Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
GENES, CHROMOSOMES & CANCER 46:118–129 (2007) RESEARCH ARTICLES Profiling Genomic Copy Number Changes in Retinoblastoma Beyond Loss of RB1 Ella Bowles,1–3{ Timothy W. Corson,1,4{ Jane Bayani,1 Jeremy A. Squire,1,3,5 Nathalie Wong,6 Paul B.-S. Lai,7 and Brenda L. Gallie1,3,4,8* 1 Division of Applied Molecular Oncology, Ontario Cancer Institute/Princess Margaret Hospital, University Health Network,Toronto,ON,Canada 2 Department of Zoology,University of British Columbia,Vancouver,BC,Canada 3 Department of Medical Biophysics,University of Toronto,Toronto,ON,Canada 4 Department of Molecular and Medical Genetics,University of Toronto,Toronto,ON,Canada 5 Department of Laboratory Medicine and Pathobiology,University of Toronto,Toronto,ON,Canada 6 Department of Anatomical and Cellular Pathology,Chinese University of Hong Kong, Shatin,N.T., SAR Hong Kong,China 7 Department of Surgery,Chinese University of Hong Kong, Shatin,N.T., SAR Hong Kong,China 8 Department of Ophthalmology,University of Toronto,Toronto,ON,Canada Loss of both RB1 alleles is rate limiting for development of retinoblastoma (RB), but genomic copy number gain or loss may impact oncogene(s) and tumor suppressor genes, facilitating tumor progression. We used quantitative multiplex polymerase chain reaction to profile "hot spot" genomic copy number changes for gain at 1q32.1, 6p22, and MYCN, and loss at 16q22 in 87 primary RB and 7 cell lines. Loss at 16q22 (48%) negatively associated with MYCN gain (18%) (Fisher’s exact P 5 0.031), gain at 1q32.1 (62%) positively associated with 6p "hot spot" gain (43%) (P 5 0.033), and there was a trend for positive association between 1q and MYCN gain (P 5 0.095). Cell lines had a higher frequency of MYCN amplification than primary tumors (29% versus 3%; P5 0.043). Novel high-level amplification of 1q32.1 in one primary tumor, confirmed by fluorescence in situ hybridization, strongly supports the presence of oncogene(s) in this region, possibly the mitotic kinesin, KIF14. Gene-specific quantitative multiplex polymerase chain reaction of candidate oncogenes at 1q32.1 (KIF14), 6p22 (E2F3 and DEK), and tumor suppressor genes at 16q22 (CDH11) and 17q21 (NGFR) showed the most common gene gains in RB to be KIF14 in cell lines (80%) and E2F3 in primary tumors (70%). The patterns of gain/loss were qualitatively different in 25 RB compared with 12 primary hepatocellular carcinoma and 12 breast cancer cell lines. Gene specific analysis of one bone marrow metastasis of RB, prechemotherapy and postchemotherapy, showed the typical genomic changes of RB pretreatment, which normalized after chemoC 2006 Wiley-Liss, Inc. therapy. V INTRODUCTION Retinoblastoma (RB) is a rare childhood eye cancer initiated by biallelic loss (mutational events M1 and M2) of the RB susceptibility gene, RB1, as predicted by Knudson’s two-hit model (Knudson, 1971). However, further genomic changes, termed M3-Mn, present in most RB, may support tumor growth by activating oncogenes and inactivating tumor suppressor genes (Gallie et al., 1999a). The most common chromosomal changes in RB tumors, detected by comparative genomic hybridization (CGH), are gains around 1q31-1q32, 2p24, and 6p22, and losses at 16q22 (Mairal et al., 2000; Chen et al., 2001; Herzog et al., 2001; Lillington et al., 2003; van der Wal et al., 2003; Zielinski et al., 2005). Previously, we used quantitative multiplex polymerase chain reaction (QM-PCR) and loss of heterozygosity analyses to narrow further three of these regions to ‘‘hot spot’’ sequence tagged sites (STS) at 1q32.1 (Corson et al., 2005), 6p22 (Chen et al., 2002), and 16q22 (Marchong et al., 2004). The 2p24 C V 2006 Wiley-Liss, Inc. region is occasionally amplified in RB, and includes the MYCN oncogene, the assumed target of the amplification (Lee et al., 1984; Squire et al., 1986; Godbout et al., 1998; Lillington et al., 2002). Subsequent expression analyses have led to the identification of candidate cancer genes at or near each genomic ‘‘hot spot.’’ KIF14 is a mitotic kinesin and candidate oncogene encoded near the 1q32.1 ‘‘hot spot’’ (Corson et al., 2005). It is highly Supported by: National Cancer Institute of Canada, the Canadian Retinoblastoma Society, funds from the Blind Ball, Canadian Genetic Diseases Network, University of Toronto Department of Medical Biophysics, University of Toronto Vision Science Research Program Studentship and Canadian Institutes for Health Research Canada Graduate Scholarship. {These authors contributed equally to this work. *Correspondence to: Brenda L. Gallie, Division of Applied Molecular Oncology, Ontario Cancer Institute/Princess Margaret Hospital, University Health Network, Room 8-415, 610 University Avenue, Toronto, ON M5G 2M9, Canada. E-mail: [email protected] Received 7 July 2006; Accepted 31 August 2006 DOI 10.1002/gcc.20383 Published online 10 November 2006 in Wiley InterScience (www.interscience.wiley.com). GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION overexpressed in RB compared with normal retina, and in medulloblastoma, breast, and lung cancer samples (Corson et al., 2005; Corson and Gallie, 2006). The 6p22 ‘‘hot spot’’ overlaps the gene for another kinesin, RBKIN (KIF13A), initially posited as a candidate oncogene (Chen et al., 2002). However, more recent work suggests that the DNA-binding protein and leukemia oncogene DEK and/or the transcription factor E2F3, both overexpressed in RB, are more promising RB oncogene candidates at 6p22 (Grasemann et al., 2005; Orlic et al., 2006). The 16q22 ‘‘hot spot’’ overlaps the CDH11 gene (Marchong et al., 2004), which encodes the cell adhesion molecule and candidate tumor suppressor cadherin-11, expression of which is lost in RB (Marchong et al., 2004). Finally, on the basis of loss of expression, we have identified the anti-apoptotic neurotrophin receptor p75NTR (encoded by NGFR at 17q21) as another candidate tumor suppressor gene in RB (Dimaras et al., 2006); the genomic status of this gene in RB is unknown. We hypothesized that patterns of genomic change may be important in the molecular progression of RB. Using the ‘‘hot spot’’ STSs defining each minimal region of gain/loss, in addition to primers for MYCN, we developed one QM-PCR assay to evaluate rapidly genomic copy number in 87 RB tumors and 7 cell lines. We show associations between specific genomic gains/losses, MYCN amplification, and a novel 1q amplification that we confirmed by interphase fluorescence in situ hybridization (FISH). In a separate assay, the qualitative patterns of gains/losses in RB for the candidate oncogenes and tumor suppressor genes were similar to those in ‘‘hot spot’’ profiles of RB, but distinct from patterns in primary hepatocellular carcinoma and breast cancer cell lines. The genomic signature characteristic of RB was detectable in bone marrow containing metastatic RB, and cleared with chemotherapy. MATERIALS AND METHODS DNA Samples DNA was extracted from 87 primary RB, 7 RB cell lines for which primary tumor was no longer available (RB247C, RB267, RB355, RB383, RB409, WERI-Rb1, Y79) (Gallie et al., 1999b), and lymphocytes of seven unaffected relatives, as described (Richter et al., 2003). All samples were received for RB1 gene diagnostic testing and were used for research when these studies were complete (Richter et al., 2003). The Research Ethics Boards of the Wellesley Hospital, the Hospital for 119 Sick Children, the University Health Network, and the University of Toronto approved research use of blood and tumor material with parental consent. Clinical details of extent of intraocular RB were not available for the majority of these samples. Liver tumors were collected from 12 patients who underwent curative surgery for hepatocellular carcinoma at the Prince of Wales Hospital, Hong Kong, China. Informed consent was obtained from all patients. DNA was extracted from these hepatocellular carcinoma specimens using the QIAamp1 DNA kit (Qiagen, Hong Kong, China) according to the manufacturer’s instructions. DNA was obtained from 12 breast cancer cell lines (BT-474, BT-483, Hs578T, MCF7, MDA-MB-231, MDAMB-415, MDA-MB-468, UACC-812, UACC-893, T-47D, ZR-75-1, and ZR-75-30) as described (Corson et al., 2005). Quantitative Multiplex Polymerase Chain Reaction We combined three previously identified ‘‘hot spot’’ STS markers of genomic gain/loss (SHGC154194 at 1q32.1, SHGC-103950 at 6p22, and WI5835 at 16q22) into a single multiplex PCR along with custom-designed primers for MYCN (Table 1; STS primer sequences are available through UniSTS, http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db ¼ unists). In the gene-specific QM-PCR we combined primers for KIF14, DEK, E2F3, CDH11, and NGFR (Table 1). As in our previous studies (Chen et al., 2002; Marchong et al., 2004; Corson et al., 2005), the internal control for two copies was STS WI-7221 (335 bp) at 10q21, a region rarely gained or lost in RB (Chen et al., 2001), breast, or liver cancer (Baudis and Cleary, 2001). The STSs used had no known polymorphisms and did not lie in duplicated regions (Bailey et al., 2001). The forward primers were 50 -Cy5-labeled (Integrated DNA Technologies, Coralville, IA). The total reaction volume of 10 ll consisted of 13 Multiplex PCR Master Mix (includes Taq polymerase, buffer, 3 mM MgCl2, and dNTPs; Qiagen, Burlington, ON, Canada) and 200 ng of genomic DNA for the ‘‘hot spot’’ assay and 100 ng in the gene-specific assay. Primers in the ‘‘hot spot’’ assay consisted of 2 pmol for SHGC-103950, WI-5835 and MYCN, and 1 pmol for SHGC-154194 and WI-7221; and in the gene-specific experiments, 4 pmol for WI-7221, and 2 pmol for all others. PCR for 20 cycles was confirmed in pilot experiments to maintain linearity of generation of the PCR products. PCR conditions, PAGE, and the QM-PCR assay spanning a 26 Mbp region of 1q were performed as described (Corson et al., Genes, Chromosomes & Cancer DOI 10.1002/gcc 120 BOWLES ET AL. TABLE 1. Gene-Specific QM-PCR Primers Used in this Study, Written 50 to 30 Gene symbola KIF14 MYCN DEK E2F3 CDH11 NGFR Band Primer location PCR product (bp) Forward PCR primer Reverse PCR primer 1q32.1 2p24.1 6p22 6p22 16q22 17q21 Intron 11-Exon 12 Intron 2 Exon 6 Intron 2-Exon 3 Intron 1 Intron 2 212 252 235 260 293 315 TGGAGGATTATGCTTACTGTGGA CCAAAGAGTGGCATTGCCTTGT GAGCATCTGTGAGGTTCTTG AACGACAGCCTTGCTTCAGA TACTGCCCATCAAATCCACCCT AGACCACCACTGTAAGGTCA TGCCATGTCTCTCTCCTGTTGA TGCCAAGAAGACAGCTGTTGGA CGTCAACAGAAGGTATAGTC AGCACTTCTGCTGCCTTGTT TGAGGCTGGATACCTACACTTG TCAGGGCAGCCAGGATTTGAT Full gene names: KIF14, kinesin family member 14; MYCN, v-myc myelocytomatosis viral related oncogene, neuroblastoma derived; DEK, DEK oncogene (DNA binding); E2F3, E2F transcription factor 3; CDH11, cadherin 11, type 2, OB-cadherin; NGFR, nerve growth factor receptor (TNFR superfamily, member 16). a 2005). Positive controls were RB of known DNA copy number for each of the STSs used in previous studies (Chen et al., 2002; Marchong et al., 2004; Corson et al., 2005). When amplification, rather than low level gain, was detected, QM-PCR assays were repeated without primers for the amplicon, to enable accurate quantitation of the other markers. Copy number for each ‘‘hot spot’’ STS or gene was calculated using the OpenGene system and Gene Objects 3.1 software (Visible Genetics) as described (Corson et al., 2005). The minimum value for gain was the mean copy number +3 SD of seven normal samples in the ‘‘hot spot’’ assay and six normal samples in the gene-specific assay. Amplification was defined as copy number 10. For loss, the cutoff was 1.2, the mean of the overlap between samples of known one copy and known two copies of the RB1 gene (Richter et al., 2003). The coefficient of variance for each mean copy number was <20%. Paraffin FISH A bacterial artificial chromosome (BAC) clone containing the KIF14 and DDX59 genes, RP1192G12, and an adjacent intergenic BAC clone, RP11-121P12 (Centre for Applied Genomics, Toronto, ON, Canada) were directly labeled with Rhodamine-dUTP (Roche, Laval, QC, Canada) by PCR as described (Stoecklein et al., 2002). FISH to normal human lymphocytes confirmed the genomic location of the BACs. Heparinized whole blood was cultured with phytohemagglutinin for 72 hr in a CO2 incubator in RPMI 1640 containing 20% FCS, 1% (v/v) L-glutamine, and 1% (v/v) penicillin/streptomycin (Invitrogen, Burlington, ON, Canada). The cells were harvested for cytogenetic preparations as previously described (Bayani and Squire, 2004). Metaphase slides were prepared and aged for at least 1 week before use. Approximately Genes, Chromosomes & Cancer DOI 10.1002/gcc 300 ng of labeled probes were precipitated in excess human Cot-1 (Invitrogen) and sonicated salmon sperm DNA (Roche) and resuspended in a 50% formamide, 10% dextran sulfate, and 23 SSC hybridization buffer (DAKO, Mississauga, ON, Canada). The probes were heat denatured and hybridized to pepsin-treated and denatured normal human metaphase slides as described previously (Bayani and Squire, 2004). Following a rapid-wash protocol of 0.43 SSC and 0.3% NP-40 at 728C for 3 min, then 23 SSC and 0.1% NP-40 at room temperature for 5 min, slides were mounted in DAPI/ Antifade (Vector Laboratories, Burlington, ON, Canada) and visualized with a Zeiss Axioskop fluorescent microscope. Unstained 6 lm paraffin embedded sections of RB2306 were dewaxed in xylene and dehydrated in 100% ethanol. The sections were incubated in 23 SSC at 758C for 20 min, followed by proteinase K (0.25 mg/ml in 23 SSC) at 458C for 30 min. The slides were passed through a dehydration series and air-dried. Rhodamine-labeled RP11-92G12 and RP11-121P12 and FITC-labeled centromere 1 probes (Roche) were added to the processed paraffin sections and codenatured using the HybRite System (Vysis, Downers Grove, IL) at 808C for 6 min and hybridized overnight at 378C. The slides were rapidly washed as described above. To confirm MYCN amplification, digoxigenin-labeled probes for MYCN (Oncor, Gaithersburg, MA) and FITC-labeled centromere 2 probes (Roche) were hybridized to paraffin sections as described above. Digoxigenin-labeled probes were detected with an anti-digoxigenin antibody conjugated to FITC (Roche). Slides were visualized using a motorizedstaged Zeiss Axioskop fluorescent microscope and imaged using Z-stacking (AxioVision software). Copy number changes were assessed by the number of signals found within a nucleus above the cell ploidy as determined by the number of centromeric signals. GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION Statistical Analysis Two-tailed Fisher’s exact test was used to analyze associations between regions of genomic change, unilateral/bilateral diagnosis, cell line versus primary tumor, and sex, for the ‘‘hot spot’’ assay. Associations between regions of genomic change and RB1 mutation type were analyzed by the v2 test. P values of 0.05 were considered significant. Only 93 of the 94 RB studied were included in the statistical analysis. RB2306 was excluded since we could not obtain accurate copy number estimates for all markers, possibly owing to high-level amplification of both 1q (KIF14) and MYCN. RESULTS "Hot Spot" QM-PCR Identifies Gain/Loss Associations To evaluate patterns of genomic gain and loss in RB, we examined copy number by QM-PCR at three gain/loss ‘‘hot spots’’ and MYCN simultaneously, in 87 primary tumors and 7 cell lines. This patient cohort included 51 females and 43 males; 24 individuals were bilaterally and 70 were unilaterally affected. The genomic changes in 20 of these samples were previously analyzed by CGH (Chen et al., 2001). Frequencies of genomic gain and loss were consistent with those shown previously by karyotype (Squire et al., 1985; Potluri et al., 1986), CGH (Mairal et al., 2000; Chen et al., 2001; Herzog et al., 2001; Lillington et al., 2003; van der Wal et al., 2003; Zielinski et al., 2005), and QM-PCR (Chen et al., 2002; Marchong et al., 2004; Corson et al., 2005). Overall, 1q32.1 gain was most common at 62%, followed by 16q22 loss (48%), 6p22 gain (43%), and MYCN gain or amplification (18%) (Fig. 1). The most common gains were low-level: marker copy numbers for 1q32.1 averaged 2.60 6 0.66 (range: 0.86–4.84); for MYCN, 2.01 6 0.91 (range: 0.96– 5.70); and for 6p22, 2.51 6 1.02 (range: 0.77–5.60). Loss at 16q22 was always single-copy loss, averaging 1.32 6 0.70 (range: 0.19–3.76) (Fig. 1A). No tumors were nullizygous for the 16q22 STS. In 15% of samples, no genomic copy number changes at any of the examined loci were detected. RB cell lines showed a higher frequency of genomic changes (gain in 71% at 1q32.1 and 6p22, 57% at 16q22 and MYCN) compared with primary tumors (Table 2), although the differences were only significant for high level MYCN amplification (Fisher’s exact P ¼ 0.043), which was present in 29% of cell lines, compared with only 3% of primary RB. In 121 comparison with the 16% of primary tumors that showed no copy number change at any of the loci examined, beyond biallelic inactivation of RB1, all cell lines showed at least two genomic changes (Fig. 1B, Table 2). Considering cell lines and primary tumors together, we observed significant associations between genomic changes (Fig. 1B). Gain at 1q and 6p positively associated (P ¼ 0.033) and 1q gain showed a trend toward association with gain at MYCN (P ¼ 0.095). MYCN gain and 16q22 loss were negatively associated (P ¼ 0.031). Genomic changes did not associate with sex, type of RB1 mutations in the tumors, or unilateral/bilateral diagnosis (data not shown). Large copy number amplification was suggested in QM-PCR by very large signals for some PCR products, and failure of other low-copy number products to amplify as expected, including the internal control. High-level amplification of MYCN was observed in 5/94 RB samples, with a mean copy number of 23.02 6 17.70 (range: 11–53) (Fig. 1A). Since QM-PCR does not provide accurate numerical estimates of high-level amplification, these copy numbers are approximations. These five samples included two RB cell lines (Y79 and RB355) with amplification at the MYCN locus previously identified by Southern blot (Squire et al., 1986), and three primary tumors. Amplification of MYCN never occurred with 16q22 loss (Fig. 1B). In one RB with MYCN amplification, RB2306, a novel amplification of SHGC-154194 at 1q32.1 was also detected, with a mean copy number of 13.55 (Figs. 1A and 4A; see below). Gene-Specific QM-PCR for Candidate Oncogenes and Tumor Suppressors To ascertain whether gains of the ‘‘hot spot’’ STSs reflect gains of candidate oncogenes and tumor suppressor genes identified after the ‘‘hot spot’’ study was initiated, we assembled a genespecific QM-PCR for KIF14 at 1q32.1, DEK and E2F3 at 6p22, CDH11 at 16q22, NGFR at 17q21 (Table 1), and the 10q21 internal control. We compared genomic copy number profiles of 25 RB (20 primary RB and five cell lines) with 12 primary hepatocellular carcinomas and 12 breast cancer cell lines. Gene-specific gain and loss fractions in 20 primary RB were similar to those determined by ‘‘hot spot’’ STSs. E2F3 and KIF14 were most commonly altered (Fig. 2A, Table 2). Percentage of primary RB showing gain was 70% for E2F3, 52% for KIF14, 40% for DEK; and showing loss was 45% for CDH11. No tumors showed gain or loss of Genes, Chromosomes & Cancer DOI 10.1002/gcc 122 BOWLES ET AL. Figure 1. Genomic copy number of "hot spots" in 94 RB by QM-PCR. (A) Gain is scored for tumors showing copy numbers 3 SD (red line) above the average (black line) of normal samples (black-filled circles). Loss is copy number 1.2 (blue line). Low-level gain (pink-filled circles) and single-copy loss (blue-filled circles) are the most common genomic changes seen in RB. Amplification (10 copies) (green-filled circles) is observed for MYCN (5/94 RB) and 1q32.1 (1/94 RB). The copy numbers for these amplicons are estimates, since QM-PCR is imprecise for quantitation of large amplicons. Cell lines are indicated by orange bordered circles. (B) Associations between genomic changes. The most common gain in RB is 1q32.1, followed by 16q22 loss, 6p22 gain, and MYCN gain. Gain at 1q32.1 occurs commonly with 6p22 gain (Fisher’s exact P 5 0.033) and MYCN gain (P 5 0.095). MYCN gain and 16q22 loss rarely occur together (P 5 0.031). Each row represents an individual tumor or cell line, while each column is the indicated marker. Pink indicates gain; green, amplification; blue, loss; white, no change; and gray, copy number not determined. Cell lines are separated and marked in orange. NGFR. Again, the 5 RB cell lines included in this assay (Fig. 2B; Table 2) showed a greater frequency of genomic gain than the primary tumors: 80% for KIF14, 60% for E2F3 or DEK. Every cell line showed some genomic gain, but no cell line showed CDH11 genomic loss. The breast cancer and liver samples showed different patterns of genomic changes for these genes than RB (Fig. 2). The percentage of breast cancer cell lines showing gain was 50% for KIF14, 42% for E2F3, 33% for NGFR, 17% for CDH11, and 8% for DEK; and showing loss was 25% for NGFR and 8% for CDH11 (Fig. 2C). Genes, Chromosomes & Cancer DOI 10.1002/gcc The percentage of hepatocellular carcinomas showing gain was 58% for KIF14, 33% for E2F3, 17% for DEK, 17% for CDH11, and 8% for NGFR; and E2F3, CDH11, and NGFR showed loss in one sample each (Fig. 2D). While the cohort of each cancer type was too small for statistical evaluation of associated changes, qualitative patterns of changes appeared different for the three types of tumor. Gain of KIF14 and E2F3 Detected in Retinoblastoma Bone Marrow Metastasis We examined bone marrow containing metastatic RB prior to and after systemic chemotherapy. 123 GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION TABLE 2. Gain/Loss Breakdown in "Hot Spot" and Gene-Specific QM-PCR Analyses of Retinoblastoma Primary Tumors and Cell Lines "Hot spot" QM-PCR Primary Cell lines Gene-specific QM-PCR Primary 53 33 62 5 2 71 11 10 52 35 51 41 5 2 71 Amp No amp % 41 45 48 4 3 57 a 3 2 60 9 11 45 8 12 40 13 73 15 3 2 60 8 12 40 3 2 60 0 5 0 4 3 57 14 73 16 4 3 57 MYCN amplification MYCN amplification overalla 2 84 2 3 84 3 73 14 84 E2F3 & DEK MYCN overalla 2 5 29 7 0 100 2 5 29 P ¼ 0.043b Any gain/loss P ¼ 0.024b Any gain/loss Gain/loss No gain/loss % Cell lines CDH11 MYCN Gain No gain % Primary DEK 14 6 70 16q Loss No loss % Cell lines 4 1 80 E3F3 6p Gain No gain % Primary KIF14 1q Gain No gain % Cell lines a 16 4 80 5 0 100 Including RB2306 data for KIF14 and MYCN, determined by FISH. Statistically significant (Fisher’s exact test) differences in proportion of gain/loss (primary tumors versus cell lines) are indicated. b Prior to treatment, histopathology and RB1 mutant allele-specific PCR (data not shown) demonstrated RB in bone marrow, and QM-PCR showed gain of both KIF14 (mean copy number 2.56) and E2F3 (mean copy number 2.85) (Fig. 3). After treatment, bone marrow copy numbers of these genes were normal (mean copy numbers 1.75 and 1.79, respectively). A Novel 1q Amplicon Detected by QM-PCR In the initial ‘‘hot spot’’ assay of primary tumor RB2306, amplification of 1q32.1 and MYCN ablated peaks for 16q22 and 6p22, and reduced the control 10q21 peak (Fig. 4A). Therefore, the observed 1q32.1 and MYCN amplifications were not artifacts of a deletion of the internal control STS at 10q21, which would lead to artificially high copy number estimates for all other STSs. Given the novelty of this finding, we examined the 1q32.1 amplicon further with QM-PCRs spanning a 26 Mbp region of chromosome arm 1q (Corson et al., 2005) (Figs. 4B–4D). The internal control peak (10q21) was vanishingly small, while amplification of 7–75 copies was seen for 13 of the 15 STSs, spaced across the 26 Mbp region. STSs located 188.9 and 192.4 Mbp from 1pter were the only STSs that did not show amplification; perhaps the larger STSs are more sensitive than smaller PCR products to depletion of reagents in a multiplex containing amplified STSs. To confirm amplification of 1q32.1 in RB2306, we performed FISH with two BAC clones spanning approximately 232 kbp of 1q31-32 containing KIF14 and one other gene, DDX59. The BAC Genes, Chromosomes & Cancer DOI 10.1002/gcc 124 BOWLES ET AL. Figure 2. Patterns of copy number change of candidate oncogenes and tumor suppressors by QM-PCR in (A) RB primary tumors, (B) RB cell lines, (C) breast cancer cell lines, and (D) primary hepatocellular carcinoma. Each tumor type shows a qualitatively distinct ‘‘signature’’ of genomic changes. Each row represents an individual tumor or cell line, while each column is the indicated gene. Color coding as in Fig. 1B. Figure 3. Clearance of metastatic RB shown by QM-PCR. The genomic changes characteristic of RB are recognizable in bone marrow which contains 10% metastatic RB, as quantitated by PCR for the specific mutant RB1 alleles of the primary RB (not shown). By QMPCR, gain of KIF14 (1q32.1) and E2F3 (6p22) is apparent. After treatment, when no signal was detectable for the RB1 mutant alleles of the RB in bone marrow and stem cells harvested for autologous stem cell rescue, genomic gain of KIF14 and E2F3 was no longer evident. clones mapped correctly to 1q31-32 in human lymphocytes (data not shown). Hybridization of the BAC probes to formalin-fixed, paraffin embedded Genes, Chromosomes & Cancer DOI 10.1002/gcc tumor revealed amplification greater than ploidy (Fig. 5). A region of normal retinal cells served as a technical control for overall hybridization effi- GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION 125 Figure 4. Tumor RB2306 shows 1q and MYCN amplification. (A) QM-PCR "hot spot" assay shows vanishingly small internal control (10q21), 6p22, and 16q22 peaks in comparison with the normal sample, while the 1q32.1 and MYCN peaks in RB2306 are large. (B–D) QM-PCR output of 15 STSs in three multiplexes spanning a 26 Mbp region of 1q (Corson et al., 2005). Amplification was identified by the vanishingly small internal control peak (10q21) in RB2306 compared with the normal sample. Peaks are indicated by the chromosome 1 location of the STS (in Mbp from 1pter, based on NCBI human genome build 36.1). Calculated copy numbers for each STS in RB2306 are given; nonamplified STSs are italicized. ciency and scoring (Fig. 5A). In all tumor cells analyzed, at least 5–7 BAC signals were evident, with centromere 1 showing 2–3 signals per cell (Fig. 5B). There was little cell-to-cell variation in copy number of BAC and centromere signals throughout the tumor specimen. To confirm MYCN amplification in RB2306, we performed FISH with a MYCN-specific probe, which showed a mean of 5–7 signals per cell (Fig. 5D) across the tumor specimen. The FISH signals in tumor cells (Figs. 5B and 5D) were of variable intensity compared with the more consistent intensity in adjacent normal retina (Figs. 5A and 5C), suggesting that there might be multiple, variably sized, contiguous amplicons. KIF14 and MYCN probes were cohybridized and localized independently, indicating that these genes do not reside within the same amplicon in RB2306 (Fig. 5D). DISCUSSION A Molecular Model of Retinoblastoma Progression Nearly all RB tumors analyzed molecularly show additional genomic changes (Squire et al., 1985; Mairal et al., 2000; Chen et al., 2001; Herzog et al., 2001; Lillington et al., 2003; van der Wal et al., 2003; Zielinski et al., 2005) besides the initiating M1 and M2 events, loss of both RB1 alleles. Our present work represents the largest series to date of RB characterized for genomic changes beyond RB1. We examined RB for gain and loss of genomic ‘‘hot spots’’ suggested by previous studies (Chen et al., 2002; Marchong et al., 2004; Corson et al., 2005), which may reflect the further mutational events necessary for RB progression (M3 to Mn). On the basis of the frequency of the genomic changes and patterns of association between Genes, Chromosomes & Cancer DOI 10.1002/gcc 126 BOWLES ET AL. Figure 5. FISH analysis on paraffin sections of RB2306. (A) Normal retinal cells hybridized with FITC-labeled (light blue) centromere 1 and rhodamine-labeled (red) KIF14 show normal diploid hybridization patterns. Truncation of the cells due to sectioning results in loss of signals in some cells. (B) Adjacent RB cells show additional copies of the KIF14-containing BAC probe (red) over ploidy established by the centromere 1 probe (light blue). (C) Normal retinal cells hybridized with FITC-detected (light blue) MYCN and rhodamine-labeled (red) KIF14 show diploid hybridization. (D) Adjacent RB cells show multiple signals for MYCN (light blue) and KIF14 (red). MYCN and KIF14 do not co-localize within the cells, suggesting independent amplifications. Inset: A single tumor cell shows at least six signals for KIF14 (red) and six signals for MYCN (light blue). Original magnification 363. Figure 6. Potential molecular model for RB progression, based on frequencies of gain/loss and associations of genomic ‘‘hot spots.’’ Loss of each RB1 allele is shown as the M1 and M2 events, leading to the benign retinoma. 1q32.1 and 6p22 gain are the most common genomic changes, shown as M3 and M4. 16q22 loss and MYCN gain negatively associate, so are shown as alternate M5 events. Note that this model only incorporates ‘‘hot spot’’ copy number changes, and not other molecular events in RB progression. genomic regions in individual tumors (Fig. 1B), we propose that the most common genomic changes in RB, which may represent the M3 and M4 events, are chromosome 1q32.1 and 6p22 gain (Figs. 1 and 6). These gains commonly occur together, further supporting their shared role as early genomic events in progression. 16q22 loss showed a negaGenes, Chromosomes & Cancer DOI 10.1002/gcc tive association with MYCN gain, suggesting that these might be alternate M5 events (Fig. 6). Interestingly, in this cohort, we did not see the positive association between 1q gain and 16q loss noted in two previous studies (Oliveros and Yunis 1995; Herzog et al., 2001), possibly since these studies considered whole chromosome arms, rather GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION than the ‘‘hot spots’’ studied here. This association is consistent with our model (Fig. 6), however. The lack of association between genomic changes and unilateral/bilateral diagnosis, RB1 mutation type, or gender (data not shown) suggests that once RB is initiated, progression follows common pathways, as previously observed (Chen et al., 2002; Marchong et al., 2004; Corson et al., 2005). We observed a novel amplification at 1q32.1 in one RB, RB2306 (Figs. 1, 4, and 5). There were no unusual clinical features of this bilateral RB in a female, although the tumor also showed MYCN amplification. FISH for KIF14 and MYCN suggested independent amplicons (Fig. 5D). Highlevel amplification on 1q has been observed by CGH in RB (Mairal et al., 2000; Lillington et al., 2003), but not around 1q32.1, and has not previously been confirmed by FISH. We confirmed that the novel 1q32.1 amplicon spans at least 26 Mbp (Fig. 4). This strongly suggests the presence of an oncogene(s) in this region; full definition of the amplified genes is important. A possible candidate oncogene in this amplicon is KIF14, which is highly overexpressed in RB and other tumors (Corson et al., 2005), and predicts outcome in breast cancer (Corson and Gallie, 2006). The 1q32.1 BAC probes used in FISH span KIF14 and one other gene, DDX59, which is not overexpressed in RB (Corson et al., 2005). The 1q32.1 oncogene could also be one of the 1q32 candidate RB oncogenes proposed by Gratias et al. (2005). Interphase FISH suggested 5–7 independent KIF14 and MYCN signals per cell for each probe, without evidence of colocalized FISH signals. QM-PCR suggested overall higher copy numbers, suggesting that there might be more than one gene copy per FISH signal. After we initiated the ‘‘hot spot’’ study, we identified candidate oncogenes and tumor suppressor genes within the minimal regions of gain/loss in RB that may be driving the genomic change. Gain/loss fractions of the putative target oncogenes or tumor suppressor genes in each genomic region (Fig. 2A) were similar to the results of the ‘‘hot spot’’ QMPCR (Fig. 1B, Table 2). Interestingly, E2F3 showed a much higher proportion of gain (70% of primary tumors) than the 6p22 ‘‘hot spot’’ (41%), supporting recent findings that the oncogene at 6p22 may not lie within the previously determined minimal region of gain (Grasemann et al., 2005; Orlic et al., 2006). E2F3 is *2 Mbp removed from the ‘‘hot spot’’ marker used in the first QM-PCR. Genomic gain/loss is not the only mechanism dysregulating gene expression in RB. Although 127 p75NTR expression is lost (Dimaras et al., 2006), we saw no gross NGFR genomic copy number changes in RB (Fig. 2A). Loss of p75NTR expression may occur by mutations that do not affect copy number, or via epigenetic or transcriptional events. Aberrant methylation of genes (Choy et al., 2002, 2004, 2005) and other unidentified genomic events may also contribute to RB progression. Our model of the molecular progression of RB (Fig. 6) only incorporates data on the copy number changes at specific ‘‘hot spots’’ presented in this paper. Pattern of Genomic Changes Reflects Tumor Types To demonstrate the utility of the QM-PCR approach beyond RB, and to investigate in other cancers the genomic status of the candidate cancer genes identified in RB, we examined breast cancer cell lines and primary hepatocellular carcinomas (Fig. 2), two cancers also characterized by 1q gain, 6p gain, and 16q loss. KIF14 and E2F3 showed high frequency gain in all three tumor types, suggesting that these might be ‘‘common’’ cancer genes, likely owing to their roles in proliferation. DEK genomic gain and CDH11 loss appeared to be more restricted to RB. Interestingly, DEK gene expression is increased in hepatocellular carcinoma (Kondoh et al., 1999; Lu et al., 2005), and CDH11 expression is increased, not decreased, in breast cancer (Pishvaian et al., 1999; Feltes et al., 2002). Our findings suggest that the increases in expression seen in these two cancers are largely not mediated by genomic copy number change. The NGFR gene shows both genomic gain and loss in breast cancer cell lines, consistent with reports of variable expression (Aragona et al., 2001; Sakamoto et al., 2001; Davidson et al., 2004; Popnikolov et al., 2005; Reis-Filho et al., 2006). Genomic Copy Number Change Suggests Proliferative Status Amplification of a genomic region strongly suggests the presence of an oncogene, and MYCN has been well characterized as such by amplification in neuroblastoma (Schwab 2004) and less commonly in RB (Lee et al., 1984; Squire et al., 1986; Sakai et al., 1988; Choi et al., 1993; Doz et al., 1996; Mairal et al., 2000; Herzog et al., 2001; Lillington et al., 2002; Zielinski et al., 2005). Based on data from 87 tumors, we provide a robust estimate that MYCN amplification occurs in only 3% of primary RB. The previous studies cited above have estimated a MYCN amplification frequency of 0–30%. MYCN copy number gain and amplification in neuroblastoma and rhabdomyosarcoma are strongly Genes, Chromosomes & Cancer DOI 10.1002/gcc 128 BOWLES ET AL. associated with aggressive disease and poor prognosis (Chan et al., 1997; George et al., 2005; Williamson et al., 2005). We cannot examine the relationship of genomic changes to clinical prognosis since all of the RB samples were obtained from eyes removed in treatment that was curative, as is usual for RB. We do not have clinical details of intraocular disease for most samples, since they were submitted primarily for RB1 mutation identification; the samples were only used for research after the clinical studies were complete. However, the rare RB that grow in tissue culture (Gallie et al., 1999b) do so because they have acquired autonomous features that may be a surrogate for more aggressive tumors, consistent with our observation of higher frequency of genomic change and MYCN amplification in cell lines. While MYCN gain or amplification were much more common in cell lines than primary RB, loss of the candidate tumor suppressor gene CDH11 (Marchong et al., 2004) was documented in 0% of 5 cell lines, compared with 45% in primary RB, suggesting that tumors with CDH11 loss do not have a propensity for growth in culture. Of the primary RB studied, 16% showed no genomic copy number changes in the regions studied. We speculate that these may be poorly proliferative tumors, akin to a distinct nonproliferative retinal tumor consistent with the benign precursor, retinoma (Gallie et al., 1982). We have recently shown that retinoma underlies proliferative RB in 15% of eyes removed for treatment (Dimaras et al., submitted). Although both alleles of RB1 were mutated in retinoma, as in RB, retinoma expressed low levels of KIF14 and E2F3 that were highly expressed in the adjacent RB. QM-PCR: A Rapid, Flexible, Screening Technique with Diagnostic Potential Genomic rearrangements and copy number changes are irreversible indicators of tumor status. Conversely, mRNA and protein expression are influenced by numerous environmental factors, such as hypoxia, in addition to cell of origin and differentiation. Our data suggest that a genomic profile might distinguish tumor types and subtypes that might be relevant to treatment and prognosis. Detection of metastatic RB in bone marrow by QM-PCR for common genomic copy number changes also indicates diagnostic potential. We have used allele-specific PCR in surveillance for RB1 mutant alleles in bone marrow and cerebrospinal fluid (Gallie et al., unpublished data), but we were surprised that the genomic copy number Genes, Chromosomes & Cancer DOI 10.1002/gcc changes of metastatic RB were readily detectable in untreated bone marrow. These genomic gains ceased to be evident after chemotherapy, suggesting that this QM-PCR could be developed for clinical use. We routinely use QM-PCR in the efficient clinical identification of RB1 mutations, where optimization of copy number detection is subjected to regular audit (Richter et al., 2003). Here, we used QM-PCR to generate a possible model of genomic progression in RB, to identify a novel amplicon, and to assess the genomic status of candidate oncogenes and tumor suppressors not only in RB but also in breast and liver cancers. QM-PCR is thus rapid, sensitive, and cost effective for simultaneous analysis of defined genomic regions in large numbers of samples. Cumulatively, our results provide insight into the order and diagnostic value of specific, key genomic changes in RB, and the potential general importance of these changes in cancer. ACKNOWLEDGMENTS The authors thank Professors Michael O’Keefe and Susan Kennedy of Dublin, Ireland, for providing DNA and sections for FISH analysis, Solutions by Sequence for DNA samples and QM-PCR equipment, and advice, and members of the Gallie laboratory for their input. REFERENCES Aragona M, Panetta S, Silipigni AM, Romeo DL, Pastura G, Mesiti M, Cascinu S, La Torre F. 2001. Nerve growth factor receptor immunoreactivity in breast cancer patients. Cancer Invest 19: 692–697. Bailey JA, Yavor AM, Massa HF, Trask BJ, Eichler EE. 2001. Segmental duplications: Organization and impact within the current human genome project assembly. Genome Res 11:1005–1017. Baudis M, Cleary ML. 2001. Progenetix.net: An online repository for molecular cytogenetic aberration data. Bioinformatics 17:1228– 1229. Bayani J, Squire J. 2004. Fluorescence in situ hybridization. In: Bonifacino JS, Dasso M, Lippincott-Schwartz J, Yamada KM, editors. Current Protocols in Cell Biology. Hoboken, NJ: Wiley, Unit 22.4. Chan HS, Gallie BL, DeBoer G, Haddad G, Ikegaki N, Dimitroulakos J, Yeger H, Ling V. 1997. MYCN protein expression as a predictor of neuroblastoma prognosis. Clin Cancer Res 3:1699–1706. Chen D, Gallie BL, Squire JA. 2001. Minimal regions of chromosomal imbalance in retinoblastoma detected by comparative genomic hybridization. Cancer Genet Cytogenet 129:57–63. Chen D, Pajovic S, Duckett A, Brown VD, Squire JA, Gallie BL. 2002. Genomic amplification in retinoblastoma narrowed to 0.6 megabase on chromosome 6p containing a kinesin-like gene, RBKIN. Cancer Res 62:967–971. Choi SW, Lee TW, Yang SW, Hong WS, Kim CM, Lee JO. 1993. Loss of retinoblastoma gene and amplification of N-myc gene in retinoblastoma. J Korean Med Sci 8:73–77. Choy KW, Lee TC, Cheung KF, Fan DS, Lo KW, Beaverson KL, Abramson DH, Lam DS, Yu CB, Pang CP. 2005. Clinical implications of promoter hypermethylation in RASSF1A and MGMT in retinoblastoma. Neoplasia 7:200–206. Choy KW, Pang CP, Fan DS, Lee TC, Wang JH, Abramson DH, Lo KW, To KF, Yu CB, Beaverson KL, Cheung KF, Lam DS. 2004. Microsatellite instability and MLH1 promoter methylation in human retinoblastoma. Invest Ophthalmol Vis Sci 45:3404–3409. GENOMIC EVENTS IN RETINOBLASTOMA PROGRESSION Choy KW, Pang CP, To KF, Yu CB, Ng JS, Lam DS. 2002. Impaired expression and promotor hypermethylation of O6-methylguanineDNA methyltransferase in retinoblastoma tissues. Invest Ophthalmol Vis Sci 43:1344–1349. Corson TW, Gallie BL. 2006. KIF14 mRNA expression is a predictor of grade and outcome in breast cancer. Int J Cancer 119:1088– 1094. Corson TW, Huang A, Tsao MS, Gallie BL. 2005. KIF14 is a candidate oncogene in the 1q minimal region of genomic gain in multiple cancers. Oncogene 24:4741–4753. Davidson B, Reich R, Lazarovici P, Ann Florenes V, Nielsen S, Nesland JM. 2004. Altered expression and activation of the nerve growth factor receptors TrkA and p75 provide the first evidence of tumor progression to effusion in breast carcinoma. Breast Cancer Res Treat 83:119–128. Dimaras H, Coburn B, Pajovic S, Gallie BL. 2006. Loss of p75 neurotrophin receptor expression accompanies malignant progression to human and murine retinoblastoma. Mol Carcinogen 45:333– 343. Doz F, Peter M, Schleiermacher G, Vielh P, Validire P, Putterman M, Blanquet V, Desjardins L, Dufier JL, Zucker JM, Mosseri V, Thomas G, Magdelenat H, Delattre O. 1996. N-MYC amplification, loss of heterozygosity on the short arm of chromosome 1 and DNA ploidy in retinoblastoma. Eur J Cancer 32A:645–649. Feltes CM, Kudo A, Blaschuk O, Byers SW. 2002. An alternatively spliced cadherin-11 enhances human breast cancer cell invasion. Cancer Res 62:6688–6697. Gallie BL, Campbell C, Devlin H, Duckett A, Squire JA. 1999a. Developmental basis of retinal-specific induction of cancer by RB mutation. Cancer Res 59:1731s–1735s. Gallie BL, Ellsworth RM, Abramson DH, Phillips RA. 1982. Retinoma: Spontaneous regression of retinoblastoma or benign manifestation of the mutation? Br J Cancer 45:513–521. Gallie BL, Trogadis J, Han L-P. 1999b. Retinoblastoma. In: Masters JR, Palsson B, editors. Human Cell Culture. Dordrecht: Springer, pp. 361–374. George RE, London WB, Cohn SL, Maris JM, Kretschmar C, Diller L, Brodeur GM, Castleberry RP, Look AT. 2005. Hyperdiploidy plus nonamplified MYCN confers a favorable prognosis in children 12 to 18 months old with disseminated neuroblastoma: A pediatric oncology group study. J Clin Oncol 23:6466–6473. Godbout R, Packer M, Bie W. 1998. Overexpression of a DEAD box protein (DDX1) in neuroblastoma and retinoblastoma cell lines. J Biol Chem 273:21161–21168. Grasemann C, Gratias S, Stephan H, Schuler A, Schramm A, KleinHitpass L, Rieder H, Schneider S, Kappes F, Eggert A, Lohmann DR. 2005. Gains and overexpression identify DEK and E2F3 as targets of chromosome 6p gains in retinoblastoma. Oncogene 24:6441–6449. Gratias S, Schuler A, Hitpass LK, Stephan H, Rieder H, Schneider S, Horsthemke B, Lohmann DR. 2005. Genomic gains on chromosome 1q in retinoblastoma: Consequences on gene expression and association with clinical manifestation. Int J Cancer 116:555–563. Herzog S, Lohmann DR, Buiting K, Schuler A, Horsthemke B, Rehder H, Rieder H. 2001. Marked differences in unilateral isolated retinoblastomas from young and older children studied by comparative genomic hybridization. Hum Genet 108:98–104. Knudson AG, Jr. 1971. Mutation and cancer: Statistical study of retinoblastoma. Proc Natl Acad Sci USA 68:820–823. Kondoh N, Wakatsuki T, Ryo A, Hada A, Aihara T, Horiuchi S, Goseki N, Matsubara O, Takenaka K, Shichita M, Tanaka K, Shuda M, Yamamoto M. 1999. Identification and characterization of genes associated with human hepatocellular carcinogenesis. Cancer Res 59:4990–4996. Lee WH, Murphree AL, Benedict WF. 1984. Expression and amplification of the N-myc gene in primary retinoblastoma. Nature 309:458–460. Lillington DM, Goff LK, Kingston JE, Onadim Z, Price E, Domizio P, Young BD. 2002. High level amplification of N-MYC is not associated with adverse histology or outcome in primary retinoblastoma tumours. Br J Cancer 87:779–782. 129 Lillington DM, Kingston JE, Coen PG, Price E, Hungerford J, Domizio P, Young BD, Onadim Z. 2003. Comparative genomic hybridization of 49 primary retinoblastoma tumors identifies chromosomal regions associated with histopathology, progression, and patient outcome. Genes Chromosomes Cancer 36:121–128. Lu ZL, Luo DZ, Wen JM. 2005. Expression and significance of tumor-related genes in HCC. World J Gastroenterol 11:3850–3854. Mairal A, Pinglier E, Gilbert E, Peter M, Validire P, Desjardins L, Doz F, Aurias A, Couturier J. 2000. Detection of chromosome imbalances in retinoblastoma by parallel karyotype and CGH analyses. Genes Chromosomes Cancer 28:370–379. Marchong MN, Chen D, Corson TW, Lee C, Harmandayan M, Bowles E, Chen N, Gallie BL. 2004. Minimal 16q genomic loss implicates cadherin-11 in retinoblastoma. Mol Cancer Res 2:495–503. Oliveros O, Yunis E. 1995. Chromosome evolution in retinoblastoma. Cancer Genet Cytogenet 82:155–160. Orlic M, Spencer CE, Wang L, Gallie BL. 2006. Expression analysis of 6p22 genomic gain in retinoblastoma. Genes Chromosomes Cancer 45:72–82. Pishvaian MJ, Feltes CM, Thompson P, Bussemakers MJ, Schalken JA, Byers SW. 1999. Cadherin-11 is expressed in invasive breast cancer cell lines. Cancer Res 59:947–952. Popnikolov NK, Cavone SM, Schultz PM, Garcia FU. 2005. Diagnostic utility of p75 neurotrophin receptor (p75NTR) as a marker of breast myoepithelial cells. Mod Pathol 18:1535–1541. Potluri VR, Helson L, Ellsworth RM, Reid T, Gilbert F. 1986. Chromosomal abnormalities in human retinoblastoma. A review. Cancer 58:663–671. Reis-Filho JS, Steele D, Di Palma S, Jones RL, Savage K, James M, Milanezi F, Schmitt FC, Ashworth A. 2006. Distribution and significance of nerve growth factor receptor (NGFR/p75NTR) in normal, benign and malignant breast tissue. Mod Pathol 19:307– 319. Richter S, Vandezande K, Chen N, Zhang K, Sutherland J, Anderson J, Han L, Panton R, Branco P, Gallie B. 2003. Sensitive and efficient detection of RB1 gene mutations enhances care for families with retinoblastoma. Am J Hum Genet 72:253–269. Sakai K, Tanooka H, Sasaki MS, Ejima Y, Kaneko A. 1988. Increase in copy number of N-myc in retinoblastomas in comparison with chromosome abnormality. Cancer Genet Cytogenet 30:119–126. Sakamoto Y, Kitajima Y, Edakuni G, Hamamoto T, Miyazaki K. 2001. Combined evaluation of NGF and p75NGFR expression is a biomarker for predicting prognosis in human invasive ductal breast carcinoma. Oncol Rep 8:973–980. Schwab M. 2004. MYCN in neuronal tumours. Cancer Lett 204: 179–187. Squire J, Gallie BL, Phillips RA. 1985. A detailed analysis of chromosomal changes in heritable and non-heritable retinoblastoma. Hum Genet 70:291–301. Squire J, Goddard AD, Canton M, Becker A, Phillips RA, Gallie BL. 1986. Tumour induction by the retinoblastoma mutation is independent of N-myc expression. Nature 322:555–557. Stoecklein NH, Erbersdobler A, Schmidt-Kittler O, Diebold J, Schardt JA, Izbicki JR, Klein CA. 2002. SCOMP is superior to degenerated oligonucleotide primed-polymerase chain reaction for global amplification of minute amounts of DNA from microdissected archival tissue samples. Am J Pathol 161:43–51. van der Wal JE, Hermsen MA, Gille HJ, Schouten-Van Meeteren NY, Moll AC, Imhof SM, Meijer GA, Baak JP, van der Valk P. 2003. Comparative genomic hybridisation divides retinoblastomas into a high and a low level chromosomal instability group. J Clin Pathol 56:26–30. Williamson D, Lu YJ, Gordon T, Sciot R, Kelsey A, Fisher C, Poremba C, Anderson J, Pritchard-Jones K, Shipley J. 2005. Relationship between MYCN copy number and expression in rhabdomyosarcomas and correlation with adverse prognosis in the alveolar subtype. J Clin Oncol 23:880–888. Zielinski B, Gratias S, Toedt G, Mendrzyk F, Stange DE, Radlwimmer B, Lohmann DR, Lichter P. 2005. Detection of chromosomal imbalances in retinoblastoma by matrix-based comparative genomic hybridization. Genes Chromosomes Cancer 43:294–301. Genes, Chromosomes & Cancer DOI 10.1002/gcc