| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Departments of Genetics and Tumor Cell Biology [Y. L., H. L. M., F. Z., M. R., P. J. M.] and Developmental Neurobiology [P. J., R. H., M. C., T. C., R. G.], Saint Jude Children’s Research Hospital, Memphis, Tennessee 38105, and Division of Pediatric Hematology/Oncology, Mayo Clinic, Rochester, Minnesota 55905 [C. W.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Signaling pathways that regulate normal cerebellar development such as the SHH3 and WNT pathways have been linked to medulloblastoma tumorigenesis (2 , 3) . SHH and WNT proteins compose two families of secreted molecules that are required for cell growth and fate determination in many tissues including the nervous system (4, 5, 6, 7) . Germ-line mutation of PTC1, the receptor for SHH, is responsible for Gorlin syndrome, a familial cancer predisposition syndrome with a high incidence of medulloblastoma (8) . Mutation of PTC1 also occurs in sporadic medulloblastoma, and germ-line and somatic mutations of Suppressor of Fused, a downstream negative regulator of the SHH pathway, were reported recently in children with medulloblastoma (9) . Up-regulation of SHH target genes, such as the transcription factor GLI1, also occur in medulloblastoma (2 , 10) . Similar to Gorlin syndrome, Ptc1 haploinsufficiency in the mouse can lead to medulloblastoma, underscoring the direct relationship between SHH/PTC1 signaling and medulloblastoma (11, 12, 13) .
The WNT pathway regulates ß-catenin, a key transcriptional activator, via a multimeric protein complex that includes the APC tumor suppressor protein, GSK-3ß (a serine-threonine kinase), and Axin (5 , 14 , 15) . WNTs bind to the Frizzled family of receptors and inhibit GSK-3ß-mediated phosphorylation and degradation of ß-catenin. This enables translocation of ß-catenin to the nucleus, where it activates the T-cell factor/lymphoid enhancer factor transcriptional complex, thereby up-regulating target genes that include cyclin D1 and c-MYC (16) . Deletion and/or mutations in APC, AXIN1, or ß-catenin have been reported in sporadic medulloblastoma, and germ-line mutation of APC predisposes to medulloblastoma in Turcot’s syndrome (17, 18, 19, 20, 21) . Recent studies suggest a strong interrelationship between SHH and WNT signaling, with the phosphorylation status of GSK-3ß acting as a signaling nexus between the two pathways (7 , 22) . Thus, during development the SHH and WNT pathways act as important tumor suppressor pathways in the cerebellum.
Defects in DNA repair, or defective responses to DNA damage, predispose to cancer (23) . For example, mismatch repair defects are associated with colorectal cancer (24) , whereas leukemia or lymphoma occurs in human syndromes associated with deficiencies in the response to DNA double-strand breaks, such as ataxia-telangiectasia and Nijmegen breakage syndrome (23) . Recent evidence indicates that brain tumors may also arise in association with defects in the DNA damage response. In this regard, mice with DNA ligase IV or Parp1 deficiency have a high incidence of medulloblastoma (25 , 26) . DNA damage responses are also important in other pathways linked to medulloblastoma; Ptc1+/- mice are hypersensitive to ionizing radiation, and the incidence of medulloblastoma in these mice is dramatically increased after ionizing radiation (27, 28, 29) . Thus, genotoxic stress may feature in the etiology of medulloblastoma.
Mouse models that recapitulate human disease have provided an enormous benefit toward molecular understanding of tumor biology. However, only recently have a number of mouse models of medulloblastoma become available (13 , 25 , 26 , 30 , 31) . The defined genetics coupled with the ready accessibility of tissues, even at early developmental stages, are major advantages of mouse models. In this study, we have used a number of different mouse models for medulloblastoma to identify genes that may be important for the genesis of medulloblastoma. We identified a cohort of genes selectively up-regulated in all of the mouse medulloblastomas examined, as well as in a large subset of human medulloblastomas, thus providing new molecular insights into this tumor.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Human Tissue.
With the approval of the St Jude Children’s Research Hospital Institutional Review Board, tumor material from 50 children (
19 years of age) undergoing surgical resection of a primary medulloblastoma was collected and snap frozen. Fixed tumor material from the same tissues was used for central pathology review to confirm the diagnosis and subtype of medulloblastoma. Frozen sections from each case were examined by light microscopy to ensure that 
80% consisted of tumor. Samples were then subdivided, and RNA or protein was extracted after direct homogenization in Trizol (Invitrogen) or sonication in protein lysis buffer, respectively. The expression of individual mRNA was measured by real-time PCR (see below). ERBB2 protein expression in each tumor was determined using Western blot analysis as described (35)
with the anti-ERBB2 mouse monoclonal antibody NCL-CB11 (Novacastra). Blots were reprobed with a monoclonal antibody for ß-actin (Sigma Chemicals) to control for protein loading. Unsupervised hierarchical cluster analysis was performed using Pearson coefficient analysis (GeneMaths version 2.01; Applied Maths, Inc.). Direct comparisons between molecular and clinical covariables were performed using Fisher’s exact test.
GeneChip Hybridization.
Total RNA was extracted using Trizol (Invitrogen) according to the manufacturer’s instructions. For comparative medulloblastoma analyses, tissues were obtained from Lig4-/-p53-/- (n = 3), Ptc1+/- (n = 2), Ptc1+/-p53-/- (n = 2), Ink4d-/-Kip1+/-p53-/- (n = 1), Kip1-/-p53-/- (n = 1), Ink4c-/-p53-/- (n = 1), and Ink4d+/-Ink4c-/-p53-/- (n = 2), and the controls were adult WT (n = 3) 9 weeks of age, p53-/- (n = 2) 9 weeks of age, WT (n = 3, one sample pooled from three animals) at P5, and p53-/- (n = 2, with one sample pooled from three animals) at P5.
RNA quality and integrity for microarray were assessed using the Agilent 2100 Bioanalyzer (Agilent). Double-stranded cDNA from total RNA was synthesized using the SuperScript Double-Stranded cDNA Synthesis kit (Life Technologies, Inc.) and T7-dT24-DNA primer GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG, according to the manufacturer’s instructions. Using this double-stranded cDNA as a template, a labeled antisense cRNA was synthesized with biotinylated UTP and CTP by in vitro transcription using the T7 RNA Transcript Labeling kit (ENZO Diagnostics, Inc.). The labeled cRNA was passed through a CHROMA SPIN-100 column (Clontech), fragmented by heating and ion-mediated hydrolysis, and then used for hybridization to murine genome MG_U74Av2 oligonucleotide arrays (Affymetrix) according to the manufacturers’ instructions. After hybridization, oligonucleotide arrays were washed and stained with phycoerythrin-conjugated streptavidin (Molecular Probes). The arrays were then scanned using a laser confocal scanner (Agilent), and expression signals of each gene were calculated using Affymetrix Microarray version 4.0 software (MAS v4).
Data Analysis.
Data analysis was done using GeneSpring version 5.0 software (Silicon Genetics). Initially, per-chip and per-gene normalizations were done for all Affymetrix array output data to control for variations between different analyses in intensity and difference in detection efficiency between spots within an array, using the 50th percentile of all measurements as a positive control for each sample. Background corrections were applied using the lower tenth percentile intensities as needed during normalization. Only genes marked as "Marginal" or "Present" were used for normalization. After normalization, data from each medulloblastoma sample could be compared directly to other medulloblastoma samples or control groups. These normalized values were filtered using the two-component Roche-Lorenzato model for estimating variation from control strength. Furthermore, data were restricted with a one-way ANOVA (cutoff, P < 0.05 in different genotypes). Finally, genes with "Present" calls in more than half of experimental samples were taken for additional statistical analyses. Using these approaches for quality control, array data with low signal intensity or genes with very low variation throughout the entire experimental groups were removed. The selected group of genes from this quality control was examined by unsupervised analysis strategies to find subgroups of genes based on similar expression patterns in an unbiased manner. K-means cluster analysis (four different groupings) and a self-organizing map (4 x 4-dimension grouping) identified two major patterns of expression profiles that showed either increased or decreased expression levels in the entire medulloblastoma samples compared with control groups. Supervised analyses were then used to identify up-regulated or down-regulated genes in the entire medulloblastoma compared with control arrays. A normalized value for individual genes of each medulloblastoma was compared with those of either adult or P5 control groups, and only genes with at least 2-fold difference were selected for hierarchical cluster analyses.
Real-Time PCR.
Real-time PCR was done using a 7900HT Sequence Detection System (ABI) and the TaqMan One Step PCR Master Mix Reagents kit (AIB). Oligonucleotide primers and a TaqMan probe for each gene were designed using Primer Express version 2.0 software (ABI). Primer/TaqMan probe sets for murine genes were: sFrp1 5'-TCCTCCATGCGACAACGA (forward), 5'-TGATTTTCATCCTCAGTGCAAACT (reverse), and 5'-TGAAGTCAGAGGCCATCATTGAACATCTCTG (TaqMan probe); Ptc2 5'-TGGAGCCACCTTGGTACAAGA (forward), 5'-GGCGCTCAGGAAAGCATGT (reverse), and 5'-CTGGCCCTGACAGATGTGGTCCCT (TaqMan probe); Math1 5'-ATGCACGGGCTGAACCA (forward), 5'-TCGTTGTTGAAGGACGGGATA (reverse), and 5'-CCTTCGACCAGCTGCGCAACG (TaqMan probe); Gli1 5'-GCTTGGATGAAGGACCTTGTG (forward), 5'-GCTGATCCAGCCTAAGGTTCTC (reverse), and 5'-CCGGACTCTCCACGCTTCGCC (TaqMan probe); and Jpo1 CTTTCCCTACAGGTGGGTCATT (forward), 5'-TAGGCACAGCTGACAGGCTACA (reverse), and 5'-TACTGGTGACGAAGTGACTGAGTCCCTCAC (TaqMan probe).
Primer/TaqMan probe sets for human genes were: sFRP1 5'-ACACGCACTGCCCTGTCA (forward), 5'-TTTGGAGCGTGGCTATGGA (reverse), and 5'-CTTGTTCTTGCAGCATTCCCGCTCC (TaqMan probe); PTC2 5'-ATCCTAGCTGGGAGCCTGAAG (forward), 5'-TCCCGCATCCCAGAGAGA (reverse), and 5'-TCCACTCTGGCTTCGTGCTTACTTCCA (TaqMan probe); MATH1 5'-CCAAATATGAGACCCTGCAGATG (forward), 5'-CCTCCGCTGGGCGTTT (reverse), and 5'-CCAAATCTACATCAACGCCTTGTCCGA (TaqMan probe); GLI1 5'-GGACCTGCAGACGGTTATCC (forward), 5'-AGCCTCCTGGAGATGTGCAT (reverse), and 5'-CCTCACCCAGCTCCCTCGTAGCTTTC (TaqMan probe); and JPO1 5'-GTGGATGGCTACATGAATGAAGAT (forward), 5'-CGGAAGGGTCACGGATGAT (reverse), and 5'-CTGCCCAGAAGCCGTCGCTCC (TaqMan probe). Real-time PCR data were analyzed using SDS version 2.0 software (ABI). Total RNA from the cerebral cortex of neonatal WT animals for murine genes and total RNA from human adult (Ambion) or fetal (Clontech) brains for human genes were used to generate standard curves for relative quantitation, and 18S rRNA assay reagents (ABI) were used as an internal standard.
Histology and in Situ Hybridization.
Mouse paraffin sections were stained with H&E according to standard procedures. For in situ hybridization, medulloblastomas from Ptc1+/-p53-/- and Lig4-/-p53-/- as well as age-matched adult and P7 WT brains were examined. Nucleotides 247-1375 of sFrp1 (GenBank accession U88566) and nucleotides 701-1251 of Gli1 (GenBank accession AB025922) were generated using standard PCR methods from mouse embryonic cDNA library as a template and were cloned into pSPORT1 (Life Technologies, Inc.). The sFrp1 primers were 5'-GGGCGGCCGCCGACGTCGCCGAGCAACATGG (forward) and 5'-CCGTCGACCACTTAAAAACAGACTGGAAGG (reverse). The Gli1 primers were 5'-GCGGCCGCAAGCTGAAGTCAG (forward) and 5'-GTCGACCGAAGGGTGCGTCT (reverse). Full-length Math1 (atoh1, GenBank accession NM_007500) was isolated using PCR and then cloned into pGEM-T Easy (Promega). Math1 primers were 5'-CTCGTCGACCAGTGCGATGTCCCGCCTGCTG (forward) and 5'-CGCGAGCTGTTGCCTTCCTAACTGGCCTCATC (reverse). IMAGE4
clones for Ptc1 (clone 3989319; nucleotides 3521–4281 and 3' untranslated region) and Ptc2 (clone 3972649; nucleotides 1891–3568 and 3' untranslated region) were obtained from Research Genetics, Inc. In situ hybridization was done using sense and antisense 33P riboprobes using T7, T3, or SP6 RNA polymerase (Promega), according to the manufacturer’s directions. Cryosections were washed in 0.1 M Tris/50 mM EDTA (pH 8.0) buffer, followed by acetylation with 0.25% acetic anhydride/0.1 M triethanolamine (pH 8.0). Slides were then rinsed in 2x SSC and dehydrated in an ethanol series. Sections were hybridized at 60°C overnight in 600 mM NaCl, 10 mM Tris (pH 8.0), 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.1% BSA, 1 mM EDTA (pH 8.0), 10% dextran sulfate, 100 µg/ml salmon sperm DNA, 50 µg/ml total yeast RNA, 50 µg/ml yeast tRNA, and 50% deionized formamide. Slides were then rinsed in 4x SSC/50% formamide, followed by washing in 2x SSC, treated with 20 µg/ml RNase A in 10 mM Tris/1 mM EDTA (pH 8.0)/500 mM NaCl to remove unbound probe, washed in 2x SSC, followed by 0.2x SSC, and then dehydrated in ethanol. The in situ signals were visualized using NTB2 emulsion (Kodak) according to the manufacturer’s recommendations. Images were captured with a Spot digital camera (Diagnostic Instruments, Inc.) and processed using Adobe Photoshop version 7.0 software.
Radiosensitivity Analysis.
Radiosensitivity experiments were done using Ptc1+/- or Ink4c-/- and their WT littermates by irradiating P5 mice with 18 Gy from a cesium irradiator (at a rate of 4.3 Gy/min). Brains were collected at 3 and 6 h postirradiation after transcardial perfusion with 4% paraformaldehyde. Two WT and two mutant P5 pups were collected at each time point. Cryosections of these brains were stained with 1% neutral red (Aldrich Chemical) in 0.1 M acetic acid (pH 4.8). Pyknotic cells present in the EGL of the first and second folia in the midsagittal and paramidsagittal sections of the cerebellum were counted using a double-blind method. The number of pyknotic cells present after irradiation was determined using the Bonferroni t test. Comparisons with P < 0.05 were considered significantly different.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
15% (12
, 13)
], Ptc1+/- with inactivation of p53 [Ptc1+/-p53-/-; medulloblastoma occurrence, 100% (30)
], and DNA ligase IV with p53 inactivation [Lig4-/-p53-/-; medulloblastoma occurrence, 100% (25)
]. We also used medulloblastoma samples from p53-/- mice with disruption of inhibitors of cyclin-dependent kinases (Kip1-/-p53-/-,Ink4d+/-Ink4c-/-p53-/-, Ink4d-/-Kip1+/-p53-/-, and Ink4c-/-p53-/-) in which medulloblastoma occurs with an incidence of 5–15% by 5 months of age.5
Inhibitors of cyclin-dependent kinases can regulate Rb function and may act in this manner to promote transformation, because disabling the Rb and p53 pathways leads to medulloblastoma (31)
. However, although p53-null mice are cancer prone, they rarely develop medulloblastoma (36, 37, 38)
. Morphologically, the mouse tumors resemble the "classic" subtype of human medulloblastoma, except Kip1-/-p53-/- medulloblastomas, which resembled the large-cell anaplastic subtype. All mouse medulloblastomas analyzed were characterized by high proliferation (Ki67 immunostaining) and apoptosis (terminal deoxynucleotidyl transferase-mediated nick end labeling-positive cells) and immunoreactivity toward synaptophysin (12
, 25)
. Initial data analysis comparing medulloblastoma and age-matched control cerebellum showed substantial differences in gene expression between 9-week-old tissue from WT or p53-null mice compared with the medulloblastoma samples. Because medulloblastoma may derive from granule cell precursors (2) , we reasoned that a more appropriate control tissue would be the developing cerebellum. The mouse cerebellum undergoes a period of postnatal neurogenesis until 3 weeks of age, whereby granule cells that are generated in an EGL migrate inward radially, undergoing differentiation to form the internal granule cell population (39) . Therefore, we tested P5 cerebellum, which contains a substantial EGL, as a potential control tissue.
We compared gene expression between medulloblastoma, P5 cerebellum, and adult cerebellum and included genes with at least 2-fold changes for hierarchical cluster analysis. We found that the overall gene expression profile of P5 cerebellum was very similar to that of medulloblastoma (Fig. 1A)
. An expanded view of select regions of the cluster analysis from Fig. 1A
is shown in Fig. 1, B and C
, and underscores the similarity in gene expression between the P5 cerebellum and the medulloblastomas. These results were obtained after normalization of the raw data to remove background noise (see "Materials and Methods"); values were filtered to eliminate nonsignificant signals compared with control signals and further sorted using a one-way ANOVA (cutoff was P < 0.05), and genes with Present calls in more than one-half of the samples were used in the final analysis. Thereby, data with low signal intensity or genes with very low variation throughout the entire sample set were removed; 3,119 genes passed the above criteria and were selected from 12,488 total genes and expressed sequence tags present on the Affymetrix GeneChips for each sample included in this study. Of the 3,119 genes analyzed, 494 genes showed increased expression, and 203 genes showed decreased expression in the P5 cerebellum and medulloblastoma samples compared with adult 9-week-old tissue.6
Thus, gene expression profiles in tumors from the mouse models of medulloblastoma under study here are similar to those of the developing cerebellum but are very different from age-matched controls. These data further strengthen the notion of medulloblastoma arising as a derivative of the external granule cell population of the cerebellum.
|
. To more thoroughly assess this cohort of genes and to identify those that are medulloblastoma specific, we reanalyzed our filtered data set to select those that averaged 2-fold differences between the medulloblastoma samples and normal tissue controls (both P5 and adult cerebellum). This analysis identified a group of 57 genes that showed increased expression in all tumor samples compared with adult and P5 controls, and 26 genes showed decreased gene expression (Fig. 2)
. Because this group contained genes that showed p53-specific changes in expression, we further refined this group to identify those that are specifically up-regulated in medulloblastoma independently of p53. We compared medulloblastoma gene expression to normal cerebellum (adult and P5) and then, independently, to p53-/- tissues (adult and P5) to generate two separate gene lists (Fig. 3A)
. We excluded genes from this list that showed any variation in expression below 2-fold changes between control and tumor groups, further reducing the size of the gene list. Notably, 55 genes were differentially regulated and specific to the medulloblastomas compared with WT cerebellum (all of these appear in the gene list of Fig. 2
), whereas in the p53-/- tissue we identified 59 genes that were differentially regulated. Within these two lists were an overlapping group of 30 genes, the expression of which was altered in the medulloblastoma compared with both WT and p53-/- tissue (Fig. 3B)
. Of this group, there were 21 unique genes (Sox18 appeared twice) that were significantly up-regulated and 9 genes that were down-regulated in all of the tumors. Analysis of gene expression between medulloblastoma samples did not identify any significant genotype-specific patterns.
|
|
-aminobutyric acid receptor subunit 
6, which is associated with mature granule cells (41)
.
Medulloblastoma-specific Genes Are Highly Expressed in Tumor Tissue and Are Present in Immature Proliferative Granule Cells.
We confirmed the array data using real-time PCR to measure expression of representative genes identified as up-regulated in the different medulloblastomas. We selected these genes for further analysis because they were linked to medulloblastoma or to pathways important for cerebellar development. Real-time PCR analysis using TaqMan probes confirmed that high levels of expression of sFrp1, Ptc2, Math1, Gli1, and the Myc-responsive target JPO1 (42)
occurred in all of the medulloblastoma samples but not in WT adult or another tumor type, pro-B-cell lymphoma, also common in Lig4-/-p53-/- animals (Ref. 43
; Fig. 4A
). Consistent with the array data, a number of other genes analyzed by real-time PCR, such as Gli3 and Ptc1, did not show differential expression in the medulloblastoma samples (Ref. 25
and data not shown).
|
. The hybridization signal in the EGL at P7 corresponds to the outer proliferative matrix layer rather than the premigratory mantle zone, indicating that expression of these genes is restricted to regions of cellular proliferation. In situ hybridization analysis of Ptc1, for comparison to Ptc2, and Math1 as a marker for proliferative EGL cells, is also shown (Fig. 4, J and K)
. Ptc2 expression was more restricted than Ptc1 and was confined to the proliferative region of the EGL, suggesting a selective involvement of Ptc2 in proliferative populations of the EGL. In all cases described above, sense probes showed no hybridization signal (data not shown). Thus, these genes are present in the proliferative cells of the EGL and at high levels in medulloblastoma, supporting a role in tumor proliferation.
Genetic Predisposition to Medulloblastomas and the Susceptibility to Genotoxic Stress.
It is intriguing that a variety of different genetic backgrounds predispose to medulloblastomas with similar histology and gene expression profiles. Therefore, we questioned possible connections that could account for the similar gene expression profiles. The loss of one allele of Ptc1 has been shown to predispose to radiosensitivity during development (28
, 29)
, and Ptc1 has been implicated in the DNA damage response via interaction with phosphorylated cyclin B (44)
. Furthermore, tumor development in Ptc1+/- mice is accelerated after ionizing radiation or on a p53-/- genetic background (30)
. Therefore, in some scenarios Ptc1 deficiency predisposes to genotoxic stress, and similar to the Lig4-/-p53-/- mice, this may be important for the development of medulloblastoma. There is also substantial evidence that genotoxic stress features prominently in the developing nervous system, particularly the cerebellum (25
, 45)
. Thus, to establish links between the different medulloblastoma models that could account for similar gene expression profiles, we examined the susceptibility of early postnatal developing Ptc1+/- and Ink4c-/- cerebellum to the effects of DNA damage induced by ionizing radiation. There were significantly more pyknotic nuclei present in the EGL of Ptc1+/- mice after 18 Gy of ionizing radiation compared with WT (P < 0.01), as indicated by arrows in Fig. 5
. Quantitation of pyknotic nuclei in the Ptc1+/- EGL revealed a 20% increase in radiation-induced apoptosis after radiation (Fig. 5D)
. This suggests that Ptc1 haploinsufficiency increases susceptibility to genotoxic stress in the developing cerebellum. The predisposition to medulloblastoma observed in Ptc1+/- and Lig4-/- mice may therefore involve inappropriate responses to genotoxic stress in granule neuron precursors within the EGL. However, analysis of Ink4c-/- mice did not show any unusual sensitivity to genotoxic stress via this apoptosis assay (P < 0.09; data not shown). The substantially lower rate of medulloblastoma occurrence in the Ink4c-/-p53-/- mice compared with the Lig4-/-p53-/- and Ptc1+/-p53-/- mice (100% incidence) possibly reflects a lower susceptibility to endogenous genotoxic stress.
|
). Group B patients also had a significantly greater incidence of ERBB2 oncogene expression (P < 0.05). ERBB2 expression has been shown previously to associate with increased mitotic index, metastasis, and poor clinical outcome in medulloblastoma (46)
. Indeed, group B patients had a shorter 5-year survival rate (54% ± 14.3) compared with group A patients (70% ± 8.5), although this difference in survival did not reach statistical significance. Analysis of Ptc1+/-p53-/- and Lig4-/-p53-/- mouse tumors by Western blot analysis and immunohistochemistry demonstrated that these also express high levels of murine Erbb2, while the levels are very low in developing cerebellum (data not shown). Therefore, the genetic mouse models of medulloblastoma analyzed in the current study may recapitulate some of the molecular characteristics of the aggressive ERBB2-expressing form of human medulloblastoma. The distribution of other clinical parameters including patient age and metastatic stage were not significantly different between groups A and B; however, there was a tendency for group B tumors to display desmoplastic morphology (P < 0.06). Expression of the myc-responsive gene JPO1 was not linked to sFRP1, PTC2, GLI1, or MATH1. The general expression of JPO1 may be related to high c- or N-MYC expression in most of the human tumors (data not shown). Analysis of more human tumors will establish whether the gene expression pattern observed among group B tumors correlates with ERBB2 expression and poor-outcome medulloblastoma and whether this pattern is associated with specific tumor histopathology.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In humans, it is generally problematic to identify and obtain appropriate normal tissue for comparative analysis with tumors. The lack of correct control tissue hampers identification of gene expression changes relevant to transformation. Therefore, in lieu of normal controls, microarray approaches to understanding human tumors often compare features that are clearly defined between tumor groups, such as response to chemotherapy, survival, metastasis, or disease reoccurrence (47, 48, 49, 50, 51, 52)
. Although these studies are valuable, they are unlikely to identify key genetic changes that may be causative for tumorigenesis. Notably, with the exception of N-MYC and GLI, analysis of human medulloblastoma (52)
did not feature the same set of genes as we identified in our study, probably because of an absence of appropriate normal tissue. Using developing cerebellum as a normal tissue control, we identified a small number of genes that are highly up-regulated in medulloblastoma. This previously unrecognized cohort of 20 genes was coordinately up-regulated in all mouse medulloblastomas studied and, therefore, define a characteristic fingerprint for these tumors. Although these genes may also be expressed in proliferating EGL, their inclusion in this list is not simply a reflection of medulloblastoma as an expanding EGL. For example, whereas 670 genes showed similar expression between the P5 cerebellum and the tumors (Fig. 1)
, only 21 genes were exclusively up-regulated in medulloblastoma compared with P5 tissue, suggesting that they may have specific relevance to the tumorigenic process.
Of particular note, we identified the overexpression of select members of the SHH and WNT pathways, PTC2 and sFRP1, in both human and mouse tumors. Although disruption of these pathways can be causal in medulloblastoma (2)
, PTC2 and sFRP1 have not been implicated previously in this tumor. However, although overexpression of PTC2 has been associated with basal cell carcinoma (53)
, mutations in PTC2 are not commonly associated with medulloblastoma (49)
. Given the connection between PTC1 and medulloblastoma, and the sequence (and probably functional) similarity between PTC1 and PTC2 (54)
, abnormally high PTC2 expression in medulloblastoma may reflect disruption of coordinated functions between PTC1 and PTC2 during SHH signaling. Generation of a Ptc2-null mouse will be valuable for investigating the role of this protein during tumorigenesis and cerebellar development. Recently, both sFrp1 and Ptc2 have been shown to be targets of SHH signaling (55)
; although in this case SHH promoted up-regulation of Ptc2 but repression of sFrp1, perhaps reflecting differences in biological context between cell lines and tumors. Moreover, in proliferating granule neuron precursors, sFrp1 and Ptc2 were detected by in situ hybridization, suggesting that these genes are functionally relevant in this proliferating population (Fig. 4)
. Except for an ability of sFrp1 to bind directly to Wnt and the modulation of metanephric development in a cell culture system (56
, 57)
, little functional data are available for the secreted Frizzled-related genes, although modulation of the Wnt pathway is likely to have substantial effects upon cerebellar development (see the "Introduction").
In addition to sFrp1 and Ptc2, other genes identified in our analyses, such as Gli1, N-Myc, and cyclin D1, are functionally important for cerebellar development and further point to a persistent activation of the SHH pathway (58, 59, 60, 61) . N-Myc has been shown to control cell cycle progression of cerebellar granule neuron precursor cells after SHH stimulation (59) , and N-Myc-null cerebellum is markedly reduced in size because of attenuation of proliferation linked to a failure to suppresses p18Ink4c and p27Kip1 (61) . Thus, overexpression of N-Myc likely provides a significant proliferative advantage to cerebellar granule cells. Previous work reported N-MYC up-regulation as a feature of desmoplastic medulloblastoma (52) , although our data suggest a more general N-MYC association with this disease. Furthermore, cyclin D is directly linked to growth promotion by SHH (58) , and inactivation of cyclin D1 (and cyclin D2) in the mouse markedly affects normal cerebellar development and function (60) . Thus, our data argue that the SHH pathway is perturbed in all mouse medulloblastomas studied, despite the different genetic lesions underpinning these tumors. Importantly, the readout of this pathway is context dependent because SHH stimulation in vitro using either cerebellar granule neurons or mesenchymal cells leads to a different spectrum of gene expression changes than those we observed in medulloblastoma (55 , 62) . Therefore, in vivo analysis of cerebellar signaling provides important additional insights into the biology of the SHH pathway.
Many of the mutant mice used in this study are p53 deficient. p53 is a transcription factor important for control of cell proliferation and apoptosis, and mutation of p53 is an important cooperating event in carcinogenesis (63)
. Published data support a strong connection between p53 mutations in human medulloblastoma; the results of several studies suggest an incidence of 10% (64, 65, 66, 67, 68, 69)
. Additionally, high p53 levels are found in up to 40% of medulloblastomas, implying that some aspect of p53 signaling is dysfunctional, because elevated levels of p53 exist without the tumor cells undergoing proliferative arrest or apoptosis (70
, 71)
. For comparison, PTC1 mutations have been reported to occur at a similar frequency of 9.5% (72, 73, 74, 75, 76)
. Thus, mutation of p53 represents one of the more frequent genetic alterations reported to occur in human medulloblastoma. All of the mouse models described here, with the exception of Ptc1+/- mice, harbor germ-line mutations in p53, further underscoring the relevance of p53 deficiency to this disease. Interestingly, the gene expression patterns observed in the Ptc1+/- tumors are remarkably similar to those in Ptc1+/-p53-/- tumors. As has been suggested for lymphoma (77)
, defective p53 responses may attenuate apoptosis, subsequently increasing the probability that granule neuron precursor cells will develop mutations that promote deregulated growth leading to medulloblastoma. Because cerebellar granule cells are the most abundant neuronal cell in the nervous system, they are a large target for the stochastic acquisition of mutations.
The importance of the mouse data for understanding human medulloblastoma is highlighted by the identification of a similar subset of genes coordinately deregulated in human tumors (16 of 50; 32%) including sFRP1 and PTC2. Thus, the mouse models of medulloblastoma studied in this report share molecular characteristics with a significant proportion of human medulloblastomas and therefore provide important models of this disease. This unique sFRP1 expression signature occurs despite up-regulation of either c- or N-MYC in most of the human tumors,7
suggesting that MYC overexpression is a general growth-promoting activity, rather than a specific primary event. The increase in MYC expression also probably accounts for the relatively high JPO1 expression throughout the human medulloblastoma samples. It will be important to identify what distinguishes the sFRP1-expressing tumors from the other tumors used in our analysis. If this group (Fig. 6
, group A) does not involve alterations in the SHH pathway, then it will be important to determine which other pathways are perturbed in this subset of medulloblastomas. A significantly greater proportion of the sFRP1-positive tumors expressed high levels of ERBB2 protein. Because high sFRP1 expressing tumors correlate with increased ERBB2 and because this is linked to a poor clinical outcome in medulloblastoma patients (35)
, it will be important to increase the human tumor sample size to better determine correlates with other tumor features such as prognosis and metastasis. The molecular fingerprint for medulloblastoma described in this report may help to define a unique subtype of tumor for which particular therapies may be more appropriate.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by NIH Grants CA-096832, NS-37956, NS-39867, and CA-21765 and by a grant from the American Lebanese and Syrian Associated Charities of Saint Jude Children’s Research Hospital. Supplemental data are available on the AACR Web site. 
2 To whom requests for reprints should be addressed, at Genetics, 332 North Lauderdale Street, Memphis, TN 38105. Phone: (901) 495-2700; Fax: (901) 526-2907; E-mail: peter.mckinnon{at}stjude.org
3 The abbreviations used are: SHH, sonic hedgehog; WNT, wingless; PTC/Ptc, Patched; sFRP/sFrp, secreted Frizzled-related protein; APC, adenomatous polyposis coli; GSK-3ß, glycogen synthase kinase-3ß; WT, wild type; P5 and P7, postnatal days 5 and 7, respectively; EGL, external germinal layer.
4 Internet address: http://est.llnl.gov/.
5 F. Zindy and M. F. Roussel, manuscript in preparation.
6 A complete list of these 697 genes is presented as Supplementary data, Fig. 1
S.
7 R. Hernan, Y. Lee, P. J. McKinnon, R. J. Gilbertson, unpublished observation.
Received 5/ 2/03. Revised 6/12/03. Accepted 6/17/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
subunit is expressed only in cerebellar granule cells. J. Mol. Biol., 214: 619-624, 1990.[Medline]
This article has been cited by other articles:
|
|
 |
|
  R. J. Ward, L. Lee, K. Graham, T. Satkunendran, K. Yoshikawa, E. Ling, L. Harper, R. Austin, E. Nieuwenhuis, I. D. Clarke, et al. Multipotent CD15+ Cancer Stem Cells in Patched-1-Deficient Mouse Medulloblastoma Cancer Res., June 1, 2009; 69(11): 4682 - 4690. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Uziel, F. V. Karginov, S. Xie, J. S. Parker, Y.-D. Wang, A. Gajjar, L. He, D. Ellison, R. J. Gilbertson, G. Hannon, et al. The miR-17~92 cluster collaborates with the Sonic Hedgehog pathway in medulloblastoma PNAS, February 24, 2009; 106(8): 2812 - 2817. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P.-O. Frappart, Y. Lee, H. R. Russell, N. Chalhoub, Y.-D. Wang, K. E. Orii, J. Zhao, N. Kondo, S. J. Baker, and P. J. McKinnon Recurrent genomic alterations characterize medulloblastoma arising from DNA double-strand break repair deficiency PNAS, February 10, 2009; 106(6): 1880 - 1885. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Kessler, H. Hasegawa, S. N. Brun, B. A. Emmenegger, Z.-J. Yang, J. W. Dutton, F. Wang, and R. J. Wechsler-Reya N-myc alters the fate of preneoplastic cells in a mouse model of medulloblastoma Genes & Dev., January 15, 2009; 23(2): 157 - 170. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. C. Johnson, M. E. Dillard, P. Baluk, D. M. McDonald, N. L. Harvey, S. L. Frase, and G. Oliver Lymphatic endothelial cell identity is reversible and its maintenance requires Prox1 activity Genes & Dev., December 1, 2008; 22(23): 3282 - 3291. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. B. Corcoran, T. Bachar Raveh, M. T. Barakat, E. Y. Lee, and M. P. Scott Insulin-like Growth Factor 2 Is Required for Progression to Advanced Medulloblastoma in patched1 Heterozygous Mice Cancer Res., November 1, 2008; 68(21): 8788 - 8795. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Fan and C. G. Eberhart Medulloblastoma Stem Cells J. Clin. Oncol., June 10, 2008; 26(17): 2821 - 2827. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Galvin, H. Ye, D. J. Erstad, R. Feddersen, and C. Wetmore Gli1 Induces G2/M Arrest and Apoptosis in Hippocampal but Not Tumor-Derived Neural Stem Cells Stem Cells, April 1, 2008; 26(4): 1027 - 1036. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Grimmer and W. A. Weiss BMPs oppose Math1 in cerebellar development and in medulloblastoma Genes & Dev., March 15, 2008; 22(6): 693 - 699. [Full Text] [PDF]
|
|
|
|
|
|
H. Zhao, O. Ayrault, F. Zindy, J.-H. Kim, and M. F. Roussel Post-transcriptional down-regulation of Atoh1/Math1 by bone morphogenic proteins suppresses medulloblastoma development Genes & Dev., March 15, 2008; 22(6): 722 - 727. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Briggs, I. M. Corcoran-Schwartz, W. Zhang, T. Harcke, W. L. Devereux, S. B. Baylin, C. G. Eberhart, and D. N. Watkins Cooperation between the Hic1 and Ptch1 tumor suppressors in medulloblastoma Genes & Dev., March 15, 2008; 22(6): 770 - 785. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Hambardzumyan, O. J. Becher, M. K. Rosenblum, P. P. Pandolfi, K. Manova-Todorova, and E. C. Holland PI3K pathway regulates survival of cancer stem cells residing in the perivascular niche following radiation in medulloblastoma in vivo Genes & Dev., February 15, 2008; 22(4): 436 - 448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kawagoe, A. Kandilci, T. A. Kranenburg, and G. C. Grosveld Overexpression of N-Myc Rapidly Causes Acute Myeloid Leukemia in Mice Cancer Res., November 15, 2007; 67(22): 10677 - 10685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Salsano, L. Croci, E. Maderna, L. Lupo, B. Pollo, M. T. Giordana, G. G. Consalez, and G. Finocchiaro Expression of the neurogenic basic helix-loop-helix transcription factor NEUROG1 identifies a subgroup of medulloblastomas not expressing ATOH1 Neuro-oncol, July 1, 2007; 9(3): 298 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Farioli-Vecchioli, M. Tanori, L. Micheli, M. Mancuso, L. Leonardi, A. Saran, M. T. Ciotti, E. Ferretti, A. Gulino, S. Pazzaglia, et al. Inhibition of medulloblastoma tumorigenesis by the antiproliferative and pro-differentiative gene PC3 FASEB J, July 1, 2007; 21(9): 2215 - 2225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Sasai, J. T. Romer, H. Kimura, D. E. Eberhart, D. S. Rice, and T. Curran Medulloblastomas Derived from Cxcr6 Mutant Mice Respond to Treatment with a Smoothened Inhibitor Cancer Res., April 15, 2007; 67(8): 3871 - 3877. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Zindy, T. Uziel, O. Ayrault, C. Calabrese, M. Valentine, J. E. Rehg, R. J. Gilbertson, C. J. Sherr, and M. F. Roussel Genetic Alterations in Mouse Medulloblastomas and Generation of Tumors De novo from Primary Cerebellar Granule Neuron Precursors Cancer Res., March 15, 2007; 67(6): 2676 - 2684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. E. Bar, A. Chaudhry, M. H. Farah, and C. G. Eberhart Hedgehog Signaling Promotes Medulloblastoma Survival via BclII Am. J. Pathol., January 1, 2007; 170(1): 347 - 355. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Pogoriler, K. Millen, M. Utset, and W. Du Loss of cyclin D1 impairs cerebellar development and suppresses medulloblastoma formation Development, October 1, 2006; 133(19): 3929 - 3937. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Nieuwenhuis, J. Motoyama, P. C. Barnfield, Y. Yoshikawa, X. Zhang, R. Mo, M. A. Crackower, and C.-c. Hui Mice with a Targeted Mutation of Patched2 Are Viable but Develop Alopecia and Epidermal Hyperplasia. Mol. Cell. Biol., September 1, 2006; 26(17): 6609 - 6622. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Zindy, P. S. Knoepfler, S. Xie, C. J. Sherr, R. N. Eisenman, and M. F. Roussel N-Myc and the cyclin-dependent kinase inhibitors p18Ink4c and p27Kip1 coordinately regulate cerebellar development PNAS, August 1, 2006; 103(31): 11579 - 11583. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Lee, H. L. Miller, H. R. Russell, K. Boyd, T. Curran, and P. J. McKinnon Patched2 modulates tumorigenesis in patched1 heterozygous mice. Cancer Res., July 15, 2006; 66(14): 6964 - 6971. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Shakhova, C. Leung, E. van Montfort, A. Berns, and S. Marino Lack of Rb and p53 Delays Cerebellar Development and Predisposes to Large Cell Anaplastic Medulloblastoma through Amplification of N-Myc and Ptch2. Cancer Res., May 15, 2006; 66(10): 5190 - 5200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. T. Yan, D. Kaushal, M. Murphy, Y. Zhang, A. Datta, C. Chen, B. Monroe, G. Mostoslavsky, K. Coakley, Y. Gao, et al. XRCC4 suppresses medulloblastomas with recurrent translocations in p53-deficient mice PNAS, May 9, 2006; 103(19): 7378 - 7383. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. Thompson, C. Fuller, T. L. Hogg, J. Dalton, D. Finkelstein, C. C. Lau, M. Chintagumpala, A. Adesina, D. M. Ashley, S. J. Kellie, et al. Genomics Identifies Medulloblastoma Subgroups That Are Enriched for Specific Genetic Alterations J. Clin. Oncol., April 20, 2006; 24(12): 1924 - 1931. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Sasai, J. T. Romer, Y. Lee, D. Finkelstein, C. Fuller, P. J. McKinnon, and T. Curran Shh pathway activity is down-regulated in cultured medulloblastoma cells: implications for preclinical studies. Cancer Res., April 15, 2006; 66(8): 4215 - 4222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Su, V. Gopalakrishnan, D. Stearns, K. Aldape, F. F. Lang, G. Fuller, E. Snyder, C. G. Eberhart, and S. Majumder Abnormal Expression of REST/NRSF and Myc in Neural Stem/Progenitor Cells Causes Cerebellar Tumors by Blocking Neuronal Differentiation. Mol. Cell. Biol., March 1, 2006; 26(5): 1666 - 1678. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Kawagoe and G. C. Grosveld Conditional MN1-TEL knock-in mice develop acute myeloid leukemia in conjunction with overexpression of HOXA9 Blood, December 15, 2005; 106(13): 4269 - 4277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kawagoe and G. C. Grosveld MN1-TEL myeloid oncoprotein expressed in multipotent progenitors perturbs both myeloid and lymphoid growth and causes T-lymphoid tumors in mice Blood, December 15, 2005; 106(13): 4278 - 4286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Uziel, F. Zindy, S. Xie, Y. Lee, A. Forget, S. Magdaleno, J. E. Rehg, C. Calabrese, D. Solecki, C. G. Eberhart, et al. The tumor suppressors Ink4c and p53 collaborate independently with Patched to suppress medulloblastoma formation Genes & Dev., November 15, 2005; 19(22): 2656 - 2667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Argenti, R. Gallo, L. Di Marcotullio, E. Ferretti, M. Napolitano, S. Canterini, E. De Smaele, A. Greco, M. T. Fiorenza, M. Maroder, et al. Hedgehog Antagonist RENKCTD11 Regulates Proliferation and Apoptosis of Developing Granule Cell Progenitors J. Neurosci., September 7, 2005; 25(36): 8338 - 8346. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Huang, C. S.W. Ho, R. Ponzielli, D. Barsyte-Lovejoy, E. Bouffet, D. Picard, C. E. Hawkins, and L. Z. Penn Identification of a Novel c-Myc Protein Interactor, JPO2, with Transforming Activity in Medulloblastoma Cells Cancer Res., July 1, 2005; 65(13): 5607 - 5619. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Romer and T. Curran Targeting Medulloblastoma: Small-Molecule Inhibitors of the Sonic Hedgehog Pathway as Potential Cancer Therapeutics Cancer Res., June 15, 2005; 65(12): 4975 - 4978. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. G. Oliver, T. A. Read, J. D. Kessler, A. Mehmeti, J. F. Wells, T. T. T. Huynh, S. M. Lin, and R. J. Wechsler-Reya Loss of patched and disruption of granule cell development in a pre-neoplastic stage of medulloblastoma Development, May 15, 2005; 132(10): 2425 - 2439. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. N. Fuller, X. Su, R. E. Price, Z. R. Cohen, F. F. Lang, R. Sawaya, and S. Majumder Many human medulloblastoma tumors overexpress repressor element-1 silencing transcription (REST)/neuron-restrictive silencer factor, which can be functionally countered by REST-VP16 Mol. Cancer Ther., March 1, 2005; 4(3): 343 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Hallahan, J. I. Pritchard, S. Hansen, M. Benson, J. Stoeck, B. A. Hatton, T. L. Russell, R. G. Ellenbogen, I. D. Bernstein, P. A. Beachy, et al. The SmoA1 Mouse Model Reveals That Notch Signaling Is Critical for the Growth and Survival of Sonic Hedgehog-Induced Medulloblastomas Cancer Res., November 1, 2004; 64(21): 7794 - 7800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Di Marcotullio, E. Ferretti, E. De Smaele, B. Argenti, C. Mincione, F. Zazzeroni, R. Gallo, L. Masuelli, M. Napolitano, M. Maroder, et al. RENKCTD11 is a suppressor of Hedgehog signaling and is deleted in human medulloblastoma PNAS, July 20, 2004; 101(29): 10833 - 10838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. T. Kho, Q. Zhao, Z. Cai, A. J. Butte, J. Y.H. Kim, S. L. Pomeroy, D. H. Rowitch, and I. S. Kohane Conserved mechanisms across development and tumorigenesis revealed by a mouse development perspective of human cancers Genes & Dev., March 15, 2004; 18(6): 629 - 640. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
,