluorescence immuno-staining

Coverslips were placed into the 24-well plate, after which 3 × 104 cells were seeded. When the cells reached 70% confluence, starvation was performed for 6 h. Then, the cells were treated with various concentrations of IL-6 for 24 h. Cells on the coverslips were placed in 4% para-formaldehyde at 4°C for 20 min, treated with 0.5% Triton X-100 for 5 min at room temperature, blocked with 10% normal goat serum for 1 h at 37°C, and incubated with fascin-1 antibody (dilution rate: 1:100, 37°C for 1 h or 4°C overnight) in a humidified chamber. The coverslips were washed in PBST three times for 5 min, and then incubated with a second antibody conjugated to FITC for 1 h at 37°C in the humidified chamber. The coverslips were washed in PBST three times for 5 min, mounted with 50% glycerol, and photographed under fluorescence microscope.

Immunoblotting

For this process, 3 × 105 cells were seeded into a 6-well plate for 24 h, starved for 6 h, and treated with various concentrations of IL-6 for a determined time. They were then lysed in ice-cold lysis buffer (150 mM NaCl, 20 mM Tris–HCl pH 7.5, 1% NP-40, 50 mM NaF, 1 mM EDTA pH 8.0, 1 mM PMSF, and 1 mM Na3VO4) for 30 min and centrifuged at 14,000g for 15 min. The supernatant was collected and the protein concentration determined. 30 μg of each protein sample was added into each lane for SDS–PAGE. The proteins were then transferred to PVDF membranes (Millipore). After blocking in 5% nonfat milk for 2 h, the membranes were probed with appropriate primary antibody at room temperature for 2 h (or at 4°C overnight), washed with TBST three times for 10 min, and incubated in horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology) at a dilution rate of 1:5,000. The proteins were visualized by an enhanced chemiluminescence (ECL) system (Pierce, Rockford, IL) using X-ray film. The films were photographed and analyzed using an AlphaEaseFC software (Version 4.0.0; Alpha Innotech).

Statistical analysis

All numerical data were presented as mean ± SD for at least three individual experiments. The Normality test, Student’s t test, and two-sample nonparametric test were used. All statistical analyses were performed using SPSS version 16.0. Statistical significance was accepted when P < 0.05.

Results

The expression of IL-6 mRNA in glioblastoma samples and cell lines

The autocrine secretion of IL-6 in glioblastoma cell lines have been investigated for almost 20 years now [6]. To validate earlier reports, we examined the expression of IL-6 in glioma samples and glioblastoma cell lines. IL-6 mRNA was detected in all three glioma samples and one para-tumor sample. However, no obvious expression was found in the normal brain tissue. In the sample obtained from a hypertensive intracranial hematoma patient, we also detected IL-6 expression (Fig. 1a). As to glioblastoma cell lines, IL-6 mRNA expression was detected in all three cell lines that were tested with U251, U87, and T98G (Fig. 1b).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig1_HTML.gif
Fig. 1

IL-6 mRNA is expressed in glioblastoma samples and cell lines. a The expression of IL-6 mRNA in clinical samples. The total mRNA of clinical samples were abstracted and IL-6 gene was amplified by RT–PCR. N Normal brain tissue, P paratumor tissue, H hypertensive intracranial hematoma invaded brain tissue, T glioblastoma tissue. b The expression of IL-6 mRNA in glioblastoma cell lines. The total mRNA of U251, U87, and T98G glioblastoma cell lines was abstracted and IL-6 gene was amplified by RT–PCR

IL-6 promotes U251 and T98G glioblastoma cell invasion and migration

In previous studies, IL-6 was thought to have promotive effects on the proliferation of glioma cell lines. Here, we investigated the influence of IL-6 on invasion and migration in glioblastoma cell lines of T98G and U251. Evaluated by the Transwell chamber system, we found that IL-6 promotes T98G and U251 glioblastoma cell invasion and migration, as can be seen in Fig. 2a, b, in a dose-dependent manner. Based on WHA, another cell migration assessment system used by this study, we obtained a different result on the T98G cell line (Fig. 2c). Since the total amount of cells may directly influence the number of invasive cells, the proliferation promotion effect of IL-6 on glioblastoma cells in this experiment should be excluded. MTT assay was employed, and no proliferation differences were observed under the experiment conditions for invasion and migration assay (Fig. 2d).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig2a_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig2b_HTML.gif
Fig. 2

IL-6 promotes T98G and U251 glioblastoma cell invasion and migration. a IL-6 promotes T98G and U251 glioblastoma cells invasion using a Martrigel–Transwell assay. Cells were starved for 6 h and suspended in serum-free DMEM containing various concentrations of IL-6; 3 × 104 cells were added to the upper chamber; 600 μl serum containing DMEM was added to the lower chamber. After incubation for 24 h, cells on the inferior surface were fixed and stained. Photographs were taken under a microscope. Results are representative of those observed in three independent experiments. Bar graphs represent the mean invasive cells of five random view field (mean ± SD) (*P < 0.05 vs untreated cells). b IL-6 promotes T98G and U251 glioblastoma cells migration using Transwell migration assay. Transwell migration assay was also employed following the methods of the Transwell invasion assay, except that there was no coating of Martrigel on the filter membrane. Results are representative of those observed in three independent experiments (original magnification ×400). Bar graphs represent the mean invasive cells of five random view field (mean ± SD) (*P < 0.05 vs untreated cells). c The effect of IL-6 on migration was measured by wound healing assay. The wound line was prepared using the tip of a 10-μl pipette, and the cells were then treated with various concentrations of IL-6. Photographs were taken and the width of the wound line was measured at 0, 12, and 24 h. Results are representative of three independent experiments (original magnification ×100). Bar graphs represent the mean migration distance at 6 sites (mean ± SD) (*P < 0.05 vs untreated cells). d The effect of IL-6 on proliferation was measured by MTT assay; 5 × 103 cells in 200 μl DMEM were seeded into a well of a 96-well plate and MTT assay was performed, as described in “Materials and methods”. Bar graphs represent the results in three independent experiments (mean ± SD) (*P < 0.05 vs untreated cells)

IL-6 promotes MMP9 but not MMP2 expression in T98G and U251 glioblastoma cell lines

Extracellular matrix proteolysis is very important in cell invasion. Here, we analyzed MMP2/9 levels in order to determine its underlying mechanism. We investigated MMP2 and MMP9 at gene, protein, and functional levels after treatments with various concentrations of IL-6 for 24 h. RT–PCR revealed that both MMP2 and MMP9 were expressed at gene level (Fig. 3a) in T98G and U251 cell lines. The expression of MMP2 was not influenced by IL-6 treatment, but MMP9 expression increased as IL-6 dose increased (Fig. 3a). Western blotting confirmed the results at the protein level (Fig. 3b). The activities of MMP2 and MMP9 secretion into the supernatant by T98G and U251 were analyzed by gelatine zymography. MMP2 activity was detected in the two cell lines, but MMP9 activity was only detected in the U251 cell line. Consistent with the gene expression pattern, the activity of MMP2 did not change in both cell lines. However, MMP9 activity increased with IL-6 concentration in the U251 cell line. MMP9 activity in the T98G cell line was not detected under our experimental conditions (Fig. 3c).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig3_HTML.gif
Fig. 3

IL-6 promotes MMP9 expression in T98G and U251 glioblastoma cell lines. a IL-6 treatment promoted MMP9 gene expression in T98G and U251. Cells were starved for 6 h and treated with various concentrations of IL-6 for 24 h. Total mRNA were abstracted and MMP2/9 and β-actin were amplified by RT–PCR. b MMP9 protein expression elevated in an IL-6-dependent manner. Cells were starved for 6 h and treated for 24 h with various concentrations of IL-6. Total proteins were abstracted and MMP2/9 expression was tested by western blot. c IL-6 treatment increased MMP9 activity by zymography. Cells were starved for 6 h and treated with IL-6 for 24 h. The supernatant was collected for zymography

IL-6 increases fascin-1 expression and modifies its distribution fashion

Cellular deformity is another important aspect of cell invasion and migration. Fascin-1, the key filopodial bundling protein, is critical in the formation of cellular protrusions that facilitate direct interaction with ECM and promote cell migration [11, 13, 14]. Firstly, we investigated fascin-1 at the protein level after being exposed to IL-6 for 24 h. It was found that fascin-1 expression increased with IL-6 stimulation in a dose-dependent manner (Fig. 4a). Secondly, the distribution pattern of fascin-1 was investigated by fluorescence immuno-staining. The result demonstrated that, after IL-6 treatment, fascin-1 was transported onto the margin of the cell. More protrusions were formed, especially in the U251 cell (Fig. 4b). The changes in distribution of fascin-1 imply that the IL-6-treated cells have a more invasive phenotype.
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig4_HTML.gif
Fig. 4

IL-6 increases fascin-1 expression and modifies its distribution fashion. a IL-6 treatment promoted fascin-1 protein expression. Cells were starved for 6 h and treated for 24 h with various concentrations of IL-6. Total proteins were abstracted and fascin-1 expression was tested using Western blot. b The distribution fashion of fascin-1 changed after IL-6 exposure. Cells were treated with IL-6 for 24 h, and then fixed and incubated in primary antibody for 1 h at 37°C. FITC-conjugated second antibody was linked. Photos were taken under a fluorescence microscope (magnification ×400). Results are representative of those observed in three independent experiments

The possible signal pathways that involved in IL-6 induced glioblastoma cell invasion

To further illustrate the mechanism of IL-6 promotion of glioblastoma cell invasion and migration, we attempted to find the signal pathways mediating its stimulation. Serum-free starved cells were treated with various concentrations of IL-6 for 20 min, after which the total proteins were extracted. Three IL-6 associated pathways were studied: STAT3, p42/44 MAPK, and PI3K/Akt. For both cell lines, when exposed to IL-6, STAT3 obtained phosphorylation at ser 727 in a dose-dependent manner, but the phosphorylation at Tyr 705 was not evident (Fig. 5). Total and phosphorylation Akt levels were stable (Fig. 5). The stimulating effect of IL-6 in p42/44 MAPK pathway was different for the two cell lines. The phosphorylation level of p42/44 MAPK increased with IL-6 stimulation in the U251 cell line, but it decreased in the T98G cell line (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig5_HTML.gif
Fig. 5

The possible pathways that mediate the IL-6-stimulating signals. Cells were starved for 6 h, treated with various concentrations of IL-6 for 20 min, and washed with ice-cold PBS. Total proteins were abstracted and the activation of three independent pathways was tested. Results are representative of those observed in the three independent experiments

IL-6 promotes tumor angiogenesis

Tumor angiogenesis is extremely important in tumor proliferation and invasion. The first step in tumor angiogenesis is the migration and formation of sprouts in the vascular endothelial cell. We tested whether the IL-6 secreted by glioblastoma cells can influence the tumor vascular endothelial cells in a paracrine mechanism. Attracted by serum containing RPMI 1640 in the lower chamber, it was observed that ECV-304 cells suspended in the glioblastoma supernatant were more invasive than those suspended in RPMI 1640 (Fig. 6a, b). Since IL-6 could be secreted into the supernatant by glioblastoma cells, we tested whether IL-6 contributes to the increasing migration ability of ECV-304. We found that IL-6 indeed promoted vascular endothelial cell migration (Fig. 6a, b). Furthermore, IL-6 McAb has reversed the invasive effect of the glioblastoma supernatant (Fig. 6a, b). RT–PCR also revealed the expression of IL-6R subunits gp130 and gp80 in the ECV-304 cell line (Fig. 6c). These results demonstrate that IL-6 partially promotes endothelial cell migration in order to facilitate tumor angiogenesis in a paracrine manner.
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-010-0158-0/MediaObjects/11060_2010_158_Fig6_HTML.gif
Fig. 6

IL-6 paracrine promotes tumor angiogenesis. a IL-6 paracrine promotes vascular endothelial cell migration. U251 cells were starved for 6 h, and then cultured in fresh serum-free RPMI 1640 with various concentrations of IL-6 for another 6 h. The supernatant was collected for succeeding experiments. Serum-free starved ECV-304 cells were suspended in different culture medium: control (RPMI 1640), RPMI 1640 + IL-6, U251 supernatant, and U251 supernatant + IL-6 McAb. 3 × 104 cells in 200 μl were seeded in the upper chamber of the Transwell system. Transwell migration assay was performed. b Bar graphs represent the mean invasive cells of five random view fields (mean ± SD) (*P < 0.05 vs untreated cells). c ECV-304 cell expresses IL-6 and IL-6R (receptor). Total mRNA of ECV-304 was abstracted. RT–PCR was performed to amplify the IL-6 and IL-6R genes. Results are representative of those observed in three independent experiments

Discussion

In this study, we initially confirmed the expression of IL-6 in glioblastoma samples and cell lines. IL-6 has been reported to promote cell invasion and migration in many solid tumors [4, 17, 23, 24]. In this study, we investigated the effect of IL-6 on the invasion and migration of glioblastoma. Results suggest that IL-6 promotes glioblastoma cell invasion and migration in a dose-dependent manner. WHA, a traditional system used to evaluate the migration ability of tumor cells, was employed. Under our experiment conditions, in contrast with the U251 cell line, we found that IL-6 decreased the wound-healing ability of the T98G cell line. This may at least be partially explained by the increasingly inhibitive effect of IL-6 on the T98G cell line. Previous studies have reported that IL-6 contributes as a growth factor in glioblastoma [8]. However, our experiments indicated dissimilar results at the T98G cell line, and, thus, regarding IL-6 as a growth promotive cytokine should be done with caution. Similar results were also obtained at the T47D breast cancer cell line [4], although in other breast cancer cells, IL-6 was thought to exert a promotive effect during proliferation. These may reflect the special biological properties of various cell lines. Meanwhile, the special biological properties between cell lines may be another explanation for the inhibitive effect of IL-6 on WHA at the T98G cell line. The Matrigel Transwell chamber system, as opposed to WHA, may be a more reasonable method in evaluating their invasion and migration ability. Yet, despite the different results at the T98G cell line using WHA, we can still conclude that IL-6 promotes glioblastoma cell invasion and migration.

Tissue invasion during metastasis requires cancer cells to induce a stromal environment dominated by cross-linked networks of collagens [9]. To penetrate this structural barrier, cancer cells use either protease-dependent or protease-independent invasion schemes [9, 19, 22, 25]. Protease-dependent invasion programs rely on MMP family members to cleave impeding collagen fibrils [9], mainly on MMP2 and MMP9. Alternatively, collagenous barriers can be broken by a protease-independent fashion, wherein cancer cells use actomyosin-based mechanical force to physically displace matrix fibrils with amoeboid-like deformities [9, 25]. During the invasion process in vivo, both protease-dependent EMC lysis and protease-independent amoeboid-like deformities exist. We opine that invasion is a complex process that includes both protease-dependent and protease-independent aspects. Enhancing any of the steps associated with invasion would promote the overall invasive ability. Fascin-1 is a 55KD actin bundling protein that contributes in the arrangement of actin bundles concentrated in cell membrane protrusions, thus affecting cell motility [14]. Removing the expression of fascin-1 in glioblastoma cell results in decreased cell migration, adhesion, and invasion ability [13]. In this study, we found that both T98G and U251 cell lines after IL-6 treatment obtained elevated MMP9 mRNA and protein expressions. Based on findings from gelatin zymography, the functional activity of MMP9 increased with IL-6 concentration at the U251 cell line, but this was not detected at the T98G cell line. An explanation for this is that the expression level of MMP9 at the T98G cell line is beyond the sensitivity of zymography according to the conditions presented by our study. This was strongly supported by the fact that 35 cycles was needed for MMP9 amplification while only 30 cycles was needed for MMP2 in RT–PCR assay. Both T98G and U251 cell lines expressed high level of MMP2, and the expression level was stable regardless of IL-6 stimulation. Similarly, not only did the amount of fascin-1 increase but changes in the distribution fashion were also observed. The cell skeleton was altered, thus leading to more invasive phenotype. We can therefore conclude that IL-6 promotes glioblastoma cell invasion and migration in both a protease-dependent and protease-independent fashion. Compared with the U251 cell line, the low expression level of MMP9 could partially explain the poor invasion ability at the T98G cell line. Furthermore, because the expression of fascin-1 correlates with the pathological grade of glial tumors, it can be concluded that IL-6 may participate in the malignant transformation of glial tumors. When cultured in serum-free DMEM, a recent study (data unpublished) has shown different results regarding the U87 glioblastoma cell line for the MMP2 expression. This may be explained by the different culture conditions and the unique biological properties of different cell lines. The different biological properties between T98G and U87 cell lines have also been proposed by other authors [26].

Several signal pathways have been reported to facilitate the invasion signals in glioblastoma. Blockage of the PI3K/Akt pathway in a glioblastoma cell line results in the suppression of cell proliferation, arrest of cell cycle, reduction of cell invasion, and promotion of cell apoptosis; in contrast, activation of this pathway results in enhanced invasion ability in vitro and in vivo [2729]. The PI3K/Akt pathway has also been reported to be involved in regulating the invasion in other malignancies [18, 3035]. A similar invasion-regulating effect has also been examined at the STAT3 and MAPK pathways [22, 34, 3644]. In this study, we tested the activation of STAT3, PI3K/Akt, and MAPK pathways. It was found that the PI3K/Akt signal pathway was constantly activated, but had no additional activation, by IL-6 stimulation. We also found that the phosphorylation of STAT3 at Ser727 increased with IL-6 stimulation in a dose-dependent manner. Phosphorylation of the MAPK pathway increased at the U251 but decreased at the T98G with IL-6 stimulation. Since the MAPK/ERK pathway also contributes to glioblastoma cell proliferation [4549], we suspect that the moderate increase in the inhibitive effect of IL-6 at the T98G cell line was mediated by the MAPK pathway. Although the PI3K/Akt pathway is involved in regulating the invasion of glioblastoma, it is not related to IL-6-induced invasion. A recent study (data unpublished) has shown that the blockage of the STAT3 pathway with JSI-124, a specific STAT3 pathway inhibitor, reversed the IL-6 induced invasion promotion effect in glioblastoma cell lines. The invasion regulatory effect of IL-6 in glioblastoma is dependent on the phosphorylation of STAT3 at ser 727.

Neovascularization is a very important aspect of tumor invasion. It facilitates the infiltration and growth of tumor cells. Glioblastoma is one of the most prominent vascularized malignances. It has been reported that IL-6 contributes to tumor angiogenesis by inducing transcription of VEGF through the STAT3 signal pathway in glioblastoma [50]. In this study, we found another action pattern of IL-6 when promoting tumor angiogenesis. In vivo, glioblastoma cells grow in a microenvironment, which at the minimum have glial cells, abnormal vascular endothelial cells, and infiltrating immunocytes, including their secreted cytokines. Cells in the tumor microenvironment inevitably interact with each other via the cytokines they secrete. Using the Transwell chamber system, we partially mimicked the tumor microenvironment in vitro and confirmed that IL-6 secreted by glioblastoma cells could promote endothelial cell migration to form new vessels. Surprisingly, we also detected IL-6 expressions at the vascular endothelial cell line ECV-304. It seems that the association between glioblastoma and vascular endothelial cells is bi-directional, and that they promote each other’s migration. IL-6 plays a critical role in the interaction.

In conclusion, our results demonstrate that IL-6 promotes glioblastoma cell invasion and migration via the STAT3 pathway in both a protease-dependent and protease-independent fashion. Tumor cells and tumor vascular endothelial cells may interact with each other to promote tumor invasion, by which, during the interaction, IL-6 could serve as an important mediator. Thus, IL-6 is an important anti-invasion target in glioblastoma.

Acknowledgments

This project was supported by the National Natural Science Foundation of China (No. 30872645). We thank Aijun Hao and Shidou Zhao from the Department of Histology and Embryology of Shandong University for their kind assistance.

Copyright information

© Springer Science+Business Media, LLC. 2010