Key Laboratory of Cancer Prevention and Therapy of Tianjin, National Clinical Research Center of Cancer, Tianjin Medical University Cancer Institute and Hospital, Tianjin, China
Address correspondence and reprint requests to Yongjie Ma, Tianjin Medical University Cancer Institute and Hospital, Huanhu West Road, Hexi District, Tianjin 300060, China. E-mail: email@example.com
Girdin, an actin-binding protein, possesses versatile functions in a multitude of cellular processes. Although several studies have shown that Girdin is involved in the cell DNA synthesis, actin cytoskeleton rearrangement, and cell motility, the molecular mechanisms of Girdin in tumor development and progression remain elusive. In this study, through over-expression and siRNA experiments, we found that Girdin increased migration of LN229 human glioblastoma cells. On the other hand, reducing Girdin impaired F-actin polymerization, which is essential for cell morphogenesis and motility. Matrix metalloproteinase 2, critical in human glioma migration and invasion, was down-regulated upon Girdin reduction and led to decreased invasion in vitro and in vivo. In addition, silencing Girdin expression impaired the phosphorylation of two important adhesion molecules, integrin β1 and focal adhesion kinase, resulting in cell adhesion defects. Our immunohistochemical study on human gliomas tissue sections indicated that Girdin expression was positively related with glioma malignancy, supporting the in vitro and in vivo results from cell lines. Collectively, our findings suggest a critical role for Girdin in glioma infiltration.
We show that reduction of Girdin, an actin-binding protein, leads to impaired F-actin polymerization and down-regulated expression of matrix metallopeptidase protein 2 (MMP-2), phosphorylated integrin β1, and phosphorylated focal adhesion kinase (FAK), which resulted in decreased migration, adhesion, and invasion of glioblastoma cells. Girdin was positively correlated with glioma malignancy and negatively associated with clinical prognosis, suggesting Girdin as a critical regulator in glioma infiltration.
Highly invasive and neurologically destructive malignant gliomas are the most common form of primary brain tumors. Glioblastoma, the biologically most aggressive tumor, preferentially invade along myelinated fibers tracts in white matter (intrafascicular growth; Drappatz et al. 2009), characterizing the most lethal feature of the brain tumors with a survival median of less than 15 months among glioblastoma patients (Brandsma and van den Bent 2007). Currently, despite modern treatments such as surgery, radiation, and chemotherapy, there is no promising therapy for recurring Glioblastoma (Wong et al. 1999; Lamborn et al. 2008).
Akt-binding protein Girdin has been found highly expressed in human malignant tissues such as breast cancer, colon cancer, lung cancer, and uterine cervical carcinomas (Jiang et al. 2008). Discovered in 2005, Girdin was first named as Akt phosphorylation enhancer (Anai et al. 2005). The human Girdin cDNA encodes 1870 amino acids with a predicted protein mass of ~ 220 kDa (Enomoto et al. 2006). The C-terminal region of Girdin includes an Akt phosphorylation site, a membrane binding domain, and a Gα-protein-binding domain.
As an Akt substrate and an actin-binding protein, Girdin participates in Akt-mediated cancer progression and angiogenesis (Anai et al. 2005; Enomoto et al. 2005; Kitamura et al. 2008). Studies have indicated that Girdin is responsible for vascular endothelial growth factor-mediated endothelial cell angiogenesis (Kitamura et al. 2008). Moreover, Girdin is located at the leading edge of activated cells, and participates in endocytosis and exocytosis, extension of lamellipodia, and regulation of metastasis (Enomoto et al. 2005; Simpson et al. 2005; Jiang et al. 2008). Recent reports have indicated that a down-regulation of Girdin can lead to decreased migration and invasion of breast cancer cells (Jiang et al. 2008). Le-Niculescu et al. (2005) and Ghosh et al. (2008) have also identified Girdin as a Gαi3-binding partner, and Gαi3 preferentially localizes to the leading edge and controls the functions of Girdin during cell migration in mammalian cells. However, the molecular mechanisms of these cellular processes remain unclear.
In this study, we tested the hypothesis that Girdin directly participates in glioblastoma cells' migration and invasion. Over-expressing Girdin in LN229 human glioblastoma cells, we found that Girdin increased cell migration. Using the small interference RNA technology, we knocked down the expression of Girdin in glioblastoma cells, resulting in inhibition of cell migration and invasion. Our data also demonstrate that integrin β1, focal adhesion kinase (FAK), and matrix metalloproteinase 2 (MMP-2) all are involved in Girdin signaling pathways, which are critical for glioblastoma cell migration and invasion. The in vivo invasion results of mice xenograft models corroborated the in vitro results. Furthermore, the results of our immunohistochemistry, using human gliomas tissue sections, also indicated that Girdin expression was closely related to glioma malignancy. This is in line with our in vivo and in vitro results from cell lines. Taken together, our observations suggest that Girdin is required in glioblastoma migration and invasion, thus, serving as a novel target for therapeutic interventions for malignant glioma infiltration.
Materials and methods
Cell culture and reagents
Human glioblastoma cell lines, LN229 and U87, were obtained from Leibniz Institute DSMZ-German Collection of Microorganisms and cell cultures (Germany). The cells were cultured in a complete medium (RPMI 1640 supplement with 10% fetal bovine serum) in a 5% CO2 37°C incubator. The cell lines have been tested and authenticated by DNA (short tandem repeat) profiling and this work was performed in Beijing Microread Genetics Co. Ltd. (Beijing, China). The recombinant human epithelial growth factor (EGF) was obtained from R&D systems (Minneapolis, MN, USA). The antibody to Girdin (Code: 18979) was from Immuno-Biologicl Laboratory Co. Ltd. (Gunma, Japan). The antibodies toward glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (sc-25778), MMP-2 (sc-20172), MMP-9 (sc-10737), integrin β1 (sc-8978), and phosph-FAK (sc-11765) were from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). The antibody for phosph-integrin β1 (Thr 788/789) (AB8123) was from Millipore Corporation (Bedford, MA, USA).
Construction of lentiviral vector expressed with full length of Girdin
Maps, sequences, and cloning information for the pCDH-CMV-MCS-EF1-Puro lentiviral vector are available online (Addgene http://www.addgene.org/). Open reading frame Clone of Homo sapiens of full-length Girdin was provided by OriGene Technologies Inc. Girdin was cloned into BamHI and NotI site of the PCDH-CMV-MCS-EF1-Puro lentiviral vector. Following bacterial transformation, three to five bacterial clones from each plasmid were picked and plasmid DNA was purified and sequenced. The sequences of the inserts from all the clones were 100% correct.
Lentivirus production and infection
Lentivirus was produced by transfection into the retroviral packaging HEK-293T cell line. The day before transfection, HEK-293T cells were cultured in 10-cm culture dish. Lentiviruses were produced by co-transfection of lentiviral plasmid, packing plasmids ∆R, and pVSVg into HEK-293T cells. After transfection, viral supernatant was collected and centrifuged. Supernatant was collected and the virus was used to infect cultured LN229 cells. After 48-h infection, lentivirus-infected cells were screened by 2 mg/mL puromycin for 2 weeks to establish stably expressing cells and verified by western blot analysis.
Transfection to LN229 cells with plasmids containing Girdin shRNA
The transfection was performed with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Girdin-specific shRNA plasmid for LN229 (sense sequence: 5′-GAAGGAGAGGCAACTGGAT-3′, nucleotides 4166–4184; Jiang et al. 2008) and the scrambled sequence inserted into pGPU6/GFP/Neo (control shRNA vector) were from GenePharma Corp. (Shanghai, China). To establish stable siGirdin/LN229 cell lines, the G418-resistant cells were screened for generating stable clones, and their expression level of Girdin protein was monitored by western blotting.
Transient Girdin siRNA transfection to U87 cells
U87 cells were transfected transiently with 1 μmol of control, or human Girdin-siRNA oligos (5′-GAAGGAGAGGCAACTGGAT-3′) and 28 μL GenePorter Transfection Reagents (Gene Therapy Systems, San Diego, CA, USA). The transfected cells were cultured in complete medium for 36 h before performing functional assays or harvesting cell lysates for protein expression analyses.
Cell lysates were applied for performing sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and then the proteins were transferred onto polyvinylidene difluoride membranes. Subsequently, the membranes were blocked with 5% skim milk solution and were incubated with the first antibodies overnight at 4°C. The horseradish peroxidase-conjugated secondary antibodies were applied for membrane incubation at 25°C. Results were visualized by enhanced chemiluminescence substrate.
Proliferation assays were done as described previously (Ma et al. 2011). Cells (3.5 × 104 cells/mL) were plated and cultured in a complete medium in 12-well plates at the same time. From the second day to the sixth day, control cells and Girdin-deficient cells were trypsinized, respectively, and the total cell number was counted each day.
Chemotaxis and chemokinesis assay
Chemotaxis assay and chemokinesis assay were done as described previously (Zhang et al. 2009). The cells (8 × 105 cells/mL) suspended in the binding medium were added to the upper chambers. EGF was loaded into the lower chemotaxis chamber with different concentrations (1, 10, 100, and 1000 ng/mL). Then, the chamber was incubated at 37°C for 3 h. The membrane was then fixed and stained. The numbers of migrating cells were counted at 200× in three separate fields by a light microscope. For the chemokinesis assay, cells were suspended in medium containing different concentrations of EGF (0, 1, 10, and 100 ng/mL) before loading to the upper chambers. The lower chambers were added to the same concentrations of EGF. The chamber was incubated at 37°C for 3.5 h. The numbers of migrating cells were counted as same as the chemotaxis assay.
Cells were plated in 35-mm dishes overnight to grow into a monolayer. Then, it was lined out with an even trace in the middle using a 10-μL pipette tip. The cells were then incubated at 37°C in 5% CO2 incubator. The wounds were photographed at intervals and the distance of the wounds was measured under a light microscope (100×).
Cellular F-actin measurement
The F-actin content was detected as described previously (Sun et al. 2005). Briefly, cultured cells were stimulated by 50 ng/mL EGF at 37°C at different time points (0, 15, 30 s, 1, and 2 min), then fixed, permeabilized, and incubated with Oregon Alexa-fluro 568 phalloidin in F-actin buffer. The labeled phalloidin that were bound to F-actin was extracted by methanol. The fluorescence was captured by a microplate fluorescence reader at excitation wavelength of 578 nm and emission wavelength of 600 nm. The relative F-actin content was calculated by the equations: F-actin Δt/F-actin 0 = fluorescence Δt/fluorescence 0.
The adhesion assay was carried out as described previously (Takino et al. 2003). 1.5 mL of cell (4 × 105 cells/mL) solution was added to a 35-mm dish containing glass coverslips. The coverslips had been pre-treated with fibronectin. After 5, 15, 30, and 45 min of incubation, the cells were washed gently twice and then fixed with 4% paraformaldehyde. The cells attached to the coverslips were counted under a light microscope at 200×.
Matrigel invasion assay
The invasion of cells in vitro was measured by the invasion of cells through Matrigel-coated transwell inserts as described previously (Takino et al. 2003). Briefly, the cells were added in triplicate wells and binding medium with 10 ng/mL of EGF was added to the lower well. After 24 h of incubation, the invading cells were fixed and stained. Cells' invasion through the membranes were counted under a microscope in five random fields at 200× magnifications.
In brief, supernatants of cultured cells were electrophoresed in sodium dodecyl sulfate–polyacrylamide gel electrophoresis under non-reducing conditions, containing 1% gelatin as MMP substrate. After electrophoresis, the gel was rinsed in 1% Triton X-100 for 1 h, washed in water, and incubated overnight in substrate buffer containing 50 mM Tris–HCl, 5 mM CaCl2, and 150 mM NaCl (pH 7.5) at 37°C with gentle shaking. Then, the gel was stained in 0.1% Coomassie Blue R-250 and destained using 10% methanol and 5% acetic acid solution. The gel was visualized using a Bio-Doc Imaging system (UVP Corporation, Upland, CA, USA).
Tumor proliferation in vivo
Male athymic Nu/Nu mice (4–5 weeks of age) were purchased from Beijing Vital River Company (Beijing, China). All animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals, prepared by the National Academy of Sciences, and published by the National Institutes of Health. A total of 3 × 106 cells were subcutaneously inoculated into one flank of each mouse. Tumor size was measured each week using a digital caliper and the mice were killed after 7 weeks. Fifteen animals were used in each group. The volume (V) of the tumors was calculated by the equation: V = 1/2 ab2, where a and b are the larger and smaller diameter of the subcutaneous tumor, respectively.
Tumor invasion in vivo
Nude mice were purchased from Beijing Weitonglihua Company. For the intracranial orthotopic human glioma model, approximately 5 × 104 glioblastoma cells were injected intracranially into the frontal cortex of 5-week-old male nude mice. All animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals, prepared by the National Academy of Sciences, and published by the National Institutes of Health. Nineteen mice in the control group and 18 mice in the experimental group received intracranial human glioblastoma cells transplants. Four weeks later, the mice were killed and the whole brain was fixed with 4% paraformaldehyde and removed. Samples were soaked in wax, and then were cut to 5-mm-thick sections with routine histological methods.
Glioma specimens were obtained from 101 patients (Table 1), who underwent surgery from 2000 to 2004 and 2008 to 2012 in Tianjin Medical University Cancer Institute & Hospital. This study was approved by the Ethical Committee of the Tianjin Medical University Cancer Institute & Hospital. All patients provided written informed consent for studies of their tumors. The sections from all cases were reviewed and classified according to current World Health Organization guidelines. None of the patients had received radiotherapy or chemotherapy prior to surgery. The histopathology was reviewed and the diagnosis in each case confirmed independently by three pathologists according to World Health Organization criteria. Among 101 patients, 90 patients were followed up. Fifty-nine patients died of tumors with a mean of 28 months (1–99 months); 31 patients were survived with a mean of 78 months (20–160 months); 67 patients developed recurrence with a mean of 22 months (1–80 months); 23 patients had no recurrence with a mean of 78 months (10–160 months).
Table 1. Girdin expression and pathological features of gliomas
Low & moderate
p values were calculated by Spearman's rank-correlation test.
Age: 46 (9–80) years, F = 0.625, p = 0.537 (anova test).
Tumor size: 4.34 ± 1.38 cm, F = 4.159, p = 0.018 (anova test).
Immunohistochemistry assay was performed using standard techniques (Zhang et al. 2009). Paraffin-embedded tumor sections were fixed in 4% paraformaldehyde and incubated with anti-Girdin antibody (1 : 60) for overnight at 4°C, and then were incubated with second antibody. Peroxidase activity was localized with diaminobenzidine. Staining was evaluated by a neuropathologist and an investigator blinded to diagnosis. For quantitative analysis, the evaluation of the immunostaining results was based on a double scoring system. The intensity score was decided as follows: 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining). The staining area was scored as follows: 0 (0%), 1 (1–10%), and 2 (11–100%). The Histo score (H score) was calculated by multiplying the intensity score and the score of staining area, producing a total range of 0–6. Results of the H score were categorized into three groups: no (score = 0), low & moderate (score = 1–3), and high (score = 4–6).
Statistical analysis was carried out using Prism 3.0 Software (GraphPad Software Inc., La Jolla, CA, USA). Data were presented as mean ± SD. Statistical significance for comparisons between groups was determined using two-way anova. Student's t-test was used for invasion in vitro analysis, and chi-square test was used for in vivo analysis.
Over-expression of Girdin increased migration ability of LN229 cells
First, we over-expressed Girdin expression in LN229 human glioblastoma cells, and the western blot results showed that there was a higher expression level of Girdin in the experimental group (Girdin/LN229) than that in control group (vetor/LN229) and parental cells (Fig. 1a). The proliferation assay results suggested an increase of proliferation upon Girdin over-expression, while there was no statistical difference between Girdin/LN229 and vetor/LN229 cells (Fig. 1b). Results of scratch assays indicated that it took the Girdin/LN229 cells a shorter time to fill the gap than the vetor/LN229 cells (Fig. 1c). Adhesion ability is another key response associated with cell movement. We detected over-expression of Girdin affected cell adhesion in the LN229 cells; the number of adherent Girdin/LN229 cells was significantly increased at 5 and 15 min compared with their control cells (Fig. 1d). The increase in the migratory ability usually leads to an increase in the invasive ability. The invasion assay showed prominent differences between the Girdin/LN229 and control cells. Significantly, higher invasion rate was observed in the Girdin/LN229 cells (Fig. 1e).
Knocking down Girdin expression in glioblastoma cell lines
To further assess the function of Girdin, we employed RNA-mediated interference to suppress the expression of Girdin in two glioblastoma cell lines, LN229 and U87. Plasmid expressing Girdin siRNA sequence was transfected into LN229 cells to obtain stable Girdin-depleted LN229 cells, which were designated as siGirdin/LN229 cells. A siRNA vector containing a scrambled sequence was also transfected to the LN229 cells to generate control cells, which were designated as scr/LN229 cells. After the G418 selection, several independent stable clones were isolated and the Girdin expression was verified by western blotting analysis. Girdin protein level was significantly reduced in siGirdin/LN229 clone 2 cells compared with scr/LN229 cells (Fig. 2a). Furthermore, Girdin siRNA was transiently transfected into U87 cells and Girdin protein expression was significantly decreased comparing to control U87 cells (Fig. 2b). To confirm the function of Girdin, all the subsequent functional assays and western blot analysis were conducted using the siGirdin/LN229 clone 2 and siGirdin/U87 glioblastoma cells, which show significant reduction of Girdin protein.
Prior study showed that suppression of endogenous Girdin could lead to a reduction of DNA synthesis in HepG2 cells, a human hepatocellular liver carcinoma cell line (Anai et al. 2005). In this study, we examined the effect of Girdin silencing on the cell proliferation in LN229 cells in vitro. The result of proliferation assay in this study indicated that reduction of Girdin led to a decreased cell proliferation compared with control cells (Fig. 2c, *p < 0.01), which is consistent with previous report. In case of U87 cells, similar observation was noticed after Girdin reduction (Fig 2e).
Reduction of Girdin impaired the migration ability of glioblastoma cells
To further confirm the effects of Girdin on LN229 cells' migration, two-dimensional scratch assays were performed. After making the scratch in the fluent monolayer cells, scr/LN229 cells migrated into the wound and resulted in wound closure within 24 h. In contrast, siGirdin/LN229 cells were significantly less motile as supported by the lagging mean distance of closure (Fig. 2d). In case of U87 cells, it showed the similar pattern upon Girdin reduction (Fig 2f).
Chemotaxis plays a central role in various biological processes, such as cellular morphogenesis, innate immunity, inflammation, and metastasis of cancer cells (Eccles 2004; Martin and Parkhurst 2004; Bottcher and Niehrs 2005; Sasaki and Firtel 2006). Cell migration in response to chemotactic stimuli is a key aspect of many physiological and pathological processes (Van Haastert and Devreotes 2004). To investigate whether Girdin silencing affects LN229 cells' migration, we conducted three-dimensional cell migration assays using a 48-well chemotaxis model. Compared with the control cells, we found that siGirdin/LN229 cells displayed a significant reduced chemotaxis ability. Meanwhile, EGF-induced chemotaxis in scr/U87 and siGirdin/U87 cells was detected, and the cells' chemotaxis ability was also significantly attenuated with Girdin reduction, consistent with the data of LN229 cells (Fig. 3a). Chemokinesis, as a chemical-induced cellular random motility detection stimulated in a gradient-independent manner, is a critical component of chemotaxis (Demuth and Berens 2004). The analysis showed that chemokinesis was significantly impaired in siGirdin/LN229 cells compared with the control (Fig. 3b). Taken together, these results clearly demonstrated that Girdin reduction by small RNA interference technology impaired the migration ability of glioblastoma cells in vitro.
Reduction of Girdin impaired the F-actin polymerization in the LN229 cells
Along the cell movement, the cellular components are rearranged leading to changes in cell morphology, which include the development of membrane protrusions and formation of lamellipodia at the leading edges. The formation of membrane anchors allows cytoskeletal contraction, which finally advances a cell forward (Demuth and Berens 2004). During this process, actin filaments (also called F-actin) regulate many aspects of dynamic cell motility (Pollard and Borisy 2003). Previous reports showed that Girdin could associate with actin filaments via its C-terminal domain in different cell types, such as human Hela uterine carcinoma and human mammary carcinoma cell line MDA-MB-231 (Jiang et al. 2008). In this study, immunofluorescent staining of the F-actin was performed. The result showed that the F-actin polymerization was impaired in siGirdin/LN229 cells and the cells' shape changed with rugged boundaries compared with the control (Fig. 3c). Because the F-actin polymerization is a quick and transient process (Ananthakrishnan and Ehrlicher 2007; Zhang et al. 2009), and chemokine-induced actin polymerization is within the first 1–2 min after stimulation (Voermans et al. 2001), our result showed that the transient actin polymerization occurred at 15 s and 1 min in the control, while the F-actin polymerization of siGirdin/LN229 cells were delayed and the peaks were lower than the control (Fig. 3d). This result suggested that Girdin reduction induced the impaired F-actin polymerization, which may partially explain the decreased migration ability.
Reduction of Girdin impaired the adhesion ability of glioblastoma cells
Cells' adhesion plays an important role in different physiological processes including hemostasis, inflammation, migration, and neuronal development (Gumbiner 1992; Hynes and Lander 1992; Alattia et al. 2002). Previous studies showed that the decreased cells' adhesion ability to substrate correlated with decreased cells' migration (Guo et al. 2009; Zhang et al. 2009). To detect whether Girdin affects cell adhesion in the LN229 and U87 cells, adhesion assays were applied. The results showed that the number of adherent siGirdin/LN229 cells (at 15 and 30 min) or siGirdin/U87 cells (at 5 and 15 min) was significantly decreased compared with their controls, respectively (Fig. 4a and b). Integrins are involved in the adhesion process between cell and extracellular substrate interactions and the β1 subunit of integrin is reported to be related to malignant behavior of gliomas (Hynes and Lander 1992). FAK can bind with the integrin β1 to promote cell attachment (Schlaepfer et al. 1994; Sieg et al. 1999; van Nimwegen and van de Water 2007). In this study, we used western blotting assay to observe the integrin-β1 and FAK activities with Girdin reduction. As shown in Fig. 4(c), both phosphorylated integrin β1 and phosphorylated FAK levels decreased in the siGirdin/LN229 cells, consistent with the reduction in adhesion.
Girdin silencing inhibits glioma invasion
The studies described above showed that Girdin could modulate cells' migration. Next, we tested whether Girdin reduction could induce impaired invasion ability. EGF (10 ng/mL) was used to stimulate the cells to penetrate through the matrigel toward the lower chamber. The number of invasive siGirdin/LN229 cells was significantly reduced compared with the control (Fig. 4e). MMPs can degrade extracellular matrix proteins and is important for gliomas invasion (Demuth and Berens 2004; Kessenbrock et al. 2010; de la Pena et al. 2010). To detect how Girdin regulates cells' invasion, we focused on the relationship between Girdin and MMPs, particularly the MMP-2 and MMP-9 expression, both are highly related to the progression of glioblastoma (Levicar et al. 2003; Kong et al. 2007). As shown in Fig. 4(d), the expression of the MMP-2 was significantly decreased in the siGirdin/LN229 cells, compared with the control by western blot assay. The expression of MMP-9 did not show significant changes. The activity of MMPs was also determined by gelatin zymography and the activity of MMP-2 was found impaired which was consistent with the western blot result. These results suggest that Girdin may be involved in the regulation of glioma cells' invasion by MMP-2.
To determine whether the in vitro assays described above have any bearing on tumorigenicity in vivo, we applied the subcutaneous mouse xenograft model to examine the role of Girdin on cell proliferation. Thirty Nu/Nu mice were used and were divided into two groups, which were injected with siGirdin/LN229 or scr/LN229 clone cells, respectively. The size of tumors was measured each week. Tumor growth in the siGirdin/LN229 group mice was significantly slower than the control (Fig. 4f, p < 0.05). To further validate the role of Girdin in cells' invasion, we applied intracranial orthotopic mouse xenograft glioma models. To measure invasiveness in vivo, the number of satellite tumors (tumor foci not connected with the main tumor) was counted in sections and determined as a semiquantitative measure of tumor invasion according to the previous report. The mean satellite tumors of the two groups are shown in Fig. 4(g) (*p < 0.05).
Expression of Girdin was associated with gliomas progression
To study the role of Girdin expression in clinical glioma development, the immunohistochemical analysis on 101 glioma tissue samples was conducted. We found that Girdin expression was positively associated with the histological grades (rs = 0.224, p =0.025). However, no significant associations were identified between the expression of Girdin and gender (rs = 0.067, p =0.503), pathological type (rs = −0.012, p =0.907), peritumoral edema (rs = 0.121, p =0.227), or tumor location (rs = −0.190, p =0.057; Table 1). The immunohistochemical staining of Girdin is shown in Fig. 5(a) and (b). Figure 5(a: ii–iv) illustrates the images of categorized H score (negative, low to moderate and high), respectively. Figure 5(a: i) shows the image of the negative control (without primary antibody). Increased expression of Girdin was correlated with worse progression-free survival (PFS) or overall survival (OS) (Fig. 5c and d).
Cell migration is an evolutionarily conserved mechanism that underlies the development and functioning of cells or tissues and take place in normal and pathogenic processes (Kurosaka and Kashina 2008). We used the scratch assay and EGF-induced chemotaxis assay to investigate the Girdin function on Glioblastoma LN229 cell migration, which showed that Girdin reduction attenuated the cells' migration ability with supporting results from another glioblastoma cell line U87. Our results were consistent with the results of breast cancer cells in the previous report (Jiang et al. 2008). Actin is one of the major ingredients in microfilament of cytoskeleton and actin reorganization is critical for cellular morphological structure. It provides the forces for cellular motility (Pollard and Borisy 2003). In the migrating cells, the leading lamellipodium is a sheetlike protrusion filled with actin filaments at high density (Small et al. 2002). Long and flexible actin filaments cannot sustain a pushing force (Pollard and Borisy 2003), therefore cells must create a dense array of short-branched filaments by utilizing cross-linking proteins, which allow nascent filaments to push against the membrane at the leading edge. Previous studies have shown that Girdin could recruit actin filaments in bundles or a meshwork, and anchor the bundles to the plasma membrane at the cortical region of the cells. In cells directional migration, the phosphorylation of Girdin by Akt causes Girdin to detach from the membrane, move to the leading edge, and cross-link newly generated short filaments to facilitate lamellar protrusion and cell migration (Enomoto et al. 2005, 2006). Moreover, Girdin-deficient cells showed multiple protrusions which limit directional migration (Enomoto et al. 2005). Our observation showed that Girdin silencing significantly impaired the F-actin polymerization in LN229 cells, partially explaining the decreased migration ability.
Except for the rearrangement of cytoskeleton (Ananthakrishnan and Ehrlicher 2007), cell migration is also closely related to the adhesion ability of cells to substrate (Demuth and Berens 2004). Our study found that the Girdin reduction not only attenuated the cells' migration but also concurrently impaired the cells' adhesion ability to extracellular substrate. Cells stably adhere to the extracellular substrate via a number of adhesion molecules such as integrins, which are transmembrane glycoproteins composed of two subunits α and β. It is reported that β1 subunit of integrins is related to malignant behavior of gliomas (Hynes and Lander 1992) and its phosphorylation can activate the FAK, a cytoplasmic-tyrosine kinase to mediate adhesion (Guan and Shalloway 1992; Natarajan et al. 2003). Results of our study demonstrated that reduction of Girdin led to the decrease in phosphorylation of integrin β1 and FAK which may, in turn, impair the adhesion ability in glioma cells.
Like migration and adhesion, cell invasion is also an important step in tumor development (Irie et al. 2005). Our matrigel invasion assay showed that the reduction of Girdin led to the decrease in invasiveness in the LN229 cells. We also established mice xenograft model to further investigate the invasive capacity of Girdin suppression in vivo, further supporting the identified Girdin function in glioma invasion.
To determine how Girdin promotes glioblastoma cell migration and invasion, we elucidated the relationship between Girdin and MMPs. It is well known that protein levels of MMP-2 and MMP-9 are highly increased upon glioblastoma progression (Vihinen et al. 2005; Kessenbrock et al. 2010; de la Pena et al. 2010). Our study showed that reduction of Girdin led to the significantly decreased expression of MMP-2 in LN229 cells, which may be a plausible reason for their impaired invasion ability.
Previous studies have reported that Girdin is highly expressed in glioma patients and may maintain the stemness of glioblastoma cells, such as the high level of cell motility or invasion ability (Natsume et al. 2012; Zhao et al. 2013). In this study, our immunohistochemistry results also indicated that Girdin expression was closely related to the malignancy of gliomas. Moreover, increased expression of Girdin was correlated with worse PFS or OS, which was consistent with the previous report. In summary, our in vitro and in vivo results demonstrated that Girdin was involved in migration and invasion of human glioblastoma cells. Further studies are required to delineate the detailed regulation mechanism and factors.
Acknowledgments and conflict of interest disclosure
This study was supported by China 863 program (2012AA020101); National Scientific Foundation of China (81272358), Foundation of National Clinical Research Center of Cancer (N14B10). We thank Dr. Ming Zhang (Assistant Professor, Department of Epidemiology and Biostatistics, University of Georgia, USA) for her English proofreading of the manuscript. The authors have no conflicts of interest to declare.
All experiments were conducted in compliance with the ARRIVE guidelines.