Abstract
Doublecortin-like kinase-long (DCLK-long) and doublecortin-like (DCL) are two splice variants of DCLK gene. DCL and DCLK-long are microtubule-associated proteins with specific expression in proliferative neural progenitor cells. We have tested the hypothesis that knockdown of DCL/DCLK-long by RNA interference technology will induce cell death in neuroblastoma (NB) cells. First, we analyzed the expression of DCL and DCLK-long in several human neuroblastic tumors, other tumors, and normal tissues, revealing high expression of both DCL and DCLK-long in NB and glioma. Secondly, gene expression profiling revealed numerous differentially expressed genes indicating apoptosis induction after DCL/DCLK-long knockdown in NB cells. Finally, apoptosis was confirmed by time-lapse imaging of phosphatidylserine translocation, caspase-3 activation, live/dead double staining assays, and fluorescence-activated cell sorting. Together, our results suggest that silencing DCL/DCLK-long induces apoptosis in NB cells.
Introduction
Neuroblastoma (NB) is a pediatric tumor arising from immature sympathetic neuroblast cells (Maris & Matthay 1999). It is the most common solid cancer in childhood and the second highest cause of cancer deaths in children (Maris et al. 2007). NB exhibits characteristics of immature sympathetic neuroblasts (Brodeur 2003). NBs contain a mixture of neuroblastic and neuroendocrine cell types that are organized in lobular structures with a central necrotic zone (Jogi et al. 2002, Poomthavorn et al. 2009). This pediatric tumor presents a broad spectrum of clinical behaviors. A subset of tumors undergo spontaneous regression, while others show relentless progression (Tang et al. 2006, Castel et al. 2007, Maris et al. 2007). About half of all cases are classified as high-risk, with overall survival rates below 40%, despite intensive multimodal therapy (Maris et al. 2007). Microtubule-destabilizing agents, such as Vinca alkaloids, are used in NB treatment. However, NB patients develop pharmacoresistance to these chemotherapeutic agents, and systemic toxicity also occurs, which make NB difficult to treat (Don et al. 2004).
Studies have shown that microtubule-destabilizing agents block mitosis primarily by inhibiting the dynamics of spindle microtubules, leading to mitotic arrest (Jordan et al. 1992, Lobert et al. 1999). This arrest induces mitochondrial permeability transition, release of pro-death molecules into the cytosol, and caspase-dependent apoptosis of neoplastic cells (Bhalla 2003). Different mechanisms have been highlighted linking mitotic arrest to the initiating events of the mitochondrial apoptosis pathway. These initiating events include either the direct inactivation of the antiapoptotic Bcl2 by phosphorylation, or the activation of the pro-apoptotic molecules Bax and Bad, which in turn inactivate Bcl2 or Bcl-xL (Konishi et al. 2002, Yamaguchi & Wang 2002, Bhalla 2003).
Since NBs derive from proliferating neuroblasts, the study of genes involved in mitotic spindle formation in neuroblast is of specific interest to find new intervention points for NB treatment. We and others have recently identified and characterized one of such genes, doublecortin-like kinase (DCLK). DCLK is a member of the doublecortin (DCX) gene family (Coquelle et al. 2006, Reiner et al. 2006) and regulates neurogenesis (Shu et al. 2006, Vreugdenhil et al. 2007), neuronal migration (Koizumi et al. 2006), retrograde transport of glucocorticoid receptors (GR; Fitzsimons et al. 2008), and mitotic spindle formation in neuroblasts (Shu et al. 2006, Vreugdenhil et al. 2007). The genomic organization of the DCLK gene is rather complex and gives rise to numerous splice variants. The main splice variants encoded by DCLK gene are doublecortin-like (DCL), DCLK-long, DCLK-short, and calcium/calmodulin-dependent protein kinase (CaMK)-related peptide (CARP; for the functional domains of these splice variants see Supplementary Figure 1, see section on supplementary data given at the end of this article). DCLK-long and DCL contain two DCX domains (Gleeson et al. 1999, Burgess & Reiner 2000, Vreugdenhil et al. 2007), whereas DCLK-long and DCLK-short contain an additional CaMK-like domain (Schenk et al. 2007). CARP lacks both DCX and CaMK-like domains (Vreugdenhil et al. 1999). The microtubule-associated proteins (MAPs), DCL and DCLK-long, exhibit high homology with DCX (Shu et al. 2006, Vreugdenhil et al. 2007). DCX is a MAP involved in the regulation of microtubule dynamics, neuronal migration, and positioning in the neocortex (Bai et al. 2003). During embryonic development, both DCL and DCLK-long are expressed specifically in areas of high neuroblast proliferation but not in other proliferative tissues (Vreugdenhil et al. 2007). Silencing of DCL/DCLK-long by RNA interference leads to disruption of the mitotic spindles and arrests the cells at prometaphase (Shu et al. 2006, Vreugdenhil et al. 2007). Interestingly, DCLK-long, DCLK-short, and CARP have been linked to apoptosis (Burgess & Reiner 2001, Kruidering et al. 2001, Schenk et al. 2007).
Here, we study the expression and the consequences of DCL/DCLK-long knockdown in NB cells. We show for the first time that there is a specific expression of DCL and DCLK-long in human NBs and profound apoptosis induction after their knockdown, suggesting that these DCLK gene splice products, which are specifically expressed in proliferative neuroblasts, are possible future therapeutic targets for the treatment of NB.
Materials and methods
Cell culture and transfection
Mouse N1E-115 and human SH-SY5Y NB cells were cultured as described (Vreugdenhil et al. 2007, Molenaar et al. 2008). Cells were grown in 24-well plates (Corning Life Sciences BV, Amsterdam, The Netherlands) coated with 100 ng/μl poly-l-lysine (Sigma–Aldrich, Inc). SH-SY5Y cells were seeded in plates coated with 100 μg/ml poly-d-lysine. For microscopy, both cell lines were grown in 24-well plates with glass bottom (Greiner Bio-One BV, Alphen aan den Rijn, The Netherlands) coated with 200 ng/μl poly-l-lysine (200 μg/ml poly-d-lysine for SH-SY5Y cells) at 80% confluence. The transfection of N1E-115 cells and the siRNAs (siDCL-1, siDCL-2, and siDCL-3) used were described by Vreugdenhil et al. 2007. SH-SY5Y cells were transfected with 200 nM siDCLK-4 (GCCCACUGCAGCUUCUACCTT-sense and GGUAGAAGCUGCAGUGGGCTT-antisense) and 200 nM siDCLK-5 (UGGAGUACACCAAGAAUGTT-sense and CAUUCUUGGUGUACUCCATT-antisense) siRNAs using lipofectamine 2000 (Invitrogen) as described in the manufacturer's protocol. AllStars Negative Control siRNA from Qiagen was used as negative control (NC). A transfection efficiency of 95±5% was obtained, which was determined by quantifying the percentage of transfected cells with a non-targeting siRNA conjugated to FITC (Qiagen).
Protein extraction and western blots
Protein extraction, SDS-PAGE, and western blotting were performed as previously (Vreugdenhil et al. 2007). A previously described anti-CAMKLK1 antibody was used to detect DCL (Kruidering et al. 2001, Vreugdenhil et al. 2007, Boekhoorn et al. 2008, Tuy et al. 2008). CPG16 CaM Kinase VI antibody (Becton Dickinson BV, Breda, The Netherlands) was used for detecting DCLK-long. Relative optical densities were analyzed and quantified using ImageJ software (http://rsbweb.nih.gov/ij/; Abramoff et al. 2004) on images obtained from three independent blots. Alkaline phosphatase assay was performed using 30 U of alkaline phosphatase for 30 min at 30 °C, followed by the addition of 10 mM sodium pyrophosphate as inhibitor (Francis et al. 1999).
Development of doxycycline-inducible stable cell line
N1E-115 cells (1×106) were transfected with 5 μg of a short hairpin RNA (shRNA) expression vector for DCL (TaconicArtemis, Cologne, Germany; Seibler et al. 2007) by cell electroporation using Amaxa Nucleofactor system (amaxa GmbH, Cologne, Germany) as described in the manufacturer's protocol. The medium was changed daily and replaced by medium containing 500 μg/ml G418 (Geneticin; Invitrogen). Two weeks after transfection, single clones were picked by standard procedures. The inducible DCL knockdown was tested by western blotting in samples from cells treated with 1 μg/ml doxycycline (Dox) for 3 days, refreshing the medium daily.
RNA isolation
RNA isolation, the concentration measurement, and the integrity determination were performed as described by Dijkmans et al. 2008.
Gene expression profiling from human samples and cell lines
The NB tumor panel used for Affymetrix HG-U133 Plus 2.0 Microarray analysis contains 88 NB samples derived from primary tumors of untreated patients. mRNA isolation and profiling methods were described previously (Molenaar et al. 2008). The expression data were normalized with the MAS5.0 algorithm within the Affymetrix's GCOS program. Target intensity was set to 100 (α1=0.04 and α2=0.06). The mRNA profiles of two other NB panels, adult tumors, and normal tissues are publicly available and taken from the National Cancer Institute (NCI) Gene Expression Omnibus (GEO) database (GSE2109 (https://expo.intgen.org/geo/listPublicGeoTransactions.do) GSE4290 (http://rembrandt-db.nci.nih.gov), GSE2658 (http://lambertlab.uams.edu)). All data were analyzed using the R2 bioinformatic tool (http://r2.amc.nl, J Koster, personal communication). To identify correlating genes with the two DCLK splice variants in the 88-NB panel, the r value is calculated for all genes in the human genome (log2-transformed data are used). The significance of finding a certain correlation (P value) was calculated by the following formula: 
Gene expression profiling from mouse NB N1E-115 cells
Sample preparation, hybridization to microarray, and detection were performed as described (Dijkmans et al. 2008). Raw signals were converted to expression values by Expression Array System Analyzer Software Version 1.1.1 (Applied Biosystems, Nieuwerkerk ad Ijssel, The Netherlands). Out of all 33 012 genes on the array, 14 076 (43%) had signal-to-noise ratio higher than 3, and were regarded as expressed in the N1E-115 cells. Subsequently, a quantile-normalization step was performed, and probe-to-gene annotation release version 12_05 was used for gene annotation. BRB-array software tools (Simon & Lam 2006) were used to identify genes that were differentially expressed among classes using a random-variance t-test (Wright & Simon 2003). Genes were considered statistically significant if their P value was <0.001 and their false discovery rate was lower than 0.015, according to Benjamini & Hochberg (1995). We performed biological pathway analysis for the genes with P<0.001. Differentially expressed GO groups of genes were identified as follows. For each GO group, the number n of genes represented on the microarray in that group was calculated, and subsequently, the Fisher (LS) statistic and Kolmogorov–Smirnov (KS) statistic were performed as described (Simon & Lam 2006). A GO category is regarded significantly differentially regulated if either significance level was <0.01. All GO categories with between 5 and 100 genes represented on the array were considered. Differentially expressed Biocarta pathways were identified using the Hotelling T-square test.
Quantitative real-time PCR
Quantitative real-time PCR (RT-qPCR) was carried out by TaqMan technology using a Universal Probe Library (Roche) following the manufacturer's protocol.
Live/dead double staining and caspase-3 activation assays
Forty-eight hours after N1E-115 and SH-SY5Y NB cell transfection with siRNAs or induction for 3 days with Dox, live/dead double staining assay (Calbiochem, San Diego, CA, USA) was performed as described in the manufacturer's protocol. To detect caspase-3 activation, N1E-115 cells were imaged in the presence of ‘Nuncview’ Alexa-488-labeled caspase-3 substrate (Biotium Inc., Hayward, CA, USA) as indicated in the manufacturer's protocol. N1E-115 cells treated with 20 nM staurosporine (STS; Sigma–Aldrich, Inc.) for 3 h were used as positive control. STS is a kinase inhibitor known to induce apoptosis (Lopez & Ferrer 2000). SH-SY5Y human NB cells incubated for 5 h with 500 nM STS were used as positive control. Differential interference contrast (DIC) and fluorescence imaging were performed on a Nikon TE-2000 E system under 37 °C and 5% CO2 controlled conditions.
Time-lapse imaging of phosphatidylserine translocation
Forty-eight hours after N1E-115 cells had been transfected, time-lapse imaging of phosphatidylserine (PS) translocation was performed as described recently (Puigvert et al. 2009, 2010) for a period of 19 h. Cells treated with 20 nM STS (Sigma–Aldrich, Inc.) for 3 h were used as positive control. DIC and fluorescence imaging were performed on a Nikon TE-2000 E system under 37 °C and 5% CO2 controlled conditions.
Image analysis and cell counting
ImageJ software (Abramoff et al. 2004) was used for image analysis and cell counting. Time-lapse images were processed with Image-Pro Plus (Version 5.1; Media Cybernetics).
Fluorescence-activated cell sorting
Fluorescence-activated cell sorting (FACS) analysis was performed as described previously with some changes (Puigvert et al. 2009). Forty-eight hours after transfection, cells seeded in 24-well plates were trypsinized, washed with PBS, resuspended in PBS/EDTA (4 mM), and fixed in 67% ethanol overnight at −20 °C. Cells were stained with FACS staining solution (1 mg/ml propidium iodide, 10 mg/ml RNase A, and PBS) for 45 min in the absence of light at room temperature. After resuspension, 5000–10 000 cells were analyzed by flow cytometry on a FACSCalibur (Becton Dickinson). The CellQuest software (Becton Dickinson) was used for data analysis.
Statistical analysis
Unless otherwise indicated, assays were carried out for three independent experiments run in triplicates. Results are expressed as mean±s.e.m. Where appropriate, Student's t-test was done, and P<0.05 was considered statistically significant.
Results
DCL and DCLK-long mRNAs are highly expressed in human NBs
To evaluate mRNA expression levels of DCL and DCLK-long in our previously described neuroblastic tumor panel (Molenaar et al. 2008), we used the bioinformatic platform R2 identifying their expression levels in different tumors, in normal tissues, and in cell lines (http://r2.amc.nl, J Koster, personal communication). It also allowed us to predict their involvement in specific signal transduction routes. This application uses a database that contains 19 438 microarrays of 20 845 tumors and normal tissue samples. First, we analyzed the expression of DCLK in neuroblastic tumors (Fig. 1). The R2 platform is linked to the TranscriptView web application (Valentijn et al. 2006), which we used to identify the position of the probesets on transcription variants (Fig. 1A). There are two Affymetrix 133U plus2 microarray probesets annotated to the DCLK gene locus. Probeset 205399_at targets exon 20 of the DCLK gene, which is only transcribed in DCLK-long and DCLK-short variants (Vreugdenhil et al. 2001; Fig. 1A). Probeset 229800_at targets exon 8, which is transcribed in DCL and CaMK-CARP transcripts (Vreugdenhil et al. 2007; Fig. 1A). DCLK-short and CARP have been detected in adult neuronal cells but not in neuroblasts (Vreugdenhil et al. 2001, Burgess & Reiner 2002, Engels et al. 2004). Since exon 8 and exon 20 are mutually exclusive, we can use the corresponding probesets to separately analyze the expression of DCL and DCLK-long transcriptional variants. We compared the expression of both transcripts in the NB datasets with expression in various other tumors and normal tissues. The bar plot of Fig. 1B shows an increased DCLK-long expression in NB compared with various adult tumor types and normal non-nervous tissues. Expression in NBs was in the same range as the expression in adult central nervous system (CNS), as expected, since the probeset also recognizes other DCLK splice variants (Fig. 1A), highly expressed in different areas of the adult brain (Vreugdenhil et al. 2001, Burgess & Reiner 2002, Engels et al. 2004). With the exon 8 probeset, we identified abundant DCL expression in NB compared with other tissue types. Interestingly, only gliomas showed comparable expression levels (Fig. 1C). To analyze a putative correlation between the two probesets, we used gene expression profiles from 88 NBs (see Materials and methods for statistical tests). No significant correlation was found between DCL and DCLK-long expression, as measured with exon 8 or exon 20 probesets, suggesting that both splice variants may have different expression profiles as expected from their embryonic expression (Lin et al. 2000, Vreugdenhil et al. 2007, Boekhoorn et al. 2008).
Expression analysis of two splice variants of DCLK gene in neuroblastomas and other tissues. (A) Analysis of the Affymetrix probesets using transcript view shows the known expressed sequence tags (ESTs) of the DCLK1 gene locus. The probes target exon 8 (probeset 229800_at) and exon 20 (probeset 205399_at). RefSeq: reference sequence; TU Current: currently known Transcriptional Units, Hg_u133p2: probesets. The source for the public available data is given in the Materials and methods section. (B and C) Average microarray mRNA expression levels of DCLK-long (B) and DCL (C) splice variants in various adult tumor types (blue) and normal tissues samples (green) compared three independent neuroblastoma tumor series (red/pink). The number in brackets for each tissue type indicates the number of samples. (D) Western blotting of DCL expression at variable levels in neuroblastoma cell lines (D3, D4, D5, and D7), in other cell lines (D1 and D2), and in human primary neuroblastomas (D8–D13). D1, COS-1 cells; D2, Hela cells; D3, NG108-15 cells; D4, NS20Y cells; D5, N1E-115 cells; D6, marker; D7, SH-SY5Y cells. (E) Confirmation of differential DCL phosphorylation isoforms in NG108-15 cells. The higher molecular weight band visible in endogenous and transfected DCL corresponds to a phosphorylated form of DCL as shown by an alkaline phosphatase assay. This band is not observed in the presence of sodium pyrophosphate (Na pyroph), a phosphatase inhibitor, added prior to the phosphatase. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
To estimate the signal transduction pathways in which DCLK variants are involved, we searched for genes with correlating expression patterns. This analysis revealed 1206 genes with a significant correlation (P<0.01) with DCLK-long. This gene set exhibits enrichment of genes involved in microtubule-based processes and axon projection (see Supplementary Table 1, see section on supplementary data given at the end of this article). The same analysis for DCL showed 880 genes with a significant correlation (P<0.01). Interestingly, this correlation was most significant for GO clusters involved in mitochondrial respiratory chain processes (see Supplementary Table 2, see section on supplementary data given at the end of this article), suggesting a link between DCL and mitochondria. In silico analysis, using PSORT II software (http://psort.ims.u-tokyo.ac.jp/; Nakai & Horton 1999), of the subcellular localization of human DCL also predicts that 17.4% of this MAP is located in mitochondria (see Supplementary Table 3, see section on supplementary data given at the end of this article).
DCL and DCLK-long proteins are expressed in human NBs and in NB cell lines
The above-mentioned experiments provided important information on the expression of DCL and DCLK-long mRNA in NBs. Western blotting showed expression of the DCL protein in different human NBs (Fig. 1D), validating the observation done at the mRNA level. Moreover, DCL and DCLK-long proteins were also detected in mouse and human NB cell lines (Figs 1D and 2), and were not observed in non-NB cell lines (Fig. 1D). In Fig. 1E, we demonstrate that the double band detected in the NG108-15 cell line represents differentially phosphorylated DCL isoforms, as described previously by other authors (Friocourt et al. 2003, Tuy et al. 2008). The molecular weight values estimated for the two DCL bands are in high correlation with those previously described in the literature.
DCLK-long and DCL silencing in transfected N1E-115 mouse neuroblastoma cells at 48 h after transfection with synthetic siRNAs siDCL-2, siDCL-3, and negative control siRNA. (A) Western blotting results of DCL and DCLK-long expression. (B) Expression of DCL and DCLK-long at protein and mRNA levels. NC, negative control. The protein expression was normalized to α-tubulin, and the mRNA was normalized to GAPDH. Columns, mean of three independent experiments (n=6); bars, s.e.m. *P<0.05. **P<0.01. ***P<0.001.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
Synthetic modified siRNAs silence DCL/DCLK-long in mouse NB cells
To study the consequences of DCL/DCLK-long knockdown, a mouse N1E-115 NB cell line that endogenously expresses these MAPs was used. Three previously described and validated synthetic siRNAs were utilized (Vreugdenhil et al. 2007); two of them, siDCL-2 and siDCL-3, effectively knocked down DCL, while the third one, siDCL-1, was not effective. In parallel, a synthetic non-targeting siRNA (AllStars Negative Control siRNA, Qiagen) was used as an independent NC. Since no significant differences were found between the two NC siRNAs (see Supplementary Figure 2, see section on supplementary data given at the end of this article), we present only the results obtained with the NC siDCL-1 (indicated as NC in the figures). Both siDCL-2 and siDCL-3 silenced DCL more effectively than they silenced DCLK-long at the protein level (Fig. 2). Nevertheless, the knockdown detected by RT-qPCR was ∼50% for both MAPs (Fig. 2B), suggesting the existence of posttranslational regulatory mechanisms.
Apoptotic pathways are affected after DCL/DCLK-long knockdown
To investigate the effect of the knockdown of DCL/DCLK-long at the molecular level, we have used gene expression profiling of N1E-115 cells. Hierarchical gene cluster analysis showed clustering of siDCL-3 and siDCL-2 samples versus samples of NC (Fig. 3A). Comparing siDCL-2 and NC, samples resulted in the identification of 1034 differentially expressed genes, while 931 differentially expressed genes were identified with siDCL-3. Of these differentially expressed genes, 663 genes were in common (Fig. 3B). The majority of these 663 genes were up-regulated (562) and 101 were down-regulated (Fig. 3B). Pathway analysis resulted in the identification of significant overrepresented pathways related to cell cycle, oxidative stress, and apoptosis (Table 1). Pax6 was one of the most up-regulated genes in our expression profiling studies. The up-regulation of Pax6 has been linked to the inactivation of neuroblast proliferation, apoptosis, and acquisition of neuronal cell fate (Berger et al. 2007). Moreover, Bax is an example of an apoptotic inducer (Nutt et al. 2002) found up-regulated in the affected pathways. Oxidative phosphorylation and ATP synthesis were among the most affected biological processes (see Supplementary Table 5, see section on supplementary data given at the end of this article). Genes such as Ndufa1 (Mamelak et al. 2005) and Cox7c (Lenka et al. 1998) were up-regulated (Fig. 3C and Supplementary Table 4, see section on supplementary data given at the end of this article), indicating an active oxidative phosphorylation process in cells with decreased DCL/DCLK-long expression. Moreover, mitochondria were among the most affected cellular components (see Supplementary Table 5). Differential expression of several selected genes was confirmed by RT-qPCR (see Supplementary Table 6, see section on supplementary data given at the end of this article). Together, these data suggest that DCL/DCLK-long knockdown leads to apoptosis.
mRNA expression profiling of N1E-115 mouse neuroblastoma cells at 48 h after transfection. Cells were transfected with siDCL-2, siDCL-3, and negative control (NC) siRNAs. (A) Hierarchic clustering of the mRNA expression profiling in the different groups; green indicates reduced expression and red indicates induced expression. (B) Venn diagram highlighting the overlap of differentially expressed genes between negative control and siDCL-2 groups (N2) and negative control and siDCL-3 groups (N3). The total number of up- and down-regulated genes is indicated. (C) Normalized log-transformed gene expressions for a selection of the overlapping 663 genes. Blue, low normalized log-transformed gene expression; red, high normalized log-transformed gene expression. Microarray analyses were performed using four biological replicates (n=4) per condition. One biological replicate of the negative control group and one of siDCL-3 group were excluded from the analysis because they did not fulfill the microarray quality control criteria. The analysis was performed for a P value lower than 0.001 and a false discovery rate (FDR) lower than 0.015. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
Examples of affected pathways in mouse N1E-115 neuroblastoma cells with doublecortin-like (DCL) and doublecortin-like kinase-long (DCLK-long) knockdown. Sixty-eight pathways (Biocarta pathways) were significant at the nominal 0.0005 level of the Hotelling T-square test
| Biocarta pathway | Hotelling's test P value | Modulated genes |
|---|---|---|
| Protein kinase A at the centrosome | 2.00×10−7 | Ppp2r1a (↓↓); Prkar2a (↓↓); Cyp51 (↓↓); Rhoa (↓↓); Akap9 (↑↑); Prkar2b (↓↓); Prkaca (↑↓); Prkce (↓↓) |
| Regulation of Bad phosphorylation | 7.40×10−6 | Ywhah (↓↓); Bax (↑↑); Mapk1 (↓↓); Igf1 (↓↑); Asah1 (↓↓); Bad (↑↑); Pik3r1 (↑↓); Mapk3 (↓↓); Akt1 (↓↓); Rps6ka1 (↓↓); Bcl2l1 (↓↓); Prkaca (↑↓); Bcl2 (↓↓) |
| AKAP95 role in mitosis and chromosome dynamics | 1.26×10−5 | Ddx5 (↓↓); Ppp2r1a (↓↓); Prkar2a (↓↓); Akap8 (↓↓); Prkar2b (↓↓); Prkaca (↑↓) |
| Control of skeletal myogenesis by HDAC and calcium/calmodulin-dependent kinase (CaMK) | 2.12×10−5 | Ywhah (↓↓); Igf1r (↓↑); Ppp3ca (↓↓); Pik3r1 (↑↓); Akt1 (↓↓); Hdac5 (↓↑); Mapk7 (↓↑); Camk2a (↑↑); Mapk14 (↓↓); Mef2a (↓↓); Calm1 (↓↑); Myod1 (↑NA) |
| Multiple antiapoptotic pathways from IGF1R signaling lead to Bad phosphorylation | 2.43×10−5 | Ywhah (↓↓); Mapk1 (↓↓); Igf1r (↓↑); Raf1 (↓↓); Asah1 (↓↓); Pik3ca (↓↓); Grb2 (↓↓); Bad (↑↑); Irs1 (↓↓); Map2k1 (↓↓); Mapk3 (↓↓); Akt1 (↓↓); Rps6ka1 (↓↓); Ppp1r13b (↓↑); Sos1 (↑↑); Hras1 (↑↑); Prkaca (↑↓); Shc1 (↓↓) |
| Oxidative stress induced gene expression via Nrf2 | 2.65×10−5 | Hmox1 (↑↓); Sirt7 (↑↑); Mapk1 (↓↓); Jun (↑↑); Atf4 (↓↓); Nfe2l2 (↓↓); Mafg (↑↑); Creb1 (↓↓); Cryz (↓↓); Keap1 (↑↓); Fos (↑↑); Prkcb1 (↓↓); Por (↑↑); Mapk14 (↓↓); Maff (↑↑); Mafk (↓↓) |
| Cell cycle: G2/M checkpoint | 3.10×10−5 | Ywhah (↓↓); Chek1 (↓↓); Cdc34 (↑↑); Myt1 (↓↓); Cdkn1a (↓↓); Wee1 (↓↓); Brca1 (↓↓); Gadd45a (↑↑); Mdm2 (↓↓); Rps6ka1 (↓↓); Atm (↑↑); Atr (↓↓); Cdc25c (↓↓); Trp53 (↓↓); Cdkn2d (↑↑); Plk1 (↓↓); Prkdc (↓↓); Ywhaq (↓↓) |
| Role of Ran in mitotic spindle regulation | 3.83×10−5 | Kpna2 (↓↓); Aurka (↓↓); Kpnb1 (↓↓); Tpx2 (↓↓); Kif15 (↑↓); Rcc1 (↓↓); Ranbp1 (↓↓); Rangap1 (↓↓) |
| P53 signaling pathway | 1.63×10−4 | Ccnd1 (↓↓); Bax (↑↑); Cdkn1a (↓↓); Gadd45a (↑↑); Mdm2 (↓↓); Cdk4 (↑↓); Ccne1 (↑↑); Atm (↑↑); Pcna (↓↓); E2f1 (↑↑); Trp53 (↓↓); Apaf1 (↓↓); Bcl2 (↓↓) |
| Regulation of MAP kinase pathways through dual specificity phosphatases | 2.05×10−4 | Dusp2 (↓↓); Dusp6 (↓↓); Dusp1 (↓↓); Mapk3 (↓↓); Dusp9 (↓↓); Mapk14 (↓↓); Dusp8 (↑↑); Dusp4 (↑↑) |
| Apoptotic signaling in response to DNA damage | 3.63×10−4 | Bax (↑↑); Bid (↓↓); Cycs (↑↑); Bad (↑↑); Akt1 (↓↓); Atm (↑↑); Casp3 (↑↑); Tln1 (↑↑); Bcl2l1 (↓↓); Prkcb1 (↓↓); Prp1 (↓↓); Casp9 (↓↑); Trp53 (↓↓); Casp7 (↓↓); Apaf1 (↓↓); Bcl2 (↓↓); Stat1 (↑↑) |
| Apoptotic DNA fragmentation and tissue homeostasis | 4.64×10−4 | Hmgb1 (↓↓); Top2a (↓↓); Dffa (↓↓); Endog (↑↑); Hmgb2 (↑↑); Casp3 (↑↑); Casp7 (↓↓); Cad (↓↓) |
(↑↑), up-regulated in both comparisons (negative control versus siDCL-2 and negative control versus siDCL-3); (↓↓), down-regulated in both comparisons; (↑↓), up-regulated in the comparison negative control versus siDCL-2 and down-regulated in the comparison negative control versus siDCL-3; (↓↑), down-regulated in the comparison negative control versus siDCL-2 and up-regulated in the comparison negative control versus siDCL-3; (↑NA), up-regulated in the comparison negative control versus siDCL-2 and not altered for the second comparison.
Silencing of DCL/DCLK-long leads to apoptosis in N1E-115 NB cells
Since the above-described microarray results suggest apoptosis induction by DCL/DCLK-long knockdown in NB cells, we performed biochemical assays to investigate this possibility.
Time-lapse imaging of PS translocation (Puigvert et al. 2009, 2010) showed a significant difference between the NC and cells transfected with the effective siRNAs at the different time points (Fig. 4A and B and Supplementary Video 1, see section on supplementary data given at the end of this article). After counting the number of cells presenting FITC-labeled Annexin-V conjugated to PS at different time points, we identified an increase of PS translocation to the outer membrane in cells with DCL/DCLK-long knockdown (Fig. 4A and B), showing an increase of apoptosis in these cells. At the beginning of the assay (48 h after transfection), 10.33±1.20% apoptotic cells were quantified for siDCL-2, and 16.71±5.07% apoptotic cells were quantified for siDCL-3, while 6.93±0.90% apoptotic cells were detected in the NC (Fig. 4B). Eighteen hours after starting the assay (66 h after transfection), 79.92±0.93% cells transfected with siDCL-2 and 89.18±5.32% cells transfected with siDCL-3 were positive for FITC-labeled Annexin-V conjugated to PS. These values were significantly higher (P<0.05) than those in the NC (51.82±3.08%; Fig. 4B). Our results indicate that DCL/DCLK-long knockdown leads to apoptosis.
Apoptosis studies in mouse N1E-115 neuroblastoma cells at 48 h after transfection. (A and B) Time-lapse imaging of phosphatidylserine translocation. Images were taken at 30 min interval (see Supplementary Video 1). (A) Time-lapse imaging 0, 9 and 18 h after starting the assay. (B) Percentage of cells with translocated phosphatidylserine at different time points for the different treatments. The initial time point of the assay (0 h) corresponds to 48 h after transfection. (C) Live/dead double staining. Viable cells are stained with a cell-permeable green fluorescent cyto-dye and dead cells are stained with both cyto-dye (green) and propidium iodide (red). (D) Caspase-3 activation assay. Bar graph shows the percentage of cells with active caspase-3. STS, staurosporine. NC, negative control. Overlap of DIC and fluorescent imaging were used in the different assays. 20× magnification. Scale bars, 50 μm. Data points and Columns, mean of two independent experiments (n=6); bars, s.e.m. *P<0.05. ***P<0.001. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
We also performed double staining assays to discriminate between live and dead cells (Fig. 4C; Balcer-Kubiczek et al. 2006). In line with our PS translocation studies, DCL/DCLK-long knockdown leads to a significantly higher (P<0.05) number of dead cells 48 h after transfection with the two effective siRNAs. 22.01±1.62% N1E-115 cells transfected with siDCL-2 and 18.43±1.31% N1E-115 cells transfected with siDCL-3 presented membrane damage, which was indicated by propidium iodide staining. Significantly less (P<0.05; 9.36±0.90%) NC cells were positive for propidium iodide (Fig. 4C). Using an Alexa-488-labeled caspase-3 substrate (Puigvert et al. 2009), caspase-3 activation was also measured. Compared with the NC (11.53±1.53%), a significant increase in percentage of cells with active caspase-3 was detected when transfected with siDCL-2 (16.83±1.37%; P<0.05) and siDCL-3 (29.61±2.41%; P<0.001; Fig. 4D).
FACS corroborated the effects of DCL/DCLK-long silencing (Fig. 5). For FACS analysis, we used cells transduced with siDCL-3 due to its higher effectiveness in inducing cell death (Fig. 4). We observed a significantly higher (P<0.05) percentage of apoptotic cells (18.45±1.00%) in cells treated with siDCL-3 than in cells treated with the NC (10.39±1.61%). At this time point, no significant differences were detected in cell-cycle progression among the different experimental groups.
Fluorescence-activated cell sorting results of neuroblastoma cells with DCL/DCLK-long knockdown. (A) Histogram representation of cell population 48 h after transfection with siDCL-3 or with the negative control siRNA (NC). Data shown are representative of four independent experiments. (B) Pie graphs of the effect of DCL and DCLK-long knockdown on the distribution of mitotic cells in different phases and on the induction of apoptosis. (C) Bar graphs of cells in apoptosis, in S phase, and in G2-M phase. A significantly higher percentage of apoptotic cells were found in the siDCL-3 group than in the negative control. No significant difference was found between NC and cells with the knockdown (siDCL-3) in the S and G2-M phases of the cell cycling. STS, staurosporine. Columns, mean of four independent experiments (n=4); bars, s.e.m. *P<0.05. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
To validate the specificity of the observed effects, an inducible stable cell line was developed to express specific shRNAs (Fig. 6). First, we attempted to develop a stable cell line with constitutive expression of shRNA against DCL. However, the cells failed to survive, in agreement with the observed effects of DCL knockdown on cell survival using synthetic siRNAs. Nevertheless, DCL knockdown was possible using a Dox-inducible expression of specific shRNAs against DCL. By western blotting, we detected DCL knockdown in cells treated with Dox (88.67±2.68% in colony 1 and 63.84±5.66% in colony 6), while cells treated with vehicle depicted DCL levels comparable to the parental cell line (Fig. 6A and B). Using these stable cell lines, we observed that DCL knockdown induced a significant increase in cell death (P<0.05; Fig. 6C and D). In Dox-treated cells, 23.44±3.39% (colony 1) and 16.82±3.13% (colony 6) of dead cells were detected. In contrast, 11.77±0.13% (colony 1, no Dox), 7.00±3.25% (colony 6, no Dox), 6.00±3.14% (parental cell line with Dox), and 7.55±0.22% (parental cell line no Dox) of dead cells were detected (Fig. 6C and D). These results showed that a higher percentage of cell death was observed in cells that presented a higher DCL knockdown (colony 1, Dox). Moreover, a similar percentage of knockdown obtained with siRNA, ∼90% (Fig. 2), leads to a comparable percentage of cell death, around 20% (Fig. 4C).
DCL knockdown leads to cell death in a neuroblastoma stable cell line with an inducible shRNA expression. (A) Western blotting results of DCL, DCLK-long, and α-tubulin expression in N1E-115 stable cell line with a doxycycline (Dox)-inducible shRNA expression against DCL. An effect on DCL but not on DCLK-long expression was detected. Description of the development of this stable neuroblastoma cell line is provided in Materials and methods. The cells were treated with 1 μg/ml of doxycycline (Dox) or with vehicle (Veh). In the presence of Dox, a specific shRNA for DCL is expressed, leading to DCL knockdown. (B) Quantification results of DCL expression normalized to α-tubulin. For colony 1 and colony 6, a significant difference in DCL expression was found between cells with the induced knockdown and the cells treated with vehicle. Moreover, compared with the cells that do not present the inducible system (NC), no leakage in DCL knockdown due to shRNA expression was detected. (C and D), Live/dead double staining assays reveal an induction of cell death when DCL knockdown is induced in both colonies 1 and 6. In the negative control, no significant difference was found between cells treated with Dox and Veh. NC, negative control (N1E-115 cells). 20× magnification. Scale bars, 50 μm. Columns, mean of two independent experiments (n=6); bars, s.e.m. *P<0.05. **P<0.01. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
Together, our results indicate that the knockdown of the MAPs DCL and DCLK-long induces apoptosis in NB cells. In addition, this process might be through a caspase-3 activity-dependent pathway.
DCL/DCLK-long knockdown in human SH-SY5Y NB cells leads to cell death
To confirm the results obtained in mouse N1E-115 NB cells, we knocked down DCL/DCLK-long in human NB cells. The expression of DCL and DCLK-long was checked in different human NB cell lines by gene expression profiling. To avoid a possible compensation of DCL/DCLK-long function by other members of the DCX family (Koizumi et al. 2006), SH-SY5Y cells were selected. This cell line presents high expression of DCL and DCLK-long and low expression of DCX (Fig. 7A). In addition, this cell line presents a high rate of cell division, allowing us to perform the studies in the same time frame as with mouse NB cells. Using two effective siRNAs, siDCLK-4 and siDCLK-5, we obtained a significant DCL/DCLK-long knockdown (Fig. 7B and Supplementary Figure 3, see section on supplementary data given at the end of this article). We got 71.04±4.42% DCL knockdown with siDCLK-4 and 65.20±0.79% DCL knockdown with siDCLK-5 (Fig. 7B). 72.56±2.08% DCLK-long knockdown was quantified using siDCLK-4 and 52.84±1.63% DCLK-long knockdown was quantified using siDCLK-5 (Fig. 7B). Using live/dead double staining as with mouse NB cells, we found 27.80±1.38% dead cells using siDCLK-4 and 26.30±2.88% dead cells using siDCLK-5 (Fig. 7C and D), which was significantly higher (P<0.01 and P<0.001 respectively) than the 10.64±2.18% detected in cells treated with NC siRNA (Qiagen; Fig. 7C and D). Thus, silencing DCL/DCLK-long by synthetic siRNAs in human SH-SY5Y NB cells induced a significant increase in cell death.
DCL/DCLK-long knockdown in human SH-SY5Y neuroblastoma cells and cell death studies at 48 h after transfection. (A) Average microarray mRNA expression levels of DCLK-long, DCL, and DCX in several human neuroblastoma cells. SH-SY5Y cells (blue) were selected for further experiments since they have high level of DCL and DCLK-long expression, but low level of DCX expression. mycn: green, single copy; red, amplification; orange: overexpression (SHEP21N) or cMyc amplified (SJNB12). (B) Quantification of DCL and DCLK-long protein expression in human SH-SY5Y neuroblastoma cells 48 h after transfection with siDCLK-4, siDCLK-5 or negative control. A visible knockdown of DCL/DCLK-long was obtained. The expression is normalized to α-tubulin. Western blotting is shown in Supplementary Figure 3. (C and D) Live/dead double staining assay showed an induction of cell death in human neuroblastoma cells with DCL/DCLK-long knockdown. (C) Quantification of dead cells at 48 h after transfection. (D) Overlap of DIC and fluorescent imaging of live/dead double stained SH-SY5Y cells. Viable cells are stained with a cell-permeable green fluorescent cyto-dye, and dead cells are stained with both cyto-dye (green) and propidium iodide (red). STS, 500 nM staurosporine. NC, cells transfected with AllStars Negative Control siRNA from Qiagen. 20× magnification. Scale bars, 50 μm. Columns, mean of two independent experiments (n=6); bars, s.e.m. **P<0.01. ***P<0.001. Full color version of this figure available via http://dx.doi.org/10.1677/ERC-09-0301.
Citation: Endocrine-Related Cancer 17, 2; 10.1677/ERC-09-0301
Discussion
In the present work, we demonstrate for the first time the expression of the two MAPs DCL and DCLK-long in human NBs, and using different experimental strategies ranging from gene expression profiling to live-imaging studies, we show that DCL and DCLK-long are crucial for the proliferation and survival of NB cells. Both DCL and DCLK-long, proteins derived from the DCLK gene, are highly expressed in human NBs. Similarly, both are expressed in mouse N1E-115 and human SH-SY5Y NB cells. We demonstrated that silencing of these MAPs by RNA interference leads to apoptosis in mouse and human NB cells. Therefore, our data suggest that DCL and DCLK-long may be a potential therapeutic target for NB.
MAPs play a role in tumor cell resistance to microtubule-destabilizing agents by regulating microtubule dynamics (Bhat & Setaluri 2007). Nevertheless, in some cases, NB patients are treated with combination chemotherapy. Therefore, the development of resistance to microtubule-destabilizing agents, such as Vinca alkaloids, cannot be studied in these patients. However, an increase in microtubule-stabilizing proteins leads to resistance to Vinca alkaloids in NB cell lines (Don et al. 2004). This, in addition to the observation that DCL/DCLK-long silencing in NB cell lines leads to microtubule destabilization (Vreugdenhil et al. 2007), suggests that targeting microtubule-stabilizing proteins, such as DCL/DCLK-long, may reduce the development of chemoresistance to microtubule-destabilizing agents.
The mechanisms by which DCL and DCLK-long stabilize microtubules seem similar to that of the highly homologous DCX (Shu et al. 2006). DCL and DCLK-long as well as DCX have been shown to be crucial for neuronal proliferation, migration, and axonal outgrowth (Deuel et al. 2006, Koizumi et al. 2006, Shu et al. 2006, Vreugdenhil et al. 2007). In agreement with this, we detected a significant correlation between DCLK-long and genes involved in microtubule-based processes and axon projection human NB samples.
Our results show higher expression of DCL and DCLK-long in human NBs compared with various other tumor types and with normal non-nervous tissue. At the protein level, different DCL phosphorylated isoforms were detected in NB cell lines. Further research is needed to confirm the presence of phosphorylated DCL isoforms in NB tumors. Interestingly, DCL was also highly expressed in gliomas, in agreement with our previous findings showing high expression of this MAP in radial glial cells (Vreugdenhil et al. 2007). DCL expression in radial glial cells was shown to be crucial for proliferation and stability of the early radial glial scaffold (Vreugdenhil et al. 2007). Consistently, gliomas are a collection of tumors that occur within the CNS and arise from astrocytes, oligodendrocytes, or their precursors, radial glial cells (Holland 2001, Anthony et al. 2004). Therefore, our results indicate that DCL might also be a target of interest for glioma therapy.
Using two different specific siRNAs, siDCL-2 and siDCL-3, targeting completely different regions of the mRNA, we silenced the expression of DCL/DCLK-long in mouse NB cells, which endogenously expressed these MAPs (Vreugdenhil et al. 2007). Gene expression profiling after knockdown revealed an extensive overlap of the gene expression using the two effective siRNAs. The genes identified in non-overlapping groups might be due to off-target effects or due to distinct potencies of the two siRNAs used in these studies. Using the two siRNAs, siDCL-2 and siDCL-3, we observed coherent phenotype. This strongly suggests that the observed effects may not be due to off-targets of the individual siRNAs. Moreover, cell death due to DCL knockdown was confirmed in a shRNA-inducible stable cell line. A similar percentage of DCL knockdown obtained with synthetic siRNA and Dox-inducible shRNA leads to a comparable percentage of cell death in N1E-115 cells. This also indicates that DCL has a more relevant role in cell survival than DCLK-long, therefore suggesting that the activation of the apoptotic process is related to the microtubule-binding domains of these MAPs and not to the kinase domain of DCLK-long.
Analysis of the genes that were modulated in common by using siDCL-2 and siDCL-3 strongly indicates a specific induction of apoptosis. Consistent with this conclusion, the activation of Pax6, one of the most up-regulated genes in our expression profiling studies, leads to inactivation of neuroblast proliferation, apoptosis, and acquisition of neuronal cell fate (Berger et al. 2007). Oxidative stress pathways were also identified, supporting the idea that silencing DCL/DCLK-long leads the cells toward apoptosis. The induction of apoptosis by oxidative stress via mechanisms that involve mitochondria has been extensively documented (Green & Reed 1998, Nazarewicz et al. 2007). Furthermore, oxidative phosphorylation was found among the most overrepresented biological processes and mitochondria were found to be one of the cell components most affected. Interestingly, down-regulation of genes involved in mitochondrial function and oxidative phosphorylation pathway has been shown to be a consistent feature of many tumors (Mamelak et al. 2005). One of those genes is NDUFA1, which was up-regulated in mouse NB cells by DCL/DCLK-long knockdown. Moreover, a relation between oxidative phosphorylation, mitochondria, and apoptosis has been suggested (Green & Reed 1998). Disruption of electron transport in the oxidative phosphorylation process has been recognized as an early feature of cell death (Green & Reed 1998).
In agreement with these observations, induction of apoptosis in NB cells by silencing of DCL/DCLK-long was confirmed by several assays. A higher PS translocation to the outer membrane and a higher percentage of cells with membrane damage were observed during the knockdown. Also, a higher percentage of apoptotic cells were detected by FACS analysis in the knockdown group. Moreover, an increase in caspase-3 activation was detected in these cells as well.
We have previously shown that cells transfected with siDCL-3 present more disruption of the mitotic spindles (Vreugdenhil et al. 2007), and we identified a higher percentage of cell death in cells transfected with siDCL-3 than in those transfected with siDCL-2. Therefore, our present and previous results suggest that the effectiveness of inducing apoptosis in NB cells may be directly correlated with the level of disruption of mitotic spindles. Previously, we and others have shown that silencing (Shu et al. 2006, Vreugdenhil et al. 2007) or overexpression (Santra et al. 2006, 2009, Shu et al. 2006, Fitzsimons et al. 2008) of MAPs of the DCX family leads to inhibition of cell proliferation. As shown previously, an imbalance in the expression levels of DCL, DCLK-long, or DCX results in aberrant spindle morphology, leading to a comparable disruption in the mitotic progression (Shu et al. 2006, Vreugdenhil et al. 2007). Consistent with this, a link between mitotic arrest due to disruption of mitotic spindles and apoptosis has been described, involving the activation of the pro-apoptotic gene Bax (Bhalla 2003). In our experiments, Bax was up-regulated after DCL/DCLK-long knockdown. In addition, survivin (Bric5) was found down-regulated, suggesting mitotic spindle catastrophe leading to apoptosis (Bhalla 2003).
An alternative explanation for our results might be that intracellular transport of signaling proteins might have been disrupted by DCL/DCLK-long knockdown, which lead the cell toward apoptosis. We have previously shown that DCL regulates GR microtubule-guided intracellular transport in mouse NB cells and in brain neuroblasts (Fitzsimons et al. 2008). GR is known to be transported to mitochondria where it activates specific responsive genes. (Solakidi et al. 2007). Interestingly, mitochondria were the most affected cell components after DCL/DCLK-long knockdown. Furthermore, we found in human NBs a significant correlation between DCL and genes related with mitochondria activity. Interestingly, the relation between DCL and mitochondria function might not be independent from the role of DCL in microtubule stabilization. Mitochondria transport is known to be along microtubules (Morris & Hollenbeck 1995), and connection between microtubules and mitochondrial apoptotic machinery has been proposed (Esteve et al. 2007).
In conclusion, we identified high expression of DCLK-derived MAPs in human NBs. We demonstrate for the first time at the gene expression level and by several cell death assays that DCL/DCLK-long knockdown induces profound apoptosis in mouse and human NB cells. Therefore, our results suggest that the MAPs DCL and DCLK-long, which are specifically expressed in proliferative neuroblasts, might be targets for the treatment of NB.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1677/ERC-09-0301.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by Top Institute Pharma, The Netherlands.
Acknowledgements
The authors thank Saskia Burm, Theo Schouten, Jessica Kamphorst, and Ethan den Boer for technical assistance.
References
Abramoff MD, Magelhaes PJ & Ram SJ 2004 Image processing with ImageJ. Biophotonics International 11 36 – 42.
Anthony TE, Klein C, Fishell G & Heintz N 2004 Radial glia serve as neuronal progenitors in all regions of the central nervous system. Neuron 41 881–890.
Bai J, Ramos RL, Ackman JB, Thomas AM, Lee RV & LoTurco JJ 2003 RNAi reveals doublecortin is required for radial migration in rat neocortex. Nature Neuroscience 6 1277–1283.
Balcer-Kubiczek EK, Attarpour M, Jiang J, Kennedy AS & Suntharalingam M 2006 Cytotoxicity of docetaxel (Taxotere) used as a single agent and in combination with radiation in human gastric, cervical and pancreatic cancer cells. Chemotherapy 52 231–240.
Benjamini Y & Hochberg Y 1995 Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society 57 289–300.
Berger J, Berger S, Tuoc TC, D'Amelio M, Cecconi F, Gorski JA, Jones KR, Gruss P & Stoykova A 2007 Conditional activation of Pax6 in the developing cortex of transgenic mice causes progenitor apoptosis. Development 134 1311–1322.
Bhalla KN 2003 Microtubule-targeted anticancer agents and apoptosis. Oncogene 22 9075–9086.
Bhat KM & Setaluri V 2007 Microtubule-associated proteins as targets in cancer chemotherapy. Clinical Cancer Research 13 2849–2854.
Boekhoorn K, Sarabdjitsingh A, Kommerie H, de Punder K, Schouten T, Lucassen PJ & Vreugdenhil E 2008 Doublecortin (DCX) and doublecortin-like (DCL) are differentially expressed in the early but not late stages of murine neocortical development. Journal of Comparative Neurology 507 1639–1652.
Brodeur GM 2003 Neuroblastoma: biological insights into a clinical enigma. Nature Reviews. Cancer 3 203–216.
Burgess HA & Reiner O 2000 Doublecortin-like kinase is associated with microtubules in neuronal growth cones. Molecular and Cellular Neurosciences 16 529–541.
Burgess HA & Reiner O 2001 Cleavage of doublecortin-like kinase by calpain releases an active kinase fragment from a microtubule anchorage domain. Journal of Biological Chemistry 276 36397–36403.
Burgess HA & Reiner O 2002 Alternative splice variants of doublecortin-like kinase are differentially expressed and have different kinase activities. Journal of Biological Chemistry 277 17696–17705.
Castel V, Grau E, Noguera R & Martinez F 2007 Molecular biology of neuroblastoma. Clinical & Translational Oncology 9 478–483.
Coquelle FM, Levy T, Bergmann S, Wolf SG, Bar-El D, Sapir T, Brody Y, Orr I, Barkai N & Eichele G et al. 2006 Common and divergent roles for members of the mouse DCX superfamily. Cell Cycle 5 976–983.
Deuel TA, Liu JS, Corbo JC, Yoo SY, Rorke-Adams LB & Walsh CA 2006 Genetic interactions between doublecortin and doublecortin-like kinase in neuronal migration and axon outgrowth. Neuron 49 41–53.
Dijkmans TF, van Hooijdonk LW, Schouten TG, Kamphorst JT, Vellinga AC, Meerman JH, Fitzsimons CP, de Kloet ER & Vreugdenhil E 2008 Temporal and functional dynamics of the transcriptome during nerve growth factor-induced differentiation. Journal of Neurochemistry 105 2388–2403.
Don S, Verrills NM, Liaw TY, Liu ML, Norris MD, Haber M & Kavallaris M 2004 Neuronal-associated microtubule proteins class III beta-tubulin and MAP2c in neuroblastoma: role in resistance to microtubule-targeted drugs. Molecular Cancer Therapeutics 3 1137–1146.
Engels BM, Schouten TG, van Dullemen J, Gosens I & Vreugdenhil E 2004 Functional differences between two DCLK splice variants. Brain Research. Molecular Brain Research 120 103–114.
Esteve MA, Carre M & Braguer D 2007 Microtubules in apoptosis induction: are they necessary? Current Cancer Drug Targets 7 713–729.
Fitzsimons CP, Ahmed S, Wittevrongel CF, Schouten TG, Dijkmans TF, Scheenen WJ, Schaaf MJ, de Kloet ER & Vreugdenhil E 2008 The microtubule-associated protein doublecortin-like regulates the transport of the glucocorticoid receptor in neuronal progenitor cells. Molecular Endocrinology 22 248–262.
Francis F, Koulakoff A, Boucher D, Chafey P, Schaar B, Vinet MC, Friocourt G, McDonnell N, Reiner O & Kahn A et al. 1999 Doublecortin is a developmentally regulated, microtubule-associated protein expressed in migrating and differentiating neurons. Neuron 23 247–256.
Friocourt G, Koulakoff A, Chafey P, Boucher D, Fauchereau F, Chelly J & Francis F 2003 Doublecortin functions at the extremities of growing neuronal processes. Cerebral Cortex 13 620–626.
Gleeson JG, Lin PT, Flanagan LA & Walsh CA 1999 Doublecortin is a microtubule-associated protein and is expressed widely by migrating neurons. Neuron 23 257–271.
Green DR & Reed JC 1998 Mitochondria and apoptosis. Science 281 1309–1312.
Holland EC 2001 Progenitor cells and glioma formation. Current Opinion in Neurology 14 683–688.
Jogi A, Ora I, Nilsson H, Lindeheim A, Makino Y, Poellinger L, Axelson H & Pahlman S 2002 Hypoxia alters gene expression in human neuroblastoma cells toward an immature and neural crest-like phenotype. PNAS 99 7021–7026.
Jordan MA, Thrower D & Wilson L 1992 Effects of vinblastine, podophyllotoxin and nocodazole on mitotic spindles. Implications for the role of microtubule dynamics in mitosis. Journal of Cell Science 102 401–416.
Koizumi H, Tanaka T & Gleeson JG 2006 Doublecortin-like kinase functions with doublecortin to mediate fiber tract decussation and neuronal migration. Neuron 49 55–66.
Konishi Y, Lehtinen M, Donovan N & Bonni A 2002 Cdc2 phosphorylation of BAD links the cell cycle to the cell death machinery. Molecular Cell 9 1005–1016.
Kruidering M, Schouten T, Evan GI & Vreugdenhil E 2001 Caspase-mediated cleavage of the Ca2+/calmodulin-dependent protein kinase-like kinase facilitates neuronal apoptosis. Journal of Biological Chemistry 276 38417–38425.
Lenka N, Vijayasarathy C, Mullick J & Avadhani NG 1998 Structural organization and transcription regulation of nuclear genes encoding the mammalian cytochrome c oxidase complex. Progress in Nucleic Acid Research and Molecular Biology 61 309–344.
Lin PT, Gleeson JG, Corbo JC, Flanagan L & Walsh CA 2000 DCAMKL1 encodes a protein kinase with homology to doublecortin that regulates microtubule polymerization. Journal of Neuroscience 20 9152–9161.
Lobert S, Ingram JW & Correia JJ 1999 Additivity of dilantin and vinblastine inhibitory effects on microtubule assembly. Cancer Research 59 4816–4822.
Lopez E & Ferrer I 2000 Staurosporine- and H-7-induced cell death in SH-SY5Y neuroblastoma cells is associated with caspase-2 and caspase-3 activation, but not with activation of the FAS/FAS-L-caspase-8 signaling pathway. Brain Research. Molecular Brain Research 85 61–67.
Mamelak AJ, Kowalski J, Murphy K, Yadava N, Zahurak M, Kouba DJ, Howell BG, Tzu J, Cummins DL & Liegeois NJ et al. 2005 Downregulation of NDUFA1 and other oxidative phosphorylation-related genes is a consistent feature of basal cell carcinoma. Experimental Dermatology 14 336–348.
Maris JM & Matthay KK 1999 Molecular biology of neuroblastoma. Journal of Clinical Oncology 17 2264–2279.
Maris JM, Hogarty MD, Bagatell R & Cohn SL 2007 Neuroblastoma. Lancet 369 2106–2120.
Molenaar JJ, Ebus ME, Koster J, van Sluis P, van Noesel CJ, Versteeg R & Caron HN 2008 Cyclin D1 and CDK4 activity contribute to the undifferentiated phenotype in neuroblastoma. Cancer Research 68 2599–2609.
Morris RL & Hollenbeck PJ 1995 Axonal transport of mitochondria along microtubules and F-actin in living vertebrate neurons. Journal of Cell Biology 131 1315–1326.
Nakai K & Horton P 1999 PSORT: a program for detecting sorting signals in proteins and predicting their subcellular localization. Trends in Biochemical Science 24 34–36.
Nazarewicz RR, Zenebe WJ, Parihar A, Larson SK, Alidema E, Choi J & Ghafourifar P 2007 Tamoxifen induces oxidative stress and mitochondrial apoptosis via stimulating mitochondrial nitric oxide synthase. Cancer Research 67 1282–1290.
Nutt LK, Pataer A, Pahler J, Fang B, Roth J, McConkey DJ & Swisher SG 2002 Bax and Bak promote apoptosis by modulating endoplasmic reticular and mitochondrial Ca2+ stores. Journal of Biological Chemistry 277 9219–9225.
Poomthavorn P, Wong SH, Higgins S, Werther GA & Russo VC 2009 Activation of a prometastatic gene expression program in hypoxic neuroblastoma cells. Endocrine-Related Cancer 16 991–1004.
Puigvert JC, Huveneers S, Fredriksson L, op het Veld M, van de Water B & Danen EHJ 2009 Cross-talk between integrins and oncogenes modulates chemosensitivity. Molecular Pharmacology 75 947–955.
Puigvert JC, de Bont H, van de Water B & Danen EHJ 2010 High throughput live cell imaging of apoptosis. Current Protocols in Cell Biology [in press].
Reiner O, Coquelle FM, Peter B, Levy T, Kaplan A, Sapir T, Orr I, Barkai N, Eichele G & Bergmann S 2006 The evolving doublecortin (DCX) superfamily. BMC Genomics 7 188.
Santra M, Zhang X, Santra S, Jiang F & Chopp M 2006 Ectopic doublecortin gene expression suppresses the malignant phenotype in glioblastoma cells. Cancer Research 66 11726–11735.
Santra M, Santra S, Roberts C, Zhang RL & Chopp M 2009 Doublecortin induces mitotic microtubule catastrophe and inhibits glioma cell invasion. Journal of Neurochemistry 108 231–245.
Schenk GJ, Engels B, Zhang YP, Fitzsimons CP, Schouten T, Kruidering M, de Kloet ER & Vreugdenhil E 2007 A potential role for calcium/calmodulin-dependent protein kinase-related peptide in neuronal apoptosis: in vivo and in vitro evidence. European Journal of Neuroscience 26 3411–3420.
Seibler J, Kleinridders A, Kuter-Luks B, Niehaves S, Bruning JC & Schwenk F 2007 Reversible gene knockdown in mice using a tight, inducible shRNA expression system. Nucleic Acids Research 35 e54.
Shu T, Tseng HC, Sapir T, Stern P, Zhou Y, Sanada K, Fischer A, Coquelle FM, Reiner O & Tsai LH 2006 Doublecortin-like kinase controls neurogenesis by regulating mitotic spindles and M phase progression. Neuron 49 25–39.
Simon R & Lam A 2006 BRB Array Tools Users Guide, version 3.2. Biometric Research Branch, National Cancer Institute. http://linus.nci.nih.gov/brb.
Solakidi S, Psarra AM & Sekeris CE 2007 Differential distribution of glucocorticoid and estrogen receptor isoforms: localization of GRbeta and ERalpha in nucleoli and GRalpha and ERbeta in the mitochondria of human osteosarcoma SaOS-2 and hepatocarcinoma HepG2 cell lines. Journal of Musculoskeletal & Neuronal Interactions 7 240–245.
Tang XX, Zhao H, Kung B, Kim DY, Hicks SL, Cohn SL, Cheung NK, Seeger RC, Evans AE & Ikegaki N 2006 The MYCN enigma: significance of MYCN expression in neuroblastoma. Cancer Research 66 2826–2833.
Tuy FP, Saillour Y, Kappeler C, Chelly J & Francis F 2008 Alternative transcripts of Dclk1 and Dclk2 and their expression in doublecortin knockout mice. Developmental Neuroscience 30 171–186.
Valentijn LJ, Koster J & Versteeg R 2006 Read-through transcript from NM23-H1 into the neighboring NM23-H2 gene encodes a novel protein, NM23-LV. Genomics 87 483–489.
Vreugdenhil E, Datson N, Engels B, de Jong J, van Koningsbruggen S, Schaaf M & de Kloet ER 1999 Kainate-elicited seizures induce mRNA encoding a CaMK-related peptide: a putative modulator of kinase activity in rat hippocampus. Journal of Neurobiology 39 41–50.
Vreugdenhil E, Engels B, Middelburg R, van Koningsbruggen S, Knol J, Veldhuisen B & de Kloet ER 2001 Multiple transcripts generated by the DCAMKL gene are expressed in the rat hippocampus. Brain Research. Molecular Brain Research 94 67–74.
Vreugdenhil E, Kolk SM, Boekhoorn K, Fitzsimons CP, Schaaf M, Schouten T, Sarabdjitsingh A, Sibug R & Lucassen PJ 2007 Doublecortin-like, a microtubule-associated protein expressed in radial glia, is crucial for neuronal precursor division and radial process stability. European Journal of Neuroscience 25 635–648.
Wright GW & Simon RM 2003 A random variance model for detection of differential gene expression in small microarray experiments. Bioinformatics 19 2448–2455.
Yamaguchi H & Wang HG 2002 Bcl-XL protects BimEL-induced Bax conformational change and cytochrome C release independent of interacting with Bax or BimEL. Journal of Biological Chemistry 277 41604–41612.
