Genomic Amplification of Orthodenticle Homologue 2 in Medulloblastomas

Priority Reports

Kathy Boon,1 Charles G. Eberhart,2 and Gregory J. Riggins1

Departments of 1Neurosurgery and 2Pathology, Johns Hopkins University School of Medicine, Baltimore, Maryland

Abstract activated in about 5% of medulloblastomas (4). Comparative To better understand the genetic basis of medulloblastoma genomic hybridization studies have identified large genomic development, we sought genomic amplifications and dele- regions of amplification and deletion, but typically only large tions in these tumors using digital karyotyping in combina- megabase regions of alteration can be found, making it difficult to tion with expression analysis. Five medulloblastoma genomes identify the target of gain or loss. However, higher-resolution were karyotyped by sequencing an average of 195,745 techniques for looking at DNA copy number have recently been genomic DNA tags for each analysis. Tags were tallied at developed, such as array-based comparative genomic hybridiza- unique positions and mapped to the human genome to tion. A recent study of 58 known oncogenes in medulloblastomas determine DNA copy numbers in high resolution along each showed genomic gains of several genes including PIK3CA, PGY1, chromosome. Genomic alterations normally associated with MET, ERBB2, and CSE1L (5). Another high-resolution method medulloblastomas, including MYC amplification and isochro- called digital karyotyping was recently developed based on serial analysis of gene expression (SAGE) technology (6), and employed to mosome 17q, were easily detected. Surprisingly, analysis of only five genomes revealed novel amplicons on chromosome implicate new genes in colon cancer (7, 8). Digital karyotyping 14q, one of which contained the orthodenticle homologue 2 works by counting 21 bp sequence tags generated from specific (OTX2) homeobox gene. DNA copy number analysis showed locations in the genome; normally the 21 bp adjacent to the first that OTX2 had undergone genomic amplification in 2 of 11 NlaIII site next to a SacI restriction enzyme site. Approximately medulloblastoma cell lines and 8 of 42 primary tumors. The 200,000 tags are counted by high-throughput sequencing and the three genes and a predicted open reading frame flanking tags are mapped to the genome and counted to identify OTX2 in the 14q amplicon were not amplified in at least one chromosomal changes such as amplifications and deletions. of the other nine amplicons, implicating OTX2 as the gene In this study, digital karyotyping was used to determine genome- target conferring a selective advantage. The degree of OTX2 wide DNA copy number for five medulloblastomas. A novel amplification ranged from 8 copies to over 50 copies of the amplification located on 14q22.3 was validated by quantitative real- gene. OTX2 transcript was highly and specifically expressed in time PCR (Q-OCR) in 2 of 11 (18%) medulloblastoma cell lines and medulloblastoma or developing cells. Serial analysis of gene in 8 of 42 (19%) primary tumors. The amplicon included the gene expression of 240 different human tumors or normal tissues orthodenticle homologue 2 (OTX2), a homeobox gene. A combina- revealed that 96% of all 783 OTX2 transcripts sequenced were tion of gene expression analysis and genomic mapping was used to in medulloblastomas or embryonic stem cells. OTX2 functions implicate only this gene. OTX2 plays an important role in to specify the fate of neuroectoderm in various regions of the specification and regionalization of the forebrain and midbrain in developing brain. This developmental role is consistent with early embryogenesis (9), suggesting its possible functional role the evidence suggesting that OTX2 is a medulloblastoma in medulloblastoma development. oncogene. (Cancer Res 2005; 65(3): 703-7) Materials and Methods Introduction Primary Tumors and Cell Lines. Medulloblastoma tissue was snap frozen in liquid nitrogen. A neuropathologist gave an unambiguous Medulloblastomas are a frequently studied brain cancer occur- diagnosis of medulloblastoma for all cases, and confirmed that the sample ring in about 1 in 200,000 children annually. These tumors are a margins were tumor cells and not normal tissue. Thirty samples from Johns high-grade embryonal tumor of the cerebellum, with a small round Hopkins Neuropathology, eight from the National Cancer Institute blue cell appearance similar to other tumors arising from primitive Cooperative Human Tissue network, and four from the Duke Brain Tumor neuroectoderm (1). Bank were used in these Institutional Review Board-approved studies. The The molecular basis of medulloblastoma formation is beginning following medulloblastoma cell lines were used in this study: D283 Med (10), to be understood. For example, the sonic hedgehog and the Wnt D341 Med, D425 Med, D556 Med, MCD1, UW228-2, D721 Med, D487 Med, pathway can be activated in a percentage of medulloblastomas by D581 Med, and MHH-Med-1 (11). mutation of critical pathway regulators (2, 3). Larger chromosomal Digital Karyotyping. Genomic DNA was isolated using a DNeasy kit alterations, most frequently isochromosome 17q with loss of a 17p, (Qiagen, Valencia, CA) according to the instructions of the manufacturer have been documented. Genomic amplification via double-minute from five medulloblastoma cell lines: D487 Med, D556 Med, D721 Med, D283 Med, and MHH-Med-1. For each library 1 Ag of genomic DNA was digested chromosomes is a mechanism where the MYC oncogene is with mapping enzyme SacI ligated to biotinylated SacI linkers (Integrated DNA Technologies, Coralville, IA) and digested with the fragmenting enzyme NlaIII. DNA fragments containing the biotinylated linkers were isolated using streptavidin-coated magnetic beads (Dynal Biotech, Request for reprints: Gregory J. Riggins, Department of Neurosurgery, 5th Floor, Brown Deer, WI) and ligated to linkers including recognition sites for Mason F. Lord Building, Center Tower, 5200 Eastern Avenue, Baltimore, MD 21224. Phone: 410-550-9686; Fax: 410-550-9689; E-mail: [email protected]. MmeI. The 21 bp sequence tags were released by digestion with MmeIas I2005 American Association for Cancer Research. has been described for Long SAGE (6). A detailed protocol can be obtained www.aacrjournals.org 703 Cancer Res 2005; 65: (3). February 1, 2005

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 2005 American Association for Cancer Research. Cancer Research at http://www.digitalkaryotyping.org (7). The isolated tags are self-ligated, An average of 104,030 filtered tags per library (Table 1) was aligned PCR amplified concatenated, cloned in pZero (Invitrogen, Carlsbad, CA), across each chromosome. Tag densities were analyzed in windows and sequenced. The SAGE 2000 software package enables the extraction of ranging from 50 to 1,000 virtual tags (7). These densities were the genomic tags from the sequence files. The virtual genomic tags were calculated for each window by summing the experimental tags extracted from the human genome sequence (UCSC Genome Bioinfor- observed and dividing the sum by the average tag count for all the matics, July 2003 assembly, http://genome.ucsc.edu/) and downloaded from http://www.digitalkaryotyping.org. DNA from plasmid inserts containing same-sized windows across the genome. serial genomic tags were purified and sequenced at Agencourt Bioscience The MYC gene was used as an internal positive control. More Corporation (Beverly, MA) as part of the Cancer Genome Anatomy Project. than 25 copies at the genomic DNA level were found for D487 and Quantitative Real-Time PCR. Copy number changes between normal D556 Med and between 2 and 6 copies for D721 and D283 Med. human DNA and medulloblastoma cell lines or primary tumors were Digital karyotyping accurately predicted the published genomic determined by Q-PCR on an iCycler apparatus (Bio-Rad, Hercules, CA). The copy numbers for all five cell lines previously determined by repetitive element Line-1, which has an equivalent number in cancer and fluorescent in situ hybridization (13). MYC diploid genome copy normal genomes, was used for the normalization of DNA content. numbers for all five cell lines were also corroborated by Q-PCR Calculations and PCR conditions were all done as previously described (Table 1). Further analysis of the digital karyotypes of D556 Med, (7, 8). All PCRs were done in triplicates and threshold cycle numbers were D721 Med, D283 Med, and MHH-Med-1 did not reveal genomic averaged. PCR primers were designed using Primer 3 http://www.genome. alterations other than those already known (14, 15). Analysis of wi.mit.edu/cgi-bin/primer/primer3_www.cgi) and are available upon request. For transcript analysis total RNA was extracted from 11 medullo- chromosome 17 provided a second internal control. D487 Med had blastoma cell lines using the RNAgents total RNA isolation system a digital karyotype that produced an average of three copies of 17q (Promega, Madison, WI). Synthesis of cDNA and quantitative PCR were and one copy of 17p. This karyotype was consistent with one done as previously described (12). isochromosome 17q and one normal chromosome 17, as Immunoblotting. Cells from various medulloblastoma cell lines previously described for D487 Med (14) and observed in many were lysed in 50 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 1% NP40, medulloblastomas (1). 0.25% Na-deoxycholate, 1 mmol/L EDTA, and protease inhibitors (Protease Of most interest were two novel chromosome 14q amplifica- cocktail, Roche Diagnostics, Indianapolis, IN). Total protein concentration tions found in D487 Med (Fig. 1A). Amplicon A, on chromosomal was assayed with a Bradford Protein assay kit (Bio-Rad). Equal amounts of band 14q22.3, extended from genomic map positions 53.7 to total protein were separated on an SDS-polyacrylamide gel and electro- 55.5 Mb from pter. Tag densities ranged from f8 to 13 copies per blotted onto nitrocellulose membrane. Blocking of membrane and incubation with antibodies was done according to standard procedures diploid genome within this region (Fig. 1A). The second amplified and visualized with the enhanced chemiluminescence plus detection region (Amplicon B) has tag densities between 8 and 12.4 and is system (Amershan Biosciences, Piscataway, NJ). Monoclonal antibodies located between 71.5 and 73.0 Mb on chromosomal band 14q24.3. raised against OTX2 (R&D systems, Minneapolis, MN) were used at a 1:50 Four known genes and one open reading frame are located in dilution overnight at 4jC. Amplicon A and 14 genes in Amplicon B. Previous cytogenetic studies of D487 Med had failed to detect this amplification and Results indicated that chromosome 14 was present with no gross rearrangements (14). Digital Karyotyping of Medulloblastomas. DNA from five OTX2 Transcript Expression. We observed that for the 19 established medulloblastoma cell lines was purified for digital genes in two 14q amplicons, only one gene, OTX2, had high karyotyping (7). Cultured cell lines were used to avoid contami- messenger RNA transcript levels specific to medulloblastoma. nating normal cells. The majority of 21 bp genomic tags produced SAGE data from 20 primary medulloblastomas and a variety of by this method contain sufficient information to identify a unique malignant and normal tissues totaling 240 different cell and tumor genomic position. We sequenced a total of 978,724 tags with an types were surveyed (refs. 12, 16; Fig. 2A, http://cgap.nci.nih.gov/ average of 195,745 tags per library (Table 1). The experimental tags SAGE). The 10 bp OTX2 SAGE tag (ACCAACTGGT) occurs 783 were filtered to remove repetitive elements and compared with times in the National Cancer Institute SAGE Genie database of virtual tag sequences extracted from the public genome sequence. 15.7 million human tags, with 96% of the OTX2 transcript tags

Table 1. Summary of digital karyotyping libraries and copy number amplifications

Digital karyotyping library Total tags Filtered tags* Digital copy numbersc Q-PCR copy numbersc

MYC OTX2 MYC OTX2

D487 Med 192,766 103,309 25.4 13.2 25.4 18.6 D556 Med 176,873 76,602 71.4 3.2 131.0 3.0 D721 Med 251,684 134,313 2.2 2.0 2.0 2.0 D283 Med 188,292 110,006 6.8 2.4 7.2 2.0 MHH1-Med-1 169,109 95,918 2.4 1.4 2.4 0.8

*Number of tags that map to unique loci in the genome. Tag sequences that map to multiple regions or no match to the genome are removed. cCopy numbers were calculated per diploid genome as described (7).

Cancer Res 2005; 65: (3). February 1, 2005 704 www.aacrjournals.org

predominant band at a slightly higher molecular weight than predicted, but thought to represent a posttranslationally modified OTX2 protein, correlated with the transcript levels with two exceptions, D721 Med and D581 Med (Fig. 2B). High protein expression was found in the two cell lines with OTX2 genomic amplification. 14q Minimal Amplified Region. The transcript expression pattern of OTX2 and its role in brain development prompted us to further analyze Amplicon A on 14q22.3. We used Q-PCR and a genomic DNA template to determine copy number for the genes in the amplicon from the five original plus six additional medulloblastoma cell lines. For OTX2 two independent sets of genomic primers were used, one set flanking exon 1 and the second set between exon 4 and the exon 4/5 intron (Fig. 1C). The OTX2 copy numbers for all five original cell lines were confirmed (Table 1) and a new OTX2 25- copy amplicon was found in D425 Med. In contrast, previous cytogenetic studies by Aldorsi et al. (15) showed for D425 Med two copies of intact chromosome 14 without amplification. The other known and/or predicted genes in this region, KTN1, PELI2, C14orf101, and SEC10L1, were also tested for genomic amplification in medulloblastoma cell lines (Fig. 3). However, in D425 Med the amplification was observed only in OTX2 and not in the flanking genes (Fig. 3). Furthermore, expression analysis by Q- PCR in D425 Med showed a very high level of OTX2 transcript and a barely detectable level of C14orf101 (Fig. 2B). Barring an unidentified gene hiding within the D425 Med amplicon, the limited amplified region implicated only OTX2 as a gene that might provide a selective advantage to tumor cells when replicated. Genomic Amplification in Primary Tumors. We extended the analysis of OTX2 gene copy number to 42 primary human medulloblastoma samples. In 8 samples (19%), OTX2 was genomi- Figure 1. A, low-resolution tag density maps of three regions in cell line D487 cally amplified (Fig. 3). The size of these amplicons was consistent Med generated by digital karyotyping using a sliding window of 300. with the 14q amplicon A found in D487 Med (Fig. 1B) and generally Top, chromosome 17 data for a genome with one normal chromosome, an isochromosome 17q, and loss of 17p. Bottom, two chromosome 14 amplicons contained flanking genes. The relative copy numbers for these eight found on 14q22.3 (A) and 14q24.3 (B). MYC amplification on chromosome 8 samples are shown in Fig. 3 and varied from 8 to 56 copies. is shown as a reference. Y axis, copy numbers per haploid genome; X axis, OTX2 Mutation Analysis. Some amplified oncogenes, such as the chromosomal position extending over the whole chromosome. B, graphic representation of the genes located within amplicon (A) on 14q22.3. epidermal growth factor receptor, contain further rearrangements or Arrows, indication whether the gene is transcribed from the positive (!) point mutations beyond genomic amplification. In an attempt to or negative (p) strand. A detailed view of the corresponding region on 14q22.3 can be obtained at http://genome.ucsc.edu/ using the coordinates 53.7 and further implicate OTX2, we sought such activating mutations in 23 55.8 Mb. C, display of messenger RNA/genomic DNA alignment for OTX2. medulloblastoma genomic DNA samples. Exons 3, 4, and 5 (Fig. 1C), The exons are indicated by the boxes along the genomic DNA (line in bold). Striped boxes, coding sequences. Solid black arrows, primer positions for which include the complete coding sequence, were amplified for se- genomic amplification detection (two sets); open gray arrows, primers used quencing. However, a wild-type sequence was observed in each case. to amplify exons for mutation detection; dotted arrows, primers used to confirm expression of the gene transcripts using Q-PCR. Discussion being found highly expressed in medulloblastomas or embryonic Digital karyotyping was used to analyze genomic copy number in stem cells. There was little or no expression of OTX2 in the large five medulloblastoma cell lines and identified two novel chromo- variety of normal adult tissues or other cancer types. Two some 14 amplicons. The amplicon on 14q22.3 included the OTX2 alternative transcripts of OTX2 are known with an identical 3V gene and was shown to be present in both medulloblastoma cell end. The longer transcript (a) contains all five exons and the lines and primary tumors at a similar incidence of f19%. Although shorter transcript (b) lacks the first two exons (Fig. 1C). Both the amplicon typically contained flanking genes, one cell line had transcripts encode for the same protein with the exception of the only OTX2 amplified, suggesting OTX2 as the target of amplifica- first eight amino acids (encoded by exon 3) that are missing in tion. Of the five possible genes in the amplicon, only OTX2 had an transcript b. Q-PCR of OTX2 transcription in the medulloblastoma expression pattern consistent with pathologic transcriptional cell line panel also showed overexpression in 7 of 11 samples activation of the gene in medulloblastoma. (Fig. 2B). In particular, the cell lines D487 and D425 showed We found the combined analysis of high-resolution copy number respectively 15- and 45-fold increase in transcript levels when and gene expression specificity to be useful in this case. OTX2 had compared with normal cerebellum. The longer transcript (a) was previously been reported to be overexpressed at the transcript level detected in 6 of 11 medulloblastoma cell lines using conventional in medulloblastomas when compared with fetal brain or cerebellum PCR (Fig. 2B). High protein expression was also found in D487 and (17, 18), but not implicated through genomic alteration. Our D425 Med using a monoclonal OTX2 antibody. A clearly identification of a genomically amplified transcription factor in www.aacrjournals.org 705 Cancer Res 2005; 65: (3). February 1, 2005

Figure 2. A, expression of OTX2 in 240 different SAGE libraries indicating specificity of expression for medulloblastoma and embryonic stem cells. These ‘‘Digital Northern’’ results for OTX2 (SAGE tag: ACCAACTGGT) were obtained from SAGE Genie (http://cgap.nci.nih.gov/SAGE). B, Q-PCR analysis of OTX2 transcript levels in medulloblastoma cell lines. The graph shows the relative expression level when compared with normal cerebellum (2). Bottom, end-point analysis of the Q-PCR, detection of the full-length transcript using PCR primers starting at exon 1 or exon 2, and a control for cDNA quality. OTX2 monoclonal antibodies identify a similar band in cell lysates from the various medulloblastoma cell lines. The protein is detectable in most cell lines with a high transcript expression like D487 and D425 Med and barely or not detectable in cell lines with low or no transcript expression like MHH-1-Med, UW228-2, and MCD1. medulloblastoma is not without precedent, as this is the case for the MYC oncogene. Whereas increased gene dosage should account for elevated OTX2 expression in f19% of medulloblastomas, a much higher percentage of medulloblastomas express the transcript and protein at high levels, suggesting additional mechanism(s) for pathway activation. We did not locate mutations in OTX2 coding sequence, and if they exist, they are a rare event. However, activation of the OTX2 promoter by mutations or upstream transcription factors remains a possibility for activating transcription. In embryogenesis OTX2 expression is tightly regulated by cis-acting elements including TCF/Lef binding sites which would place the Wnt signaling pathway upstream of OTX2 (9). Wnt signaling is activated in approximately one quarter of medulloblastomas, and it may drive OTX2 expression in a significant fraction of cases. Development of the forebrain, midbrain, and anterior hindbrain depend on proper OTX2 levels (19). OTX2 controls the forebrain and midbrain development in a dose-dependent manner by regulating Fgf8 (along the anterior-posterior axis) and SHH (along the dorsal-ventral axis; ref. 20). Ectopically Figure 3. Extent of genomic amplification in 10 medulloblastomas with OTX2 expressing the OTX2 gene in the anterior hindbrain of developing amplification and three controls. Relative genomic amplification levels are indicated as copies per diploid genome and generated by Q-PCR at each mice using a knock-in strategy prevented development of the gene where a bar is shown. Copy numbers were not determined for KTN1 cerebellar vermis, but produced an enlarged inferior colliculus and SEC10L1 in primary tumor samples Med 6, 7, and 8. Also PELI2 was not (21). OTX2 also physically interacts with the translational determined in samples Med 7 and Med 8. The relative copy number of the predicted open reading frame C14orf101 was only determined in the three initiation factor 4E, providing another mechanism by which depicted medulloblastoma cell line samples. D425 Med amplification is OTX2 could modulate tumor cell behavior (22). The role of limited to OTX2 and flanking gene Sec10L1 is excluded in D487 Med.

Cancer Res 2005; 65: (3). February 1, 2005 706 www.aacrjournals.org

OTX2 in the development of the brain is consistent with its Grant support: Ludwig Trust, the Cancer Genome Anatomy Project (NIH 23X- S073), and NIH grants (U01 CA88128 and K08 NS43279). proposed role as a medulloblastoma oncogene. Further studies The costs of publication of this article were defrayed in part by the payment of page are required to determine if inhibition of OTX2 function can charges. This article must therefore be hereby marked advertisement in accordance repress medulloblastoma growth. with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Victor E. Velculescu and Jordan M. Cummins (Johns Hopkins University) for digital karyotyping technical support and Daniela Gerhard for CGAP administra- Acknowledgments tive support. C.G. Eberhart is recipient of a Burroughs Wellcome Fund Career Award. G.J. Riggins is the recipient of the Irving J. Sherman, M.D. Research Professorship in Received 9/22/2004; revised 11/23/2004; accepted 12/3/2004. Neurosurgery.

metastatic colorectal cancer patients. Proc Natl Acad somal abnormalities in human medulloblastoma cell References Sci U S A 2004;101:3089–94. lines and xenograft. Neuro-oncol 2002;4:75–85. 1. Kleihues P, Burger PC, Plate KH, et al. Pathology & 9. Kurokawa D, Kiyonari H, Nakayama R, et al. 16. Boon K, Osorio EC, Greenhut SF, et al. An anatomy of Genetics: Tumours of the Nervous System. Lyon: IARC; Regulation of Otx2 expression and its functions in normal and malignant gene expression. Proc Natl Acad 2000. mouse forebrain and midbrain. Development 2004;131: Sci 2002;99:11287–92. 2. Raffel C, Jenkins RB, Frederick L, et al. Sporadic 3319–31. 17. Michiels EM, Oussoren E, Van Groenigen M, et al. medulloblastomas contain PTCH mutations. Cancer 10. Friedman H, Burger P, Bigner S, et al. Establishment Genes differentially expressed in medulloblastoma and Res 1997;57:842–5. and characterization of the human medulloblastoma fetal brain. Physiol Genomics 1999;1:83–91. 3. Zurawel RH, Chiappa SA, Allen C, et al. Sporadic cell line and transplantable xenograft D283. Med. 18. Yokota N, Mainprize T, Taylor M, et al. Identification medulloblastomas contain oncogenic h-catenin muta- J Neuropathol Exp Neurol 1985;44:592–605. of differentially expressed and developmentally regu- tions. Cancer Res 1998;58:896–9. 11. Pietsch T, Scharmann T, Fonatsch C, et al. Charac- lated genes in medulloblastoma using suppression 4. Bigner SH, Friedman HS, Vogelstein B, et al. Ampli- terization of five new cell lines derived from human subtraction hybridization. Oncogene 2004;23:3444–53. fication of the c-myc gene in human medulloblastoma primitive neuroectodermal tumors of the central 19. Simeone A, Puelles E, Acampora D. The Otx family. cell lines and xenografts. Cancer Res 1990;50:2347–50. nervous system. Cancer Res 1994;54:3278–87. Curr Opin Genet Dev 2002;12:409–15. 5. Tong C, Hui A, Yin X, et al. Detection of oncogene 12. Boon K, Edwards J, Siu I, et al. Comparison of 20. Puelles E, Acampora D, Lacroix E, et al. Otx dose- amplifications in medulloblastomas by comparative medulloblastoma and normal neural transcriptomes dependent integrated control of antero-posterior and genomic hybridization and array-based comparative identifies a restricted set of activated genes. Oncogene dorso-ventral patterning of midbrain. Nat Neurosci 2003; genomic hybridization. J Neurosurg Spine 2004;100: 2003;22:7687–94. 6:453–60. 187–93. 13. Siu IM, Lal A, Blankenship JR, et al. c-Myc promoter 21. Broccoli V, Boncinelli E, Wurst W. The caudal limit of 6. Saha S, Sparks A, Rago C, et al. Using the transcriptome activation in medulloblastoma. Cancer Res 2003;63: Otx2 expression positions the isthmic organizer. Nature to annotate the genome. Nat Biotechnol 2002;20:508–12. 4773–6. 1999;401:164–8. 7. Wang T-L, Maierhofer C, Speicher MR, et al. Digital 14. Bigner SH, McLendon RE, Fuchs M, et al. Chromo- 22. Nedelec S, Foucher I, Brunet I, et al. Emx2 karyotyping. Proc Natl Acad Sci 2002;99:16156–61. somal characteristics of childhood brain tumors. homeodomain transcription factor interacts with eu- 8. Wang TL, Diaz LA Jr, Romans K, et al. Digital Cancer Genet Cytogenet 1997;97:125–34. karyotic translation initiation factor 4E (eIF4E) in the karyotyping identifies thymidylate synthase amplifica- 15. Aldosari N, Wiltshire RN, Dutra A, et al. Comprehen- axons of olfactory sensory neurons. Proc Natl Acad Sci tion as a mechanism of resistance to 5-fluorouracil in sive molecular cytogenetic investigation of chromo- 2004;101:10815–20.

www.aacrjournals.org 707 Cancer Res 2005; 65: (3). February 1, 2005

Kathy Boon, Charles G. Eberhart and Gregory J. Riggins

Cancer Res 2005;65:703-707.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/65/3/703

Cited articles This article cites 21 articles, 10 of which you can access for free at: http://cancerres.aacrjournals.org/content/65/3/703.full#ref-list-1

Citing articles This article has been cited by 14 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/65/3/703.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/65/3/703. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.