Advertisement
Cell and Tumor Biology

Inhibition of a New Differentiation Pathway in Neuroblastoma by Copy Number Defects of N-myc, Cdc42, and nm23 Genes

Linda J. Valentijn, Arjen Koppen, Ronald van Asperen, Heather A. Root, Franciska Haneveld and Rogier Versteeg
DOI: 10.1158/0008-5472.CAN-04-2469 Published April 2005

Abstract

The best studied oncogenic mechanisms are inactivating defects in both alleles of tumor suppressor genes and activating mutations in oncogenes. Chromosomal gains and losses are frequent in human tumors, but for many regions, like 1p36 and 17q in neuroblastoma, no mutated tumor suppressor genes or oncogenes were identified. Amplification of N-myc in neuroblastoma is strongly correlated with loss of 1p36 and gain of 17q. Here we report that N-myc down-regulates the mRNA expression of many genes with a role in cell architecture. One of them is the 1p36 gene Cdc42. Restoring the Cdc42 expression in neuroblastoma cells strongly induced differentiation. N-myc also inhibited Cdc42 functioning at the protein level. This was mediated by nm23-H1 and nm23-H2, which are located in the amplified 17q region. Nm23-H1 and nm23-H2 are strongly up-regulated downstream targets of N-myc. Nm23-H1 was shown to bind Cdc42 and prevented the induction of differentiation. Overexpression of Nm23 due to gain of 17q and induction by N-myc combined with weak expression of Cdc42 due to loss of 1p36 and down-regulation by N-myc can thus block differentiation. Although this marks Cdc42 as a candidate tumor suppressor gene, no mutations were found. Further silencing of Cdc42 by small interfering RNA induced massive apoptosis, indicating that tumor cell survival requires a minimal Cdc42 activity. Three regions of chromosomal gain and loss thus affect genes functioning in one pathway in neuroblastoma. They converge to bring the pathway out of balance and prevent Cdc42 mediated differentiation.

  • N-myc
  • Cdc42
  • nm23
  • differentiation
  • neuroblastoma

Introduction

Neuroblastoma is a childhood tumor originating from neural crest–derived cells. Approximately 20% of neuroblastomas have N-myc amplification, which is usually accompanied by loss of heterozygosity of distal chromosome 1p and additional copies of chromosome 17q ( 1). Deletions of the short arm of chromosome 1 in N-myc amplified neuroblastomas are typically very large. The shortest 1p deletion we observed in a series of N-myc amplified neuroblastomas was 24.4 Mb, from the telomere till the RUNX3 gene (ref. 2; http://genome.cse.ucsc.edu). The region is deleted in 90% to 95% of all N-myc amplified tumors, whereas about 5% of N-myc amplified neuroblastomas have no apparent 1p deletions. In addition, N-myc single copy neuroblastomas can have 1p deletions. These deletions are of variable length as well, but can be smaller with a more telomeric SRO ( 3, 4). The different SROs in N-myc-amplified neuroblastomas have raised the idea that more than one tumor suppressor gene maps on distal chromosome 1p ( 5).

N-myc-amplified neuroblastomas follow a very aggressive course ( 6). Overexpression of transfected N-myc genes in neuroblastoma cell lines strongly increased proliferation rates ( 7, 8). Recent mRNA expression profiling experiments, serial analysis of gene expression (SAGE) and microarray analysis, identified many myc-regulated genes ( 9– 13). We used SAGE to identify genes regulated by N-myc in neuroblastoma. Genes up-regulated by N-myc were involved in rRNA processing and protein synthesis ( 9). Two recently identified N-myc targets, nm23-H1 and nm23-H2, map to chromosomal band 17q21, in the chromosome 17q region overrepresented in neuroblastoma ( 14).

Additional copies of the distal chromosome 17q arm are present in about 70% of neuroblastoma tumors, including all N-myc amplified cases. The copy number ranges between two and eight extra alleles and the minimal region of gain is 17q21-17qter. The map position of the nm23-H1 and nm23-H2 genes at 17q21 makes them interesting candidates for a role in neuroblastoma pathogenesis. The nm23 proteins encode nucleoside diphosphate kinases, the suppliers of nucleoside triphosphates. In several human tumors, like breast cancer and melanoma, decreased expression was correlated with poor prognosis ( 15, 16). Up-regulation of nm23-H1 reverted melanoma cells from metastatic to nonmetastatic ( 17). In contrast, in several other tumors including neuroblastoma and osteosarcoma, elevated expression of nm23-H1 and nm23-H2 was associated with a poor prognosis ( 18, 19). The extra copies of 17q and the induction by N-myc result in a strongly increased expression of the nm23 genes in neuroblastoma cell lines and tumors ( 14). Besides the function of the nm23 proteins as dinucleoside kinases, growing evidence exists for additional functions in signal transduction ( 20– 22).

In this article, we identified genes down-regulated by N-myc in neuroblastoma. One of the key genes is Cdc42. The Cdc42 gene maps on 1p36, and one copy is consistently deleted in N-myc amplified neuroblastoma. In addition, we provide evidence that the activity of the Cdc42 protein is down-regulated by nm23-H1 and nm23-H2, which are induced by N-myc.

Materials and Methods

Cell culture. The SHEP-2 and SHEP-21N cell lines were cultured in RPMI, the SKNAS-NmycER cells in DMEM. Both media were supplemented with 10% FCS, 2 mmol/L glutamine, 50 units/mL penicillin, and 50 units/mL streptomycin. N-myc expression in SHEP-21N was repressed by tetracycline (final concentration, 50 ng/mL; Sigma, St. Louis, MO). The NmycER protein in the SKNAS-NmycER cells was activated upon addition of 4-hydroxytamoxifen (final concentration, 50 nmol/L; Sigma). The ER domain has a point mutation, which enables transactivation of the chimearic protein by 4-hydroxytamoxifen and not estrogens ( 23).

Plasmids. The pCdc42flag plasmid was constructed by amplification of cDNA with primers Cdc42Efor (5′-TATATAGAATTCATTTCAGCAATGCAGACAATTAAG-3′) and Cdc42Xrev (5′-TATATACTCGAGTAGCAGCACACACCTGCG-3′). The amplified product was digested with EcoRI and XhoI and cloned into the EcoRI and XhoI sites of pCMV-Tag4 vector (Stratagene, La Jolla, CA). The pCdc42G12Vflag and pCdc42T17Nflag constructs were produced by amplification of Cdc42 from vector pCdc42flag using primer T3 (5′-AATTAACCCTCACTAAAGGG-3′) in combination with mutant primers CdcG12Vrev (5′-GTGTAGGATATCAGGAGACATGTTTTACCAACAGCAACATCGC-3′) or CdcT17Nrev (5′-GTGTAGGATATCAGGAGACAGTTTTTACC-3′). The PCR products were digested with BamHI and EcoRV and ligated in the BamHI and EcoRV sites of pCdc42flag vector.

Northern blot analysis. Total RNA (20 μg per lane) was separated on a 1% agarose gel in the presence of 6.7% formaldehyde and blotted on Hybond-N membranes (Amersham, Piscataway, NJ) in 10× SSC. Hybridization was carried out overnight in 0.5 mol/L NaHPO4 (pH 7.0), 7% SDS, 1 mmol/L EDTA at 65°C. Filters were washed in 40 mmol/L NaHPO4, 1% SDS at 65°C. Probes were labeled by random priming of sequence-verified PCR products.

Western blotting. Primary antibodies: anti-N-myc (PharMingen, San Diego, CA), anti-Cdc42-B8 (Santa Cruz Biotechnology, Santa Cruz, CA), anti--tubulin (Roche, Nutley, NJ), anti-Caspase-3 (Cell Signaling Technology, Beverly, MA), and anti-poly(ADP-ribose) polymerase (Biomol Research Labs, Plymouth Meeting, PA). Western blotting was done according to standard procedures, secondary antibody anti-mouse linked to horseradish peroxidase (Amersham), detected with Enhanced Chemiluminescence KIT (Amersham).

Transient transfection and immunofluoresence microscopy. For transient expression, the cells were grown for 48 hours on glass slides in medium supplemented with 1% FCS. Four micrograms of plasmid DNA and 4 μL DAC-30 (Eurogentec, Seraing, Belgium) in 1 mL medium, without FCS and antibiotics, was incubated for 30 minutes at room temperature. For double transfections, 4 μg of each plasmid were incubated with 8 μg DAC-30. The DNA/DAC-30 mix was added dropwise to the cell cultures. After 16 hours, the medium was replaced by fresh medium with 1% FCS. Forty hours after transfection, the slides were fixed with 4% paraformaldehyde in PBS for 25 minutes. The cells were permeabilized with PBS/0.05% Triton X-100. The slides were washed thrice with PBS/TX (PBS, 0.01% Triton X-100). The slides were blocked for 30 minutes in Abdil (PBS, 0.01% Triton X-100, 1% bovine serum albumin). The antibodies were diluted in Abdil each incubated for 30 minutes each incubation followed by four washes in PBS/TX. Antibodies: 400 times diluted mouse-anti-flag M2 (Stratagene), 100 times diluted TRITC- or FITC-conjugated anti-mouse (Sigma), 200 times diluted goat-anti-NFL (Santa Cruz Biotechnology), and 200 times diluted TRITC-conjugated anti-goat (Sigma). The slides were rinsed with H2O, dried and mounted in Vectashield (Vector Laboratories, Burlingame, CA) with 4′,6-diamidino-2-phenylindole (nuclear staining).

Glutathione S-transferase pulldown. Nm23H1 was amplfied with adaptor oligonucleotides (5′-TATATAGGATCCATGGCCAACTGTGAGCGTAC-3′ and 5′-TATATGAATTCTCTGCCCTCCTGTCATTCAT-3′). The product was digested with BamHI/EcoRI and cloned in the pGEX2TK vector (Pharmacia, Piscataway, NJ) to yield GST-nm23H1.

GST-nm23H1 transformed bacteria were lysed in PBS-0.01%Triton X-100. Cell lysis and pulldown were done as described by Sander et al. ( 24).

RNA interference. The sequences of the single stranded RNA molecules (Isogen, Maarsen, The Netherlands) are for Cdc42-a 5′-CUAUGCAGUCACAGUUAUGTT-3′and 5′-CAUACCUGUGACUGCAUAGTT-3′, for Cdc42-b 5′-CUCACCACUGUCCAAAGACTT-3′ and 5′-GUCUUUGGACAGUGGUGAGTT-3′, for N-myc-a 5′-CACCAAGGCUGUCACCACATT-3′ and 5′-UGUGGUGACAGCCUUGGUGTT-3′, for N-myc-b 5′-CCCAGACCUCGAGUUUGACTT-3′ and 5′-GUCAAACUCGAGGUCUGGGTT-3′, and for GFP 5′-GACCCGCGCCGAGGUGAAGTT-3′ and 3′-CUUCACCUCGGCGCGGGUCTT-3′.

Cells were grown to 30% confluency in 6-cm plates. RNA interference was done as described by Elbashir et al. ( 25). In short, 20 μmol/L per single stranded RNA molecule in 30 μL annealing buffer [0.1 mol/L KAc, 30 mmol/L HEPES/KOH (pH 7.4), and 2 mmol/L MgAc] was heated at 95°C for 5 minutes, followed by incubation at 37°C for 1 hour. The small interfering RNA (siRNA) mixture was transfected serum-free with 30 μL Lipofectamine (Invitrogen, Carlsbad, CA) as described by the manufacturer. After 4 hours, the medium was supplemented with FCS to a final concentration of 10%.

For immunofluorescence analysis cells were grown on glass slides in 6-well plates and transfected with 15 μL siRNA and 15 μL Lipofectamine. After 72 hours, the cells were fixed with formaldehyde and stained with TRITC-phalloidin (Sigma) and 4′,6-diamidino-2-phenylindole.

Results

Genes down-regulated by N-myc. We have applied the SAGE technique to identify downstream target genes of N-myc. The neuroblastoma cell line SHEP has no N-myc amplification and expression. A tetracycline-regulated N-myc expression vector has been introduced into these cells, resulting in the SHEP-21N clone ( 8). SAGE libraries were constructed from N-myc expressing SHEP-21N cells and from SHEP-2 (empty vector control) cells ( 9). We identified 34,443 tags for SHEP-21N and 44,450 tags for SHEP-2, each representing one mRNA expressed in these cells. 112 Tags were significantly (P < 0.01) down-regulated in the N-myc expressing cells. We could reliably assign 78 of these tags to a total of 64 genes (http://bioinfo.amc.uva.nl/HTMseq/controller; Table 1 ). We analyzed 14 genes by Northern blot, which confirmed the SAGE data ( Fig. 1A ). The overall expression pattern showed that N-myc down-regulates many genes involved in the architecture of the cell and the extracellular matrix. Cytoskeletal components like β-actin and vimentin are down-regulated, but also genes which regulate the structures, like transgelin (TAGLN). The adhesion properties of cells depend on the extracellular matrix and membrane molecules. Genes of both groups are repressed by N-myc. Several collagens are regulated by N-myc, but also connective tissue growth factor (CTGF) and SPARC, which regulate the build-up of the extracellular matrix. In addition, cadherin 11 (CDH11) and other genes encoding membrane proteins are down-regulated.

View this table:
Table 1.

Down-regulation of gene expression by N-myc

Figure 1.
Figure 1.

Gene repression by N-myc. A, confirmation of the SAGE data. Northern blot analysis of 14 genes in SHEP-2 and SHEP-21N cells. Blot was also hybridized with N-myc (confirmation N-myc expression) and cofilin-1 (CFL1, an equally expressed gene). The 28S ribosomal band as loading control. B, Northern blot analysis of a time course experiment of N-myc. SHEP-21N cells were treated for 0 to 7 days with tetracycline, which switches off the ectopic N-myc expression. After 7 days, the cells were grown for a week in the absence of tetracycline (7+/7−). Right, analyzed genes.

To investigate whether the genes were indeed down-regulated by N-myc, we did Northern blot analysis of a time course experiment in which N-myc was switched off and on in the SHEP-21N cell line. The N-myc expression was switched off by tetracycline for 7 days, after which N-myc was switched on for another 7 days ( Fig. 1B). Two distinct patterns of N-myc regulation emerged. Expression of a limited number of genes increased within 24 hours after N-myc was turned off. This group includes galectin-1 (LGALS), S100A10, and caveolin-1 (CAV1; Fig. 1B). Reexpression of N-myc decreased the expression of these genes ( Fig. 1B). The majority of genes follow a slow N-myc dependent regulation. The expression of COL1A1, TAGLN, SPARC, FN1, and SDC2 increased 4 days after N-myc was switched off ( Fig. 1B). These analyses show that genes involved in cell architecture and adhesion are important targets of down-regulation by N-myc. Most of these genes are slowly regulated and thus most likely represent secondary targets.

Cdc42 expression is down-regulated by N-myc. One of the genes identified as an N-myc downstream target is Cdc42 ( Table 1). The SAGE-tag frequencies show that Cdc42 is expressed at a level of 12.6 mRNAs per 20,000 mRNAs in SHEP-2 and 2.9 per 20,000 in SHEP-21N. Cdc42 is a member of the Rho GTPase subfamily and functions in a series of signaling pathways, which are particularly important in cytoskeletal remodeling ( 26, 27). As many genes with a function in these processes are down-regulated by N-myc, Cdc42 might be a key target of N-myc to control cytoskeletal remodeling. Northern blot analysis confirmed the decreased expression of Cdc42 mRNA in SHEP-21N compared with SHEP-2 ( Fig. 2A ). When N-myc expression in SHEP-21N was down-regulated by tetracycline, Cdc42 mRNA expression increased within 24 hours ( Fig. 2A). Western blot analysis showed that also the Cdc42 protein level increased after N-myc was switched off ( Fig. 2B). The regulation of Cdc42 was also investigated in cell line SKNAS-NmycER, which expresses a hybrid N-myc-estrogen receptor protein that can be activated by 4-hydroxytamoxifen. The SKNAS-NmycER cells were treated with 4-hydroxytamoxifen and the Cdc42 mRNA expression decreased within 24 hours after N-myc activation ( Fig. 2C). The Cdc42 reduction was moderate (3-fold). In addition, we analyzed the expression of Cdc42 mRNA in a panel of neuroblastoma tumors by Northern blot. Cdc42 expression was reduced in 10 tumors with N-myc amplification compared with 10 N-myc single copy tumors ( Fig. 2D). The tumors show no expression of c-myc ( Fig. 2D). We conclude that Cdc42 mRNA expression in neuroblastoma cells is down-regulated by N-myc, resulting in decreased Cdc42 protein levels.

Figure 2.
Figure 2.

Cdc42 is down-regulated by N-myc in cell lines and tumors. A, Northern blot analysis of a time course experiment of N-myc. SHEP-21N cells were treated for 0 to 6 days with tetracycline, which switches off the ectopic N-myc expression. Lane 1, total RNA of SHEP-2 cells. The 28S ribosomal band as loading control. B, Western blot analysis of N-myc and Cdc42 proteins in a similar time course experiment in SHEP-21N. Part of the Coomassie-stained SDS-PAGE as loading control. C, Northern blot of SKNAS-N-mycER cells treated for 0 to 3 days with 4-hydroxytamoxifen (4OHT), which activates the N-mycER protein. Expression of Cdc42 and nm23-H1 mRNA. D, Northern blot analysis of N-myc, nm23-H1, Cdc42, and c-myc expression of 10 N-myc-amplified and 10 N-myc single-copy tumors.

The Cdc42 gene maps at 21.5 Mb from the 1p telomere, within the 24.4 Mb region consistently deleted in N-myc amplified neuroblastomas ( 2). The net result is that Cdc42 is 50% reduced by deletion of one allele in N-myc amplified neuroblastoma tumors, and the expression is further down-regulated by N-myc. We therefore analyzed whether Cdc42 could function as a tumor suppressor gene in neuroblastoma.

Activated Cdc42 induces neuronal differentiation in SHEP-21N cells. Cdc42 is a G-protein that is active in the GTP-bound state and inactive in the GDP-bound state. We investigated the role of Cdc42 in neuroblastoma by transient expression of the Cdc42 protein coupled to a COOH-terminal FLAG tag (Cdc42flag). We first transfected a mutant Cdc42 form that is permanently active (Cdc42-G12V). After 40 hours, 42% of the SHEP-21N cells expressing the Cdc42-G12Vflag protein showed neuronal differentiation, as evident by long neurite extensions of at least twice the length of the cell body ( Fig. 3A and B ). The extensions were analyzed for expression of neurofilament, a marker for neuronal differentiation. The extensions in cells transfected with Cdc42-G12V show formation of neural fibers with strong expression of the neurofilament light chain ( Fig. 3C). The differentiation was not observed in untransfected cells. In addition, we transfected SHEP-21N with a wild-type Cdc42flag construct and with a dominant-negative Cdc42 mutant (Cdc42-T17N). Wild-type Cdc42flag and the dominant-negative Cdc42-T17Nflag-transfected cells showed only 3% and 6% differentiation, respectively ( Fig. 3A and B), suggesting that the wild-type Cdc42 protein in SHEP-21N cells is in the inactive GDP-bound state. However, when N-myc was switched off with tetracycline, also wild-type Cdc42flag-transfected cells displayed neuronal differentiation (28%), whereas the dominant-negative Cdc42 construct remained inactive ( Fig. 3B). This strongly suggests that N-myc not only down-regulates the Cdc42 mRNA expression level but also keeps the Cdc42 protein in its inactive GDP-bound form.

Figure 3.
Figure 3.

Neuronal differentiation is mediated by active Cdc42 and counteracted by nm23-H1. SHEP-21N cells were grown without and with tetracycline (tet) and transfected with flag-tagged Cdc42, Cdc42-G12V, or Cdc42-T17N. Cells were fixed 40 hours after transfection and stained with an anti-FLAG antibody to identify transfected cells. A, examples of an undifferentiated (Cdc42-T17N transfected) and differentiated (Cdc42-G12V transfected) cell. B, differentiation frequencies in over 500 flag-positive cells per transfected construct. Bars, SE. *, P < 0.01 (χ2 test). C, expression of neurofilament (NFL) upon differentiation in SHEP-21N cells positive for Cdc42-G12V-flag expression. Note the nodules of the neural structures are negative for neurofilament. D, Cdc42 and nm23-H1 physically interact. SHEP-21N cells were transfected with Cdc42. The lysates were incubated with fusion protein GST-nm23H1 (lanes 1 and 2) or GST (lanes 3 and 4) linked to gluthatione-sepharose beads and analyzed for bound Cdc42 protein (pull-down) on Western blot. Part of the lysate was analyzed for Cdc42 expression. Blots were stripped and incubated with anti-N-myc. E, analysis of neuronal differentiation by Cdc42 in the presence or absence of nm23. The SHEP-21N cells were grown in the absence or presence of tetracycline and transfected with flag-tagged Cdc42 in combination with empty vector, nm23-H1 or nm23-H2. Cells were fixed 40 hours after transfection and stained with an anti-flag antibody to identify transfected cells. Approximately 1,000 flag-positive cells were scored for neuronal differentiation. Bars, SE. *, P < 0.01 (χ2 test).

Nm23 suppresses the activity of Cdc42. Cycling between the inactive GDP-bound and active GTP-bound state of Cdc42 is regulated by GTPase-activating proteins (GAP) and guanine nucleotide exchange factors. The inhibition of Cdc42-induced differentiation by N-myc indicated that N-myc controls a GAP or guanine nucleotide exchange factor of Cdc42. Previous experiments have shown that N-myc strongly up-regulates the expression of the nm23-H1 and nm23-H2 genes in SHEP-21N ( 14). We confirmed the up-regulation of the nm23-H1 gene in the SKNAS-NmycER cell line treated with 4-hydroxytamoxifen ( Fig. 2C). In addition, in tumors with N-myc amplification, nm23-H1 is also highly expressed ( Fig. 2D).

Nm23 proteins are primarily known to function as dinucleoside kinases, but a possible role for nm23-H1 in GDP and GTP exchange of Rad1 has been reported ( 20). We investigated whether the nm23-H1 and Cdc42 proteins can physically interact with each other. For this purpose, we constructed the glutathione S-transferase (GST) fusion protein GST-nm23-H1. The GST-nm23-H1 protein was immobilized on glutatione-sepharose beads and incubated with a lysate of SHEP-21N cells transfected with Cdc42. The beads were analyzed for bound proteins by Western blot. Cdc42 was pulled down by the GST-nm23-H1 protein and not by the GST protein alone ( Fig. 3D). As a control for unspecific protein binding in the pull-down assay, the Western blots were incubated with an N-myc antibody. N-myc did not bind to the GST or GST-nm23-H1 protein ( Fig. 3D). We concluded that nm23-H1 and Cdc42 can physically interact.

We also analyzed the functional interaction of nm23-H1 and nm23-H2 with Cdc42. Transient transfection of a wild-type Cdc42-flag construct into SHEP-21N cells resulted in 8% differentiation of the transfected cells ( Fig. 3E). When N-myc was switched off in this experiment, the Cdc42flag-induced differentiation increased to 20% of the transfected cells. As N-myc down-regulation resulted in a decrease of nm23-H1 and nm23-H2, we reconstituted these proteins by cotransfections. SHEP-21N cells in which N-myc expression was switched off by tetracycline were cotransfected with wild-type Cdc42flag and nm23-H1 and/or nm23-H2 expression vectors. Cotransfection of either nm23-H1 or nm23-H2 could completely replace the effect of N-myc, as the differentiation reduced from 20% to 6% ( Fig. 3E). The percentage of differentiated cells is not further reduced by transfection of both nm23-H1 and nm23-H2. Therefore, nm23-H1 and nm23-H2 prevent Cdc42-induced differentiation. As N-myc strongly up-regulates nm23-H1 and nm23-H2, they are likely to mediate the effect of N-myc.

Cdc42 RNA interference: Cdc42 is essential for cell division. Neuroblastomas with N-myc amplification still have one Cdc42 allele. The classic paradigm for tumor suppressor gene inactivation is deletion of one copy and mutation of the second allele. We screened the coding region of Cdc42 for mutations in over 50 neuroblastoma tumors and cell lines but no mutations were found. The Cdc42 expression in N-myc amplified tumors suggests haplo-insufficiency as a potential mechanism for tumorigenesis. We further decreased the expression of Cdc42 by RNA interference. The effect of two Cdc42-specific siRNAs was compared with two N-myc-siRNAs, and siRNA for green fluorescent protein, which has no target in the SHEP-21N cells. Western blot and Northern blot analysis showed that 48 hours after transfection the expression of N-myc or Cdc42 was strongly reduced by N-myc-siRNA and Cdc42-siRNA, respectively ( Fig. 4A and B ). Cell growth of the N-myc-siRNA treated cells was strongly reduced compared with the gfp-siRNA control ( Fig. 4C). Silencing of Cdc42 had an even more dramatic effect on cell growth. The majority of SHEP-21N cells treated with Cdc42-siRNA died within 72 hours after transfection ( Fig. 4C). In contrast, cells treated with gfp- or N-myc-siRNA survived the treatment. The experiments were repeated on glass slides and the actin cytoskeleton and nuclei were stained 72 hours after transfection. In the absence of Cdc42, the actin was organized in a peripheral membrane-associated ring, indicating that these cells were unable to rearrange the cytoskeleton after cell division ( Fig. 4D). The cells transfected with Cdc42-siRNA showed 16% apoptotic nuclei, compared with 2% and 5% in the gfp-siRNA and N-myc-siRNA–treated cells, respectively ( Fig. 4E). The apoptosis was further analyzed on Western blot. The cells were treated for 48 hours with siRNAs and the lysates were screened for cleaved proteins. In cells treated with Cdc42-siRNAs, both caspase3 and poly(ADP-ribose) polymerase showed the apoptosis-induced cleavage products ( Fig. 4F). No cleaved products were visible after treatment with gfp- or N-myc-siRNA. Therefore, Cdc42 is an essential gene for the growth and survival of neuroblastoma cells. This provides an explanation for the absence of Cdc42 mutations: a minimal amount of Cdc42 is required for cell survival.

Figure 4.
Figure 4.

RNA interference of Cdc42 inhibits cell division and actin remodeling. Equal numbers of SHEP-21N cells were treated with siRNA for gfp (negative control), N-myc, and Cdc42. For Cdc42 and N-myc, two siRNAs were used (a and b), directed against different regions. A, Western blot of cell lysates isolated 48 hours post-transfection. Right, incubated anti-bodies. Controls: anti-β-tubulin (tub); Coomassie-stained SDS-PAGE gel (loading control). B, Northern blot of siRNA-treated SHEP-21N cells. Right, hybridized probes. Controls: cofilin (CFL1), 18S (loading control). C, light microscopy picture of cells (40× magnification), 72 hours post-transfection. D, cytoskeleton staining with phalloidine-TRITC of the cells 72 hours post-transfection (1,000× magnification). The nuclei were stained with 4′,6-diamidino-2-phenylindole and scored for apoptosis. Arrowhead, apoptotic cell with apoptotic bodies. E, quantification of apoptosis. Bars, SE. F, Western blot of cell lysates 48 hours post-transfection. Blots were incubated with anti-caspase3 and anti-poly(ADP-ribose) polymerase (PARP). Fraction of cleaved protein (bottom of lanes).

Discussion

Cellular transformation involves increased proliferation, alterations in cell shape and adhesion. The N-myc protein regulates genes involved in these processes. Here, we have shown that N-myc represses genes determining the cell structure. The N-myc targets include structural genes but also one of the key regulators Cdc42. Cdc42 is a member of the Rho GTPase subfamily and functions in a series of signaling pathways, which are particularly important in cytoskeletal remodeling ( 26, 27). It has a major role in actin reorganization, as Cdc42-GTP binds to WASP, which in turn binds and activates the Arp2/3 complex ( 28). This induces the branching of actin filaments and the formation of the cytoskeleton. Several genes functioning in this process are regulated by N-myc. N-myc not only down-regulates Cdc42, but also actin (ACTG1 and ACTB) and ARPC1B ( Table 1). Furthermore, Cdc42 has been shown to mediate laminin-dependent differentiation in PC12 cells and murine neuroblastoma cells ( 29– 31). Here, we have shown that reconstitution of active Cdc42 allowed differentiation of human neuroblastoma cells. The Cdc42 expression and activity is down-regulated at three different levels in N-myc amplified neuroblastoma cells ( Fig. 5 ). First, the expression is 50% reduced by loss of heterozygosity of 1p36, as it maps in the 1p region consistently deleted in N-myc amplified neuroblastomas. Second, a further 4- to 5-fold down-regulation of Cdc42 mRNA is caused by the overexpression of N-myc. Together, this results in an estimated 8- to 10-fold down-regulation at the mRNA level. Third, N-myc inhibits the differentiation induced by the Cdc42 protein. This effect of N-myc is fully reconstituted by transfection of nm23-H1 and nm23-H2 constructs, which strongly suggests that the N-myc effect is mediated by the nm23 proteins.

Figure 5.
Figure 5.

Model for the down-regulation of Cdc42 activity by N-myc, loss of heterozygosity of chromosomal band 1p36, and gain of the chromosome 17q arm.

Regulation of the nm23-H1 and nm23-H2 genes in neuroblastoma is a mirror image of the Cdc42 regulation. The nm23 genes map in the minimal region of 17q gain and are present with two to eight extra copies. The expression of these nm23 copies is 6- to 10-fold up-regulated by N-myc, resulting in a 12- to 80-fold overexpression of nm23-H1 and -H2 ( 14). Therefore, the ratio between Cdc42 and nm23 protein levels is probably at least 100-fold out of balance. The differentiation-blocking effect of N-myc is thus amplified by chromosomal imbalances of the downstream mediators, Cdc42 and nm23-H1 and nm23-H2. Moreover, neuroblastomas frequently have an unbalanced translocation between chromosomes 1p and 17q, causing loss of heterozygosity of 1p and gain of 17q ( 32, 33). This would, by a single hit, disturb the balance between Cdc42 and the nm23 genes. The translocation event is often followed by duplication of the derivative chromosome 1 with the extra 17q material, further increasing the imbalance. Amplification of N-myc would boost the effect of the imbalance, by down-regulation of the remaining Cdc42 copy and up-regulation of the overrepresented nm23 genes. The down-regulation of Cdc42 and up-regulation of the nm23 genes by N-myc is very consistent in neuroblastoma cell lines and tumors. In the two cell lines in which we can manipulate N-myc activity, Cdc42 is down-regulated and the nm23 genes are up-regulated. Furthermore, comparison of a panel of 10 N-myc amplified tumors and 10 N-myc single copy neuroblastomas strongly supports an identical effect of N-myc on Cdc42 and the nm23 genes in vivo (this article and ref. 14).

The nm23 proteins could function as a GAP for Cdc42, keeping Cdc42 in the inactive, GDP-bound state. Several lines of evidence support this role for nm23-H1. Direct GAP activity of nm23-H1 was shown for the Rad1 GTPase in melanoma cells ( 20). Involvement in the regulation of Cdc42-family member Rac1 was established as well. Nm23-H1 is recruited toward adhesive junctions by ARF6-GTP ( 34). At these junctions, nm23-H1 negatively regulates TIAM1 and subsequently Rac1 is inactivated ( 21). In this study, we showed that nm23-H1 can bind to the Cdc42 protein and nm23 can inhibit the Cdc42-induced differentiation. These data suggest that nm23-H1 and nm23-H2 also exerts GAP-activity for Cdc42.

Is Cdc42 the elusive neuroblastoma suppressor gene at distal chromosome 1p? If Cdc42 indeed functions as a tumor suppressor gene, it would be unlikely to be the only 1p suppressor gene in neuroblastoma. Deletions of 1p in N-myc amplified neuroblastomas are at least 24.4 Mb. If Cdc42, which maps at 21.5 Mb from the telomere, would be the one and only suppressor gene, a certain frequency of interstitial deletions encompassing Cdc42 would be expected. The length of the deletions has previously raised the idea that more than one suppressor gene is involved ( 5). Thus far, only one homozygous 1p36 deletion was identified in a neuroblastoma cell line ( 35). The scarcity of homozygous 1p36 deletions in N-myc amplified neuroblastomas has suggested that the postulated tumor suppressor gene(s) may function by haploinsufficiency. Our data suggest that Cdc42 may represent the first example of such a gene. The chromosome 17q region gained in neuroblastoma tumors is also very large. Therefore, in addition to the nm23 genes, other 17q genes may contribute to neuroblastoma pathogenesis as well.

Neuroblastomas are embryonal tumors that originate from neural crest–derived cells. During normal embryogenesis N-myc is expressed in migrating neuroblasts ( 36) and probably prevents premature differentiation. When neuroblasts arrive at their destination in the trunk, N-myc expression ceases and differentiation proceeds, in which the Cdc42/nm23 pathway may play an important role. The strong differentiation-inducing potential of Cdc42 and the suppression of this effect by N-myc and nm23 suggest that this is a major mechanism in neuroblast differentiation. We therefore consider Cdc42 as a strong candidate tumor suppressor gene for neuroblastoma.

The classic proof of principle for a tumor suppressor gene is inactivation of both gene copies. One copy of Cdc42 is deleted, but we detected no mutations in the remaining copy. Our siRNA experiments showed that a minimal Cdc42 activity is required for viability of neuroblastoma cells, explaining why Cdc42 cannot be homozygously inactivated. Down-regulation of Cdc42 by siRNA caused a major reduction in cell number and dislocation of actin to the inner surface of rounded cells. Normally, this is only observed during cell division and should be followed by actin rearrangent and the appearance of new stress fibers. Absence of Cdc42 causes a collapse of the cytoskeleton and might thus inhibit cell cycle progression. Alternatively, the observed cell death may result from the involvement of Cdc42 in spindle formation and cell cycle progression by the Cdc42-PAR6-PKC3 pathway, as observed in Caenorhabditis elegans and rat astrocytes ( 37, 38). Surprisingly, SHEP-21N cells treated with siRNA still show many metaphases after 72 hours. This phenotype resembles the yeast Cdc42-V44A mutant, which cannot interact with multiple effector proteins ( 39). Half of these mutant cells have two or more nuclei, indicating a G2-M-phase delay and a defect in cytokinesis and cell separation.

The mechanism of inactivation of Cdc42 might represent a new paradigm for tumor suppressor gene inactivation. Chromosomal regions that consistently show hemizygosity or moderate copy number gains are frequent in human tumors. For many of them, no tumor suppressor genes or oncogenes have been identified as yet. Here we show that these relatively limited changes in copy numbers can dramatically amplify the oncogenic effect of an activated pathway, in this case the N-myc pathway. Importantly, in this model the actual tumor suppressor gene does not have to be completely inactivated. Cdc42 is not fully destroyed like classic tumor suppressor genes but can be reactivated. This offers a clinical perspective, as innovative drugs may interfere in the pathway and restore the function of the remaining Cdc42 allele.

Acknowledgments

Grant support: Dutch Cancer Society (L.J. Valentijn and A. Koppen).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank K. Berns for providing the pBSK-ER plasmid and J. Collard for the GST-plasmid.

Footnotes

    • Received July 10, 2004.
    • Revision received January 4, 2005.
    • Accepted January 31, 2005.

    References

    1. ↵
      Bown N. Neuroblastoma tumour genetics: clinical and biological aspects. J Clin Pathol 2001; 54: 897–910.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      Spieker N, Beitsma M, Van Sluis P, Chan A, Caron H, Versteeg R. Three chromosomal rearrangements in neuroblastoma cluster within a 300-kb region on 1p36.1. Genes Chromosomes Cancer 2001; 31: 172–81.
      OpenUrlCrossRefPubMed
    3. ↵
      Bauer A, Savelyeva L, Claas A, Praml C, Berthold F, Schwab M. Smallest region of overlapping deletion in 1p36 in human neuroblastoma: a 1 Mbp cosmid and PAC contig. Genes Chromosomes Cancer 2001; 31: 228–39.
      OpenUrlCrossRefPubMed
    4. ↵
      Caron H, Spieker N, Godfried M, et al. Chromosome bands 1p35-36 contain two distinct neuroblastoma tumor suppressor loci, one of which is imprinted. Genes Chromosomes Cancer 2001; 30: 168–74.
      OpenUrlCrossRefPubMed
    5. ↵
      Versteeg R, Caron H, Cheng NC, et al. 1p36: every subband a suppressor? Eur J Cancer 1995; 3: 538–41.
    6. ↵
      Schwab M, Alitalo K, Klempnauer KH, et al. Amplified DNA with limited homology to myc cellular oncogene is shared by human neuroblastoma cell lines and a neuroblastoma tumour. Nature 1983; 305: 245–8.
      OpenUrlCrossRefPubMed
    7. ↵
      Bernards R, Dessain SK, Weinberg RA. N-myc amplification causes down-modulation of MHC class I antigen expression in neuroblastoma. Cell 1986; 47: 667–74.
      OpenUrlCrossRefPubMed
    8. ↵
      Lutz W, Stohr M, Schurmann J, Wenzel A, Lohr A, Schwab M. Conditional expression of N-myc in human neuroblastoma cells increases expression of α-prothymosin and ornithine decarboxylase and accelerates progression into S-phase early after mitogenic stimulation of quiescent cells. Oncogene 1996; 13: 803–12.
      OpenUrlPubMed
    9. ↵
      Boon K, Caron HN, van Asperen R, et al. N-myc enhances the expression of a large set of genes functioning in ribosome biogenesis and protein synthesis. EMBO J 2001; 20: 1383–93.
      OpenUrlAbstract
    10. Menssen A, Hermeking H. Characterization of the c-MYC-regulated transcriptome by SAGE: identification and analysis of c-MYC target genes. Proc Natl Acad Sci U S A 2002; 99: 6274–9.
      OpenUrlAbstract/FREE Full Text
    11. Coller HA, Grandori C, Tamayo P, et al. Expression analysis with oligonucleotide microarrays reveals that MYC regulates genes involved in growth, cell cycle, signaling, and adhesion. Proc Natl Acad Sci U S A 2000; 97: 3260–5.
      OpenUrlAbstract/FREE Full Text
    12. Guo QM, Malek RL, Kim S, et al. Identification of c-myc responsive genes using rat cDNA microarray. Cancer Res 2000; 60: 5922–8.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      Schuhmacher M, Kohlhuber F, Holzel M, et al. The transcriptional program of a human B cell line in response to Myc. Nucleic Acids Res 2001; 29: 397–406.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      Godfried MB, Veenstra M, van Sluis P, et al. The N-myc and c-myc downstream pathways include the chromosome 17q genes nm23-H1 and nm23-H2. Oncogene 2002; 21: 2097–101.
      OpenUrlCrossRefPubMed
    15. ↵
      Hennessy C, Henry JA, May FE, Westley BR, Angus B, Lennard TW. Expression of the antimetastatic gene nm23 in human breast cancer: an association with good prognosis. J Natl Cancer Inst 1991; 83: 281–5.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      Florenes VA, Aamdal S, Myklebost O, Maelandsmo GM, Bruland OS, Fodstad O. Levels of nm23 messenger RNA in metastatic malignant melanomas: inverse correlation to disease progression. Cancer Res 1992; 52: 6088–91.
      OpenUrlAbstract/FREE Full Text
    17. ↵
      Steeg PS, Bevilacqua G, Kopper L, et al. Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 1988; 80: 200–4.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      Oda Y, Naka T, Takeshita M, Iwamoto Y, Tsuneyoshi M. Comparison of histological changes and changes in nm23 and c-MET expression between primary and metastatic sites in osteosarcoma: a clinicopathologic and immunohistochemical study. Hum Pathol 2000; 31: 709–16.
      OpenUrlCrossRefPubMed
    19. ↵
      Leone A, Seeger RC, Hong CM, et al. Evidence for nm23 RNA overexpression, DNA amplification and mutation in aggressive childhood neuroblastomas. Oncogene 1993; 8: 855–65.
      OpenUrlPubMed
    20. ↵
      Zhu J, Tseng YH, Kantor JD, et al. Interaction of the Ras-related protein associated with diabetes rad and the putative tumor metastasis suppressor NM23 provides a novel mechanism of GTPase regulation. Proc Natl Acad Sci U S A 1999; 96: 14911–8.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      Otsuki Y, Tanaka M, Yoshii S, Kawazoe N, Nakaya K, Sugimura H. Tumor metastasis suppressor nm23H1 regulates Rac1 by interaction with Tiam1. Proc Natl Acad Sci U S A 2001; 98: 4385–90.
      OpenUrlAbstract/FREE Full Text
    22. ↵
      Fournier HN, Dupe-Manet S, Bouvard D, et al. Integrin cytoplasmic domain-associated protein 1α (ICAP-1α) interacts directly with the metastasis suppressor nm23-H2, and both proteins are targeted to newly formed cell adhesion sites upon integrin engagement. J Biol Chem 2002; 277: 20895–902.
      OpenUrlAbstract/FREE Full Text
    23. ↵
      Littlewood TD, Hancock DC, Danielian PS, Parker MG, Evan GI. A modified oestrogen receptor ligand-binding domain as an improved switch for the regulation of heterologous proteins. Nucleic Acids Res 1995; 23: 1686–90.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      Sander EE, van Delft S, ten Klooster JP, et al. Matrix-dependent Tiam1/Rac signaling in epithelial cells promotes either cell-cell adhesion or cell migration and is regulated by phosphatidylinositol 3-kinase. J Cell Biol 1998; 143: 1385–98.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, Tuschl T. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature 2001; 411: 494–8.
      OpenUrlCrossRefPubMed
    26. ↵
      Price LS, Collard JG. Regulation of the cytoskeleton by Rho-family GTPases: implications for tumour cell invasion. Semin Cancer Biol 2001; 11: 167–73.
      OpenUrlCrossRefPubMed
    27. ↵
      Etienne-Manneville S, Hall A. Rho GTPases in cell biology. Nature 2002; 420: 629–35.
      OpenUrlCrossRefPubMed
    28. ↵
      Welch MD. The world according to Arp: regulation of actin nucleation by the Arp2/3 complex. Trends Cell Biol 1999; 9: 423–7.
      OpenUrlCrossRefPubMed
    29. ↵
      Van Leeuwen FN, Kain HE, Kammen RA, Michiels F, Kranenburg OW, Collard JG. The guanine nucleotide exchange factor Tiam1 affects the neuronal morphology; opposing roles for the small GTPases Rac and Rho. J Cell Biol 1997; 139: 797–807.
      OpenUrlAbstract/FREE Full Text
    30. Sarner S, Kozma R, Ahmad S, Lim L. Phosphatidylinosinositol 3-kinase, Cdc42, and Rac1 act downstream of Ras in integrin-dependent neurite outgrowth in N1E-115 neuroblastoma cells. Mol Cell Biol 2000; 20: 158–72.
      OpenUrlAbstract/FREE Full Text
    31. ↵
      Weston CA, Anova L, Rialas C, Prives JM, Weeks BS. Laminin-1 activates Cdc42 in the mechanism of laminin-1-mediated neurite outgrowth. Exp Cell Res 2000; 260: 374–8.
      OpenUrlCrossRefPubMed
    32. ↵
      Caron H, van Sluis P, van Roy N, et al. Recurrent 1;17 translocations in human neuroblastoma reveal nonhomologous mitotic recombination during the S/G2 phase as a novel mechanism for loss of heterozygosity. Am J Hum Genet 1994; 55: 341–7.
      OpenUrlPubMed
    33. ↵
      Savelyeva L, Corvi R, Schwab M. Translocation involving 1p and 17q is a recurrent genetic alteration of human neuroblastoma cells. Am J Hum Genet 1994; 55: 334–40.
      OpenUrlPubMed
    34. ↵
      Palacios F, Schweitzer JK, Boshans RL, D'Souza-Schorey C. ARF6-GTP recruits Nm23-H1 to facilitate dynamin-mediated endocytosis during adherens junctions disassembly. Nat Cell Biol 2002; 4: 929–36.
      OpenUrlCrossRefPubMed
    35. ↵
      Ohira M, Kageyama H, Mihara M, et al. Identification and characterization of a 500-kb homozygously deleted region at 1p36.2-p36.3 in a neuroblastoma cell line. Oncogene 2000; 19: 4302–7.
      OpenUrlCrossRefPubMed
    36. ↵
      Wakamatsu Y, Watanabe Y, Nakamura H, Kondoh H. Regulation of the neural crest cell fate by N-myc: promotion of ventral migration and neuronal differentiation. Development 1997; 124: 1953–62.
      OpenUrlAbstract
    37. ↵
      Gotta M, Abraham MC, Ahringer J. CDC-42 controls early cell polarity and spindle orientation in C. elegans. Curr Biol 2001; 11: 482–8.
      OpenUrlCrossRefPubMed
    38. ↵
      Etienne-Manneville S, Hall A. Cdc42 regulates GSK-3β and adenomatous polyposis coli to control cell polarity. Nature 2003; 421: 753–6.
      OpenUrlCrossRefPubMed
    39. ↵
      Richman TJ, Sawyer MM, Johnson DI. The Cdc42p GTPase is involved in a G2/M morphogenetic checkpoint regulating the apical-isotropic switch and nuclear division in yeast. J Biol Chem 1999; 274: 16861–70.
      OpenUrlAbstract/FREE Full Text
    40. ↵
      Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. Serial analysis of gene expression. Science 1995; 270: 484–7.
      OpenUrlAbstract/FREE Full Text
    Back to top
    Cancer Research: 65 (8)
    April 2005
    Volume 65, Issue 8

    Sign up for alerts

    View this article with LENS

    Open full page PDF
    Article Alerts
    Email Article
    Citation Tools
    Inhibition of a New Differentiation Pathway in Neuroblastoma by Copy Number Defects of N-myc, Cdc42, and nm23 Genes
    Linda J. Valentijn, Arjen Koppen, Ronald van Asperen, Heather A. Root, Franciska Haneveld and Rogier Versteeg
    Cancer Res April 15 2005 (65) (8) 3136-3145; DOI: 10.1158/0008-5472.CAN-04-2469
    del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

    Jump to section

    Advertisement