(LASP1) in the Metastatic Dissemination of Medulloblastoma

Published OnlineFirst October 5, 2010; DOI: 10.1158/0008-5472.CAN-10-0592

Therapeutics, Targets, and Chemical Biology Cancer Research Role of LIM and SH3 Protein 1 (LASP1) in the Metastatic Dissemination of Medulloblastoma

Christopher Traenka1, Marc Remke2,5, Andrey Korshunov4,6, Sebastian Bender2,5, Thomas Hielscher3, Paul A. Northcott7, Hendrik Witt2,5, Marina Ryzhova8, Jörg Felsberg9, Axel Benner3, Stephanie Riester2, Wolfram Scheurlen10, Thomas G.P. Grunewald11, Andreas von Deimling6, Andreas E. Kulozik5, Guido Reifenberger9, Michael D. Taylor7, Peter Lichter2, Elke Butt1, and Stefan M. Pfister2,5

Abstract Medulloblastoma is the most common malignant pediatric brain tumor and is one of the leading causes of cancer-related mortality in children. Treatment failure mainly occurs in children harboring metastatic tumors, which typically carry an isochromosome 17 or gain of 17q, a common hallmark of intermediate and high-risk medulloblastoma. Through mRNA expression profiling, we identified LIM and SH3 protein 1 (LASP1) as one of the most upregulated genes on chromosome 17q in tumors with 17q gain. In an independent validation cohort of 101 medulloblastoma samples, the abundance of LASP1 mRNA was significantly associated with 17q gain, metastatic dissemination, and unfavorable outcome. LASP1 protein expression was analyzed by immunohistochemistry in a large cohort of patients (n = 207), and high protein expression levels were found to be strongly correlated with 17q gain, metastatic dissemination, and inferior overall and progression-free survival. In vitro experiments in medulloblastoma cell lines showed a strong reduction of cell migration, increased adhesion, and decreased proliferation upon LASP1 knockdown by small interfering RNA–mediated silencing, further indicating a functional role for LASP1 in the progression and metastatic dissemination of medulloblastoma. Cancer Res; 70(20); 8003–14. ©2010 AACR.

Introduction fluid cytology, is a powerful prognostic marker in medulloblastoma (2, 3). Brain tumors and leukemias comprise the most common Molecular signaling pathways involved in the patho- cause of cancer-related mortality in children with medullo- genesis of standard-risk medulloblastoma such as the blastoma being the most common malignant brain tumor WNT and sonic hedgehog (SHH) signaling cascades are in this age group (1). The presence of leptomeningeal dissem- well-established (4–6), whereas the molecular biology ination at diagnosis detected either macroscopically by underlying intermediate and high-risk medulloblastoma craniospinal imaging or microscopically by cerebrospinal remains elusive. Aberrant activation of WNT signaling, typically caused by activating CTNNB1 mutations (and rarely APC or AXIN), followed by nuclear accumulation of CTNNB1 Authors' Affiliations: 1Institute of Clinical Biochemistry and Pathobiochemistry, protein as first described by Ellison and colleagues (7), is University of Wuerzburg, Wuerzburg, Germany; Divisions of 2Molecular now commonly accepted as the hallmark genetic event for Genetics and 3Biostatistics, 4Clinical Cooperation Unit Neuropathology, – German Cancer Research Center (DKFZ); Departments of 5Pediatric standard (or low) risk medulloblastoma (8 10). The WNT Oncology, Hematology & Immunology and 6Neuropathology, University subgroup of tumors may be cytogenetically identified by of Heidelberg, Heidelberg, Germany; 7Division of Neurosurgery, and Program the presence of monosomy 6, which is exclusively found in in Developmental and Stem Cell Biology, Hospital for Sick Children, Toronto, Ontario, Canada; 8NN Burdenko Neurosurgical Institute, Moscow, Russia; tumors of this subset (4, 6, 8, 11). A second distinct sub- 9Department of Neuropathology, Heinrich-Heine-University, Dusseldorf, group of standard-risk patients identified by transcriptome Germany; 10Cnopf'sche Kinderklinik, Nürnberg Children's Hospital, – Nuremberg, Germany; and 11Department of Pediatrics and Childrens' studies (4 6) is molecularly characterized by aberrant SHH Cancer Research Center, Klinikum rechts der Isar, Technische Universität activation, and is typically driven by underlying mutations München, Munich, Germany in key regulators of the pathway such as PTCH1 or SUFU Note: Supplementary data for this article are available at Cancer Research (12–14). Collectively, these two subgroups account for ap- Online (http://cancerres.aacrjournals.org/). proximately 35% to 40% of medulloblastomas. Significantly, C. Traenka and M. Remke contributed equally to this work. these subgroups typically lack chromosome 17 aberrations Corresponding Authors: Elke Butt, Grombuehlstr. 12, D-97080 Wuerzburg, (4, 5), which however is the most frequent cytogenetic event Germany. Phone: 49-9313-293630; Fax: 49-9318-8883174; E-mail: butt@ – klin-biochem.uni-wuerzburg.de or Stefan M. Pfister, Im Neuenheimer Feld in medulloblastoma (11, 15 17). 280, D-69120 Heidelberg, Germany. Phone: 49-6221-424593; Fax: 49- DNA copy number aberrations of chromosome 17, includ- 6221-424639; E-mail: [email protected]. ing deletion of 17p and gain of 17q, are typically found in the doi: 10.1158/0008-5472.CAN-10-0592 presence of isochromosome 17q (i17q), the hallmark cyto- ©2010 American Association for Cancer Research. genetic abnormality of high-risk medulloblastoma (4, 5, 11).

www.aacrjournals.org 8003

Traenka et al.

However, the driver gene(s) targeted by this cytogenetic Transcriptome analysis and microarray data analysis aberration remain unidentified. Expression profiling and microarray image analysis were We previously showed in a large cohort of patients carried out as outlined in Supplementary Materials and (n = 340) that isolated gain of chromosome 17q (without con- Methods. comitant loss of 17p as observed in i17q cases) is associated with a poor prognosis, whereas isolated 17p deletion is not cDNA synthesis and quantitative real-time PCR (11). We therefore decided to pursue the biological role of Oligo-d(t) primed cDNA synthesis of 1 μg total RNA was 17q gain in medulloblastoma. When investigating 28 pairs performed by using Superscript II Reverse Transcriptase of primary and recurrent tumors, we found that 5 out of 13 (Invitrogen) as described in the protocols of the manufacturer. tumors (38%), the primaries of which were characterized by a Each cDNA sample was analyzed in triplicate by quantitative balanced chromosome 17, showed acquired gain of 17q in real-time PCR on an ABI PRISM 7700 PCR System (Applied the relapsed tumors (18). These findings suggest that gain Biosystems) using ABsolute SYBR Green ROX Mix (ABgene) of chromosome 17q is a critically important aberration for according to the instructions of the manufacturer. Gene ex- medulloblastoma progression including metastatic dissemi- pression was normalized to three endogenous housekeeping nation in a large subset of these relapsed tumors. genes (SDHA, HPRT1,andLMNB1), which were constantly Prior attempts to identify medulloblastoma oncogenes in expressed in several expression experiments comparing sub- 17q have suggested ERBB2 (19), BIRC5 (20), and PPM1D (16). sets of primary medulloblastoma samples, medulloblastoma As none of these genes were highly ranked in our screening cell lines, and control tissues. Amplification of genomic DNA approach comparing medulloblastomas with 17q gain was excluded by intron-spanning primers and adequate test against 17q balanced tumors, we hypothesized that additional PCRs. Relative gene expression levels were calculated as candidate oncogenes in 17q should be considered. described by Pfaffl (27) with the adaptation of three used In the present study, we identified LASP1 as one of the most housekeeping genes. Primer sequences are available in differentially expressed and functionally relevant genes on 17q Supplementary Table S1. when comparing tumors with and without 17q copy number gains. The LIM and SH3 domain protein LASP1 (previously Fluorescence in situ hybridization named MLN50) was initially identified from a cDNA library Interphase fluorescence in situ hybridization analysis for of breast cancer metastases (21). LASP1 mRNA is detected chromosome 17 was performed on a medulloblastoma tissue ubiquitously at low basal levels in all normal human tissues microarray containing 207 tumor samples and 10 samples of (22), but the protein is highly overexpressed in more than nontumoral cerebellar tissues as controls (11). A commercial 55% of metastatic human breast cancer (23) and ovarian can- probe set delineating the loci of interest was used: 17p13.3/ cer (24). In a recent case control study, LASP1 expression in LIS1 (spectrum orange) and 17q21/RARA (spectrum green; mammary carcinomas correlated significantly with tumor size Vysis). Evaluation criteria and cutoffs were applied as and nodal positivity (25). The genomic locus of LASP1 at 17q12 described (11). is targeted by copy number gain or amplification in 20% to 30% of human breast cancers (26). Silencing of LASP1 by Immunohistochemistry RNA interference in various cell lines resulted in a strong A primary antibody to LASP1 (28) was applied at a concen- inhibition of cell migration and proliferation, with cell cycle tration of 10 μg/mL. Evaluation of immunostaining was per- arrest in G2-M phase (23, 24). These observations imply a dual formed in a semiquantitative manner and cases with clear function of LASP1 as both a structural scaffold at focal adhe- evidence of (cytoplasmic as well as nuclear) immunostaining sions, and as a signal transducer conveying information from were considered immunopositive. Therefore, the analysis of the cell surface to the nucleus (22). our immunohistochemistry experiments was only performed To assess the role of LASP1 in medulloblastoma, we ana- by a positive versus negative evaluation of stained cores. lyzed the prognostic value of LASP1 protein expression by immunohistochemistry on a tissue microarray representing Cell lines and cell culture conditions sections from 207 clinically and histologically well-annotated Medulloblastoma cell lines DAOY (purchased from Ameri- samples from uniformly treated medulloblastoma patients. can Type Culture Collection) and UW228-2 (kindly provided Our findings in the primary tumors, together with functional by S. Clifford, Newcastle, United Kingdom) were grown at 5 studies in medulloblastoma cell lines, provide strong evi- 1×10 cells/mL at 37°C under 5% CO2 atmosphere in DMEM dence for an important role of LASP1 in the progression (Invitrogen) containing 10% heat-inactivated fetal bovine and metastatic dissemination of medulloblastoma. serum and 1% streptomycin/ampicillin (Invitrogen). D283 medulloblastoma cell line (purchased from American Type Materials and Methods Culture Collection) was grown in α-MEM (Invitrogen) containing 10% heat-inactivated fetal bovine serum and 1% Tumor material and patient characteristics streptomycin/ampicillin. Cells were cultured until homoge- Nucleic acid isolation. Nucleic acids were extracted from neous morphology of cells was reached and up to a maximum tumor samples and cell lines as previously described (11). of four passages because LASP1 belongs to a group of several A detailed description is given in the Supplementary Materials differentially expressed proteins that are upregulated after and Methods. later passages (29).

8004 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LASP1 in Metastatic Medulloblastoma

Immunofluorescence AQ Non-Radioactive Cell Proliferation Assay (Promega). Details For immunofluorescence microscopy, cells were grown are given in the Appendix. until homogenous morphology at a maximum of 70% confluence on glass chamber slides (for D283, cells were seeded on Adhesion assay collagen-coated glass chamber slides) fixed in 4% (w/v) para- To assess cell adhesion, 48-well plates were coated with formaldehyde in PBS, permeabilized with 0.1% (w/v) Triton X- 10 μg/mL of fibrinogen (Sigma) diluted in PBS and 0.1% bo- 100 in PBS, and then stained with affinity-purified LASP1 vine serum albumin overnight at 4°C. Cells were washed antibody (diluted 10 μg/mL; ref. 25) followed by secondary in serum-free medium, resuspended at a concentration of Cy3-labeled anti-rabbit antibody (1:500; Dianova). 5×105 cells/mL and 200 μL were seeded. Cells were allowed to attach for 2 hours at 37°C. Nonadherent cells were re- Small interfering RNA preparation and transfection moved by gentle washing with PBS. Attached cells were fixed LASP1 knockdown was done using a mix of two small in 4% (w/v) paraformaldehyde for 10 minutes and then interfering RNA (siRNA) constructs targeting the LASP1 stained with 0.5% (w/v) crystal violet (in 2% ethanol, filtered sequences 5′-AAG GTG AAC TGT CTG GAT AAG-3′ (bases with 0.45 μmol/L pore size) for 20 minutes followed by wash- 49–69; named A in Supplementary Fig. S1) and 5′-AAG ing three times with PBS. The blue dye was eluted in 10% CAT GCT TCC ATT GCG AGA-3` (bases 80–100; named C acetic acid for 10 minutes, and the absorbance was measured in Supplementary Fig. S1; ordered from Dharmacon). A total at 595 nm on a plate reader. Adhesion assays were performed of three different LASP1 siRNAs were tested alone or in in 3 independent experiments, each with 6 replicates. combination in all cell lines. Knockdown efficiency of siRNA constructs was best for constructs A and C and independent Cell migration of 17q status in cell lines without (DAOY) and with 17q gain The migration assay was performed using a modified (D283; Supplementary Fig. S1). Similar results were obtained Boyden chamber assay (Transwell chambers, Corning Star). for UW228-2. Nontargeting siRNA no. 5 from Dharmacon Details are given in the Supplementary Materials and Methods. was used as a control. Cells were plated at a density of 2×105 cells/flask, grown for 24 hours at a confluence of Statistical analysis 30% to 50% and transfected with 2.5 μg of siRNA in reduced Median duration of follow-up was calculated according to serum medium OPTI-MEM-I (Life Technologies, Paisley, Korn (30). Estimation of survival time distributions was per- United Kingdom) according to the manufacturer's protocol. formed using the Kaplan-Meier method. For comparisons of After 4 hours of incubation at 37°C, the transfection medium two survival curves, the log rank test was used. Comparisons LASP1 LASP1 was replaced with 5 mL of routine cell culture medium and of patient characteristics between high and low LASP1 LASP1 incubation was continued for 42 hours. For control cells, mRNA expression and between high and low pro- either 20 μL of HiPerfect or 2.5 μg of control siRNA were tein expression were performed using Fisher's exact test. Stu- used. For adhesion and migration experiments, the cells were dent's t test was applied to determine the effects of siRNA synchronized by starving overnight in basal medium with knockdown on cell proliferation, adhesion, and migration. 0.5% FCS. At least three independent experiments were Signaling pathway impact analysis (SPIA) algorithm was used performed for each cell line, and representative results are to identify relevant pathways upon LASP1 knockdown ex- shown. LASP1 knockdown was controlled by Western blots periments (31). Spearman's rank correlation coefficient was in all experiments. assessed to describe the relationship between 17q gain and LASP1 protein expression. Gender, age, M-stage, histology, Western blot analysis 17q gain, and LASP1 protein expression were included in For Western blotting, cells were lysed in Laemmli sample the multivariate Cox proportional hazards regression model buffer. Equal amounts of protein, according to cell count, as possible prognostic factors for overall survival. A result were resolved by 10% SDS-PAGE. After blotting on a nitro- was considered significant at P ≤ 0.05. Statistical computa- cellulose membrane (Schleicher & Schuell) and blocking tions were performed with the statistical software environ- with 3% nonfat dry milk (Bio-Rad) in 10 mmol/L Tris ment R, version 2.9.0. Exact confidence intervals for (pH 7.5), 100 mmol/L of NaCl, and 0.1% (w/v) Tween 20, Spearman's rank correlation coefficient were calculated us- the membrane was first incubated with the self-generated ing StatXact software version 6. antibody against LASP1 (diluted 1 μg/mL; ref. 25) or with filamentous actin (diluted 1:5,000, Sigma-Aldrich) followed Results by incubation with horseradish peroxidase–coupled goat anti-rabbit IgG (Bio-Rad; diluted 1:5,000), and visualized Differentially expressed genes in tumors with and by enhanced chemiluminescence (Amersham Biosciences). without 17q gain Quantification of autoradiography signals was carried out To identify novel oncogenes on chromosome 17q, the by densitometry using the ImageJ software (NIH). most frequent region of chromosomal gain in medulloblastoma, we hybridized total RNA from 10 equally repre- Cell proliferation sented medulloblastoma samples with 17q gain against Cell viability was determined by the 3-(4,5-dimethylthiazol- 10 medulloblastoma samples with diploid 17q. Likely due 2-yl)-2,5-diphenyltetrazolium bromide (MTT)–based CellTiter96 to gene-dosage effects, the mean log 2 ratio of all genes

www.aacrjournals.org Cancer Res; 70(20) October 15, 2010 8005

Traenka et al.

Figure 1. A, upregulation of LASP1 mRNA in medulloblastomas with gain of chromosome 17q. Differential expression of mRNA transcripts localized on chromosome 17 as assessed by microarray analysis in 20 primary medulloblastomas. LASP1 transcripts are indicated by red dots. All data points are shown as log 2 ratios comparing pooled mRNA expression from 10 medulloblastoma samples carrying a 17q gain with 10 medulloblastoma samples showing a balanced 17q copy number status. Clones are plotted in order of chromosome position. B, association of LASP1 expression with 17q gain (left) and metastatic disease (right) in an independent cohort of 101 patients. LASP1 mRNA expression determined by quantitative real-time PCR.

on 17p was negative, whereas the mean expression ratio of found several candidate genes of potential interest with 17q genes was positive (Fig. 1A). The top 50 upregulated respect to medulloblastoma biology, including CDK5R1,a genes on 17q in tumors harboring 17q gain are listed in brain-specific activator of cyclin-dependent kinase 5 and Supplementary Table S2. Among the top 10 genes, we GNGT2, a guanine-nucleotide exchange factor known to

8006 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LASP1 in Metastatic Medulloblastoma

Figure 2. A, representative LASP1 immunohistochemistry with negative, predominantly nuclear or cytoplasmic staining in medulloblastoma (×200). B, high protein expression in tumors with 17q gain and metastatic disease (M1–3). Immunohistochemistry was performed on a tissue microarray with 207 medulloblastoma samples of uniformly treated children. C, progression-free (PFS) and overall survival (OS) rates estimated from the time of diagnosis using the Kaplan-Meier method for LASP1high and LASP1low expressing tumors as assessed by immunohistochemistry. The numbers at the bottom indicate the number of patients at risk in each molecular subgroup at a certain time point.

www.aacrjournals.org Cancer Res; 70(20) October 15, 2010 8007

Traenka et al.

be involved in retinal phototransduction signaling and cohort of 101 medulloblastoma samples for which the known to be activated in high-risk medulloblastoma DNA copy number status of 17q was available. Patient (4, 5). Most interestingly, one probe covering the trans- data in association with LASP1 mRNA expression are criptional start site of LASP1, a well-established driver shown in Supplementary Table S3. As anticipated from of metastatic dissemination (21), showed a very strong our microarray data, we found a very strong association upregulation (log 2 ratio = 3.1). A second clone for LASP1 between LASP1 expression, 17q status (P < 0.001; Fig. 1B, mapping to the 3′ part of the 3′-untranslated region also left), and metastatic disease at the time of diagnosis showed a marked upregulation; albeit at lower levels (log 2 (P = 0.002; Fig. 1B, right). To follow an unbiased approach ratio = 1.1) than observed for the first LASP1 clone (Fig. 1A). for correlation with survival data, median LASP1 expres- Notably, expressing profiling data in an independent cohort sion was used as a cutoff between LASP1high and LASP1low of 103 medulloblastomas performed on a different platform expressing tumors. Doing so, we found a statistically signif- (Affymetrix) in a different laboratory (Hospital for Sick icant difference between LASP1high and LASP1low expressing Children, Toronto, Canada) confirmed that LASP1 mRNA tumors for overall survival (P = 0.05; data not shown) us- expression was significantly associated with 17q status ing the Kaplan-Meier method, whereas progression-free (P < 0.001; Supplementary Fig. S2). survival (PFS) was not significantly different between these two groups (P = 0.21). LASP1 mRNA expression correlates with 17q status, M stage, and patient outcome in medulloblastoma LASP1 protein expression in medulloblastoma Because we had previously established a role for 17q LASP1 protein expression was subsequently evaluated in gain in medulloblastoma dissemination (11), we decided 207 medulloblastoma samples on a medulloblastoma tissue to primarily focus on LASP1 as a candidate oncogene on microarray by immunohistochemistry. Among them, 100 chromosome 17q based on its known role in metastatic (48%) samples were immunonegative (Fig. 2A, left), whereas dissemination. Therefore, we validated our microarray the remaining 107 (52%) tumors showed strong and results by quantitative real-time PCR in an independent homogeneous LASP1 expression, with a predominantly

Table 1. Patient characteristics in the cohort investigated by immunohistochemistry and fluorescence in situ hybridization

Variable No. of patients LASP1 negative LASP1 positive P*

Age (y) 0.12 <4 24 14 (58%) 10 (42%) 4–18 165 74 (45%) 91 (55%) >18 18 12 (67%) 6 (33%) Gender 0.38 Female 76 40 (53%) 36 (47%) Male 131 60 (46%) 71 (54%) Metastatic stage <0.001 M0 144 83 (58%) 61 (42%) M1 20 10 (50%) 10 (50%) M2 13 2 (15%) 11 (85%) M3 30 5 (17%) 25 (83%) Histologic subtype 0.002 Desmoplastic/MBEN 9 9 (100%) 0 (0%) Classic 176 85 (48%) 91 (52%) Large cell/anaplastic 22 6 (27%) 16 (73%) Level of resection 0.68 Gross total resection 93 43 (46%) 50 (54%) Subtotal resection 114 57 (50%) 57 (50%) 17q status <0.001 Balanced 108 69 (64%) 39 (36%) Copy number gain 99 31 (31%) 68 (69%) Total 207 100 (48%) 107 (52%)

Abbreviation: MBEN, medulloblastoma with extensive nodularity. *Fisher's exact test.

8008 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LASP1 in Metastatic Medulloblastoma

six were strongly immunopositive for LASP1 in the primary Table 2. Multivariate Cox proportional hazards tumor and all six maintained immunopositivity in the recur- regression model for overall survival rent tumor. From five patients, however, in which no LASP1 protein expression was detectable in the primary tumor as as- P Variable HR for OS (95% CI) sessed by immunohistochemistry, two patients showed strong LASP1 expression in the recurrent tumor, further strengthen- Gender ing the role of LASP1 in medulloblastoma recurrence and Female disease progression (Supplementary Fig. S3). Male 1.70 (0.87) 0.12 Age (y) In vitro inhibition of LASP1 attenuates the malignant <4 phenotype of medulloblastoma cell lines 4–18 0.43 (0.19–0.99) <0.05 To functionally assess the role of LASP1 in medulloblas- >18 0.70 (0.17–2.85) 0.62 toma, we used three well-established medulloblastoma cell M stage lines: DAOY, UW228-2, and D283. The latter cell line is com- M0 monly used as an in vitro model for metastasized medullo- M1 0.58 (0.20–1.71) 0.33 blastoma and is known to harbor an i(17q). The adherent M2 0.66 (0.20–2.14) 0.49 cell lines, DAOY and UW228-2, showed a strong accumula- M3 1.02 (0.44–2.37) 0.97 tion of LASP1 in the focal adhesion contacts and an intense Histologic subtype colocalization with filamentous actin (Fig. 3A). D283 cells, an Classic adherent/suspension cell line with a known abnormal Large cell/anaplastic 5.53 (2.52–12.11) <0.001 expression of neurofilaments (32), showed minor spreading Desmoplastic/MBEN 1.01 (0.12–8.33) 0.99 and reduced focal contacts. LASP1 localization with fila- 17q status mentous actin was less definite but clearly observed at the Balanced status leading edges of cells (Fig. 3A). Copy number gain 5.98 (2.52–14.19) <0.001 All cell lines were transfected with two combined siRNAs LASP1 immunoreactivity against LASP1, a control siRNA, or just treated with the Negative transfection reagent HiPerfect. Interestingly, a significant Positive 2.24 (1.06–4.75) 0.03 inhibition of proliferation was observed visually (Fig. 3B) Abbreviations: HR, hazard ratio; OS, overall survival; CI, and by MTT assay in these cell lines upon LASP1 silencing confidence interval; MBEN, medulloblastoma with exten- (Fig. 3C, left). siRNA-mediated knockdown was confirmed sive nodularity. to cause a profound reduction of LASP1 protein abundance in all cell lines with maximum silencing observed after 48 hours (Fig. 3C, right). membranous and cytoplasmic distribution (Fig. 2A, middle). Because LASP1 has been shown to have a functional role Only a small proportion of tumors exhibited nuclear positiv- in cell motility and tumor dissemination in other tumor ity (Fig. 2A, right). All desmoplastic medulloblastomas were entities, we investigated cell migration and adhesion in immunonegative for LASP1 (Table 1). Samples positive for medulloblastoma cell lines before and after LASP1 silencing LASP1 were significantly overrepresented in tumors with by using a modified Boyden chamber and a fibronectin 17q gain detected by fluorescence in situ hybridization adhesion assay. We observed a strong reduction in migrato- analysis (Fig. 2B, left; Spearman's correlation coefficient = ry potential (Fig. 4A) and an enhanced adhesion in all 0.32; 0.20–0.45; P < 0.001) and among patients with metastatic medulloblastoma cell lines upon LASP1 silencing (Fig. 4B). disease at diagnosis (Fig. 2B, right; Spearman's correlation Similar results regarding adhesion, migration, and prolifera- coefficient = 0.31; 0.21–0.43; P < 0.001). Interestingly, MYC or tion were obtained with single siRNA treatment (Supple- MYCN amplified tumors with extremely poor prognosis also mentary Fig. S4). show higher LASP1 protein expression as compared with In an attempt to characterize the molecular networks molecular standard risk tumors (P = 0.02). This finding further influenced by LASP1, we performed transcriptome analysis strengthens the role of LASP1 in molecular high-risk tumors. comparing differentially expressed mRNAs upon LASP1 Survival analysis by Kaplan-Meier estimate revealed a knockdown using two independent siRNA clones (A, C) in significant association of LASP1 immunoreactivity with two medulloblastoma cell lines with (D283) and without diminished progression-free survival (P = 0.001) and overall (DAOY) copy number gain of 17q. Notably, LASP1 mRNA survival (P < 0.001; Fig. 2C). Most notably, in a multivariate was the most strongly downregulated transcript, thus Cox proportional hazards regression model including confirming the specificity of knockdown experiments. variables known to be most strongly associated with overall Based on previously reported roles in cell adhesion, up- survival, LASP1 protein expression was an independent regulation of CEACAM1, NRCAM, KLHL28,andVASP upon prognostic marker besides 17q gain and anaplastic histology LASP1 knockdown was of particular interest (Supplementary (Table 2; P = 0.03). Table S4). By contrast, downregulation of the following tran- In addition, we analyzed tissues from 11 pairs of primary scripts coding for proteins implicated in invasion and mi- and recurrent tumors from the same patients. From these, gration was notable: MMP19, ADAM9, TPM3,andCAPZA2

www.aacrjournals.org Cancer Res; 70(20) October 15, 2010 8009

Traenka et al.

Figure 3. A, localization of LASP1 at focal contacts, at the leading edge, and in the nucleus of medulloblastoma cell lines. For immunofluorescence staining, cells were fixed, and stained using a polyclonal LASP1 antibody (red), and Oregon green phalloidin filamentous actin (green). Microscopic examination (B) and MTT analysis (C) of medulloblastoma cells after LASP1 knockdown show inhibition of cell proliferation (×10). LASP1 siRNA-transfected medulloblastoma cells (DAOY, UW228-2, and D283) and control cells were seeded in 48-well plates, incubated with MTT/dye solution for 4 h and the absorbance was recorded. Bars, SEM; *, P < 0.001 versus control; Student's t test. LASP1 knockdown efficiency was controlled by Western blots.

(Supplementary Table S5). By using the top 100 differentially most frequently affected cellular functions (Supplementary expressed genes for SPIA pathway annotation, “actin Table S6). cytoskeleton regulation” (GNA13, ITGB1, and PPP1CC) and “fo- These findings establish a critical role for LASP1 in the cal adhesion” (VEGF, ITGB1,andPPP1CC)wereamongthe proliferation and migration of medulloblastoma cells and

8010 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LASP1 in Metastatic Medulloblastoma

suggest that LASP1 is a functionally relevant 17q oncogene dissemination via the surrounding cerebrospinal fluid, and upregulated in intermediate and high-risk medulloblastomas one of the unifying molecular characteristics observed in carrying an isochromosome 17q. these tumors are chromosome 17 aberrations classifying them as high-risk medulloblastomas (9–11). Discussion In a comparison of primary and recurrent medulloblastoma samples from 28 patients, we previously made the in- Medulloblastomas had previously been considered a sin- teresting observation that acquired isolated gains of 17q are gle entity. However, molecular studies conducted over the a common genetic feature in relapsed medulloblastoma (18). last 5 years have shown that medulloblastoma is a molec- Based on this observation, we attempted to identify novel ularly heterogeneous disease (4–6). Genome-wide transcrip- oncogenes on 17q that could account for the selective growth tome analysis has consistently revealed at least four distinct advantage of tumor cells carrying this aberration, and biological subentities with unique underlying pathophysio- explain their promigratory and proproliferative phenotype. logic mechanisms, demographics, and clinical outcomes. A In this study, we identified LASP1 as a very promising unifying clinical feature of the two (5, 6) or three (4) molec- candidate oncogene in medulloblastoma based on its ular subgroups that are not characterized by WNT or SHH overexpression in tumors harboring gain of 17q, as com- pathway activation is the high incidence of metastatic pared with medulloblastomas with a balanced karyotype

Figure 4. A, LASP1 knockdown inhibits cell migration. Medulloblastoma cells transfected with LASP1 siRNA and control siRNA were seeded in modified Boyden chambers and incubated for 4 h. Migrating cells were fixed with paraformaldehyde and stained with crystal violet. The absorbance was measured. Bars, SEM; *, P < 0.001 versus control; Student's t test. B, increased adhesion in LASP1-depleted cells. Medulloblastoma cells were seeded in 48-well plates coated with fibronectin and incubated for 2 h. Cells were fixed with paraformaldehyde and stained with crystal violet. Absorbance was measured at 595 nm. Bars, SED; ***, P < 0.001 versus control. LASP1 knockdown efficiency was controlled by Western blot.

www.aacrjournals.org Cancer Res; 70(20) October 15, 2010 8011

Traenka et al.

on chromosome 17. This is mechanistically very intriguing, Interestingly, EZRIN (EZR), a cytoskeletal cross-linker that con- as LASP1 overexpression has been previously found to nects the oncogenic stemness marker CD44 with the cytoske- drive metastasis in breast cancer and ovarian cancer leton (43–45), has recently been proposed to participate in (23–25), both of which are also characterized by frequent medulloblastoma cell migration in concert with other cytoske- gain of 17q in late stage disease (33, 34). letal proteins. Notably, EZR silencing in medulloblastoma cell The clinical implication of our findings in medulloblastoma lines showed a similar phenotype with suppressed migration is further underlined by the strong association between and invasion as LASP1 knockdown. However, in contrast with LASP1 mRNA and protein expression within the presence of LASP1 silenced medulloblastoma cells, EZR knockdown led to metastatic disease, and its role as an excellent prognostic reduced cell adhesion (46). marker as identified by immunohistochemistry in more than Since indirect interaction of LASP1 and EZR via the 200 primary medulloblastoma samples. As anticipated from promigratory scaffolds KRP1 and PALADIN was recently our previous finding that 17q gain is frequently acquired dur- reported (47, 48) and because LASP1 physically interacts with ing medulloblastoma progression, we noted strong LASP1 the important transcription factors ZYXIN and LPP that in protein expression in two out of five recurrent medulloblasto- turn promote prometastatic transcriptional signatures, it is ma samples in which no expression was observed in the pa- tempting to speculate that LASP1 exerts part of its function tient matched primary tumor. These findings are consistent as a signal transducer embedded in a novel signaling path- with the attributed biological functions of LASP1 in primary way and as a structural focal adhesion protein (22, 49). tumors, including the regulation of cell migration and prolif- We do not yet understand whether LASP1 expression is eration (21, 23–26, 35). regulated by oncogenes or whether LASP1 itself acts as a Supporting our observations in primary tumors, we found transcription factor regulating oncogenes and proteins in- astrikingphenotypeuponsiRNA-mediated silencing of volved in cell migration. However, compared with normal tis- LASP1 in three well-established medulloblastoma cell lines. sue, tumor cells display an increased rate of LASP1-positive In addition to a marked reduction in cell proliferation, we nuclei (24, 39), supporting the hypothesis that, next to its observed a profoundly impaired propensity of medulloblasto- physiologic role as a scaffolding protein involved in cell mo- ma cells to migrate through a filter membrane in combina- tility, LASP1 may serve as a transcriptional cofactor involved tion with a substantially increased tumor cell to matrix in processes of cell proliferation and metastasis in malignant adhesion upon LASP1 knockdown. These in vitro observa- cells. Further studies will help to define the role of LASP1 in tions were nicely complemented by the identification of de- cancer development and metastasis. regulated genes and signaling pathways upon LASP1 In conclusion, we have identified LASP1 on 17q12 as a novel knockdown. Notably, KEGG pathways such as “actin cyto- candidate oncogene with effects on metastatic dissemination skeleton regulation” and “focal adhesion” were among the and disease progression of medulloblastoma with high poten- most frequently affected cellular functions as determined tial to serve as powerful biomarker for outcome prediction in by SPIA analysis. future prospective studies. Attempts to tackle signal transduc- Since its identification in breast tumor cells (21), LASP1 was tion pathways that are driving tumor cell motility and inva- found to be upregulated in a number of other cancers [ovarian siveness with targeted therapies will be an important step (24); kidney (36); and prostate (37)] and was suggested to be a towards disease control in metastatic medulloblastoma, and negative prognostic marker. A recent study identified LASP1 as LASP1 represents an excellent novel candidate oncogene for a novel p53 transcriptional target in hepatocellular carcinoma future targeted therapy approaches. (38). Inactivating p53 mutations led to increased LASP1 expres- Disclosure of Potential Conflicts of Interest sion and to a more aggressive hepatocellular carcinoma phe-

notype. Nevertheless, not all p53-mutated cell lines showed No potential conflicts of interest were disclosed. increased LASP1 expression (39). Interestingly, mRNA expression levels of p53, downstream intermediates and prominent Acknowledgments target genes were not altered upon LASP1 knockdown in medulloblastoma cell lines (Supplementary Table S4–6). We gratefully thank Leonore Senf for the collection and maintenance of the brain tumor bank in the department of W. Scheurlen. Steven Clifford In invasive breast cancer cells, LASP1 expression was in- (Newcastle, United Kingdom) is acknowledged for providing the versely correlated with prostate-derived Ets factor, a transcrip- medulloblastoma cell line UW228-2. Petra Thalheimer and Frauke Devens tion factor known to repress a variety of genes that are possibly are acknowledged for excellent technical assistance. involved in oncogenesis (40). However, interdependence of Grant Support prostate-derived Ets factor and LASP1 expression in breast cancer cell lines remains controversially discussed (39, 41). Research grant 107706 from the Deutsche Krebshilfe (E. Butt), and by grants from the Deutsche Kinderkrebsstiftung, Landesstiftung Baden-Württemberg, None of the hallmark genes involved in metastatic cascades and the “Hella-Bühler Award” (S. Pfister), and a “Young Investigator Fellowship” in breast cancer such as NM23, DLC1, NDRG1, KAI1, RHOGDI2, of the Medical Faculty of Heidelberg (M. Ryzhova). KISS1, and RKIP (reviewed in ref. 42) seem to be deregulated The costs of publication of this article were defrayed in part by the payment upon LASP1 knockdown in medulloblastoma. Thus, underlying of page charges. This article must therefore be hereby marked advertisement in molecular mechanisms might be essentially different in meta- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. static dissemination of medulloblastoma cells through the Received 02/17/2010; revised 07/07/2010; accepted 07/20/2010; published cerebrospinal fluid with LASP1 playing a central role in both. OnlineFirst 10/05/2010.

8012 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LASP1 in Metastatic Medulloblastoma

References 1. Louis D, Ohgaki H, Wiestler O, et al. The 2007 WHO classification of four novel human genes amplified and overexpressed in breast tumours of the central nervous system. Acta Neuropathol 2007;114: carcinoma and localized to the q11-21.3 region of chromosome 97–109. 17. Genomics 1995;28:367–76. 2. Zeltzer PM, Boyett JM, Finlay JL, et al. Metastasis stage, adjuvant 22. Grunewald TG, Butt E. The LIM and SH3 domain protein family: treatment, and residual tumor are prognostic factors for medullo structural proteins or signal transducers or both? Mol Cancer blastoma in children: conclusions from the Children's Cancer 2008;7:31. Group 921 randomized phase III study. J Clin Oncol 1999;17: 23. Grunewald TG, Kammerer U, Schulze E, et al. Silencing of LASP-1 832–45. influences zyxin localization, inhibits proliferation and reduces migra- 3. Zerbini C, Gelber R, Weinberg D, et al. Prognostic factors in medullo- tion in breast cancer cells. Exp Cell Res 2006;312:974–82. blastoma, including DNA ploidy. J Clin Oncol 1993;11:616–22. 24. Grunewald TG, Kammerer U, Winkler C, et al. Overexpression of 4. Thompson MC, Fuller C, Hogg TL, et al. Genomics identifies LASP-1 mediates migration and proliferation of human ovarian medulloblastoma subgroups that are enriched for specific genetic cancer cells and influences zyxin localisation. Br J Cancer 2007; alterations. J Clin Oncol 2006;24:1924–31. 96:296–305. 5. Kool M, Koster J, Bunt J, et al. Integrated genomics identifies five 25. Grunewald TG, Kammerer U, Kapp M, et al. Nuclear localization medulloblastoma subtypes with distinct genetic profiles, pathway and cytosolic overexpression of LASP-1 correlates with tumor size signatures and clinicopathological features. PLoS One 2008;3: and nodal-positivity of human breast carcinoma. BMC Cancer 2007; e3088. 7:198. 6. Northcott PA, Korshunov A, Witt H, et al. Medulloblastoma com- 26. Tomasetto C, Moog-Lutz C, Regnier CH, Schreiber V, Basset P, Rio prises four distinct molecular variants. J Clin Oncol. Epub 2010 MC. Lasp-1 (MLN 50) defines a new LIM protein subfamily charac- Sep 7. terized by the association of LIM and SH3 domains. FEBS Lett 1995; 7. Ellison DW, Onilude OE, Lindsey JC, et al, The United Kingdom 373:245–9. Children's Cancer Study Group Brain Tumour Committee. β-Catenin 27. Pfaffl MW. A new mathematical model for relative quantification in status predicts a favorable outcome in childhood medulloblastoma. real-time RT-PCR. Nucleic Acids Res 2001;29:e45. J Clin Oncol 2005;23:7951–7. 28. ButtE,GambaryanS,GottfertN,GallerA,MarcusK,MeyerHE. 8. Clifford S, Lusher M, Lindsey J, et al. Wnt/wingless pathway activa- Actin binding of human LIM and SH3 protein is regulated by cGMP- tion and chromosome 6 loss characterise a distinct molecular sub- and cAMP-dependent protein kinase phosphorylation on serine 146. group of medulloblastomas associated with a favourable prognosis. J Biol Chem 2003;278:15601–7. Cell Cycle 2006;5:2666–70. 29. Sun HJ, Bahk YY, Choi YR, Shim JH, Han SH, Lee JW. A pro- 9. Gajjar A, Chintagumpala M, Ashley D, et al. Risk-adapted cranio- teomic analysis during serial subculture and osteogenic differen- spinal radiotherapy followed by high-dose chemotherapy and tiation of human mesenchymal stem cell. J Orthop Res 2006;24: stem-cell rescue in children with newly diagnosed medulloblastoma 2059–71. (St Jude Medulloblastoma-96): long-term results from a prospective, 30. Korn EL. Censoring distributions as a measure of follow-up in survival multicentre trial. Lancet Oncol 2006;7:813–20. analysis. Stat Med 1986;5:255–60. 10. Pizer BL, Clifford SC. The potential impact of tumour biology 31. Tarca AL, Draghici S, Khatri P, et al. A novel signaling pathway on improved clinical practice for medulloblastoma: progress to- impact analysis. Bioinformatics 2009;25:75–82. wards biologically driven clinical trials. Br J Neurosurg 2009;23: 32. Trojanowski JQ, Friedman HS, Burger PC, Bigner DD. A rap- 364–75. idly dividing human medulloblastoma cell line (D283 MED) ex- 11. Pfister S, Remke M, Benner A, et al. Outcome prediction in pediatric presses all three neurofilament subunits. Am J Pathol 1987; medulloblastoma based on DNA copy-number aberrations of chro- 126:358–63. mosomes 6q, 17q, and the MYC/MYCN loci. J Clin Oncol 2009;27: 33. Hislop RG, Pratt N, Stocks SC, et al. Karyotypic aberrations of 1627–36. chromosomes 16 and 17 are related to survival in patients with 12. Raffel C, Jenkins RB, Frederick L, et al. Sporadic medulloblastomas breast cancer. Br J Surg 2002;89:1581–6. contain PTCH mutations. Cancer Res 1997;57:842–5. 34. Dimova I, Orsetti B, Negre V, et al. Genomic markers for ovarian can- 13. PietschT,WahaA,KochA,etal.Medulloblastomasofthe cer at chromosomes 1, 8 and 17 revealed by array CGH analysis. desmoplastic variant carry mutations of the human homologue of Tumori 2009;95:357–66. Drosophila patched. Cancer Res 1997;57:2085–8. 35. Neo S, Leow C, Vega V, et al. Identification of discriminators of 14. Taylor MD, Liu L, Raffel C, et al. Mutations in SUFU predispose to hepatoma by gene expression profiling using a minimal dataset medulloblastoma. Nat Genet 2002;31:306–10. approach. Hepatology 2004;39:944–53. 15. Scheurlen WG, Schwabe GC, Joos S, Mollenhauer J, Sorensen N, 36. Viney RL, Morrison AA, van den Heuvel LP, et al. A proteomic inves- Kuhl J. Molecular analysis of childhood primitive neuroectodermal tigation of glomerular podocytes from a Denys-Drash syndrome tumors defines markers associated with poor outcome. J Clin Oncol patient with a mutation in the Wilms tumour suppressor gene WT1. 1998;16:2478–85. Proteomics 2007;7:804–15. 16. Mendrzyk F, Radlwimmer B, Joos S, et al. Genomic and protein ex- 37. Li Z, Szabolcs M, Terwilliger JD, Efstratiadis A. Prostatic intrae- pression profiling identifies CDK6 as novel independent prognostic pithelial neoplasia and adenocarcinoma in mice expressing a marker in medulloblastoma. J Clin Oncol 2005;23:8853–62. probasin-Neu oncogenic transgene. Carcinogenesis 2006;27: 17. Northcott PA, Nakahara Y, Wu X, et al. Multiple recurrent genetic 1054–67. events converge on control of histone lysine methylation in medullo- 38. Wang B, Feng P, Xiao Z, Ren EC. LIM and SH3 protein 1 (Lasp1) is a blastoma. Nat Genet 2009;41:465–72. novel p53 transcriptional target involved in hepatocellular carcinoma. 18. Korshunov A, Benner A, Remke M, Lichter P, von Deimling A, Pfister J Hepatol 2009;50:528–37. S. Accumulation of genomic aberrations during clinical progression 39. Frietsch JJ, Grunewald TG, Jasper S, et al. Nuclear localisation of of medulloblastoma. Acta Neuropathol 2008;116:383–90. LASP-1 correlates with poor long-term survival in female breast 19. Gilbertson RJ, Perry RH, Kelly PJ, Pearson ADJ, Lunec J. Prognostic cancer. Br J Cancer 2010;102:1645–53. significance of HER2 and HER4 coexpression in childhood medullo- 40. Turner DP, Findlay VJ, Kirven AD, Moussa O, Watson DK. Global blastoma. Cancer Res 1997;57:3272–80. gene expression analysis identifies PDEF transcriptional networks 20. Fangusaro J, Jiang Y, Holloway M, et al. Survivin, survivin-2B, and regulating cell migration during cancer progression. Mol Biol Cell survivin-deItaEx3 expression in medulloblastoma: biologic markers 2008;19:3745–57. of tumour morphology and clinical outcome. Br J Cancer 2005;92: 41. Sood AK, Saxena R, Groth J, et al. Expression characteristics of 359–65. prostate-derived Ets factor support a role in breast and prostate 21. Tomasetto C, Regnier C, Moog-Lutz C, et al. Identification of cancer progression. Hum Pathol 2007;38:1628–38.

www.aacrjournals.org Cancer Res; 70(20) October 15, 2010 8013

Traenka et al.

42. Smith SC, Theodorescu D. Learning therapeutic lessons from membrane cytoskeleton cross-linker ezrin in medulloblastoma cells. metastasis suppressor proteins. Nat Rev Cancer 2009;9:253–64. Neuro-oncol 2009;11:381–93. 43. Liu R, Wang X, Chen GY, et al. The prognostic role of a gene 47. Gray CH, McGarry LC, Spence HJ, Riboldi-Tunnicliffe A, Ozanne signature from tumorigenic breast-cancer cells. N Engl J Med BW. Novel β-propeller of the BTB-Kelch protein Krp1 provides a 2007;356:217–26. binding site for Lasp-1 that is necessary for pseudopodial extension. 44. Martin TA, Harrison G, Mansel RE, Jiang WG. The role of the J Biol Chem 2009;284:30498–507. CD44/ezrin complex in cancer metastasis. Crit Rev Oncol Hematol 48. Rachlin AS, Otey CA. Identification of palladin isoforms and charac- 2003;46:165–86. terization of an isoform-specific interaction between Lasp-1 and 45. Wagner W, Horn P, Castoldi M, et al. Replicative senescence of mes- palladin. J Cell Sci 2006;119:995–1004. enchymal stem cells: a continuous and organized process. PLoS 49. Grunewald TG, Pasedag SM, Butt E. Cell adhesion and trans- One 2008;3:e2213. criptional activity—defining the role of the novel protooncogene 46. Osawa H, Smith CA, Ra YS, Kongkham P, Rutka JT. The role of the LPP. Transl Oncol 2009;2:107–16.

8014 Cancer Res; 70(20) October 15, 2010 Cancer Research

Role of LIM and SH3 Protein 1 (LASP1) in the Metastatic Dissemination of Medulloblastoma

Christopher Traenka, Marc Remke, Andrey Korshunov, et al.

Cancer Res 2010;70:8003-8014. Published OnlineFirst October 5, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-10-0592

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2010/10/04/0008-5472.CAN-10-0592.DC1

Cited articles This article cites 48 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/70/20/8003.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/70/20/8003.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/70/20/8003. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.