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

The Notch signaling pathway has been implicated in both developmental processes and tumorigenesis. Aberrant Notch signaling has been repeatedly demonstrated to facilitate the proliferation and survival of glioma cells by regulating downstream effectors or other signaling pathways. In glioblastoma multiforme specimens from 59 patients, Notch1 was highly expressed in tumor tissues compared with normal brain tissues, and this expression was correlated with elevated AKT phosphorylation and Snail expression. Increased nuclear localization of β-catenin and p50 as well as enhanced IKKα/AKT interaction were also observed in glioma tissues. In U87MG cells, the activation of Notch1 by DLL4 stimulation or by the overexpression of Notch intracellular domain (NICD) resulted in AKT activation and thereby promoted β-catenin activity and NF-κB signaling. Inhibition of EGFR partially blocked the β-catenin and NF-κB signaling stimulated by Notch1 activation. Furthermore, NICD overexpression in U87MG cells led to the upregulated expression of several metastasis-associated molecules, which could be abrogated by the knockdown of either β-catenin or p50. In U87MG and U251 cells, DLL4-induced cellular migration and invasion could be inhibited by either β-catenin or a p50 inhibitor. Collectively, these results indicate that Notch activation could stimulate β-catenin and NF-κB signaling through AKT activation in glioma cells. Thus, Notch activation-stimulated β-catenin and NF-κB signaling synergistically promote the migratory and invasive properties of glioma cells. (Cancer Sci 2012; 103: 181–190)

Gliomas are the most common malignant brain tumors derived from astrocytes, oligodendrocytes, or ependymal cells.(1,2) According to the World Health Organization (WHO) classification for gliomas based on histology and prognosis, grade IV glioma, or glioblastoma multiforme (GBM), is the most aggressive and prevalent brain malignancy.(3–5) Despite the continuous development of new clinical therapies, the prognosis and survival of GBM patients remain dismal,(6) and the deregulated growth and the migration and invasion of glioma cells remain the most important obstacles to the development of effective GBM treatments.

Notch signaling plays a critical role in embryonic and postnatal development.(7) In mammals, the Notch receptor family consists of four highly homologous members. These members are designated Notch1 to Notch4 and have both unique and redundant functions.(8) Following the binding of a Delta or Jagged ligand family member, the cleaved Notch receptors release the Notch intracellular domain (NICD), which can then translocate into the nucleus and interact with its corresponding co-activators to promote the transcription of downstream target genes.(9) Dysregulated Notch signaling has been implicated in several types of human cancers including glioma.(10–13) Furthermore, several studies have indicated that Notch signaling can induce epithelial–mesenchymal transition (EMT) in certain types of cancer.(14,15)

Notch signaling promotes the survival and proliferation of glioma cells through Akt-mTOR signaling and partially through the regulation of epidermal growth factor receptor (EGFR) expression by p53.(10,16) Hyperactive AKT activation, resulting from PI3K mutation or additional mechanisms, is often detected in malignant tumors.(17,18) The invasive and metastatic behavior of malignant tumor cells is often accompanied by the upregulated expression of EMT inducers, such as Snail, mesenchymal markers, such as vimentin, and the loss of adhesion molecules in epithelial cells, such as E-cadherin.(19–22) Recently, many studies have found that AKT activity could be correlated with tumor metastasis.(23,24) Enhanced AKT activity was found to be associated with increased concentrations of β-catenin in tumor cells, which could be due to decreased degradation of β-catenin or additional mechanisms.(25–27) Increased nuclear translocation of β-catenin can result in accelerated tumor cell invasion.(28–31) It has also been shown that AKT can induce IKK activation through an interaction with IKK that leads to the translocation of NF-κB to the nucleus.(32,33) The NF-κB signaling pathway was shown to be indispensable for metastasis induced by activated TGF-β signaling.(34,35)

Because both AKT and Notch1 signaling affect the invasiveness of epithelial tumor cells, we hypothesized that Notch1 could promote the migration and invasion of glioma cells through AKT signaling. In the present study, we found that the synergistic interaction between AKT and Notch1 signaling in glioma cells plays a critical role in glioma cell migration and invasion.

Materials and Methods

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

Tumor samples and plasmids.  Fifty-nine GBM specimens with adjacent normal brain tissues were selected from consenting patients with primary gliomas who had been operated on at the Department of Neurosurgery of Renji Hospital School of Medicine at Shanghai Jiaotong University in Shanghai, China. The samples were harvested from the tumors at the time of surgery and were then snap frozen and stored at −80°C. The present study was approved by the institutional ethical review board of Renji Hospital, School of Medicine, Shanghai Jiaotong University (approval number: 2010-AN-2). Informed consent was obtained from the patients and/or guardians. Full-length human Notch1 cDNA was obtained from HEK293T cells by RT-PCR and confirmed by DNA sequencing. For full-length Notch1, DLL4 and NICD cDNA were subcloned into the p3XFlag-CMV-10 expression vector. The NICD cDNA was also subcloned into the pEGFP-C1 vector (Clontech, Mountain View, CA, USA) to construct the recombinant plasmid pEGFP-C1-NICD. The transcriptional activity of β-catenin was measured using the TCF/LEF reporter kit (SABiosciences, Frederick, MD, USA).

Reagents, cell culture and transfections.  AKT inhibitor VIII (Calbiochem, Gibbstown, NJ, USA), EGFR inhibitor AG1478 (Sigma, St Louis, MO, USA), NF-κB p50 (NLS) inhibitory peptide set (IMGENEX, San Diego, CA, USA), β-Catenin/Tcf inhibitor FH535 (Merck, Whitehouse Station, NJ, USA) and recombinant human DLL4 (rDLL4) (R&D Systems, Minneapolis, MN, USA) were used in the present study. U87MG (human glioblastoma–astrocytoma, epithelial-like cell line) and U251 human glioma cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and grown in DMEM supplemented with 10% fetal calf serum (Life Technologies, Gaithersburg, MD, USA), penicillin G (100 U/mL) and streptomycin (100 μg/mL) at 37°C in an atmosphere with 5% CO2. Transfections of the U87MG cells were performed with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA).

Immunofluorescence and immunohistochemistry.  For immunofluorescence, the cells were washed with cold phosphate-buffered saline (PBS) and fixed in a paraformaldehyde solution (4% in PBS, pH 7.4) with a permeabilizing agent (0.1% TritonX-100) for 10 min at room temperature. Following incubation in blocking buffer (3% BSA in tris-buffered saline tween-20 [TBST]) for 1 h, the cells were incubated with the primary antibodies in TBST containing 3% BSA overnight at 4°C. Next, the cells were washed with PBS and then incubated with the indicated fluorescently labeled secondary antibody in the dark at 37°C for 1 h. The cells were then examined using fluorescence microscopy. Confocal images were captured with a confocal microscope (TCS SP5, Leica, Mannheim, Germany). For immunohistochemistry, tissues were fixed with paraformaldehyde (4%), embedded in paraffin and sectioned at a thickness of 3 μm. Following the incubation with primary and biotinylated secondary antibodies, Notch1, phospho-AKT and Snail expression were examined using the ABC reagent (Vector Labs, Burlingame, CA, USA) and 3,3′-diaminobenzidine hydrochloride (Sigma).

Immunoprecipitation and immunoblotting.  The tissue or cell lysate was pre-cleared with protein A-Sepharose for 1 h at 4°C and then immunoprecipitated with 2 μg of the appropriate antibody with gentle rocking at 4°C overnight. Next, protein A-Sepharose was added and the mixture was incubated for 4 h at 4°C. The beads were washed five times with lysis buffer. Immunoprecipitations were performed using the indicated antibodies. For immunoblotting, SDS–PAGE-separated proteins were transferred onto PVDF membranes (Millipore, Billerica, MA, USA). The membranes were developed using enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, Buckinghamshire, UK).

Antibodies.  The antibodies for Notch1, total ERK1/2, GAPDH, Histone H1, vimentin, HRP-labeled goat anti-mouse IgG and HRP-labeled goat anti-rabbit IgG were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The antibodies for p50, β-catenin, β-actin, total AKT, phospho-AKT (Ser473), phospho-ERK1/2, Snail and EGFR were manufactured by Cell Signaling Technology (Danvers, MA, USA). The Alexa Fluor 546-labeled goat anti-mouse antibody and Hoechst 33342 were purchased from Molecular Probes (Eugene, OR, USA).

RNA interference.  To knockdown the expression of Notch1, p50 and β-catenin, p50 siRNA (sc-29407; Santa Cruz Biotechnology), Notch1 siRNA (sc-36095; Santa Cruz Biotechnology) and β-catenin siRNA (sc-29209; Santa Cruz Biotechnology) were used. The control siRNA (sc-37007; Santa Cruz Biotechnology) was used as a control treatment. The cell lysates were prepared and Notch1, p50 and β-catenin protein levels were determined using western blot to determine the efficiency of the RNAi.

Real-time PCR.  Total RNA was extracted with the TRIzol Reagent (Invitrogen) and was then reverse transcribed using M-MLV reverse transcriptase (Promega, Madison, WI, USA). Real-time PCR was performed using an ABI 7500 fast sequence detection system (Applied Biosystems, Carlsbad, CA, USA) with the SYBR green fluorescent dye. All samples were analyzed in triplicate and in optically clear 96-well plates (Corning, New York, NY, USA). The cycling parameters were as follows: 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 1 min and an elongation step at 72°C for 30 s. The human β-actin transcript was used as an internal reference to control for variations in the total mRNA quantity of each sample. Each RNA sample was analyzed in triplicate and the following primers were used:

  • 1
     Human Notch1 primers: forward 5′-GAGGCGTGGCAGACTATGC-3′; reverse 5′-CTTGTACTCCGTCAGCGTGA-3′.
  • 2
     Human AKT primers: forward 5′-TGTAACCAGAGAGCGGGATGT-3′; reverse 5′-TTTTGGCATAACTAAGGCCGAA-3′.
  • 3
     Human Snail-1 primers: forward 5′-AATCGGAAGCCTAACTACAGCG-3′; reverse 5′-GTCCCAGATGAGCATTGGCA-3′.
  • 4
     Human Zeb1 primers: forward 5′-GATGATGAATGCGAGTCAGATGC-3′; reverse 5′-ACAGCAGTGTCTTGTTGTTGT-3′.
  • 5
     Human β-actin primers: forward 5′-CACTCTTCCAGCCTTCCTTC-3′; reverse 5′-GGATGTCCACGTCACACTTC-3′.

Luciferase assay and migration assay.  For the luciferase assay, U87MG cells were seeded in 24-well plates and then transfected with the indicated plasmids according to the manufacturer’s protocol for the TCF/LEF reporter kit (SABiosciences). Triplicate transfections were performed for each treatment. After transfection, the cells were harvested in lysis buffer containing 1 mM PMSF and 0.1% Triton X-100 in PBS. For wound healing migration assays, U251 cells were seeded and grown in 12-well plates coated with DLL4. Next, the U251 monolayers were scratched with 200 μL pipette tips and then washed with PBS. The wound images, which were captured with a digital camera (OLYMPUS, Center Valley, PA, USA) attached to an inverted microscope at 0 and 36 h, were analyzed using the Image Pro Plus software.

In vitro migration and invasion assays.  For the migration assays, the U87MG cells were transiently co-transfected with plasmids expressing DLL4 and GFP using Lipofectamine 2000 according to the manufacturer’s instructions. Aliquots of 2 × 105 cells/mL were plated onto the inserts of the 8-μm pore-sized trans-well chambers (Millipore-Chemicon, Billerica, MA, USA). An aliquot of the cells from each transfection was examined for GFP expression to verify consistent transfection efficiency between 70% and 80%. For the trans-well invasion assays, 1 × 106 cells were plated onto inserts containing a polycarbonate membrane with a pore size of 8 μm and the cells were then coated with a thin layer of ECMatrix (Millipore-Chemicon). Both migration and invasion were assayed 24 h later by measuring DNA content with a fluorescent dye, which was analyzed using a fluorescence plate reader and a 480/520 nm filter set (Perkin Elmer Life and Analytical Science, Turku, Finland).

Statistical analysis.  The Chi-squared test was used to analyze the correlations between NICD of Notch1 and p-AKT-S473, and Snail expression was used in the glioma specimens. All cell line-derived data were evaluated using Student’s t-test.

Results

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

Activation of Notch1 and AKT signaling and increased Snail-1 levels in glioma.  We investigated the crosstalk between Notch and other signaling pathways in glioma cells by examining the abundance and phosphorylation status of Notch1, AKT and ERK in 59 specimens of high-grade gliomas.

First, the level of Notch1, AKT and ERK phosphorylation in the gliomas was analyzed by immunohistochemistry. As shown in Figure 1(A), we found that the levels of phosphorylated Notch1, AKT and ERK were elevated in gliomas compared with the adjacent normal tissue. Next, we tabulated the levels of Notch1 expression in tumor tissues that had high levels of AKT phosphorylation or ERK phosphorylation, and these results are shown in Table 1. Thirty-six out of the 44 tumor specimens with enhanced AKT activation also demonstrated high Notch1 expression. Among the 15 tumor specimens with little AKT activation, 11 were negative for Notch1. In comparison, we only detected high levels of Notch1 expression in 23 of the 39 tumor specimens with high ERK phosphorylation. In contrast, Notch1 was negative in only three out of 20 tumor specimens with low ERK activation. These results indicate that the enhanced Notch1 signaling was closely related to AKT activation but not to ERK activation. Generally, levels of NICD are positively correlated with activation of Notch1 signaling.(9) As shown in Figure 1(B), we also found that NICD and AKT levels, but not ERK phosphorylation levels, were correspondingly increased in tumor tissues, as assessed using western blot. Tumor cell migration and invasion often occur in high-grade gliomas. Thus, we examined the expression of Snail, a transcription factor that accelerates the invasive capabilities of tumors.(19,22) As shown in Figure 1(B), both upregulated expression of Snail and hyperactive AKT signaling were found in glioma tissues but not in normal tissues. These results implied that the activation of Notch1/AKT signaling facilitates the invasiveness of tumor cells. Furthermore, abnormal activation of EGFR signaling is implicated in glioma tumorigenesis. As shown in Figure 1(B), EGFR levels and EGFR phosphorylation levels were elevated in some but not all tumor specimens compared with the adjacent normal tissues. However, the pattern of EGFR expression was different from that of NICD and AKT phosphorylation. In addition, as shown in Figure 1(C), elevated Notch1 and Snail-1 mRNA levels were detected in glioma tissues compared with normal tissues. Furthermore, the AKT mRNA levels remained unchanged in the tumor tissues. The results from these tumor tissues indicate that Notch and AKT signaling might work together to promote the metastasis of glioma cells. Furthermore, as the activities of β-catenin and NF-κB are related to tumor invasiveness(32–35) and can be enhanced by AKT activation, we examined nuclear expression as well. These results are shown in Figure 1(D). We detected increased β-catenin in tumor specimens. We also found that endogenous AKT bound to IKKα in tissues and that the amount of the AKT/IKKα complex was increased in glioma tissues compared with normal brain tissues. Furthermore, increased levels of nuclear p50 were observed in glioma tissues (Fig. 1D), which indicated that NF-κB signaling was enhanced in the gliomas.

image

Figure 1.  Expression of Notch1, p-AKT and p-ERK in glioma tissues. (A) Notch1, p-AKT and p-ERK expression was detected in glioblastoma multiforme using immunohistochemistry. (B) The protein expression levels of Notch intracellular domain (NICD), p-AKT, p-ERK, Snail-1, epidermal growth factor receptor (EGFR) and p-EGFR (Tyr1068) in gliomas. Tissue samples T1–T4 were from glioma patients and tissue samples N1–N4 were the adjacent normal brain tissues of the same patients. GAPDH was used as the internal control. (C) mRNA levels of Notch1, AKT and Snail-1 in gliomas. Tissue samples T1–T4 were glioma specimens and the tissue samples N1–N4 were the adjacent normal brain tissues. Data were collected from triplicate experiments and are shown as the means ± SD. * and ** indicate < 0.05 and < 0.01, respectively, in comparison with normal brain tissue. β-actin mRNA was used as the internal control. (D) Nuclear p50 and β-catenin levels in gliomas. Whole-tissue lysates were immunoprecipitated with anti-AKT antibody and analyzed using immunoblotting with anti-IKKα and anti-AKT antibodies. Nuclear extracts were immunoblotted with the antibodies as indicated. The samples were from gliomas and the corresponding adjacent normal brain tissues.

Download figure to PowerPoint

Table 1.   Relationship between Notch1 expression and AKT or ERK kinase activation in tumor specimens from 59 glioblastoma multiforme patients
Notch1 expressionKinase activation
ERK activity*AKT activity**
High (%) 39Low (%) 20High (%) 44Low (%) 15
  1. *P > 0.05; **P < 0.05. High expression: more than 50%; low expression: 50% or less.

High23 (59)17 (85)36 (82)4 (27)
Low16 (41)3 (15)8 (18)11 (73)

Activation of Notch1 enhanced β-catenin activity through AKT in glioma cells.  Because Notch1 signaling can relate to AKT and β-catenin signaling, we next investigated the relationship between these signaling pathways in U87MG cells. Overexpression of DLL4 in U87MG cells only slightly enhanced the nuclear translocation of β-catenin, although the simultaneous overexpression of DLL4 and Notch1 in U87MG cells heightened this effect. This result suggests that there is a low basal level of endogenous Notch1 expression in U87MG cells and that DLL4 might induce greater Notch signaling under conditions of Notch1 overexpression (Fig. 2A). It is known that the overexpression of NICD can constitutively activate downstream mediators of the Notch signaling pathway. As shown in Figure 2(A), the overexpression of NICD in U87MG cells greatly increased the level of β-catenin in the nucleus. By immunofluorescence staining, we consistently observed that NICD induced increased nuclear translocation of β-catenin (Fig. 2C). Treatment with the AKT inhibitor VIII eliminated the NICD-induced increase of nuclear β-catenin levels (Fig. 2B,C), thereby demonstrating that NICD affected nuclear β-catenin levels through the AKT signaling pathway. Next we determined the effects of NICD on β-catenin transcriptional activity using a gene reporter assay. As shown in Figure 2(D), β-catenin transcriptional activity was enhanced approximately 12-fold by NICD compared with the control groups, which is consistent with the effects of NICD on the nuclear translocation of β-catenin. Furthermore, NICD-induced β-catenin transcriptional activity was abrogated by treatment with the AKT inhibitor VIII (Fig. 2D). We also tested the effects of Notch activation on β-catenin signaling in U251 cells. As shown in Figure 2(E), the overexpression of NICD resulted in dramatically increased levels of β-catenin in the nuclei of U251 cells and this effect was blocked by AKT inhibition.

image

Figure 2.  Notch1 activation led to increased β-catenin activity. (A) Notch1 activation increased the nuclear translocation of β-catenin in U87MG cells. U87MG cells were transfected with the indicated plasmids. At 72 h after transfection, the whole cell lysate or the nuclear fraction was used for the detection of the expression of the indicated proteins. (B) The increased nuclear translocation of β-catenin caused by Notch1 was inhibited by the AKT inhibitor VIII. U87MG cells were transfected with or without the Notch intracellular domain (NICD) plasmid. The cells were then treated with or without the AKT inhibitor VIII (10 μM). The nuclear fractions were used for the detection of the expression of the indicated proteins. (C) The AKT inhibitor VIII blocked the Notch1-induced nuclear translocation of β-catenin, as assessed using immunofluorescence staining. U87MG cells were transfected with or without the plasmids as indicated. The cells were treated with or without the AKT inhibitor VIII (10 μM). All cells transfected were GFP positive. The nuclei were stained with Hoechst 33342. (D) Notch1 activation elevated the activity of β-catenin. U87MG cells were transfected with or without the LEF/TCF reporter gene, Notch1, DLL4 and NICD plasmids. The cells were then treated with or without the AKT inhibitor VIII (10 μM). At 72 h after transfection, the cell lysates were used to detect luciferase. ** indicates that data were collected from triplicated experiments and are shown as the means ± SD. (E) Notch1 activation increased the nuclear translocation of β-catenin in U251 cells. U251 cells were transfected with the indicated plasmids. At 72 h after transfection, the whole cell lysates or nuclear fractions were used for detecting the expression of the indicated proteins. +, with specific plasmid (or inhibitor) treatment; −, without specific plasmid (or inhibitor) treatment.

Download figure to PowerPoint

Activation of Notch1 increased NF-κB signaling through AKT in glioma cells.  It has been reported that AKT plays an important role in the regulation of NF-κB signaling.(32,33) Because AKT activation was induced by the stimulation of Notch1, we further investigated the changes in NF-κB signaling caused by Notch1 activation in U87MG cells. We found that overexpression of DLL4 might cause the degradation of IκB and a corresponding increase in nuclear p50, especially in cells that overexpress Notch1 (Fig. 3A). In cells transfected with NICD, nuclear p50 was dramatically increased (Fig. 3A,B). Conversely, the AKT inhibitor completely abrogated the effects of NICD on NF-κB signaling (Fig. 3B). To determine how AKT affected the NF-κB signaling pathway, we examined the interaction between AKT and IKKα in glioma cells. Abundance of the AKT/IKKα complex was increased after overexpression of NICD compared with controls (Fig. 3C). Furthermore, increased translocation of p50 into the nucleus was also induced by overexpression of NICD (Fig. 3C). We also tested the effects of Notch activation on NF-κB signaling in U251 cells. Overexpression of NICD led to increased nuclear p50 in this cell line (Fig. 3D). The inhibition of AKT activity completely abrogated the effects of NICD overexpression on the nuclear translocation of p50 (Fig. 3D).

image

Figure 3.  Notch1 activation induced increased NF-κB signaling. (A) Notch1 activation enhanced nuclear p50 levels in U87MG cells. U87MG cells were transfected with the indicated plasmids. At 72 h after transfection, the whole cell lysates or nuclear fractions were used for detecting the expression of the indicated proteins. (B) The AKT inhibitor VIII (10 μM) inhibited the Notch1-induced nuclear translocation of p50. U87MG cells were transfected with or without the Notch intracellular domain (NICD) plasmid. The cells were then treated with or without the AKT inhibitor VIII (10 μM). The nuclear fractions were used for detecting the expression of the indicated proteins. (C) Notch1 activation promoted the interaction of AKT and IKKα. U87MG cells were transfected with the indicated plasmids. The cells were treated with or without the AKT inhibitor VIII (10 μM). Total cell lysates were subjected to immunoprecipitation (IP) and immunoblotting using the indicated antibodies. Rabbit IgG was used as a negative control for IP. (D) Notch1 activation enhanced nuclear p50 levels in U251 cells. U251 cells were transfected with the indicated plasmids. At 72 h after transfection, the whole cell lysates or nuclear fractions were used for detecting the expression of the indicated proteins. +, with specific plasmid (or inhibitor) treatment; −, without specific plasmid (or inhibitor) treatment.

Download figure to PowerPoint

EGFR partially mediated Notch1-induced β-catenin and NF-κB signaling.  It has been shown that Notch1 can upregulate EGFR expression in glioma;(16) therefore, we investigated whether this relationship is AKT dependent. We consistently observed elevated EGFR expression induced by DLL4 overexpression in U87MG cells that was blocked by the knockdown of Notch1 with repression of AKT activation (Fig. 4A). We also investigated whether EGFR could regulate the β-catenin and NF-κB signaling that was stimulated by Notch activation. The DLL4 overexpression-stimulated degradation of IκB and the nuclear translocation of p50 were partially abrogated following treatment with the EGFR inhibitor AG1478 (Fig. 4A). The increase in nuclear β-catenin stimulated by DLL4 was partially suppressed by treatment with AG1478 (Fig. 4A). In contrast, the DLL4-induced increases in nuclear p50 and β-catenin were completely abrogated by the knockdown of Notch1 (Fig. 4A,B). In addition, DLL4-induced β-catenin transcriptional activity was also blocked to some extent by AG1478 and was completely suppressed by Notch1 knockdown (Fig. 4C). Finally, we found that treatment with AG1478 could partially interfere with the binding between AKT and IKKα induced by DLL4 (Fig. 4D), whereas knockdown of Notch1 by siRNA almost completely suppressed formation of the AKT/IKKα complex (Fig. 4D).

image

Figure 4.  Epidermal growth factor receptor (EGFR) partially regulated Notch1-induced β-catenin and NF-κB signaling. (A) Notch1-stimulated nuclear translocation of β-catenin and p50 was partially inhibited by the EGFR inhibitor AG1478. U87MG cells were transfected with the indicated plasmids. The cells were then treated with EGFR inhibitor AG1478 (100 nM). The whole cell lysates or nuclear fractions were used for detecting the expression of the indicated proteins. (B) Efficiency of Notch1 knockdown. U87MG cells were transfected with Notch1 siRNA or control RNA. Total cell lysates were subjected to immunoblotting using the indicated antibodies. (C) The Notch1-stimulated formation of the AKT/IKKα complex was partially inhibited by EGFR inhibitor AG1478. U87MG cells were transfected with the indicated plasmids. The cells were then treated with EGFR inhibitor AG1478 (100 nM). Total cell lysates were subjected to immunoprecipitation (IP) and immunoblotting using the indicated antibodies. Rabbit IgG was used as a negative control for the IP. Data were collected from triplicate experiments and are shown as the means ± SD. * and ** indicates P < 0.05 and P < 0.01 respectively, in comparison with control. (D) Notch1-stimulated β-catenin transcriptional activity was partially inhibited by EGFR inhibitor AG1478. U87MG cells were transfected with or without the LEF/TCF reporter gene, Notch1 and DLL4 plasmids. The cells were then treated with EGFR inhibitor AG1478 (100 nM). Whole cell lysates were used to detect luciferase. +, with specific plasmid (or inhibitor) treatment; −, without specific plasmid (or inhibitor) treatment.

Download figure to PowerPoint

Synergic effects of β-catenin and NF-κB signaling induced by Notch1 activation on the migration and invasion of glioma cells.  The upregulation of Notch1 signaling in tumor specimens was found to be positively correlated with expression of the EMT mediator Snail, as shown in Figure 1(A–C). β-catenin and NF-κB signaling in U87MG cells was enhanced by Notch1 activation (Figs 2A,3A), which has been shown to increase the invasiveness of cancer cells.(27,28,30,34,35) These results implied that Notch1-stimulated β-catenin and NF-κB signaling could potentially affect mediators of migration and invasion, such as Snail and vimentin. As shown in Figure 5(A,B), the transfection of either β-catenin or p50 siRNA partially blocked NICD-induced elevation of Snail-1 and vimentin protein expression in U87MG cells. However, β-catenin or p50 knockdown had no effect on AKT phosphorylation levels (Fig. 5A,B). Figure 5(C,D) shows the effects of β-catenin and p50 siRNA in glioma cells, respectively. Furthermore, real-time PCR and immunofluorescence staining were used to examine the mRNA and protein levels, respectively, of a series of molecules, including Zeb1, vimentin and Snail-1, involved in cell migration and invasion.(20–22,28) The overexpression of NICD led to elevated Zeb1, vimentin and Snail-1 expression (Fig. 5E,F). The NICD-induced upregulation of Zeb1, vimentin and Snail-1 mRNA levels was partially prevented by the knockdown of either β-catenin or p50. Furthermore, this upregulation was completely abrogated by the simultaneous knockdown of β-catenin and p50 or by treatment with the AKT inhibitor VIII. Next we performed assays in glioma cell lines to further confirm the effects of β-catenin and NF-κB signaling on migration and invasion. In the U251 and U87MG cell lines, blocking AKT activation completely abolished the DII4-stimulated migration of both cell lines, while the inhibition of NF-κB and β-catenin partially abrogated the enhanced migration of U87MG cells induced by DII4 (Fig. 6A–C). We also tested the effects of Notch signaling on cell invasion using the U251 and U87MG cell lines. As shown in Figure 6(C,D), DII4 stimulation enhanced the invasiveness of both cell lines. Similar to the findings of the migration assay, the results of the invasion assay indicated that AKT signaling had a prominent effect on DII4-induced cell invasion, and the inhibition of NF-κB and β-catenin activity abrogated this effect (Fig. 6D,E). In conclusion, enhanced Notch1 signaling promoted both the AKT-dependent migratory abilities and invasive capabilities of glioma cells.

image

Figure 5.  Effects of elevated β-catenin and NF-κB signaling induced by Notch1 activation in glioma cells. (A) Notch intracellular domain (NICD)-induced elevation of the vimentin and Snail protein expression levels was partially inhibited by the knockdown of β-catenin. U87MG cells were transfected with the indicated plasmids. Total cell lysates were subjected to immunoblotting using the indicated antibodies. (B) The NICD-induced elevated level of the vimentin and Snail proteins was partially inhibited by the knockdown of p50. U87MG cells were transfected with the indicated plasmids. Total cell lysates were subjected to immunoblotting using the indicated antibodies. (C) Efficiency of p50 knockdown. U87MG cells were transfected with p50 siRNA or control RNA. Total cell lysates were subjected to immunoblotting using the indicated antibodies. (D) Efficiency of β-catenin knockdown. U87MG cells were transfected with p50 siRNA or control RNA. Total cell lysates were subjected to immunoblotting using the indicated antibodies. (E) Inhibition of β-catenin and NF-κB signaling partially blocked the NICD-induced increase in Snail-1 expression, as assessed using immunofluorescence staining. U87MG cells were transfected with or without the indicated plasmids. The cells were treated with or without the AKT inhibitor VIII (10 μM). All cells transfected were GFP positive. The nuclei were stained with Hoechst 33342. (F) The NICD-induced elevated mRNA levels of vimentin, Snail-1, zeb1 and N-cadherin were partially abrogated by inhibition of β-catenin or NF-κB signaling. The data from triplicate experiments are shown as the means ± SD. * and ** indicate < 0.05 and < 0.01, respectively, in comparison with the control group. β-actin mRNA was used as the internal control. +, with specific plasmid (or inhibitor) treatment; −, without specific plasmid (or inhibitor) treatment.

Download figure to PowerPoint

image

Figure 6.  Notch1 signaling promoted glioma cell migration and invasion through the AKT signaling pathway. (A) Trans-well migration assay in U251 cells. U251 cells transfected with a plasmid expressing DLL4 were seeded in trans-well assay chambers. The cells were then treated with or without the p50 inhibitor (NF-κB p50 [NLS] inhibitory peptide set [50 μg/mL]), the β-catenin inhibitor (β-catenin/TCF inhibitor, FH535 [15 μM]) or the AKT inhibitor VIII (10 μM). After 24 h, the migrated cells were detached from the membrane with trypsin, lysed and quantified using a DNA dye. The data from triplicate experiments are shown as the means ± SD. (B) Trans-well migration assay in U87MG cells. U87MG cells transfected with a plasmid expressing DLL4 were seeded in trans-well assay chambers. The cells were then treated with or without the p50 inhibitor (NF-κB p50 [NLS] inhibitory peptide set [50 μg/mL]), the β-catenin inhibitor [β-catenin/TCF inhibitor, FH535 (15 μM)] or the AKT inhibitor VIII (10 μM). After 24 h, the migrated cells were detached from the membrane with trypsin, lysed and quantified using a DNA dye. The data from triplicate experiments are shown as the means ± SD. (C) Scratch wound healing assay for U251 cells. Recombinant human DLL4 (rDLL4) was immobilized on a plate. U251 cells were treated with or without the p50 inhibitor (NF-κB p50 [NLS] inhibitory peptide set [50 μg/mL]), the β-catenin inhibitor (β-catenin/TCF inhibitor, FH535 [15 μM]) or the AKT inhibitor VIII (10 μM). The graph shows the results for the relative closure of the scratch wound compared with the control. The images (original magnification, ×100) were taken at 0 and 36 h after the wound was made. The dotted lines indicate the width of the wound at these two time points. (D) Trans-well invasion assay in U251 cells. U251 cells stably transfected with the plasmid expressing DLL4 were seeded in trans-well assay chambers. The cells were then treated with or without the p50 inhibitor (NF-κB p50 [NLS] inhibitory peptide set [50 μg/mL]), the β-catenin inhibitor (β-catenin/TCF inhibitor, FH535 [15 μM]) or the AKT inhibitor VIII (10 μM). After 24 h, the migrated cells were detached from the membrane with trypsin, lysed and quantified using a DNA dye. The data from triplicate experiments are shown as the means ± SD. (E) Trans-well invasion assay in U87MG cells. U87MG cells stably transfected with the plasmid expressing DLL4 were seeded in trans-well assay chambers. The cells were then treated with or without the p50 inhibitor (NF-κB p50 [NLS] inhibitory peptide set [50 μg/mL]), the β-catenin inhibitor (β-catenin/TCF inhibitor, FH535 [15 μM]) or the AKT inhibitor VIII (10 μM). After 24 h, the migrated cells were detached from the plate with trypsin, lysed and quantified using a DNA dye. The data from triplicate experiments are shown as the means ± SD. (F) Model of Notch1-induced glioma cell migration and invasion. In glioma cells, the stimulation of Notch1 led to enhanced NF-κB and β-catenin signaling through AKT activation. Increased NF-κB and β-catenin activity resulted in an upregulation of Snail, Zeb1 and vimentin, thereby promoting cell migration and invasion in glioma. +, with specific plasmid (or inhibitor) treatment; −, without specific plasmid (or inhibitor) treatment; RFU, relative fluorescence unit.

Download figure to PowerPoint

Discussion

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

Glioblastoma multiforme, which is a grade IV glioma, is the most malignant tumor of the central nervous system.(1,3) Previous studies have reported that Notch1 signaling is dysregulated in glioma and is related to glioma progression.(10–13) However, the role of the Notch1 signaling pathway in glioma migration and invasion as well as the mechanisms associated with these processes are largely unknown. In the present study, we found that Notch1 signaling facilitated glioma cell migration and invasion by enhancing β-catenin and NF-κB signaling via AKT activation (Fig. 6D).

The Notch signaling pathway affects AKT signaling in different contexts, including glioma,(10,16) and this was validated by our data from glioma tissues. Malignant tumor cells are capable of migration and invasion, and increased expression of Snail is considered a promoter of tumor invasiveness.(19,22) An increasing number of studies has shown that hyperactivation of AKT is involved in the migration and invasion of tumor cells.(17,18) The relationship between AKT activation and Snail expression in glioma specimens implied the existence of such mechanisms. Furthermore, enhancement of β-catenin activity by AKT activation accelerated the migration and invasiveness of tumor cells.(25–27) NF-κB signaling, implicated in tumor cell migration and invasion, was also enhanced by AKT activation.(32–35) As shown in Figure 1(D), the levels of nuclear p50 and β-catenin in glioma tissues were elevated compared with normal tissues, and the interaction between IKKα and AKT was enhanced in glioma tissues, thereby suggesting that the activation of NF-κB signaling might be regulated by AKT.

We found that Notch signaling enhanced β-catenin activity and NF-κB signaling via the activation of AKT in glioma cells (Figs 2,3). Blockade of NICD-induced nuclear translocation of β-catenin and p50 following treatment with the AKT inhibitor VIII suggested that AKT was the critical mediator in this process. Furthermore, the overexpression of NICD promoted the interaction of IKK and AKT, and this effect was completely inhibited by treatment with the inhibitor. Consistent with previous reports,(32,33) repression of the IKK/AKT interaction by AKT inhibition indicated that AKT affects NF-κB signaling by directly regulating IKKα in glioma cells.

Previous studies have reported that upregulated EGFR expression contributes to AKT activation as a result of Notch signaling in glioma.(16) In U87MG cells, elevated EGFR expression was detected following the activation of Notch1 (Fig. 4A). Treatment with an EGFR inhibitor partially abrogated Notch1-induced AKT activation as well as β-catenin and NF-κB signaling. These observations suggest that increased EGFR expression or activation might have only been one of many aspects involved in the Notch1-induced AKT activation. However, enhanced EGFR activity in glioma can be caused by various factors.(36,37) Many previous studies have shown that increased Notch signaling could stimulate AKT activation through the downregulation of PTEN levels,(38,39) and PTEN is referred to as the most critical suppressor of PI3K. Therefore, we hypothesized that Notch1 activation might lead to AKT activation through multiple signaling pathways in glioma, including p53/EGFR and HES1. The potential effects of various other signaling pathways on AKT activity in glioma are worthy of further investigation.

The results from the analyzed glioma tissues implied that there might be a connection between Notch1 and Snail. Overexpression of NICD in U87MG cells led to elevated levels of vimentin, Snail and zeb1 expression, thereby suggesting that increased Notch signaling results in glioma cells that are more capable of migration and invasion. Furthermore, the knockdown of β-catenin and p50 indicated that β-catenin and NF-κB signaling independently mediated the migration and invasion of Notch1-induced glioma cells. It was also noted that AKT signaling seemed to play a prominent role in glioma cell migration and invasion that was induced by Notch1 activation, as these effects were completely blocked by inhibiting AKT activity. Furthermore, because neither the knockdown of β-catenin nor that of p50 had any effect on NICD-induced AKT activation, we concluded that β-catenin and NF-κB signaling were downstream effectors of AKT for glioma cell migration and invasion.

In summary, dramatic elevation of Notch1 signaling was related to enhanced AKT phosphorylation, and the increased signaling events leading to tumor metastasis in glioblastoma multiforme might indicate synergistic effects among these signaling pathways in glioma tumorigenesis (Fig. 6F). Cell migration and invasion in glioma cells were promoted by Notch1 activation through the β-catenin and NF-κB signaling pathways via AKT activation (Fig. 6F). The findings of the present study might help guide the development of potential therapeutic targets for the treatment of glioma patients.

Acknowledgments

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

This work was supported by grants from the Shanghai Science and Technology Development Fund (10JC1409802, 09ZR1417800), the Scientific Research and Innovation of Shanghai Municipal Education Commission (11YZ50), Key Discipline Project of Renji Hospital, School of Medicine, Shanghai Jiaotong University (RJ4101307) and the Wu Jieping Medical Foundation (320.6750.11092).

References

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