Cancer Research Cell Death Mechanisms and Cancer Therapy EMT and Cancer Progression and Treatment
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
Cancer Research 67, 408, January 1, 2007. doi: 10.1158/0008-5472.CAN-06-1317
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gratias, S.
Right arrow Articles by Lohmann, D. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gratias, S.
Right arrow Articles by Lohmann, D. R.

Clinical Research

Allelic Loss in a Minimal Region on Chromosome 16q24 Is Associated with Vitreous Seeding of Retinoblastoma

Sandrine Gratias1, Harald Rieder4, Reinhard Ullmann5, Ludger Klein-Hitpass2, Stephanie Schneider6, Réka Bölöni3, Martin Kappler7 and Dietmar R. Lohmann1

1 Institut für Humangenetik, 2 Institut für Zellbiologie, and 3 Augenklinik, Universitätsklinikum Essen, Essen, Germany; 4 Institut für Humangenetik und Anthropologie, Universitätsstraße 1, Universität Düsseldorf, Düsseldorf, Germany; 5 Max-Planck Institute for Molecular Genetics, Berlin, Germany; 6 Institut für Klinische Genetik, Universitätsklinikum Marburg, Marburg, Germany; and 7 Berufsgenossenschaftliches Forschungsinstitut für Arbeitsmedizin, Ruhr-Universität Bochum, Bochum, Germany

Requests for reprints: Dietmar R. Lohmann, Institut für Humangenetik, Universitätsklinikum Essen, Hufelandstrasse 55, D-45122 Essen, Germany. Phone: 49-201-7234562; Fax: 49-201-7235900; E-mail: dr.lohmann{at}uni-essen.de.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
In addition to RB1 gene mutations, retinoblastomas frequently show gains of 1q and 6p and losses of 16q. To identify suppressor genes on 16q, we analyzed 22 short tandem repeat loci in 58 patients with known RB1 mutations. A subset of tumors was also investigated by conventional and matrix comparative genomic hybridization. In 40 of 58 (69%) tumors, we found no loss of heterozygosity (LOH) at any 16q marker. LOH was detected in 18 of 58 (31%) tumors, including five with allelic imbalance at some markers. In one tumor LOH was only observed at 16q24. As the parental origin of allele loss was unbiased, an imprinted locus is unlikely to be involved. Analysis of gene expression by microarray hybridization and quantitative RT real-time PCR did not identify a candidate suppressor in 16q24. Cadherin 13 (CDH13), CBFA2T3, and WFDC1, which are candidate suppressors in other tumor entities with 16q24 loss, did not show loss of expression. In addition, mutation and methylation analysis showed no somatic alteration of CDH13. Results in all tumors with chromosome 16 alterations define a single minimal deleted region of 5.7 Mb in the telomeric part of 16q24 with the centromeric boundary defined by retention of heterozygosity for a single nucleotide variant in exon 10 of CDH13 (Mb 82.7). Interestingly, clinical presentation of tumors with and without 16q alterations was distinct. Specifically, almost all retinoblastomas with 16q24 loss showed diffuse intraocular seeding. This suggests that genetic alterations in the minimal deleted region are associated with impaired cell-to-cell adhesion. [Cancer Res 2007;67(1):408–16]


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Retinoblastoma, a rare childhood eye tumor, has served as a model to elucidate the genetic events underlying the development of sporadic and hereditary cancer (reviewed in refs. 1, 2). For initiation of this tumor, mutations in both alleles of the RB1 gene, a tumor suppressor on chromosome 13q14, are required. Retinoblastomas often show gains on chromosomes 1q and 6p and losses of chromosome 16q (summarized in ref. 3). Therefore, it is reasonable to assume that mutations in additional genes contribute to promotion and progression of this tumor (4). The minimum regions of genomic gains on 1q and 6p have been defined and this has led to the identification of putative target oncogenes (5–10).

Loss of all or parts of chromosome 16 is observed in 31% of retinoblastomas (51 of 162; summarized in ref. 6). In most of these tumors, the whole long arm of one homologue is lost (40 of 51, 78%). A survey of comparative genomic hybridization (CGH) analyses has indicated that most partial deletions on chromosome 16q include chromosome band 16q22 (7 of 11, 64%; ref. 3). To study alterations of this region in further detail, Marchong et al. (6) did LOH analysis of seven microsatellite markers located on 16q21-23.3 (Mb 60.9–81.5, Ensembl v368) and quantitative multiplex PCR of five sequence-tagged sites in Mb 61.5 to 75.1 (Ensembl v36) and of six exons of the cadherin 11 (CDH11) gene. They found frequent allelic loss at D16S398 (located at Mb 64.7, observed in 11 of 28 tumors, 39%) and at D16S422 (Mb 81.5, observed in 9 of 23 tumors, 39%). Quantitative multiplex PCR showed that sequences located within the CDH11 gene (Mb 63.5–63.7) are most frequently lost (41 of 71 tumors, 58%). The long arm of chromosome 16 is a frequent target of deletions in various cancers. Three candidate regions, one in 16q22.1 and two in 16q24.3, have been identified. In some tumors with 16q22.1 loss, E-cadherin (CDH1) is inactivated by mutations or silenced by epigenetic mechanisms (11, 12). This provides good evidence for a tumor-suppressor role of CDH1 in these cancers. The tumor suppressors underlying loss in the telomeric candidate regions are not defined yet. Here, we used LOH analysis and microarray expression analysis to identify candidate tumor suppressors on 16q in retinoblastoma. Moreover, we investigated if clinical manifestation is distinct depending on the presence of alterations on 16q.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Samples and extraction of DNA and RNA. Tumor and blood samples were obtained from retinoblastoma patients at the time of operative treatment. In addition, blood samples were received from parents of most patients. Tissues were snap frozen in liquid nitrogen. Informed consent was obtained from all patients or their parents. Storage and processing of samples as well as extraction of nucleic acids was done as previously described (10). Material from 58 patients (54 unilateral and 4 bilateral cases) with known mutations in the RB1 gene was used for the present study.

Microsatellite analysis of markers on chromosome 16. A total of 22 short tandem repeat loci with high polymorphic information content were analyzed (Fig. 1A ; primer sequences are available on request). Intermarker distances were even (1–2 Mb along 16q) except a gap of 15 Mb between markers D16S3080 and D16S3050 (at 16q12.1 and 16q21, respectively). PCR with labeled forward primers (FAM, PET, or NED fluorescent dyes at the 5'-end, Applied Biosystems, Weiterstadt, Germany) was done in multiplexed assays and analyzed as described (10). If loss of one allele in the tumor was incomplete, the allele ratio was determined as follows: (PI allele1 tumor / PI allele2 tumor) / (PI allele1 blood / PI allele2 blood), where PI is the peak integral. To obtain allele ratio values >1, the allele with the larger peak area in the tumor was defined as allele1. We used the criteria established in a previous study (13) to categorize the results as follows: LOH for values of allele ratio >2.5; allelic imbalance for 1.3 ≤ allele ratio ≤ 2.5; normal for values of allele ratio < 1.3.


Figure 1
View larger version (45K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Pattern of chromosome 16 alterations in retinoblastoma samples. A, results of microsatellite analysis (MSA) in 40 tumors with no LOH (black squares) at any informative marker. Allelic imbalance (blue squares) is scored for makers with allele ratios between 1.3 and 2.5. Gray squares, noninformative markers. Headers of columns with microsatellite analysis results give a summary of the results of conventional CGH and matrix CGH (matrix CGH-a, reported in ref. 17; matrix CGH-b, original): n, no copy number changes on chromosome 16; L, losses; gL, gains and losses; —, not done. B, results of microsatellite analysis in 18 tumors with LOH at some or all informative markers. Parental origin of allele losses for markers with LOH on chromosome 16: m, maternal allele retained; p, paternal allele retained. C, synopsis of the results of microsatellite analysis and matrix CGH in five tumors with complex copy number changes. Green bars, gains; orange bars, losses. Scale (left), map position of chromosome 16 markers according to Ensembl v36.

 
Microarray expression analysis. Synthesis of double-stranded cDNA was done with ~2.5 µg of total RNA and anchored T7-oligo-d(T)21-V primer [5'-GCATTA-GCGGCCGCGAAATTAATACGACTCACTATAGGGAGA(T)21V-3', MWG Biotech, Ebersberg, Germany] for first-strand synthesis as previously described (14, 15). The cDNAs were purified by phenol/chloroform/isoamyl alcohol/phase lock gel (Eppendorf, Hamburg, Germany) extraction, precipitated, and used to generate biotinylated cRNA by in vitro transcription for 16 h at 37°C (Bioarray High Yield RNA Transcript Labeling kit, Enzo Life Science, Farmingdale, NY). Purification of cRNA was done using RNeasy mini columns (Qiagen, Hilden, Germany). Fragmentation of cRNA, hybridization to HG-U133A oligonucleotide arrays (Affymetrix, Inc., Santa Barbara, CA), washing, staining, and scanning (GeneArray scanner 2500, Agilent, Palo Alto, CA) were done following standard Affymetrix protocols (Technical Manual). Signal intensities and detection calls for further analysis were determined using GeneChip Microarray Suite 5.0 Software (Affymetrix). Scaling across all probe sets of a given array to an average intensity of 1,000 units compensated for variations in the amount and quality of cRNA samples and other experimental variables. Further processing of the signal values and gene information was done with standard spreadsheet software (Excel, Microsoft Corporation).

Quantitative reverse transcription real-time PCR. Reverse transcription (RT) and quantitative real-time PCR were done using Assays on Demand (CDH13 assay ID Hs00169908_m1, Applied Biosystems) as previously described (8). For relative quantification, the expression of ß-actin (Human ACTB Endogenous Control, Part No. 4352935E, Applied Biosystems) was analyzed.

Comparative genomic hybridization. Conventional CGH was done as previously described (16). CGH results of 16 of the tumors analyzed in this study had been reported previously (16). Results of matrix CGH from 16 tumors included in this study had been reported (17). In addition, samples M5715 and M5450 were analyzed by matrix CGH using a submegabase resolution tiling path bacterial artificial chromosome array, comprising the human 32k Re-Array set9 [DNA kindly provided by Pieter de Jong (BACPAC Resources, Children's Hospital Oakland, Oakland, CA); refs. 18–21], the 1 Mb Sanger set (clones kindly provided by Nigel Carter, Wellcome Trust Sanger Centre, Cambridge, United Kingdom; ref. 18) and a set of 390 subtelomeric clones (assembled by members of the European Cooperation in the Field of Scientific and Technical Research B19 initiative: Molecular cytogenetics of solid tumors). Hybridizations were done as described by Erdogan et al. (22). Detailed step-by-step protocols are also available online.10 Further analysis and visualization of matrix CGH data was done using the software package CGHPRO (23). Data were normalized by subgrid Lowess. No background subtraction was done. Circular Binary Segmentation (24), in combination with a threshold of ±0.2 log2 ratio, was used for the objective determination of presence and size of chromosomal imbalances. Results of matrix CGH of these samples are available at Gene Expression Omnibus (GEO submission no. GSE5359).11

CDH13 methylation and mutation analysis. Bisulfite treatment of DNA was done as described (25). Primers were modified from Toyooka et al. (26). Primer sequences are available on request. Reactions at 25 µL containing 3 µL bisulfite-treated DNA, 0.2 mmol/L of each deoxynucleotide triphosphate, 2 µmol/L of each primer, 2.5 µL of 10x PCR buffer, and 1 unit Taq Polymerase (AmpliTaq Gold, Applied Biosystems) were exposed to a thermal profile starting with 95°C for 10 min followed by 35 cycles of 95°C/15 s, 65°C/30 s, and 72°C/15 s, and ending with 72°C for 7 min. Products were separated by agarose gel electrophoresis and visualized with ethidium bromide. Constitutional and M.Sss I in vitro methylated DNAs (NEB, Frankfurt, Germany) were used as positive controls. Primers for PCR and sequencing of all 14 exons of CDH13 (Vega gene ID OTTHUMG00000072884) were chosen from intronic and untranslated regions using Primer3 software (27). Primer sequences are available on request. Templates for sequencing were generated by PCR in 25 µL reactions with 80 ng DNA; 10 pmol of each primer (GoTaq Green Mastermix, Promega, Madison, WI); and a thermal profile starting with 95°C/2 min followed by 35 cycles of 95°C/30 s, 51°C/30 s, and 72°C/1 min, and ending with 72°C for 5 min. For sequence reactions, a BigDye Terminator v1.1 kit was used according to the kit protocol. Products were analyzed on an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems).

Statistical analysis of genetic findings and patient data. The Wilcoxon test (normal approximation) and Significance Analysis of Microarray (SAM, Stanford University, Stanford, CA) were used to identify genes that were consistently differentially expressed between tumors with and without LOH. Detailed data on clinical presentation, treatment, and follow-up of patients were obtained. We used data warehouse software (Cognos Series 7.1; Cognos, Inc., Ottawa, ON, Canada) to link all clinical and genetic data and to set the stage for data mining, which was done using the tools provided by the software environment.

The influence of each aberration on seeding was tested by Fisher's exact test. To study the combined effect of the aberrations on seeding, a proportional odds model was done using the cumulative logit as link function in a logistic regression. Therefore, this variable was sorted in descending order to model the probabilities of diffuse seeding and of at least local seeding. Effect estimates (expressed as odds ratios, OR) >1 denote a higher probability of diffuse seeding and at least local seeding for the studied aberration. Additionally, age was included in the model as covariate. Calculations were done using SAS version 9.1 (SAS Institute, Inc., Cary, NC), and tests were conducted two-sided with a significance level of {alpha} = 5%. Statistical evaluation of the findings was also done using JMP 5.1 software (SAS Institute).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Identification of a minimal deleted region 16q. Forty of 58 (69%) tumors showed no LOH at any informative short tandem repeat marker (Fig. 1A). Results of conventional CGH (16)12 and matrix CGH (17) were available from 34 and 11 of these tumors, respectively. In all but three tumors, results of microsatellite analysis were accordant with those of CGH. Results were discordant in tumor M24733, which showed losses at all 16q probes in matrix CGH but had normal copy number in conventional CGH and no allele loss in microsatellite analysis, and in tumors M22136 and M23818, which showed loss of 16q in conventional CGH but not in microsatellite analysis. However, for both M22136 and M23818, signal losses observed in conventional CGH were just below the threshold, which was set at 80% of the mean fluorescent signal intensity. Also, it is to be noted that microsatellite analysis of tumor M23818 showed allelic imbalance at one marker D16S398 (Fig. 1A, blue square). Threshold values for classification by microsatellite analysis, conventional CGH, and matrix CGH are similar but not identical. As a consequence, classification of samples with low-level changes may come to different results. The different classification of tumors M24733, M22136, and M23818 suggests that the sensitivity of microsatellite analysis to detect low-level copy number imbalances is lower than that of matrix CGH and conventional CGH.

Microsatellite analysis detected LOH in 18 of 58 (31%) tumors (Fig. 1B). In five of these tumors (M6301, M19484, M5715, M22641, and M22731), some markers on 16q had LOH, whereas other markers showed reduced but not missing signals of one of the alleles (allelic imbalance). In conventional CGH, three of the tumors with LOH and allelic imbalance (M6301, M19484, and M5715) showed losses along the entire long arm. In matrix CGH, tumor M5715 showed gains in a small region on 16q (Mb 69.44–69.75; Fig. 1C). In this tumor, the two informative markers next to this region have allele loss (D16S3067 at 67.6 Mb and D16S3118 at 74.8 Mb), whereas all other informative 16q loci in this tumor had allelic imbalance. The remaining two of the five tumors with LOH and allelic imbalance also revealed complex changes in matrix CGH that were not detected by conventional CGH (Fig. 1C). For example, in tumor M22641, markers from 16q11.2 to 16q22.1 had allelic imbalance, whereas all informative markers toward the telomere showed clean loss of one allele. The matrix CGH results of this tumor showed gains in the region with allelic imbalance and losses in the proximal part of the region with LOH (Fig. 1C).

In two tumors, M5450 and M24794, LOH was detected only at markers in telomeric parts of 16q. In conventional CGH, both tumors had no copy number changes on chromosome 16, whereas in matrix CGH, both tumors showed DNA losses (Fig. 1C). Notably, tumor M24794 had losses only in a part of the region with LOH. This suggests that LOH in the other regions is due to isodisomy. Results of microsatellite analysis in all tumors with chromosome 16 alterations are in line with a single minimal deleted region. The boundary of this minimal deleted region toward the centromere is defined by retention of heterozygosity at D16S422 (Mb 81.5) in tumor M24794. All tumors with LOH showed alterations in the region of D16S3026 (Mb 88), which is the short tandem repeat polymorphism with the most distal location known on this chromosome. The parental origin of allele loss on 16q was determined by genotyping of parental blood DNA in 15 patients. Tumors from nine patients showed loss of alleles of paternal origin; in six tumors, the alleles of maternal origin were lost (Fig. 1B).

RNA expression data. Microarray expression data were available from 12 retinoblastomas, eight without LOH at any informative marker on 16q and four with LOH, including M22641 with a mixed pattern of alterations (GEO submission no. GSE5222). SAM was done to identify genes that, regardless of the absolute RNA expression level, are consistently differentially expressed between tumors with and without LOH. Of the top 20 genes with differential expression, 13 are located on 16q. These include three genes that are located in the minimal deleted region: MBTPS1 (membrane-bound transcription factor protease, site 1), which belongs to the sutiliase (subtilisin-like serine proteases) family, and two genes that code for zinc finger proteins, ZCCHC14 and ZDHHC7. To the best of our knowledge, these three genes and the remaining 10 that are located elsewhere on chromosome 16q have no known role in tumorigenesis.

According to Ensembl (v36), the minimal deleted region on 16q identified here contains 104 genes. Sixty-four of these genes are represented by probe sets on the Affymetrix chip Hu133A that we used for microarray expression analysis. Of these, 14 genes were not expressed in any tumor analyzed here, and 26 genes showed no significant difference in expression between tumors with and without LOH on 16q. Specifically, we found no significant difference in expression for CBFA2T3, a candidate suppressor gene in breast cancer (28, 29), and for WFDC1, which shows allele loss in liver cancer (30) and is down-regulated in prostate cancer (31). Eighteen genes had significant lower expression in tumors with LOH (Wilcoxon rank sum test) but none of these genes were completely shutdown in any tumor (Table 1 ).


View this table:
[in this window]
[in a new window]

  		Table 1. List of genes in 16q24 differentially expressed between tumors with and without LOH at chromosome 16q

 
CDH13, a putative tumor-suppressor gene in several cancer types (32–34), is located in the minimal deleted region identified here but did not show detectable expression levels on our array. We used quantitative RT real-time PCR to measure CDH13 transcripts because this method can detect lower levels of expression compared with array hybridization. We used ß-actin (ACTB) as a control because this gene was expressed at high levels in normal retina and in all retinoblastoma samples tested by microarray analysis (lower, median, and upper quartile of signal values are 32,470; 37,880; and 46,601, respectively). We found expression of CDH13 in 21 of 22 retinoblastomas tested and in RNA from normal retina (Human Retina Total RNA, BD Biosciences Clontech). Expression levels varied between tumors; however, there was no correlation with the LOH status on 16q.

According to matrix CGH data, which were available from 18 tumors, three tumors had DNA gains (M22641, M5450, and M5715) in a minimal region that ranged from Mb 69.44 to 69.75 (Fig. 1C). This region contains three genes: HYDIN, JGI-931, and HYDNI.1. Only the first of these genes is represented by a probe set on the Affymetrix array (accession no. 220098_at). We found no expression of this gene in any retinoblastoma including tumor M22641 that showed gains in matrix CGH.

Results of CDH13 methylation and sequencing analysis. The methylation status of the CDH13 promoter region was tested in 17 tumors. All samples were found to be unmethylated (Table 2 ). Sequencing of the 14 exons of CDH13 in 19 tumors and normal controls showed a few rare single nucleotide variants [homozygous: c335T>A (V112D) tumor M1727, c1038G>A (T346T) tumor M22590, c1953C>G (N651K) tumor M6302; heterozygous: c1416C>T (N472N) tumor M24794, c1731G>A (D577D) tumor M1324]. Sequence analysis of corresponding DNA from blood showed heterozygosity for all variant alleles and, therefore, none of these alterations is a somatic mutation. It is to be noted that retention of heterozygosity in tumor M24794 helps to define the centromeric boundary of the minimal deleted region at Mb 82.7 (CDH13, exon 10).


View this table:
[in this window]
[in a new window]

  		Table 2. Summary of genetic findings in DNA from tumor samples

 
Correlation of 16q alterations to RB1 mutations and to clinical variables. We found no significant association between the presence of 16q LOH and the origin of the causative mutation in the RB1 gene [15 of 44 (34%) and 3 of 13 (23%) patients with somatic and germ-line mutations, respectively, P = 0.52, Fisher's exact test two-tailed; Table 2]. Tumors with 16q loss had a higher frequency of RB1 LOH [16 of 43 (37%) compared with 2 of 15 (13%), without RB1 LOH] but this did not reach statistical significance (P = 0.11). Interestingly, the distribution of patient's age at diagnosis was different depending on the chromosome 16q status in that tumors with LOH on 16q were diagnosed significantly later (median age at diagnosis 33 months) compared with retinoblastomas without LOH (median age at diagnosis 11 months, Wilcoxon P < 0.001; Fig. 2 ). Previously, we found that age at diagnosis is distinct depending on the presence of gains on 1q and 6p (8, 10). Analyzing the age distribution of the tumors studied here, we found that tumors without gains on 1q are diagnosed at a median age of 10 months (first and third quartiles 4 and 14 months, respectively) and tumors without gains on 6p at an age of 10 months (first and third quartiles 4 and 15 months, respectively; Table 2). Tumors with gains on 1q or 6p are diagnosed at a median age of 25 months (first and third quartiles 15 and 33 months and 14 and 34 months, respectively). We found no significant association between the presence of LOH on 16q and partially or completely differentiated histology, tumor necrosis, or tumor calcification (data not shown).


Figure 2
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Age at diagnosis of retinoblastoma (in months) in relationship to chromosome 16 alterations as detected by microsatellite analysis. Retinoblastomas with LOH are diagnosed significantly later than tumors with no LOH on 16q (Wilcoxon P < 0.001).

 
Some patients with retinoblastoma show tumor cells floating within the vitreous cavity (vitreous seeding). The presence of vitreous seeding was ascertained by indirect ophthalmoscopy at the first presentation of the patient. Eyes with local seeding show few tumor cells in the vitreous that are confined to the immediate vicinity of the bulk tumor. Diffuse seeding is diagnosed if tumor cells are scattered in the vitreous cavity. We analyzed if vitreous seeding is associated with the presence of genetic alterations on 16q, 1q, and 6p. For each aberration, there was a shift to the higher categories, i.e., local and diffuse seeding (Table 3 ). In all cases, Fisher's exact test rejected the hypothesis of independence between the aberration and seeding. The highest shift was observed for LOH 16q (P = 0.002). A similar result was obtained when applying the proportional odds model (Table 4 ). Again, LOH 16q showed the highest effect on seeding with an OR of 14.3 [95% confidence interval (95% CI), 1.60–127.8]. The wide range of the 95% CI is due to the small number of patients. In contrast, gain 1q and gain 6p showed no influence on seeding in the combined analysis (OR, 1.53; 95% CI, 0.27–8.65 and OR, 3.09, 95% CI, 0.56–17.0, respectively). These results changed only slightly when age was included in the model as continuous covariate (Table 4). The OR of LOH 16q increased to 23.3 with a larger 95% CI (1.05–519), whereas the results of gain 1q and 6p did not show major changes (OR, 1.98; 95% CI, 0.31–12.5 and OR, 3.48; 95% CI, 0.59–20.6). Age had no significant effect on seeding (OR, 0.67; 95% CI, 0.23–1.92). Similar results were obtained with other age transformations like logarithm or classification of age in groups (data not shown).


View this table:
[in this window]
[in a new window]

  		Table 3. Cross-tabulation and Fisher's exact test of LOH 16q, gain 1q, and gain 6p versus seeding

 

View this table:
[in this window]
[in a new window]

  		Table 4. Results of the proportional odds model with seeding as ordinal outcome and LOH 16q, gain 1q, and gain 6p as predictors

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Our results define a minimal deleted region on 16q that extends from Mb 82.7 to the telomere of this chromosome arm. This particular region has also been identified as a candidate location of one or more tumor suppressor genes in other tumor entities, namely gallbladder carcinoma (35), lung cancer (36), prostate cancer (31), ovarian cancer (32), and, most notably, breast cancer (37) and hepatocellular cancer (38). Copy number analyses have shown that most tumors with 16q alterations have a physical deletion of this region (39), and this is also true for most of our samples with allele loss. However, we identified some retinoblastomas with LOH (M22641 and M24794) or allelic imbalance (M22731) at markers located in the minimal deleted region that showed no copy number changes (isodisomy). Isodisomy at 16q, which was also reported in some breast cancer samples (40), suggests a classic two-step mutation inactivation of putative tumor-suppressor genes in this region (41). Taking into account that 16q was implicated in parent-of-origin effects (42), an alternative explanation might be that loss of the active copy of an imprinted gene in this region contributes to tumor progression. However, we could show that the parental origin of 16q losses is not limited to only one sex and, therefore, our data do not support the observation that loss of the active allele of an imprinted locus in this region plays a role in the progression of retinoblastoma. A two-step inactivation of a tumor suppressor may result in loss of gene expression. By analysis of microarray expression data, we found that 18 of 64 genes located in the retinoblastoma minimal deleted region and represented on the microarray show lower median expression in tumors with LOH at 16q. However, none of these genes showed loss of expression in any sample.

Lung cancer is a second tumor in patients with hereditary retinoblastoma (43). This indicates that this tumor may develop along similar mutational pathways. The CDH13 gene, which is located in the retinoblastoma minimal deleted region defined here, is a candidate suppressor in lung cancer (33). Using quantitative RT real-time PCR, we found that RNA levels of this gene are variable but show no correlation with allele loss in the minimal deleted region. Moreover, Marchong et al. (6) previously showed that there is no differential expression of CDH13 between retina and retinoblastoma. To find out if a mutant form of CDH13 is expressed in retinoblastoma, we sequenced the coding region in 19 retinoblastomas but identified no somatic mutation. Therefore, it is unlikely that inactivation of CDH13 is the target of 16q loss in retinoblastoma.

All retinoblastoma samples analyzed in our study are either homozygous or compound heterozygous for mutations at the RB1 locus (44–46),13 which is to be expected if inactivation of the RB1 gene initiates the development of retinoblastoma. In addition, this indicates that the tumor cell content of the samples used for DNA preparation is high. Most retinoblastomas with LOH on 16q showed almost complete loss of one allele at every informative marker on this chromosome arm. Interestingly, 3 of 18 tumors showed allelic imbalance at several adjacent informative markers on 16q. According to matrix CGH, the genetic changes in these particular tumors are complex (Fig. 1C). Specifically, matrix CGH shows that allelic imbalance can correspond to copy number gains (tumor M22641), losses (tumor M5715), or may go without detectable changes (M22731). It is understandable that allelic imbalance is observed if there are relative gains of one chromosome homologue. It is less obvious why tumors M5715 and M22731 show allelic imbalance in regions without gains. To explain similar findings on chromosome 8q in bladder cancer, it was proposed that allelic imbalance can reflect the presence of more than one evolving subclone with allele loss (47). In fact, a pattern with LOH in some regions, and allelic imbalance in other regions, on 16q may result if clonal selection favors loss of the entire long arm of chromosome 16 in a retinoblastoma that primarily had LOH in only some regions of 16q, such as tumors M5450 and M24794.

Several tumor entities show allele loss on 16q22.1 in addition to alterations on 16q24.3 (37, 48). Therefore, it is plausible that growth of retinoblastoma is enhanced by loss of a second region on 16q. Marchong et al. identified genomic loss of chromosome 16q22 in retinoblastoma with the highest frequency of genomic loss (22 of 41 samples, 54%) for a sequence-tagged site located in the CDH11 gene (Mb 63.5). Interestingly, using immunoblot analyzes, Marchong et al. have shown loss or decrease of intact CDH11 and expression of a variant form in many retinoblastomas. In our study, the proportion of tumors with allelic imbalance or loss at markers in the region of the CDH11 gene is also high [D16S3080 and D16S3050 with LOH in 15 of 29 (52%) and 14 of 31 (45%) samples, respectively]. At the RNA level, we found lower expression of CDH11 in tumors compared with normal retina but no tumor showed loss of expression. Further studies are needed to determine if mutation of CDH11 contributes to progression of retinoblastoma (6).

Marchong et al. (6) found a second hotspot of alteration at D16S422 (16q23), the most distal marker that was analyzed in this study. Nine of 23 (39%) samples showed LOH, a proportion that compares well with our findings for the identical marker (11 of 35, 31%). In our study, all tumors with LOH at D16S422 also had LOH at further distal markers (16q24). Moreover, one tumor without LOH at D16S422 showed LOH in 16q24 (M24794). It would be interesting to test the tumor samples investigated by Marchong et al. for LOH at 16q24 markers.

In previous analyses, we found that distribution of age at diagnosis of unilateral retinoblastoma varies with the presence of genomic alterations (8, 10). Here, we found that tumors with LOH on 16q are diagnosed later than tumors without LOH. To account for the correlation of genomic alterations and age at diagnosis, it has been suggested that distinct mutational pathways can result in the development of a retinoblastoma after mutational inactivation of the RB1 locus (8, 10, 16). Probably, depending on the mutational path taken, different time periods pass until a tumor focus with mutations in both RB1 alleles reaches the size that allows a clinical diagnosis to be made. It is remarkable that the age distribution is not distinct between tumors with LOH only at some 16q loci (M5715, M22641, M22731, M5450, and M24794) and tumors with LOH at all informative 16q loci. This suggests that the minimal deleted region on 16q24 contains the target that is critical for the difference in biological behavior. An alternative explanation is that progression of retinoblastoma might be characterized by a stepwise accumulation of genetic changes (4). Our data support such a model in two ways. First, co-occurrence of genetic alterations is not random. Forty-five of 53 tumors show one of only four combinations of genetic alterations: Fifteen tumors have no alteration; 6 tumors show 6p gains only; 12 tumors have gains at both 6p and 1q; and 12 tumors show alterations at all three genomic regions. Second, median age at diagnosis is increasing with the number of genetic alterations: 8, 16, 17, and 35 months for tumors with no alterations, 6p gains only, gains at 6p and 1q, and alterations at all three regions, respectively. This also suggests an order of genetic events with 6p gains occurring first, followed by 1q gains and, finally, 16q LOH. However, data from more retinoblastomas are needed to test this model.

We found that the presence of 16q alterations is strongly associated with diffuse vitreous seeding, which reflects the ability of retinoblastoma cells to detach from the bulk tumor and proliferate into small cell clusters. Loss of cell-to-cell contacts and single cell invasion of the surround has been seen in tumors such as diffuse-type gastric cancer, which is associated with mutations in E-cadherin, CDH1 (12, 49). It is understandable that tumors with E-cadherin loss show a diffuse growth pattern because cadherin genes code for cell adhesion molecules that mediate cell-to-cell adhesion. CDH1 and several other cadherin genes are located on 16q. The presence of diffuse vitreous seeding in tumor M24794, which has no copy number changes on 16q outside of the minimal deleted region, suggests that alteration of this part of 16q only is associated with diffuse vitreous seeding. A cadherin gene, CDH13, is located in this region but shows neither differential expression (6) nor mutations (this study) in retinoblastoma. Further studies are needed to determine if other protein coding genes in the minimal deleted region have acquired mutations and to find out if these alterations explain association with vitreous seeding. Alternatively, loss in 16q24 may not target a protein coding gene but a micro-RNA (miRNA). Progression of some cancers is accompanied by alterations in miRNA genes, including gene loss (reviewed in ref. 50). However, as of now, no miRNA gene is mapped to 16q24 (Ensembl v40). From a clinical point of view, diffuse vitreous seeding is important because its presence has a grave effect on therapeutic decisions in patients with retinoblastoma. The strong association to loss of a region on 16q24 identified here will help to identify the mechanisms that underlie this adverse tumor growth pattern.


   Acknowledgments
 
Grant support: Deutsche Forschungsgemeinschaft grants Lo 530/6-1, Ri 1123/1-1, and KFO 109 (Klinische Forschergruppe Ophthalmologische Onkologie und Genetik). The Kulturstiftung Essen has supported the medical data warehouse.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Susanne Weber and Thomas Lehnert of Kulturstiftung Essen; Pieter de Jong and the BACPAC Resources Centre (http://bacpac.chori.org) for providing DNA of the human 32k BAC Re-Array Set; Nigel Carter and the Mapping Core and Map Finishing groups of the Wellcome Trust Sanger Institute for initial clone supply and verification of the 1Mb array; the COST B19 Action "Molecular Cytogenetics of Solid Tumours" for the assembly of the subtelomeric array; and Inga Nowak, Saskia Seeland, and Michael Zeschnigk for their assistance in mutation and methylation analysis of the CDH13 gene.


   Footnotes
 
8 http://www.ensembl.org. Back

9 http://bacpac.chori.org/pHumanMinSet.htm. Back

10 http://www.molgen.mpg.de/~abt_rop/molecular_cytogenetics/. Back

11 http://www.ncbi.nlm.nih.gov/geo/. Back

12 In preparation. Back

13 D.R. Lohmann, unpublished data. Back

Received 4/12/06. Revised 10/16/06. Accepted 11/ 3/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. DiCiommo D, Gallie BL, Bremner R. Retinoblastoma: the disease, gene and protein provide critical leads to understand cancer. Semin Cancer Biol 2000;10:255–69.[CrossRef][Medline]
  2. Lohmann DR, Gallie BL. Retinoblastoma: revisiting the model prototype of inherited cancer. Am J Med Genet C Semin Med Genet 2004;129:23–8.[Medline]
  3. Chen D, Gallie BL, Squire JA. Minimal regions of chromosomal imbalance in retinoblastoma detected by comparative genomic hybridization. Cancer Genet Cytogenet 2001;129:57–63.[CrossRef][Medline]
  4. Gallie BL, Campbell C, Devlin H, Duckett A, Squire JA. Developmental basis of retinal-specific induction of cancer by RB mutation. Cancer Res 1999;59:1731–5s.
  5. Chen D, Pajovic S, Duckett A, Brown VD, Squire JA, Gallie BL. Genomic amplification in retinoblastoma narrowed to 0.6 megabase on chromosome 6p containing a kinesin-like gene, RBKIN. Cancer Res 2002;62:967–71.[Abstract/Free Full Text]
  6. Marchong MN, Chen D, Corson TW, et al. Minimal 16q genomic loss implicates cadherin-11 in retinoblastoma. Mol Cancer Res 2004;2:495–503.[Abstract/Free Full Text]
  7. Orlic M, Spencer CE, Wang L, Gallie BL. Expression analysis of 6p22 genomic gain in retinoblastoma. Genes Chromosomes Cancer 2006;45:72–82.[CrossRef][Medline]
  8. Grasemann C, Gratias S, Stephan H, et al. Gains and overexpression identify DEK and E2F3 as targets of chromosome 6p gains in retinoblastoma. Oncogene 2005;24:6441–9.[Medline]
  9. Corson TW, Huang A, Tsao MS, Gallie BL. KIF14 is a candidate oncogene in the 1q minimal region of genomic gain in multiple cancers. Oncogene 2005;24:4741–53.[CrossRef][Medline]
  10. Gratias S, Schüler A, Hitpass LK, et al. Genomic gains on chromosome 1q in retinoblastoma: consequences on gene expression and association with clinical manifestation. Int J Cancer 2005;116:555–63.[CrossRef][Medline]
  11. Tamura G, Yin J, Wang S, et al. E-Cadherin gene promoter hypermethylation in primary human gastric carcinomas. J Natl Cancer Inst 2000;92:569–73.[Abstract/Free Full Text]
  12. Becker KF, Atkinson MJ, Reich U, et al. E-cadherin gene mutations provide clues to diffuse type gastric carcinomas. Cancer Res 1994;54:3845–52.[Abstract/Free Full Text]
  13. Tschentscher F, Prescher G, Zeschnigk M, Horsthemke B, Lohmann DR. Identification of chromosomes 3, 6, and 8 aberrations in uveal melanoma by microsatellite analysis in comparison to comparative genomic hybridization. Cancer Genet Cytogenet 2000;122:13–7.[CrossRef][Medline]
  14. Baugh LR, Hill AA, Brown EL, Hunter CP. Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 2001;29:E29.
  15. Durig J, Nuckel H, Huttmann A, et al. Expression of ribosomal and translation-associated genes is correlated with a favorable clinical course in chronic lymphocytic leukemia. Blood 2003;101:2748–55.[Abstract/Free Full Text]
  16. Herzog S, Lohmann DR, Buiting K, et al. Marked differences in unilateral isolated retinoblastomas from young and older children studied by comparative genomic hybridization. Hum Genet 2001;108:98–104.[CrossRef][Medline]
  17. Zielinski B, Gratias S, Toedt G, et al. Detection of chromosomal imbalances in retinoblastoma by matrix-based comparative genomic hybridization. Genes Chromosomes Cancer 2005;43:294–301.[CrossRef][Medline]
  18. Fiegler H, Carr P, Douglas EJ, et al. DNA microarrays for comparative genomic hybridization based on DOP-PCR amplification of BAC and PAC clones. Genes Chromosomes Cancer 2003;36:361–74.[CrossRef][Medline]
  19. Krzywinski M, Bosdet I, Smailus D, et al. A set of BAC clones spanning the human genome. Nucleic Acids Res 2004;32:3651–60.[Abstract/Free Full Text]
  20. Osoegawa K, Mammoser AG, Wu C, et al. A bacterial artificial chromosome library for sequencing the complete human genome. Genome Res 2001;11:483–96.[Abstract/Free Full Text]
  21. Ishkanian AS, Malloff CA, Watson SK, et al. A tiling resolution DNA microarray with complete coverage of the human genome. Nat Genet 2004;36:299–303.[CrossRef][Medline]
  22. Erdogan F, Chen W, Kirchhoff M, et al. Impact of low copy repeats on the generation of balanced and unbalanced chromosomal aberrations in mental retardation. Cytogenet Genome Res 2006;115:247–53.[CrossRef][Medline]
  23. Chen W, Erdogan F, Ropers HH, Lenzner S, Ullmann R. CGHPRO—a comprehensive data analysis tool for array CGH. BMC Bioinformatics 2005;6:85.[CrossRef][Medline]
  24. Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics 2004;5:557–72.[Abstract]
  25. Zeschnigk M, Böhringer S, Price EA, Onadim Z, Masshöfer L, Lohmann DR. A novel real-time PCR assay for quantitative analysis of methylated alleles (QAMA): analysis of the retinoblastoma locus. Nucleic Acids Res 2004;32:e125.[Abstract/Free Full Text]
  26. Toyooka KO, Toyooka S, Virmani AK, et al. Loss of expression and aberrant methylation of the CDH13 (H-cadherin) gene in breast and lung carcinomas. Cancer Res 2001;61:4556–60.[Abstract/Free Full Text]
  27. Rozen S, Skaletsky HJ. Primer3 on the WWW for general users and for biologist programmers. In: Krawetz S, Misener S, editors. Bioinformatics methods and protocols: methods in molecular biology. Totowa (NJ): Humana Press; 2000. p. 365–86.
  28. Powell JA, Gardner AE, Bais AJ, et al. Sequencing, transcript identification, and quantitative gene expression profiling in the breast cancer loss of heterozygosity region 16q24.3 reveal three potential tumor-suppressor genes. Genomics 2002;80:303–10.[CrossRef][Medline]
  29. Kochetkova M, McKenzie OL, Bais AJ, et al. CBFA2T3 (MTG16) is a putative breast tumor suppressor gene from the breast cancer loss of heterozygosity region at 16q24.3. Cancer Res 2002;62:4599–604.[Abstract/Free Full Text]
  30. Larsen M, Ressler SJ, Gerdes MJ, et al. The WFDC1 gene encoding ps20 localizes to 16q24, a region of LOH in multiple cancers. Mamm Genome 2000;11:767–73.[CrossRef][Medline]
  31. Watson JE, Doggett NA, Albertson DG, et al. Integration of high-resolution array comparative genomic hybridization analysis of chromosome 16q with expression array data refines common regions of loss at 16q23-qter and identifies underlying candidate tumor suppressor genes in prostate cancer. Oncogene 2004;23:3487–94.[CrossRef][Medline]
  32. Kawakami M, Staub J, Cliby W, Hartmann L, Smith DI, Shridhar V. Involvement of H-cadherin (CDH13) on 16q in the region of frequent deletion in ovarian cancer. Int J Oncol 1999;15:715–20.[Medline]
  33. Sato M, Mori Y, Sakurada A, Fujimura S, Horii A. The H-cadherin (CDH13) gene is inactivated in human lung cancer. Hum Genet 1998;103:96–101.[CrossRef][Medline]
  34. Toyooka S, Toyooka KO, Harada K, et al. Aberrant methylation of the CDH13 (H-cadherin) promoter region in colorectal cancers and adenomas. Cancer Res 2002;62:3382–6.[Abstract/Free Full Text]
  35. Matsuo K, Kuroki T, Kitaoka F, Tajima Y, Kanematsu T. Loss of heterozygosity of chromosome 16q in gallbladder carcinoma. J Surg Res 2002;102:133–6.[CrossRef][Medline]
  36. Sato M, Mori Y, Sakurada A, et al. Identification of a 910-kb region of common allelic loss in chromosome bands 16q24.1-q24.2 in human lung cancer. Genes Chromosomes Cancer 1998;22:1–8.[CrossRef][Medline]
  37. Cleton-Jansen AM, Callen DF, Seshadri R, et al. Loss of heterozygosity mapping at chromosome arm 16q in 712 breast tumors reveals factors that influence delineation of candidate regions. Cancer Res 2001;61:1171–7.[Abstract/Free Full Text]
  38. Riou P, Saffroy R, Comoy J, et al. Investigation in liver tissues and cell lines of the transcription of 13 genes mapping to the 16q24 region that are frequently deleted in hepatocellular carcinoma. Clin Cancer Res 2002;8:3178–86.[Abstract/Free Full Text]
  39. Knuutila S, Autio K, Aalto Y. Online access to CGH data of DNA sequence copy number changes. Am J Pathol 2000;157:689.[Free Full Text]
  40. Cleton-Jansen AM, Buerger H, Haar N, et al. Different mechanisms of chromosome 16 loss of heterozygosity in well- versus poorly differentiated ductal breast cancer. Genes Chromosomes Cancer 2004;41:109–16.[CrossRef][Medline]
  41. Cavenee WK, Dryja TP, Phillips RA, et al. Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 1983;305:779–84.[CrossRef][Medline]
  42. Karason A, Gudjonsson JE, Upmanyu R, et al. A susceptibility gene for psoriatic arthritis maps to chromosome 16q: evidence for imprinting. Am J Hum Genet 2003;72:125–31.[CrossRef][Medline]
  43. Fletcher O, Easton D, Anderson K, Gilham C, Jay M, Peto J. Lifetime risks of common cancers among retinoblastoma survivors. J Natl Cancer Inst 2004;96:357–63.[Abstract/Free Full Text]
  44. Lohmann DR, Gerick M, Brandt B, et al. Constitutional RB1-gene mutations in patients with isolated unilateral retinoblastoma. Am J Hum Genet 1997;61:282–94.[Medline]
  45. Klutz M, Horsthemke B, Lohmann DR. RB1 gene mutations in peripheral blood DNA of patients with isolated unilateral retinoblastoma. Am J Hum Genet 1999;64:667–8.[CrossRef][Medline]
  46. Yilmaz S, Horsthemke B, Lohmann DR. Twelve novel RB1 gene mutations in patients with hereditary retinoblastoma. Mutations in brief no. 206. Online. Hum Mutat 1998;12:434.[Medline]
  47. Adams J, Williams SV, Aveyard JS, Knowles MA. Loss of heterozygosity analysis and DNA copy number measurement on 8p in bladder cancer reveals two mechanisms of allelic loss. Cancer Res 2005;65:66–75.[Abstract/Free Full Text]
  48. Matsuo K, Satoh S, Okabe H, et al. SIAH1 inactivation correlates with tumor progression in hepatocellular carcinomas. Genes Chromosomes Cancer 2003;36:283–91.[CrossRef][Medline]
  49. Guilford P, Hopkins J, Harraway J, et al. E-cadherin germline mutations in familial gastric cancer. Nature 1998;392:402–5.[CrossRef][Medline]
  50. Calin GA, Croce CM. MicroRNA-cancer connection: the beginning of a new tale. Cancer Res 2006;66:7390–4.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 CarcinogenesisHome page
S. Madar, R. Brosh, Y. Buganim, O. Ezra, I. Goldstein, H. Solomon, I. Kogan, N. Goldfinger, H. Klocker, and V. Rotter
Modulated expression of WFDC1 during carcinogenesis and cellular senescence
Carcinogenesis, January 1, 2009; 30(1): 20 - 27.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gratias, S.
Right arrow Articles by Lohmann, D. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gratias, S.
Right arrow Articles by Lohmann, D. R.

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