SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References

Wnts are secreted ligands that consist of 19 members in humans, regulate cell proliferation, differentiation, motility and fate in many stages including the embryonic stage and tumorigenesis. Wnts bind to cell surface receptors named Frizzleds and LRPs, and transduce their signals through β-catenin-dependent and -independent intracellular pathways. Gliomas are one of the most common intracranial tumors. Gliomas exhibit a progression associated with widespread infiltration into surrounding neuronal tissues. However, the molecular mechanisms that stimulate the invasion of glioma cells are not fully understood. We established two cell lines from human glioma cases and analyzed the expression of all Wnt and Frizzled members in these cell lines and other well-known glioma cell lines by real-time PCR study. The mRNA of Wnt-5a and -7b and Frizzled-2, -6 and -7 were overexpressed in glioma cells. The elevation of Wnt-5a expression was most remarkable. Although Wnt-5a is reported to have oncogenic and antioncogenic activity in several cancers, the role of Wnt-5a signaling in human glioma cells remains unclear. Immunohistochemical study also revealed high expression of Wnt-5a in 26 (79%) of 33 human glioma cases. The positivity of Wnt-5a expression was correlated with the clinical grade. Knockdown of Wnt-5a expression suppressed migration, invasion and expression of matrix metalloproteinase-2 of glioma cells. Reciprocally, treatment with purified Wnt-5a ligand resulted in stimulation of cell migration and invasion. MMP-2 inhibitor suppressed the Wnt-5a-dependent invasion of U251 cells. These results suggested that Wnt-5a is not only a prognostic factor but also a therapeutic target molecule in gliomas for preventing tumor cell infiltration. (Cancer Sci 2011; 102: 540–548)

Gliomas of astrocytic, oligodendroglial and ependymal origin account for more than 70% of all brain tumors. The most frequent (65%) and most malignant histological type is glioblastoma. Despite advances in surgical and clinical neuro-oncology, their prognosis remains poor. In a meta-analysis of 12 randomized clinical trails, the overall survival rate of patients with glioblastomas (grade IV) and anaplastic astrocytomas (grade III) was 40% at 1 year and only slightly higher (46%) after combined radiotherapy and chemotherapy.(1) Glioblastomas may develop de novo (primary glioblastoma) or by progression from low-grade or anaplastic astrocytoma (secondary glioblastoma). Various genetic alterations were reported in glioblastoma.(2) Furthermore, glioblastomas are characterized by marked invasiveness and vascularity. Glioma tumor cells often invade beyond the main tumor mass at diagnosis and this makes surgical treatment difficult. Clarification of the molecular mechanism of invasion could be useful in understanding the pathogenesis of glioma.

Wnt proteins are a large family of cysteine-rich secreted molecules consisting of at least 19 members in mammals to date. Wnt regulates cell growth, cell motility and cell fate.(3,4) There are at least three distinct intracellular pathways activated by Wnt proteins. The canonical pathway is identified as a β-catenin-dependent pathway that is highly conserved among species.(5) When Wnt acts on its cell-surface receptors, which consist of Frizzled (Fz) and LRP5/6, β-catenin is stabilized by release from the Axin-mediated degradation complex. The accumulated β-catenin is translocated to the nucleus where it binds to the translocation factor T-cell factor/lymphoid enhancer factor and thereby stimulates the expression of various target genes.(6) Some Wnts, including Wnt-1, -3a and -7a, activate the β-catenin-dependent pathway. Furthermore, accumulation of β-catenin in the cytosol has frequently been observed in many malignant tumors.(7,8)

Another class of Wnts, including Wnt-2, -4, -5a, -5b, -6 and -11, activate β-catenin-independent pathways including planar cell polarity and the Ca2+ pathway. These pathways partially overlap and primarily regulate cell motility and polarity.(9) In these pathways, Wnt binds to Fz and activates downstream effectors including Ca2+/calmodulin-dependent protein kinase II, protein kinase C, c-Jun N-terminal kinase (JNK), nuclear factor of activated T cells and small GTPases.(5)

Wnt-5a is known to activate β-catenin-independent pathways(9–11) in various cell lines and shows contrary functions. Previous studies indicated that Wnt-5a stimulates cell motility, invasiveness and aggressiveness in some cancer cells.(9,11–15) These reports suggest that Wnt-5a has oncogenic properties. On the other hand, some reports indicate that Wnt-5a acts as a tumor suppressor due to its ability to inhibit proliferation, motility and invasiveness in thyroid tumor and colorectal cancer cell lines.(16,17)

Taken together, these findings indicate that the action of Wnt-5a should be analyzed in each cell type. Marked expression of Wnt-5a was already reported but its roles in glioma cells are still unclear.

In the present study we found that Wnt-5a is highly expressed in advanced stages of glioma and that its expression is involved in invasive activities and the transcription of matrix metalloproteinase-2 (MMP-2), which is highly expressed in glioblastoma and responsible for the degradation of surrounding tissues.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References

Materials and chemicals.  The dsRNA for Wnt-5a and scramble were purchased from Santa Cruz Biotechnology, Inc (Santa Cruz, MA, USA). The universal probes for real-time PCR were provided by Roche Diagnostics GmbH (Mannheim, Germany). Neo conditioned medium (CM) and Wnt-3a or Wnt-5a CM was prepared as previously described.(10,18) Mouse Wnt-5a protein was purified to homogeneity from CM produced by mouse L overexpressing mouse Wnt-5a as previously described.(10) Anti-Wnt-5a antibody was purchased from LIFESPAN BIOSCIENCE (Seattle, WA, USA). Anti-β-catenin and -GSK-3β antibodies were purchased from BD Biosciences (Bedford, MA, USA). Anti-Dvl antibody was prepared in rabbits by immunization with recombinant protein of Dvl-1-(1-140).(19) Normal human brain total RNA was purchased from Clontech Laboratories, Inc (Mountain View, CA, USA). MMP-2 inhibitor (MMP-2 inhibitor III) was purchased from Calbiochem (Darmstadt, Germany).

Tissue samples.  Patients were treated at Kagoshima University Hospital. In total, 33 primary tumors were collected from patients diagnosed with glioma. Patients provided written informed consent. The research protocol and consent form were approved by the Institutional Review Board of Kagoshima University in advance of the study.

Immunohistochemistry and immunocytometry.  The tissue samples were sectioned and stained with hematoxylin and eosin, anti-Wnt-5a or -β-catenin antibody as previously described.(20) The stainings for Wnt-5a were visually classified into four grades according to the percentage of Wnt-5a-positive tumor cells (positive tumor cells/total tumor cells) by two authors (HH and KA) as follows: score 0, Wnt-5a-positive tumor cells were less than 1%; score 1, Wnt-5a-positive tumor cells were 1–9%; score 2, Wnt-5a-positive tumor cells were 10–49%; score 3, Wnt-5a-positive tumor cells were more than 50%. The procedures were completed by the two authors independently and there was no discrepancy between their decisions.

Glioma-derived cell lines.  U87MG and U251 glioblastoma-derived cell lines were purchased from ECACC and RIKEN BIORESOURCE CENTER (Tsukuba Science City, Ibaraki, Japan), respectively. U105 cells were provided by Dr Kuratsu Jun-ichi (Department of Neurosurgery, Kumamoto University, Japan). To establish the G24M and G5M cell lines, primary culture tumor cells from glioblastoma cases were prepared as previously described.(21)

Real-time RT-PCR.  Total RNA was extracted from cells as previously described(22,23) and reverse-transcribed using a First-strand cDNA Kit (Roche Diagnostics). Expressions of Wnt and Fz member mRNA were estimated by real-time PCR using LightCycler TaqMan Master, LightCycler 1.5 and LightCycler software version 3.5 (Roche Diagnostics) according to the manufacturer’s instructions. Expression of each mRNA was normalized using β-actin as a loading control. Each mRNA expression was compared with that of normal human brain. The series of primers used in the present study are listed in Table 1.

Table 1.   Primer sets used in the present study for real-time RT-PCR
  1. #Numbers included in the primer name refer to the name of the Universal Probe (Roche Diagnostics) used for real-time RT-PCR.

Wnt1-#81-1F5′-cgctggaactgtcccact-3′Fz1-#69-1F5′-cggcaagaccctcaactc-3′
Wnt1-#81-2R5′-aacgccgtttctcgacag-3′ Fz1-#69-2R5′-ccttgtttgctgttggtgag-3′
Wnt2-#4-1F5′-ccagagccctgatgaatctt-3′Fz2-#1-1F5′-ggtgtcggtggcctacat-3′
Wnt2-#4-2R5′-ttgtttcaagaaccgcttaca-3′ Fz2-#1-2R5′-gagaagcgctcgttgcac-3′
Wnt2b-#34-1F5′-gccgtgtcatgctcagaa-3′Fz3-#75-1F5′-acagcaaagtgagcagctacc-3′
Wnt2b-#34-2R5′-gtggactacccctgctgatg-3′Fz3-#75-2R5′-ctgtaactgcagggcgtgta-3′
Wnt3-#81-1F5′-ctcgctggctacccaattt-3′Fz4-#19-1F5′-ttcacaccgctcatccagta-3′
Wnt3-#81-2R5′-gcccagagatgtgtactgctg-3′Fz4-#19-2R5′-tgcacattggcacataaaca-3′
Wnt3a-#5-1F5′-catgaaccgccacaacaac-3′Fz5-#2-1F5′-tggagctgcgctaatcct-3′
Wnt3a-#5-2R5′-tggcacttgcacttgaggt-3′Fz5-#2-2R5′-ttggtgtgtgatccatgagg-3′
Wnt4-#4-1F5′-gcagagccctcatgaacct-3′Fz6-#23-1F5′-tgggttggaagcaaaaagac-3′
Wnt4-#4-2R5′-cacccgcatgtgtgtcag-3′Fz6-#23-2R5′-tcttcgactttcactgattgga-3′
Wnt5a-#48-1F5′-attgtactgcaggtgtaccttaaaac-3Fz7-#54-1F5′-gccagcttgtgcctaatagaa-3′
Wnt5a-#48-2R5′-cccccttataaatgcaactgttc-3′Fz7-#54-2R5′-agccgggagaaactcacag-3′
Wnt5b-#56-1F5′-ctgctgctgctgttcacg-3′Fz8-#18-1F5′-cgctggtcatctacctcttca-3′
Wnt5b-#56-2R5′-caccgggttcaaagctaatg-3′Fz8-#18-2R5′-gatgcggaacagggacac-3′
Wnt6-#5-1F5′-cagttccagttccgcttcc-3′Fz9-#44-1F5′-cctgcatcccctagagacag-3′
Wnt6-#5-2R5′-gaacacgaaggccgtctc-3′Fz9-#44-2R5′-tcacttgcctgaccttgaca-3′
Wnt7a-#75-1F5′-cttcgggaaggagctcaaa-3′Fz10-#25-1F5′-ttggttttccagcgaagg-3′
Wnt7a-#75-2R5′-gcaatgatggcgtaggtga-3′Fz10-#25-2R5′-ccacaaattagttacacaagaggcta-3′
Wnt7b-#4-1F5′-cgcctcatgaacctgcata-3′MMP1-#7-1F5′-gctaacctttgatgctataactacga-3′
Wnt7b-#4-2R5′-gctgcatccggtcctcta-3′MMP1-#7-2R5′-tttgtgcgcatgtagaatctg-3′
Wnt8a-#75-1F5′-gggtcaaacaatggaaaaaca-3′MMP2-#70-1F5′-ataacctggatgccgtcgt-3′
Wnt8a-#75-2R5′-tccccaaattccacattgtc-3′MMP2-#70-2R5′-aggcacccttgaagaagtagc-3′
Wnt8b-#75-1F5′-tgtgatgactcccgcaac-3′MMP9-#6-1F5′-gaaccaatctcaccgacagg-3′
Wnt8b-#75-2R5′-tccgaagcccacattgtc-3′MMP9-#6-2R5′-gccacccgagtgtaaccata-3′
Wnt9a-#8-1F5′-tccagttccgctttgagc-3′MMP13-#73-1F5′-ccagtctccgaggagaaaca-3′
Wnt9a-#8-2R5′-agccgaggagatggcatag-3′MMP13-#73-2R5′-aaaaacagctccgcatcaac-3′
Wnt9b-#55-1F5′-ccccaaccttgttgaggac-3′Ror2-#17-1F5′-cccctcattaaccagcacaa-3′
Wnt9b-#55-2R5′-tcccttatttccctccatgtc-3′Ror2-#17-2R5′-ttcccaaaccggtcctct-3′
Wnt10a-#16-1F5′-atccacgcgagaatgagg-3′β-actin-#64-1F5′-ccaaccgcgagaagatga-3′
Wnt10a-#16-2R5′-ccgcatgttctccatcact-3′β-actin-#64-2R5′-ccagaggcgtacagggatag-3′
Wnt10b-#27-1F5′-aatgcgaatccacaacaaca-3′  
Wnt10b-#27-2R5′-tccagcatgtcttgaactgg-3′  
Wnt11-#81-1F5′-tgtgctatggcatcaagtgg-3′  
Wnt11-#81-2R5′-cagtgttgcgtctggttcag-3′  
Wnt16-#8-1F5′-caatgaacctacataacaatgaagc-3′  
Wnt16-#8-2R5′-cagcggcagtctactgacat-3′  

RNA interference.  Eighty pmol of dsRNA for Wnt-5a or scramble was transiently transfected into U87MG, U105 and U251 cells (60-mm-diameter dish) using Lipofectamine2000 (Invitrogen, Carlsbad, CA, USA). After 24–72 h transfection, cellular responses were analyzed.

Preparation of soluble fractions.  For isolation of the soluble fractions of U251 cells, the cells were suspended in an isotonic buffer (20 mM Tris/HCl [pH 7.5], 150 mM NaCl, 1 mM DTT, 2 μg/mL leupeptin, 1 mM PMSF and 50 mM sodium β-glycerophosphate) and homogenized by sonication. The homogenates were centrifuged at 100 000g for 30 min at 4°C.

Western blot analysis.  Samples were subjected to SDS-PAGE, probed with antibodies and visualized with an ECL Western blotting system (GE Healthcare, Buckinghamshire, England).

Cell migration and invasion activity.  To measure cell migration activity, a modified Boyden-chamber (tissue culture treated, 6.5 mm in diameter, 10 μm thickness, 8 μm pores; Transwell [Corning, Inc., Cambridge, MA, USA]) was used as previously described.(24,25) The lower surface of the filters was coated with 10 μg/mL fibronectin. Cells (2.5 × 104 cells) were suspended in RPMI1640 containing 0.1% BSA and applied to the upper chamber. The same medium was applied to the lower chamber. After several hours, the cells on the upperside of the filters were wiped off and the cells that migrated to the lower side of the upper chamber were fixed with 4% paraformaldehyde in PBS, stained with propidium iodide (1 μg/mL) and counted.

The invasive potential of the cells was analyzed using a Matrigel-coated modified Boyden-chamber as described above.

To perform the wound healing assay, the cells were plated onto 35-mm-diameter dishes and the monolayer cells were manually scratched with a plastic tip; then, after being washed with PBS, the wounded monolayers of the cells were allowed to heal for 6–24 h in RPMI1640 containing 10% FBS. The width of the wounds is expressed as a percentage of the initial width at zero time.(26) When necessary, cells were transfected with dsRNA for Wnt-5a or scramble in advance or simultaneously treated with mouse Wnt-5a (500 ng/mL) or control buffer (1 mM Tris/HCl (pH7.5), 10 mM NaCl and 0.05% [w/v] CHAPS) in migration, invasion and wound healing assays. The effect of MMP-2 inhibitor on the invasion activity was analyzed using a Matrigel-coated modified Boyden-chamber as described above in the presence or absence of MMP-2 inhibitor III (50 nM). These experiments were performed at least four times.

In-gel-gelatinase assay.  Isolation of membrane vesicles from cell-CM and in-gel-gelatin zymography was performed as previously described.(27) In brief, after 24 h of incubation with subconfluent cells in RPMI1640 containing 1% FBS, CM was collected and centrifuged at 100 000g for 1 h at 4°C. Precipitated fractions (1.3 or 2.6 μg of protein/lane) were developed on SDS-polyacrylamide gel containing gelatin (1 mg/mL) from swine (Bio-Rad Laboratories, Inc., Hercules, CA, USA). After electrophoresis, gels were rinsed with distilled water twice, immersed in 2.5% (w/v) Triton-X100 for 2 h, distilled in water three times, incubated in 50 mM Tris/HCl (pH 7.5), 150 mM NaCl and 10 mM CaCl2 at 37°C for 48 h, and stained with 1% Coomassie Brilliant Blue and destained.

Statistical analyses.  Statistical analyses were carried out with the nonparametric Wilcoxon rank sum test using R software (The R Foundation for Statistical Computing [Vienna, Austria]) or the linear-by-linear association test (exact P value, two-sided) using StatXact-4 (Cytel Software Corporation, Cambridge, MA, USA).

Other assays.  Protein concentrations were determined using a BCA protein assay (Thermo SCIENTIFIC, Waltham, MA, USA) with BSA as a standard.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References

Expression of Wnt and Fz family genes in glioma-derived cells.  Real-time RT-PCR studies revealed that Wnt-5a was predominantly and commonly overexpressed among all 19 Wnt families in glioma-derived cell lines. Wnt-7b were moderately expressed. Wnt-1, -2, -5b, -9a and -11 were weakly expressed (Fig. 1a). Expressions of the remaining Wnts were not increased (data not shown).

image

Figure 1.  The mRNA expression of core components of Wnt pathways. (a) The mRNA levels of various Wnts in U87MG, U105, U251, G5M and G24M cells were quantified by real-time RT-PCR. (b) The mRNA levels of various Frizzled (Fz) components in the cells were quantified by real-time RT-PCR. Relative mRNA levels are expressed as proportions of the mRNA levels of various Wnts or Fzs in normal human brain tissues.

Download figure to PowerPoint

Fz-2, -6 and -7 were dominantly expressed among 10 Fz members in glioma-derived cell lines (Fig. 1b). Expressions of Fz-1, -3, -4, -5 and -8 were not increased. Expressions of Fz-9 and -10 were not detected in glioma cells (data not shown). Furthermore, expression of Ror2, a novel Wnt-5a receptor, was very low (data not shown). These findings suggest that signaling pathways could be activated in glioma cells through overexpression of Wnts or Fz.

Immunohistochemical analysis of Wnt-5a and β-catenin in human glioblastoma tissue.  The present study focused on Wnt-5a because its expression was the greatest and it was the most commonly elevated Wnt (Fig. 1a). Normal brain and glioma tissues were stained with hematoxylin and eosin, anti-Wnt-5a or anti-β-catenin antibody (Fig. 2a–i).

image

Figure 2.  Imaging analyses of Wnt-5a and β-catenin in glioma. Representative images of glioma cases and human brain tissue. (a,d,g) Normal human brain tissue. (b,e,h) Astrocytoma (grade II). (c,f,i) Glioblastoma (grade IV). (j–m) Giant cell glioblastoma (grade IV). (a–c,k) Normal human brain tissue or glioma regions stained with hematoxylin and eosin. (d–f,l) Microscopic images stained with anti-Wnt-5a antibody. (g–i,m) Microscopic images stained with anti-β-catenin antibody. (j) Computer tomography image. Bar, 50 μm (a–i,l,m); 200 μm (k). The arrow indicates the margin between giant cell glioblastoma tissue and normal neuronal tissue. Insets, high-magnification images of the fields indicated by the arrowheads. Note that some cells showing Wnt-5a antibody-reactive staining in Figure 2d were small blood vessels.

Download figure to PowerPoint

Expression of Wnt-5a was elevated in higher grade glioma (Fig. 2d–f). There was a significant positive relationship between clinical grading and Wnt-5a staining positivity (Table 2, < 0.05) in accordance with a previous report.(28) The stainings of β-catenin were weak or localized mainly on the plasma membrane. Although the stainings of β-catenin were slightly enhanced in accordance with clinical grades, apparent cytosolic or nuclear accumulation of β-catenin staining were not observed in these cases (Fig. 2g–i).

Table 2.   Relationship between WHO stage and Wnt-5a protein expression score in human glioma cases
 Grade IIGrade IIIGrade IVSubtotal
  1. Statistical significance was analyzed by linear-by-linear association test (exact P value, two-sided, < 0.05).

Score 05 (15%) 4 (12%) 2 (6%)11 (33%)
Score 11 (3%) 2 (6%) 0 (0%) 3 (9%)
Score 22 (6%) 5 (15%) 2 (6%) 9 (27%)
Score 31 (3%) 1 (3%) 8 (24%)10 (30%)
Total9121233

Furthermore, we found an exceptional case of giant cell glioblastoma (grade IV) without infiltration, Wnt-5a positivity or nuclear β-catenin accumulation (Fig. 2j–m). Generally, giant cell glioblastoma cases showed atypical growth with a clear margin between tumors and normal tissues, and could have the potency to have different characteristics and genetic backgrounds from normal type glioblastomas.

It was reported that clinical grades of glioma correlated with infiltrative activity.(29) Taken together, we hypothesized the possibility that Wnt-5a is involved in regulating not only cell growth, as previously reported,(28) but also infiltration in human glioma cells through β-catenin-independent pathways.

Response of U251 glioma cells to Wnt-5a protein.  Wnt-5a CM stimulated the phosphorylation of Dvl of U251 cells in the presence of 2% (v/v) FBS as well as Wnt-3a CM (Fig. 3a). Wnt-3a could stabilize the soluble fraction of β-catenin but Wnt-5a could not stabilize β-catenin (Fig. 3b). These data suggest that the canonical Wnt/β-catenin pathway was present in U251 cells and that Wnt-5a did not stimulate this pathway. Transfection of dsRNA for Wnt-5a into U251 cells resulted in 75% knockdown of the expression of Wnt-5a (Fig. 3c). Similar efficiency of Wnt-5a knockdown was performed in U87MG and U105 cells (data not shown). We also examined the cell proliferation of U87MG, U105 and U251 cells in a condition with or without knockdown of Wnt-5a. No significant differences in cell growth were observed with Wnt-5a knockdown up to 96 h (Fig. 3d).

image

Figure 3.  Activity of Wnt-5a. (a) Phosphorylation of Dvl stimulated by Wnt-5a conditioned medium (CM). After U251 cells were treated with various concentrations of control (lanes 1,2), Wnt-3a (lanes 3,4) or Wnt-5a (lanes 5,6) CM for 1 h, the lysates were subjected to western blotting and probed with the indicated antibodies. p-Dvl, phosphorylated Dvl. (b) Accumulation of soluble β-catenin induced by Wnt-3a but not by Wnt-5a. After U251 cells were treated with the control, Wnt-3a or Wnt-5a CM for 6 h, the lysates were subjected to the procedure described above. The blotting of GSK-3β was used as a loading control. (c) Efficiency of Wnt-5a knockdown. U251 cells were transfected with Wnt-5a or scramble dsRNA. After 48 h, the amounts of Wnt-5a mRNA were analyzed by real-time RT-PCR. The results shown are means ± SE from five independent experiments. *< 0.05. (d) Effect of Wnt-5a knockdown on the growth of glioblastoma cells. Squares, U87MG cells; triangles, U105 cells; circles, U251 cells. Open symbols, scramble dsRNA transfected; closed symbols, Wnt-5a knockdown.

Download figure to PowerPoint

Wnt-5a-dependent cellular migration of glioma cells.  To analyze the role of Wnt-5a in glioma cells, we performed wound healing and transwell assays. When Wnt-5a was knocked down in U251 cells, the cells were allowed to migrate in scratch-wound cultures for 12 h. The migration of Wnt-5a knockdown U251 cells was retarded compared with that in the control experiments (Fig. 4). Furthermore, purified Wnt-5a protein stimulated the migration of U251 cells in scratch-wound cultures (Fig. 5). When Wnt-5a was knocked down in U87MG, U105 and U251 cells, the migration of cells was reduced compared with the control cells (Fig. 6a,b). In contrast, when purified Wnt-5a protein was added to the lower chamber, migration of U87MG, U105 and U251 cells was stimulated (Fig. 6c,d). Taken together, these results suggest that Wnt-5a is associated with the cell motility of glioma cells.

image

Figure 4.  Effect of Wnt-5a knockdown on glioma motility. (a) Inhibition of cell migration by Wnt-5a dsRNA. U251 cells were transfected with scramble or Wnt-5a dsRNA for 72 h and then the cells were wounded. The culture was continued for 12 h. Bars, 400 μm. (b) Migration distances were measured and are expressed as percentages of the migration transfected with scramble dsRNA at 12 h. The results shown are means ± SE from at least five independent experiments. *< 0.05.

Download figure to PowerPoint

image

Figure 5.  Wnt-5a-dependent cell motility in glioma cells. (a) Stimulation of cell motility by Wnt-5a (500 ng/mL) treatment. U251 cells treated with growth medium containing control buffer or Wnt-5a were wounded. The culture was continued for 12 h. Bars, 400 μm. (b) Migration distances were measured and are expressed as percentages of the migration in the presence of the control buffer at 12 h. The results shown are means ± SE from seven independent experiments. **< 0.01.

Download figure to PowerPoint

image

Figure 6.  Wnt-5a is involved in the migration of U87MG, U105 and U251 glioblastoma cells. (a) Inhibition of cell migration by Wnt-5a knockdown. The cells were allowed to migrate for 2.5 h. These images are representative. (b) Relative migration activities of U251 glioblastoma cells in Figure 6a are expressed as proportions of migration of the control cells. (c) Stimulation of cell migration by Wnt-5a (500 ng/mL) treatment. The cells were allowed to migrate for 2.5 h. These images are representative. (d) Relative migration activities of U251 glioblastoma cells in Figure 6c are expressed as proportions of migration of the control cells. Bars, 100 μm. The results shown are means ± SE from five independent experiments. **< 0.01.

Download figure to PowerPoint

Wnt-5a-dependent MMP-2 induction and invasion activity of human glioblastoma cells.  Several gene families including Wnts and MMPs are reported to be associated with aggressiveness of cancer. Previous reports indicated the existence of MMP-2 and MMP-9 activity in glioblastoma.(30) It was reported that MMP-2 is extremely upregulated in malignant gliomas and correlated with the malignant progression of human gliomas in vivo.

To estimate the role of Wnt-5a in the induction of MMPs, we transfected dsRNA for Wnt-5a into U251 cells. The expression and activity of MMP-2 were downregulated in response to the knockdown of Wnt-5a. Gelatinase activity due to MMP-9 (90 K) was not detected (Fig. 7a,b). Reciprocally, MMP-2 activity was increased by the treatment of purified Wnt-5a (Fig. 7c). Expressions of MMP-1 and MMP-9 were low and MMP-13 was not detected (data not shown) in U251 cells with/without treatment of purified Wnt-5a compared with that for MMP-2.

image

Figure 7.  Wnt-5a-dependent MMP-2 activity in U251 glioblastoma cells (a) The decrease of mRNA levels of MMP-2 in Wnt-5a-knockdown U251 cells were quantified by real-time RT-PCR. The results shown are means ± SE from five independent experiments. *< 0.05. (b) In-gel-gelatinase assay. Membranous vesicles of conditioned medium (CM) (1.3 or 2.6 μg of protein) from control or Wnt-5a-knockdown U251 cells were subjected to an in-gel-gelatinase assay. The arrow indicates the position of MMP-2 activity. (c) Effect of Wnt-5a (500 ng/mL) treatment on mRNA level of MMP-1, -2 and -9. White bars indicate each mRNA level in normal human brain as controls. Shaded bars, untreated U251 cells; black bars, Wnt-5a treated U251 cells.

Download figure to PowerPoint

These findings suggest that Wnt-5a stimulated the expression of MMP-2 but not MMP-1, -9 or -13 in terms of their transcription level.

We then performed matrigel invasion assays. Invasion of Wnt-5a knockdown U87MG, U105 and U251 cells were reduced compared with the control cells, while treatment with purified Wnt-5a protein stimulated the invasion of U87MG, U105 and U251 cells (Fig. 8a–c).

image

Figure 8.  Wnt-5a and MMP-2 are involved in the invasion of glioma cells. (a) Invasion activity of U87MG, U105 and U251 cells by Wnt-5a. The cells were allowed to invade in the Wnt-5a-knockdown condition or the Wnt-5a (500 ng/mL)-treated condition. The cells were allowed to invade for 12–36 h. The images are representatives. Bars, 100 μm. (b) Relative invasion activities of Wnt-5a-knockdown U251 glioblastoma cells are expressed as proportions of the invasion of control cells. (c) Relative invasion activities of Wnt-5a-treated U251 glioblastoma cells are expressed as proportions of the invasion of control cells. The results shown are means ± SE from four or five independent experiments. *< 0.05. **< 0.01. (d) MMP-2-dependent stimulation of invasion activity by Wnt-5a treatment in U251 cells. Invasion activity was measured under similar conditions to Figure 8c with or without MMP-2 inhibitor (50 nM). The results shown are means ± SE from four independent experiments. *< 0.05.

Download figure to PowerPoint

Treatment with MMP-2 inhibitor suppressed the Wnt-5a-dependent invasive activity of U251 cells (Fig. 8d).

These results suggest that Wnt-5a stimulated the invasive activity of glioblastoma cells through MMP-2.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References

Glioma is one of the commonly diagnosed brain tumors. Surgical therapy of glioma is difficult owing to its infiltration into surrounding neuronal tissue. Although mutations in glioma have been well studied, most of them are correlated with gene instability, cell cycle regulation and abnormal cell growth.(2)

MMPs are zinc-containing endopeptidases that mediate cell surface-associated dissociation and degradation of the extra-cellular matrix (ECM); this could cause differences in morphogenic tissue interactions during the growth, invasion, metastasis of malignant tumors and tumor-induced angiogenesis.(31)

MMP-2 and MMP-9 were reported to be expressed in glioma and thought to be important for the invasive characteristics of glioma.(32) Activities of MMP were suppressed by the tissue inhibitor of metalloproteinase (TIMP) family, soluble inhibitors of MMP, and TIMP are released from MMP in response to biological events such as inflammation, ECM turnover, tissue remodeling and cell behavior.(33) Other potential biochemical functions of TIMP have also been elucidated, including a role for TIMP-2 in binding to membrane type-1 MMP (MMP-14) on cell surfaces, which is critical for the activation of proMMP-2.(34) The molecular mechanisms that induce activation or production of MMPs remains to be clarified.

In this study, we observed predominant expression of Wnt-5a in glioma cells. Wnt-5a showed high expression in 30% of 237 gastric cancer cases and 28% of 98 prostate cancer cases.(14,25) Reportedly, Wnt-5a stimulated proliferation of human glioblastoma cells and immunohistochemical studies have shown a correlation between glioma stages and Wnt-5a positivity.(28) We also found a positive correlation between tumor grading and Wnt-5a staining in 33 human glioma cases (Table 2). β-Catenin is known to accumulate in the cytosol or nuclei of cancer cells.(18) However, histological analyses also revealed that β-catenin was not overexpressed or accumulated in glioma cells in the present study.

Since several reports have mentioned the relationship between the Wnt/β-catenin pathways and medulloblastoma(35) or glioblastoma,(36,37) neuro-oncologists seem to be interested in this issue. We have shown that U251, U87MG and U105 cells (data not shown) did not accumulate cytosolic or nuclear β-catenin in the endogenous level and in response to Wnt-5a, and focused our analyses on the β-catenin-independent role of Wnt-5a in glioma cells.

We found no publication in the Pubmed database using the key word of “Wnt-5a and giant cell glioma” and only one publication in the Pubmed database using the key word of “Wnt and giant cell glioma”.(38) Jozwiak et al.’s report was about giant cell astrocytoma concomitant with hereditary disease and should not be applied to other sporadic giant cell glioblastoma cases. Although the expression or physiological function of Wnt-5a and β-catenin expression remains to be explored, our giant glioblastoma case (grade IV) with low staining of Wnt-5a and invasion prompted us to study the role of Wnt-5a in cell motility and the infiltrative activity of glioma cells.

Wnt-5a is reported to have tumor suppressive activity(16,17) and oncogenic activity, as well as regulating several cellular responses including convergent extension, cell polarity, inhibition of β-catenin signaling cell motility and axonal repulsion.(9,11–15,25) Wnt-1 and -3a activate MMP expressions and migration through the canonical β-catenin pathway in T cells.(39) Upregulation of MMP-2 in human glioma has been reported and suggested to deteriorate tumor invasiveness.(32) However, it was not clear whether other Wnts could induce MMP production.

In the present study we have shown that knockdown of Wnt-5a reduced cell motility and the invasion activity of glioblastoma cells. The knockdown of Wnt-5a resulted in a mild decrease of the MMP-2 mRNA level and strong suppression of MMP-2 activity. MMP-2 activity is thought to be dependent not only on the amount of mRNA, but also on the speed of translation of mRNA, protein degradation and secretion of MMP-2s, and with its activation by being released from TIMP. TIMP are an inhibitor family of MMP. We could not rule out the possibility that knockdown of Wnt-5a influences these steps in a complicated manner by not suppressing only MMP-2 mRNA expression. Judging from the result that the MMP-2 inhibitor suppressed the Wnt-5a-dependent invasion activity of U251 cells, Wnt-5a could regulate the infiltration activity of glioblastoma cells through MMP-2 activity.

Wnt-5a is known to bind Fz-2, -5, -8 and Ror2 but not Fz-6 and -7.(40,41) Yamamoto et al. reported that binding of Wnt-5a to Fz-2 or Ror2 induces MMP-1 expressions by recruitment of JunD to the AP-1 binding site of the promoter region of the MMP-1 gene through the activation of PKC and JNK in prostate cancers, and that Wnt-5a promoted cell motility through activation of Rac.(25) Enomoto et al. have shown that Wnt-5a promotes MMP-13 transcription through Ror2 in sarcoma cells.(42)

In this study, elevated expressions of Fz-2, -6 and -7 were observed but no remarkable expressions of Ror2, MMP-1, -9 and -13 (data not shown) were detected in human glioma cell lines.

Fz-2 was reported to transduce Wnt-5a signals, while Fz-6 and Fz-7 did not.(40,41) These reports make us speculate that Wnt-5a might promote MMP-2 by activation of PKC and JNK through Fz-2, and that Wnt-5a might increase cell motility through activation of Rac in U251 glioblastoma cells. Given that infiltration is the major therapeutic problem of the gliomas, the Wnt-5a-mediated pathway could not only be a marker for prognosis but also a therapeutic target in glioma.

In the present study, we analyzed the β-catenin-independent role of Wnt-5a in the invasion of glioma. Further studies will be necessary to understand the physiological and pathological roles of Wnt-5a in glioma cells.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References

The authors would like to thank the Joint Research Laboratory, Kagoshima University Graduate School of Medical and Dental Sciences, for the use of their facilities, and Dr K. Kobayashi (Associate Professor, Department of Biochemistry and Genetics, Kagoshima University Graduate School of Medical and Dental Sciences). This work was supported by Grants-in-Aid for Scientific Research (C) from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by grants from the Naito Foundation.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure statement
  8. References