Contribution of Notch signaling activation to human glioblastoma ...

J Neurosurg 106:417–427, 2007

Contribution of Notch signaling activation to human glioblastoma multiforme MASAYUKI KANAMORI, M.D.,1,2 TOMOHIRO KAWAGUCHI, M.D.,1,2 JANICE M. NIGRO, PH.D.,2 BURT G. FEUERSTEIN, M.D., PH.D.,2 MITCHEL S. BERGER, M.D.,1,2 LUCIO MIELE, PH.D.,3 AND RUSSELL O. PIEPER, PH.D.1,2 Department of Neurological Surgery and 2Brain Tumor Research Center, UCSF Cancer Center, University of California, San Francisco, California; and 3Department of Biopharmaceutical Sciences and Cancer Center, University of Illinois at Chicago, Illinois 1

Object. Because activation of Notch receptors has been suggested to be critical for Ras-mediated transformation, and because many gliomas exhibit deregulated Ras signaling, the authors measured Notch levels and activation in primary samples and cell lines derived from glioblastoma multiforme (GBM) as well as the contribution of Notch pathway activation to astrocytic transformation and growth. Methods. Western blot analysis of Notch 1 expression and activation showed that Notch 1 protein was overexpressed and/or activated in Ras-transformed astrocytes, in three of four GBM cell lines, and in four of five primary GBM samples. Expansion of these studies to assess mRNA expression of components of the Notch signaling pathway by cDNA expression array showed that cDNAs encoding components of the Notch signaling pathway, including the Notch ligand Jagged-1, Notch 3, and the downstream targets of Notch (HES1 and HES2), were also overexpressed relative to nonneoplastic brain controls in 23, 71, and 51% of 35 primary GBMs, respectively. Furthermore, inhibition of Notch signaling by genetic or pharmacological means led to selective suppression of the growth and expression of markers of differentiation in cells exhibiting Notch pathway deregulation. Conclusions. Notch activation contributes to Ras-induced transformation of glial cells and to glioma growth, survival, or both and as such may represent a new target for GBM therapy.

KEY WORDS • Notch pathway • Notch 1 • glioblastoma multiforme • differentiation • glial cell

H

UMAN Notch genes encode heterodimeric transmem-

brane receptors that play a key role in a variety of cellular processes, including proliferation, differentiation, survival, and apoptosis.1 The four human Notch genes encode single precursor peptides that are each cleaved into an extracellular peptide and a transmembrane peptide.15 The extracellular peptide subunit contains epidermal growth factor receptor–like repeats, noncovalently associates with the transmembrane peptide, and binds cell membrane–associated ligands of the Delta or Jagged families.25 Ligand binding sensitizes the heterodimeric receptor to cleavage events mediated by members of the ADAM and gsecretase families of proteases. These cleavage events release the intracellular domain (NICD) of the transmembrane peptide from the membrane-bound receptor. The NICD then translocates to the nucleus and forms a complex Abbreviations used in this paper: BrdU = bromodeoxyuridine; EGFP = enhanced green fluorescent protein; FACS = fluorescenceactivated cell sorting; GBM = glioblastoma multiforme; GFAP = glial fibrillary acidic protein; Ig = immunoglobulin; MAML = Mastermind-like; MTS = 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; NICD = Notch intracellular domain; PBS = phosphate-buffered saline; SD = standard deviation; siRNA = small interfering RNA.

J. Neurosurg. / Volume 106 / March, 2007

(or complexes) with the transcription factor CBF-1 and with transcriptional coactivators of the MAML family, which in turn stimulate the transcription of multiple target genes.2,33,35 The nuclear translocation of NICD has been shown to be sufficient for the transforming activity of Notch and to be necessary for the transforming ability of mutant H-RasV12 in fibroblasts and kidney epithelial cells.34 These and other studies suggest that deregulated activation of Notch can contribute to the development of various malignancies. In addition to playing a role in tumorigenesis, Notch signaling is also important to normal development.23 Delay of terminal differentiation by Notch signaling has been reported in some cases, although in other cases Notch seems to be necessary for differentiation.7,16,19,24,26,36 In the central nervous system, Notch 1 activation is necessary to maintain neural progenitor identity and to prevent terminal differentiation.10,14,27 On the other hand, activation of Notch 1 and Notch 3 in progenitor cells leads to glial differentiation.31 This two-way potential of Notch signaling is dependent on cell type, phase of development, and interaction with other factors. The complicated functions of Notch have made it difficult to fully understand how Notch signaling might contribute to tumor development. Although a number of tumor types have been shown to exhibit deregulated Notch signal417

M. Kanamori et al. ing (as reviewed elsewhere13,18,20,30,37), the role of Notch signaling in human brain tumors remains controversial.4,22 The potential for Notch to play a role in human glioma development is of particular interest because Ras pathway activation cooperates with Akt signaling to transform rodent and human astrocytes12,28,29 and because Ras activation in the absence of mutation has been reported to be proportional to glioma grade.5 Given the importance of the Ras pathway in glioma development, the necessity of Notch signaling for Ras-induced transformation, and the generalized role Notch is thought to play in cellular growth and differentiation, we examined Notch signaling in primary samples and cell lines derived from GBM (the most advanced form of human glioma) as well as the contribution of Notch pathway activation to astrocytic transformation and growth. We report that Notch signaling is deregulated in a majority of the GBM cell lines and primary GBM samples we examined, and that inhibition of Notch signaling by genetic or pharmacological means leads to cell cycle exit and suppression of the growth of cells exhibiting Notch pathway deregulation. The results of these studies support the idea that the Notch signaling pathway may represent a new target in the treatment of human GBM. Materials and Methods

(MAML1), 37393_at (HES1), and 36352_at (HES2). Probe intensity values were derived from captured array images using Microarray Suite, version 5.0 (Affymetrix). For data analysis, fold induction was determined relative to the average of five nonneoplastic samples, or to 100 if this average was less than 100 units. Only samples exhibiting a greater than twofold signal induction were scored as positive for target overexpression. Creation of Cells Expressing Antisense RNA or siRNA Targeted to NOTCH1 Human antisense NOTCH1 (base pairs 5225–7270) was cloned into a pMXI-EGFP retroviral vector. Then NHA E6/E7/hTERT/HRasV12, U251, and U87 cells were retrovirally infected as previously described29 with a blank pMXI-EGFP construct or the construct encoding antisense NOTCH1. Three days after infection, EGFP-positive cells were sorted on a FACS Vantage (Becton Dickinson), collected, and expanded for further study. For siRNA studies, cells were transfected with siRNA targeting NOTCH1 (SMARTPool, Dharmacon) or a negative control siRNA (cyclophilin B, Dharmacon) using Oligofectamine (Invitrogen) according to the manufacturer’s protocol. Briefly, cells were plated 48 hours prior to siRNA transfection and were incubated with Oligofectamine transfection reagent with or without siRNA in a serum-free medium. Following 4 hours of exposure to the siRNA, the medium was replaced with fresh growing media. The cells were cultured for another 48 hours and analyzed for individual assay. Transfection efficiency was more than 90%, determined by using Cy3-labeled nontargeting siRNA before using siRNA targeting Notch 1. The final concentration of siRNA was 120 nM.

Cell Lines and Reagents

Colony Formation Efficiency, Soft Agar Growth, and Cell Viability and Growth Assays

Human astrocytes containing constructs encoding HPV16 E6, HPV16 E7, hTERT, and/or V12 H-Ras were created as previously described.29 All GBM cell lines and primary GBM samples used were obtained from the University of California–San Francisco Brain Tumor Research Center tissue bank. Normal brain samples were derived from temporal lobe tissue removed during epilepsy surgery. The g-secretase inhibitor (LLNle-CHO) was obtained from Calbiochem. Jagged-1–Fc chimera protein and anti–human IgG Fc– specific antibody were obtained from R&D Systems and Sigma, respectively.

For colony formation assays, preplated cells (1,000 cells per sixwell plate) were incubated with the g-secretase inhibitor LLNleCHO (0–4 mM, 24 hours), washed, and allowed to form colonies for 10 days. Cells were then stained with methylene blue (Sigma), and colonies containing more than 50 cells were counted. Soft agar colony formation assay was carried out as previously described.28 Cell viability and/or growth was determined using the inner salt MTS assay 96 hours after siRNA transfection according to the instructions of the manufacturer (Promega). All experiments were repeated three times. Values are expressed as means 6 SDs.

Western Blot Analysis

Analysis of Cell Cycle Distribution by Flow Cytometry

Cell lysate (50 mg) from cultured cells or primary tumors was prepared as previously described,29 subjected to 7.5 or 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and electroblotted onto an Immobilon-P membrane (Millipore). The membrane was blocked in 5% nonfat milk and incubated with antibodies against C-terminal Notch 1 (Santa Cruz Biotechnology), cleaved (activated) Notch 1 (NICD) (Cell Signaling Technology), GFAP (Promega), vimentin (Santa Cruz Biotechnology) or a-tubulin (Santa Cruz Biotechnology) overnight at 4˚C. Bound antibody was detected with horseradish peroxidase–conjugated secondary IgG or IgM (Santa Cruz Biotechnology) using enhanced chemoluminescence Western blotting detection reagents (Amersham).

Cells were trypsinized and collected 96 hours after siRNA transfection. These cells were washed in PBS and fixed in 70% ethanol at 220˚C. The cells were then washed again with PBS and incubated in PBS containing 40 mg/ml propidium iodide (Sigma) and 200 mg/ml RNase A (Sigma) for 1 hour at room temperature in the dark. Stained nuclei were then analyzed on FACScan machine (Becton Dickinson) with 10,000 events per determination as described previously.9 ModFit LT software (Verity Software House) was used to assess cell cycle distribution. Bromodeoxyuridine incorporation was performed and analyzed at the same time point. Bromodeoxyuridine (10 mM) was added to the medium, and the cells were incubated for 1 hour at room temperature before trypsinization. The cells were then fixed and stained according to the manufacture’s instruction (BD Biosciences Pharmingen).

Expression Array Analysis Total RNA was extracted from primary GBM using Trizol reagent (Invitrogen). Briefly, total RNA was prepared from tumor samples that had been immediately stored in liquid nitrogen following surgical resection. We ground 100 mg of tissue into a fine powder with a mortar and pestle on dry ice and isolated the RNA using Trizol reagent according to the manufacturer’s protocol. We passed 50 mg of total RNA from each tissue sample over an RNeasy column (Qiagen) for further purification. Preparation of cRNA, hybridization, and scanning of the oligonucleotide microarray (U95Av2, human GeneChip, Affymetrix) was performed as described elsewhere.21 The probe set included 33178_at (JAG1), 32173_at (JAG2), 38083_at (NOTCH2), 38750_at (NOTCH3), 39048_at (NOTCH4), 1226_at (ADAM17), 36269_at (ADAMTS3), 40797_at (ADAM10), 37292_at

418

Immunohistochemical Analysis Cells were grown on glass coverslips for 72 hours after each treatment and then fixed in 4% paraformaldehyde for 15 minutes at room temperature. After permeabilization with acetone (20 min, 220˚C), the cells were blocked with 1% bovine serum albumin and 2% goat serum, then incubated with primary antibody (rabbit anti-GFAP antibody, 1:1000 dilution overnight at 4˚C). After incubation with a biotinylated secondary antibody (goat anti–rabbit, 1:200 dilution; Santa Cruz Biotechnology) for 60 minutes at room temperature, antigens were revealed with streptavidin-conjugated horseradish peroxidase and diaminobenzamide (Vector Laboratories). The nuclei were counterstained with methyl green.


Notch signaling in glioblastoma multiforme Jagged-1–Fc Treatment Jagged-1–Fc (10 mg) was incubated for 1 hour on ice with anti– human Fc antibody (5 mg) to yield a 2:1 stoichiometric ratio and a final concentration of Jagged-1–Fc of 100 mg/ml. The Jagged-1–Fc/ anti-Fc complex or the anti-Fc antibody alone (negative control) was incubated with 5 3 105 cells at room temperature for 1 hour, after which cells were cultured as usual for further analysis. Statistical Analysis For array data, analysis of association was performed using the Fisher exact test. For all other data, statistical significance was assessed using the Student t-test.

Results Overexpression and Activation of Notch 1 in GBM Cell Lines and Primary GBM Samples

To begin to address a possible role for Notch deregulation in the development of astrocytic tumors, we first examined Notch 1 expression in cell lysates derived from non–Rastransformed astrocytes, Ras-transformed astrocytes, and GBM cell lines. As shown in the Western blot in Fig. 1A, human astrocytes transformed by coexpression of E6, E7, hTERT, and mutant V12 H-Ras (Lane 2) had a level of Notch 1 expression that was four- to fivefold greater than

FIG. 1. Overexpression/activation of Notch 1 in Ras-transformed astrocytes, human GBM cell lines, and primary human GBM. A: Western blot analysis with antibody to Notch 1 C-terminal portion and a-tubulin in NHA E6, E7, and hTERT (Lane 1), NHA E6, E7, hTERT, and H-RasV12 (Lane 2), U373 (Lane 3), A172 (Lane 4), U87 (Lane 5), and U251 (Lane 6) cells. B: Western blot analysis with antibody to Notch 1 Cterminal portion, cleaved/activated Notch 1 (NICD), and a-tubulin in normal brain tissue (Lane 1) and human primary GBMs (Lanes 2–6). Values derived from three independent experiments are normalized to a-tubulin controls and are expressed as the mean fold expression relative to NHA E6, E7, and hTERT (A) or normal brain (on overexposure, B). Asterisks designate values differing significantly (p , 0.05, Student t-test) from the controls. C: Heat map demonstrating activation of the Notch signaling pathway in primary human GBM. Values from probe sets covering 11 members of the Notch signaling pathway in an expression array analysis of 35 primary GBM and five nonneoplastic human brain samples are shown. Values for genes are in rows and case numbers are in columns. The intensity of whiteness is proportional to the fold induction of expression of the target relative to that observed in controls. Average probe intensity values of less than 100 were assigned a value of 100. For this representation, an induction of less than twofold was not assigned a color, and the maximum intensity represented a 7.3-fold induction. D: Bar graphs showing the absolute signal intensity values from the same array. The black bar on the bottom of each graph identifies normal brain sample.


419

M. Kanamori et al.

FIG. 2. Graph showing selective inhibition effect of g-secretase on the colony-forming ability of Notch 1–overexpressing (NHA E6, E7, hTERT, and H-RasV12, U251) compared with non–Notch 1– overexpressing (NHA E6, E7, and hTERT, U87) cells. Cells were exposed to 0 to 4 mM of g-secretase inhibitor for 24 hours, washed with PBS, then incubated for 10 days in a drug-free culture medium, after which colonies of more than 50 cells were counted. Colony formation efficiency is expressed as a percentage of the number of colonies formed after exposure to g-secretase inhibitor relative to the number of colonies formed after exposure to vehicle. Values are the means 6 SDs of three experiments.

that noted in astrocytes immortalized by coexpression of E6, E7, and hTERT in the absence of mutant H-Ras (Lane 1). An expansion of the analysis to human GBM cell lines revealed that three of four GBM lines (U373, A172, U251; Lanes 3, 4, and 6, respectively; Fig. 1A) also had elevated levels of Notch 1 expression while a 4th GBM cell line (U87, Lane 5) had Notch 1 levels not significantly different from those in non–Ras-transformed astrocytes. Further analysis of Notch 1 expression and of activated Notch 1 (NICD) levels in cell lysates derived from primary GBM (Fig. 1B) showed that two of five tumors exhibited elevated levels of Notch 1 and four of five primary GBMs, including both tumors with elevated Notch 1 levels, exhibited a greater than twofold increase in Notch 1 activation relative to normal brain tissue. Activation of Notch Signaling Pathways in Primary GBM

To further expand our analysis of Notch pathway activation in a larger set of GBMs, we used expression arrays to measure mRNA levels of Notch 2, Notch 3, and Notch 4 (Notch 1 could not be analyzed on this array), the Notch ligands Jagged 1 (JAG1) and Jagged 2 (JAG2), proteases that cleave Notch receptors (ADAM17, ADAMTS3, ADAM10), 420


FIG. 3. Notch 1 downregulation by siRNA blocks cell growth and induces a differentiation-like response in Notch 1–overexpressing (U251 and NHA E6, E7, hTERT, and H-RasV12) but not non–Notch 1–overexpressing (U87 and NHA E6, E7, and hTERT) cells. A–D right: Western blot analysis of cleaved Notch 1 (NICD), GFAP, vimentin, and a-tubulin expression in four cell lines 96 hours after exposure to vehicle, negative control siRNA, or Notch 1–specific siRNA (overexposure in B and D right). A–D left: Bar graphs showing cell growth of the same cell lines 96 hours after siRNA transfection as assessed by the MTS assay. E–H: Results of a FACS analysis of BrdU incorporation (left) and cell cycle distribution (right) 1 hour after exposure to BrdU (0 or 10 mM). In the flow cytometric data presented in the right panels, the y-axis values refer to anti-BrdU staining (reflecting BrdU incorporation), while the x-axis values reflect propidium iodide uptake. I: Bright-field photomicrographs showing untreated U251, U87, NHA E6, E7, hTERT, and H-RasV12, and NHA E6, E7, and hTERT cells and the same cells infected with negative control or siRNA targeting Notch 1. Arrowheads demonstrate the differentiated phenotypes with long cellular processes and elongated stellate shapes. Original magnification 3 400. All values expressed are the means 6 SDs of three experiments. Asterisks designate values that differ significantly (p , 0.05) from the controls.

Notch signaling in glioblastoma multiforme


421

M. Kanamori et al.

FIG. 4. The inhibition of the Notch pathway blocks cell growth and induces a differentiation-like response in Notch 1– overexpressing (U251 and NHA E6, E7, hTERT, and H-RasV12) but not non–Notch 1–overexpressing (U87 and NHA E6, E7, and hTERT) cells. A: Western blot analysis of cleaved Notch 1 (NICD), GFAP, vimentin, and a-tubulin in four cell lines 72 hours after exposure to control vehicle or soluble Jagged-1–Fc antibody complex. B: Graphs showing cell growth of the same cell lines 72 hours after treatment with soluble Jagged-1–Fc antibody complex as assessed by MTS assay.

422


Notch signaling in glioblastoma multiforme

C: Bright-field photomicrographs of control cells (left) of U251, U87, NHA E6, E7, hTERT, and H-RasV12, and NHA E6, E7, and hTERT and of the same cells treated with negative control (center) or soluble Jagged-1–Fc antibody complex (right). Arrowheads demonstrate the differentiated phenotypes with long cellular processes and elongated stellate shapes. Values expressed are means 6 SDs of three experiments. Asterisks designate values that differ significantly from the controls. Ab = antibody.


423

M. Kanamori et al. a Notch coactivator (MAML1), and the Notch downstream targets HES1 and HES2 in a set of 35 primary GBMs. Expression was then compared with that found in five control samples of nonneoplastic brain tissue. Of the probe sets and genes analyzed, only JAG1 and JAG2 were consistently expressed in the controls. Expression of JAG1, however, was elevated more than twofold in eight (23%) of the 35 GBMs (Fig. 1C). The Notch 3 mRNA levels were also elevated more than twofold above the background levels in 25 (71%) of 35 of the GBMs analyzed. Furthermore, at least one of the transcriptional targets of Notch (HES1 or HES2) was more than twofold overexpressed relative to controls in 18 (51%) of 35 GBMs analyzed. In 40% of the GBMs, HES1 was overexpressed, and 13 of 14 HES1-overexpressing tumors also overexpressed Notch 3 (p , 0.05), suggesting a statistically significant association between these variables. These results show that the Notch pathway is deregulated in a significant percentage of primary human GBMs. Consequences of Notch Signaling Inhibition

To assess the consequences of Notch pathway deregulation and the possibility that Notch receptor pathways might serve as a therapeutic target in gliomas, the consequences of pharmacological or genetic inhibition of Notch signaling were examined. Incubation with LLNle-CHO, inhibitor of g-secretase,17 the enzyme that cleaves and activates Notch receptors, had a significantly greater suppressive effect on the colony-forming ability of cells overexpressing Notch 1 (U251 GBM cells, IC50 1.3 6 0.3 mM, and Ras-transformed human astrocytes, IC50 1.1 6 0.3 mM) than on cells that did not overexpress Notch 1 (non–Ras-transformed astrocytes, IC50 2.4 6 0.3 mM, and U87 GBM cells, IC50 2.0 6 0.15 mM) (p , 0.05) (Fig. 2). Because LLNle-CHO is a relatively nonspecific and indirect inhibitor of Notch 1 signaling,34 however, we also inhibited Notch signaling genetically using Notch 1–targeted antisense RNA and siRNA and then monitored the effects on activated cleaved Notch 1 levels, clonogenicity, and cell growth. Transfection of U251 cells, U87 cells, Ras-transformed astrocytes or non–Rastransformed astrocytes with a Notch 1 siRNA resulted in the reduction of NICD expression relative to that noted in cells transfected with a negative control siRNA or transfection reagent 2 days after siRNA removal (Fig. 3A–D, upper right). The siRNA-mediated suppression of NICD in turn suppressed the transformed phenotype because exposure to Notch 1 siRNA also significantly decreased the number of viable U251 cells and Ras-transformed astrocytes present 4 days after siRNA introduction as measured by MTS assay (Fig. 3A left and C left). This growth-suppressive effect, however, was limited to cells that overexpressed Notch 1 because the number of viable U87 cells and non–Ras-transformed astrocytes that expressed minimal levels of activated Notch 1 was unaffected by Notch 1 siRNA (Fig. 3B left and D left). Similar results were noted in studies in which U251, U87, Ras-transformed astrocytes, and non–Rastransformed astrocytes were infected with a Notch 1 antisense-encoding retroviral construct. Expression of Notch 1 antisense significantly suppressed levels of NICD in all four cell lines, although antisense-mediated growth suppression was observed in only Notch 1 overexpressing U251 and Ras-transformed astrocytes (data not shown). 424

The Notch 1 siRNA-mediated growth suppression in U251 and Ras-transformed astrocytes was not associated with increased apoptosis, because FACS analysis performed in all four cell lines showed that Notch 1 siRNA exposure did not result in any increase in the percentage of cells exhibiting sub-G1 DNA content (data not shown). Notch 1 siRNA exposure also did not alter the percentage of 2n (G1) or 4n (G2) DNA content (data not shown), suggesting that siRNA-mediated Notch 1 inhibition of U251 cell growth was not associated with activation of the G1 and G2 cell cycle checkpoints. As shown in Fig. 3E through H, siRNA exposure did, however, result in the reduction of BrdUincorporated cells in U251 cells and Ras-transformed astrocytes, but not in U87 cells and non–Ras-transformed astrocytes. These results indicate that siRNA treatment induced general cell cycle slowing or cell cycle exit. Notch 1 siRNA exposure also resulted in selective changes in cell morphology, with U251 cells and Ras-transformed astrocytes exhibiting a differentiated phenotype (long cellular processes, elongated stellate shape, indicated by arrowheads in Fig. 3I) relative to the flattened, spread phenotype noted in cells treated with negative control siRNA and untreated control cells, while the spindlelike morphology of U87 cells and non–Ras-transformed astrocytes that do not express high levels of Notch 1 was largely unaffected by transfection of Notch 1 siRNA. Consistent with this finding, exposure to siRNA targeting Notch 1 significantly increased expression of GFAP and significantly decreased expression of vimentin in U251 cells and Ras-transformed astrocytes, but not in U87 cells and non–Ras-transformed astrocytes (Fig. 3A–D, right). Increased expression of GFAP and decreased expression of vimentin are hallmarks of differentiation in glial cells,11 suggesting that suppression of Notch 1 leads to selective cellular differentiation in those GBM cells that overexpress Notch 1. To verify results observed with siRNA, we also inhibited Notch 1 activation using soluble Jagged-1–Fc protein. Jagged-1–Fc, a fusion protein of the extracellular portion of human Jagged-1 and the Fc region human IgG, is known to block Notch 1 activation.3,8 This protein significantly suppressed levels of NICD in U251, U87, Ras-transformed astrocytes and non–Ras-transformed astrocytes after 48 hours of incubation (Fig. 4A, top panel). As shown in Fig. 4B, exposure to Jagged-1–Fc significantly decreased the number of viable U251 cells and Ras-transformed astrocytes, but not U87 cells or non–Ras-transformed astrocytes, as measured by MTS assay. Furthermore, exposure to Jagged1–Fc altered morphology in U251 and Ras-transformed astrocytes (long cellular processes, elongated stellate shape, indicated by arrowheads in Fig. 4C), but not in U87 or non–Ras-transformed astrocytes. In addition to the resulting NICD suppression, exposure to Jagged-1–Fc also resulted in GFAP induction and reduction of vimentin levels in U251 cells and Ras-transformed astrocytes, but not in U87 cells or non–Ras-transformed astrocytes (Fig. 4A). These results show that suppression of Notch 1 activation by either of these two methods led to selective cellular differentiation of GBM cells that overexpress Notch 1. Taken together, these results show that Notch pathways are deregulated in a significant percentage of human GBMs, and that activation of Notch 1 can play a key role in maintaining the growth and undifferentiated state of glioma cells. J. Neurosurg. / Volume 106 / March, 2007

Notch signaling in glioblastoma multiforme Discussion The role that Notch signaling plays in cellular transformation and cancer is complex and in some cases controversial. Notch 1 signaling has been shown to be important in Ras transformation of fibroblasts and kidney epithelial cells,34 and a component of the Notch signaling pathway, MAML, has been shown to be activated by translocation in a subset of salivary gland carcinomas.32 Conversely, the Notch pathway suppresses skin carcinogenesis, suggesting that in some settings Notch can act as a tumor suppressor.20 In light of the involvement of Notch signaling in Ras transformation, the known activation of the Ras pathway in gliomas, and potential involvement of Notch in cellular differentiation, proliferation, and survival, we considered the possible role that Notch signaling may play in the development and growth of gliomas. The results of the present study show that the Notch signaling pathway is deregulated at multiple points in nearly three fourths of human GBMs, that inhibition of one member of the Notch receptor family can selectively suppress the growth of GBMs and Rastransformed astroctyes, and that Notch signaling may represent a reasonable therapeutic target for the treatment of GBM. The data presented in this study clearly show that the Notch pathway is deregulated in a significant proportion of GBM cell lines and primary GBMs. Notch 1 itself was overexpressed in the majority of GBM cell lines and primary GBMs examined, and although the sample size was relatively small, elevated levels of cleaved, activated NICD were observed in 80% of primary GBMs assayed relative to nonneoplastic brain tissue, in which NICD activation was minimal. In addition, Notch 3 was overexpressed at the mRNA level in 71% of a larger set of primary GBMs. Because the two sets of primary GBMs analyzed in this study did not overlap and because Notch 1 was not present on the array used for cDNA expression analysis, it was not possible to assess the relationship between Notch 1 protein overexpression or activation and Notch 3 mRNA overexpression. Because the majority of the GBMs we examined, however, exhibited activated Notch 1 and overexpressed Notch 3 at the mRNA level, it seems reasonable to think that a significant percentage of GBMs have activated both receptors. A recent study also showed that in GBM, relative to nonneoplastic brain tissue, both Notch 1 mRNA and Jagged-1 mRNA were upregulated.22 Although Notch 1 was not represented in our cDNA array data, Jagged-1 mRNA overexpression was observed, and a correlation between Notch 1 expression and expression of its ligand, Jagged-1, is possible in this data set. On the other hand, no correlation existed between mRNA expression of the Notch ligand Jagged-1 and Notch 3 overexpression (overexpression of Jagged-1 was found in only five of 22 Notch 3 overexpressing GBMs). This lack of correlation might, however, be expected, because there are multiple Notch ligands, at least some of which may not be regulated at the transcriptional level. Nevertheless, there was a statistically significant association between overexpression of Notch 3 and overexpression of the Notch target gene HES1 (13 of 14 HES1-overexpressing GBMs also overexpressed Notch 3), suggesting that not only Notch itself, but the entire Notch signaling pathway is activated in the majority of human GBMs. As the number of potential associations examined J. Neurosurg. / Volume 106 / March, 2007

increases, however, the likelihood of finding associations by chance also increases, necessitating more direct studies as were performed with genetic and pharmacological inhibition of the Notch pathway. The genetic and pharmacological inhibition data presented in this paper support the idea that the Notch pathway may be a novel therapeutic target in GBM. Exposure of cells to a relatively nonspecific g-secretase inhibitor or to two different suppressors of Notch 1 activation (Notch 1– targeted inhibitory RNAs or soluble Jagged-1–Fc protein) inhibited the growth of cells with elevated Notch 1 expression and/or activation relative to cells that did not exhibit Notch 1 activation. While these studies suggest that Notch 1 inhibition alone can suppress the growth of Ras-transformed astrocytes and GBM cells in anchorage-dependent and anchorage-independent settings, the situation is likely to be more complicated in the in vivo setting in which highgrade tumors consist of a heterogeneous mix of differentiated and undifferentiated cells. Furthermore, because the role of Notch signaling in cell growth and differentiation often depends upon context, the present results need to be extended through in vivo studies. The present results do, however, suggest that Notch signaling can be inhibited at different levels with growth-suppressive consequences for glioma cells and that the Notch signaling pathway may therefore serve as a potential target for glioma therapy. The results presented here do differ somewhat from the results of a recent study by Purow and colleagues,22 in which the effects of Notch suppression on glioma growth were similarly examined. In that study, siRNA targeting Notch 1 suppressed the growth of cells that overexpressed Notch 1 and cells that did not overexpress Notch 1. Although there are several differences in methods between that study and the present one (for example, transfection frequency and transfection reagent), none of the four approaches used to suppress Notch 1 activation suppressed the growth of cells that did not overexpress Notch 1, including U87 cells. The study by Purow and colleagues also showed that Notch 1 knockdown by siRNA induced apoptosis in Notch 1–overexpressing cells, whereas in the present study, suppression of Notch 1 by any of three different techniques led to differentiation but not apoptosis. Although these differences may be a consequence of the degree of Notch 1 suppression (which was perhaps larger in the study by Purow et al.), the results of the two studies are in agreement that Notch 1 suppression has deleterious effects on glioma cells. Although the results of the present study demonstrate Notch signaling deregulation in GBM and identify Notch 1 as a therapeutic target in GBM, the exact means by which Notch signaling contributes to glioma growth and/or gliomagenesis remains unclear. The ability of siRNA targeting Notch 1 to suppress the growth of Ras-transformed astrocytes suggests that, as in fibroblasts and kidney epithelial cells, Notch 1 plays a key role in the Ras-transformation process. In related studies, however, overexpression of NICD in immortalized astrocytes did not lead to transformation, nor did overexpression of NICD in Ras-transformed astrocytes lead to enhanced cell growth in soft agar (data not shown). Additionally, although the Ras transformation in astrocytes clearly induced Notch 1 expression, the relationship between Ras pathway and Notch 1 activation levels does not appear to be absolute. In fact, both the U251 and U87 cell lines, which differ in Notch activation, are 425

M. Kanamori et al. both reported to have high levels of Ras–guanosine triphosphate.6 The present data therefore suggest that although Notch 1 itself cannot cause cellular transformation in astrocytes, it can, at least under the conditions used in this study, play a key role in the expression and/or maintenance of the transformed phenotype initiated by other alterations. The identity of these other alterations is unknown, although one possibility is that activation of other forms of Notch, in particular Notch 3, may work synergistically with Notch 1 to bring about glial transformation. Alternatively, Notch signaling in some cell types (for example, neural stem cells), is thought to promote proliferation and to block differentiation, thereby promoting a less differentiated state. Because inhibition of Notch signaling in the present study led to suppression of cell growth in Notch-overexpressing cells, it may also be possible that Notch may function alone or in concert with other pathways to promote glioma growth by suppressing differentiation. This idea is consistent with our observation that inhibition of Notch 1 in Notch 1–overexpressing cells not only suppressed cell growth, but also altered expression of markers of glial differentiation toward a more differentiated state and led to the appearance of cells with a more stellate, differentiated appearance. The relative importance of Notch in driving proliferation and suppressing differentiation, however, may also vary by tumor type, and tumors may not all rely on or require the same degree of Notch activation. Recent observations suggest that there are higher levels of expression of Notch 1 in low-grade gliomas and oligodendrogliomas than in GBMs,22 and although the basis for this disparity is unclear, it may simply reflect the differences in the differentiation states of these tumors. Although the exact means by which Notch signaling suppresses differentiation in GBM cells remains to be defined, the results of the present study clearly identify Notch deregulation as a common alteration in human GBM and define suppression of differentiation as a key function of Notch activation in gliomas. A better understanding of how Notch mediates effects on growth, transformation, and differentiation of glial cells is likely to contribute to the development of Notch 1 as a potential therapeutic target in the treatment of human gliomas. Conclusions The results of the present study show that the Notch signaling pathway is dysregulated in human GBM cell lines and primary GBM samples and that inhibition of Notch signaling leads to growth arrest and differentiation of those GBM cells in which the Notch pathway is activated. These results suggest that Notch signaling plays a role in GBM growth and that the Notch pathway represents a new target in the therapy of human GBM. Acknowledgments We thank Kathleen Lamborn and Ivan Smirnov for statistical advice and Anita Lal and Joe Costello for helpful comments on the manuscript. Dr. Kanamori and Dr. Kawaguchi contributed equally to this study. References 1. Artavanis-Tsakonas S, Rand MD, Lake RJ: Notch signaling: cell

426

2.

3. 4. 5.

6.

7.

8.

9.

10.

11. 12.

13. 14. 15. 16. 17. 18. 19.

20.

fate control and signal integration in development. Science 284: 770–776, 1999 Brou C, Logeat F, Gupta N, Bessia C, LeBail O, Doedens JR, et al: A novel proteolytic cleavage involved in Notch signaling: the role of the disintegrin-metalloprotease TACE. Mol Cell 5: 207–216, 2000 Conboy IM, Conboy MJ, Smythe GM, Rando TA: Notch-mediated restoration of regenerative potential to aged muscle. Science 302:1575–1577, 2003 Fan X, Mikolaenko I, Elhassan I, Ni X, Wang Y, Ball D, et al: Notch1 and notch2 have opposite effects on embryonal brain tumor growth. Cancer Res 64:7787–7793, 2004 Feldkamp MM, Lala P, Lau N, Roncari L, Guha A: Expression of activated epidermal growth factor receptors, Ras-guanosine triphosphate, and mitogen-activated protein kinase in human glioblastoma multiforme specimens. Neurosurgery 45:1442–1453, 1999 Feldkamp MM, Lau N, Roncari L, Guha A: Isotype-specific Ras.GTP-levels predict the efficacy of farnesyl transferase inhibitors against human astrocytomas regardless of Ras mutational status. Cancer Res 61:4425–4431, 2001 Garces C, Ruiz-Hidalgo MJ, Font de Mora J, Park C, Miele L, Goldstein J, et al: Notch-1 controls the expression of fatty acidactivated transcription factors and is required for adipogenesis. J Biol Chem 272:29729–29734, 1997 Hicks C, Ladi E, Lindsell C, Hsieh JJ, Hayward SD, Collazo A, et al: A secreted Delta1-fc fusion protein functions both as an activator and inhibitor of Notch1 signaling. J Neurosci Res 68: 655–667, 2002 Hirose Y, Katayama M, Stokoe D, Haas-Kogan DA, Berger MS, Pieper RO: The p38 mitogen-activated protein kinase pathway links the DNA mismatch repair system to the G2 checkpoint and to resistance to chemotherapeutic DNA-methylating agents. Mol Cell Biol 23:8306–8315, 2003 Hitoshi S, Alexson T, Tropepe V, Donoviel D, Elia AJ, Nye JS, et al: Notch pathway molecules are essential for the maintenance, but not the generation, of mammalian neural stem cells. Genes Dev 16:846–858, 2002 Ho CL, Liem RK: Intermediate filaments in the nervous system: implications in cancer. Cancer Metastasis Rev 15:483–497, 1996 Holland EC, Celestino J, Dai C, Schaefer L, Sawaya RE, Fuller GN: Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice. Nature Genet 25:55–57, 2000 Jang MS, Zlobin A, Kast WM, Miele L: Notch signaling as a target in multimodality cancer therapy. Curr Opin Mol Ther 2: 55–65, 2000 Kimble J, Simpson P: The LIN-12/Notch signaling pathway and its regulation. Annu Rev Cell Dev Biol 13:333–361, 1997 Logeat F, Bessia C, Brou C, LeBail O, Jarriault S, Seidah NG, et al: The Notch 1 receptor is cleaved constitutively by a furin-like convertase. Proc Natl Acad Sci U S A 95:8108–8112, 1998 Lowell S, Jones P, Le Roux I, Dunne J, Watt FM: Stimulation of human epidermal differentiation by delta-notch signalling at the boundaries of stem-cell clusters. Curr Biol 10:491–500, 2000 McLendon C, Xin T, Ziani-Cherif C, Murphy MP, Findlay KA, Lewis PA, et al: Cell-free assays for gamma-secretase activity. FASEB J 14:2383–2386, 2000 Nickoloff BJ, Osborne BA, Miele L: Notch signaling as a therapeutic target in cancer: a new approach to the development of cell fate modifying agents. Oncogene 22:6598–6608, 2003 Nickoloff BJ, Qin JZ, Chaturvedi V, Denning MF, Bonish B, Miele L: Jagged-1 mediated activation of notch signaling induces complete maturation of human keratinocytes through NF-kappaB and PPARgamma. Cell Death Differ 9:842–855, 2002 Nicolas M, Wolfer A, Raj K, Kummer JA, Mill P, van Noort M, et al: Notch 1 functions as a tumor suppressor in mouse skin. Nat Genet 33:416–421, 2003


Notch signaling in glioblastoma multiforme 21. Nigro JM, Misra A, Zhang L, Smirnov I, Colman H, Griffin C, et al: Integrated array-comparative genomic hybridization and expression array profiles identify clinically relevant molecular subtypes of glioblastoma. Cancer Res 65:1678–1686, 2005 22. Purow BW, Haque RM, Noel MW, Su Q, Burdick MJ, Lee J, et al: Expression of Notch-1 and its ligands, Delta-like-1 and Jagged-1, is critical for glioma cell survival and proliferation. Cancer Res 65:2353–2363, 2005 23. Radtke F, Raj K: The role of Notch in tumorigenesis: oncogene or tumor suppressor? Nat Rev Cancer 3:756–767, 2003 24. Rangarajan A, Syal R, Selvarajah S, Chakrabarti O, Sarin A, Krishna S: Activated Notch1 signaling cooperates with papillomavirus oncogenes in transformation and generates resistance to apoptosis on matrix withdrawal through PKB/Akt. Virology 286: 23–30, 2001 25. Rebay I, Fleming RJ, Fehon RG, Cherbas L, Cherbas P, Artavanis-Tsakonas S: Specific EGF repeats or Notch mediate interactions with Delta and Serrate: implications for Notch as a multifunctional receptor. Cell 67:687–699, 1991 26. Robey E: Regulation of T cell fate by Notch. Annu Rev Immunol 17:283–295, 1999 27. Schuurmans C, Guillemot F: Molecular mechanisms underlying cell fate specification in the developing telencephalon. Curr Opin Neurobiol 12:26–34, 2002 28. Sonoda Y, Ozawa T, Aldape KD, Deen DF, Berger MS, Pieper RO: Akt pathway activation converts anaplastic astrocytoma to glioblastoma multiforme in a human astrocyte model of glioma. Cancer Res 61:6674–6678, 2001 29. Sonoda Y, Ozawa T, Hirose Y, Aldape KD, McMahon M, Berger MS, et al: Formation of intracranial tumors by genetically modified human astrocytes defines four pathways critical in the development of human anaplastic astrocytoma. Cancer Res 61: 4956–4960, 2001 30. Sriuranpong V, Borges MW, Ravi RK, Arnold DR, Nelkin BD, Baylin SB, et al: Notch signaling induces cell cycle arrest in small cell lung cancer cells. Cancer Res 61:3200–3205, 2001


31. Tanigaki K, Nogaki F, Takahashi J, Tashiro K, Kurooka H, Honjo T: Notch1 and Notch3 instructively restrict bFGF-responsive multipotent neural progenitor cells to an astroglial fate. Neuron 29:45–55, 2001 32. Tonon G, Modi S, Wu L, Kubo A, Coxon AB, Komiya T, et al: t(11;19)(q21;p13) translocation in mucoepidermoid carcinoma creates a novel fusion product that disrupts a Notch signaling pathway. Nat Genet 33:208–213, 2003 33. Weihofen A, Binns K, Lemberg MK, Ashman K, Martoglio B: Identification of signal peptide peptidase, a presenilin-type aspartic protease. Science 296:2215–2218, 2002 34. Weijzen S, Rizzo P, Braid M, Vaishnav R, Jonkheer SM, Zlobin A, et al: Activation of Notch-1 signaling maintains the neoplastic phenotype in human Ras-transformed cells. Nat Med 8:979–986, 2002 35. Wu L, Aster JC, Blacklow SC, Lake R, Artavanis-Tsakonas S, Griffin JD: MAML1, a human homologue of Drosophila mastermind is a transcriptional co-activator for NOTCH receptors. Nat Genet 26:484–489, 2000 36. Yasutomo K, Doyle C, Miele L, Fuchs C, Germain RN: The duration of antigen receptor signalling determines CD4+ versus CD8+ T-cell lineage fate. Nature 404:506–510, 2000 37. Zlobin A, Jang M, Miele L: Toward the rational design of cell fate modifiers: notch signaling as a target for novel biopharmaceuticals. Curr Pharm Biotechnol 1:83–106, 2000

Manuscript submitted February 17, 2005. Accepted August 7, 2006. This work was supported by National Institutes of Health Grants Nos. CA94989 and CA 100011 (R.O.P.) and CA97257 (R.O.P. and M.S.B.). Address reprint requests to: Russell O. Pieper, Ph.D., UCSF Cancer Center, 2340 Sutter Street, Room N219, San Francisco, California 94115–0875. email: [email protected].

427