Loss of heterozygosity of 1p in uveal melanomas with monosomy 3
Article first published online: 22 APR 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 116, Issue 6, pages 909–913, 10 October 2005
How to Cite
Häusler, T., Stang, A., Anastassiou, G., Jöckel, K.-H., Mrzyk, S., Horsthemke, B., Lohmann, D. R. and Zeschnigk, M. (2005), Loss of heterozygosity of 1p in uveal melanomas with monosomy 3. Int. J. Cancer, 116: 909–913. doi: 10.1002/ijc.21086
- Issue published online: 28 JUL 2005
- Article first published online: 22 APR 2005
- Manuscript Accepted: 12 JAN 2005
- Manuscript Received: 12 AUG 2004
- Deutsche Forschungsgemeinschaft. Grant Number: LO530/3-5
- Klinische Forschergruppe Ophthalmologische Onkologie und Genetik, Teilprojekt I-1. Grant Number: KFO109
- uveal melanoma;
- smallest region of deletion overlap (SRO);
- monosomy 3;
- chromosome 1
Gains and losses of chromosomes 1, 3, 6 and 8 are nonrandom chromosomal aberrations in uveal melanoma. Monosomy 3 is the most frequent abnormality and is associated with poor prognosis. To identify regions of allelic loss on the short arm of chromosome 1 and to investigate if these alterations contribute to uveal melanoma progression, we performed microsatellite analysis of 10 loci in 70 uveal melanomas. A total of 51 tumors were obtained from patients with clinical follow-up data, 19 tumors were from recent patients without follow-up. Loss of heterozygosity (LOH) of at least 1 marker was more frequent in tumors with monosomy 3 (40%) than in tumors with disomy 3 (10%). In particular, loss of the entire short arm of chromosome 1 was only observed in tumors with monosomy 3 (p = 0.0001). By comparing the extent of 1p LOH in all tumors with monosomy 3, we were able to define a smallest region of overlap (SRO) of approximately 55 Mb, which is flanked by markers D1S507 and D1S198. On the basis of our data and published cytogenetic data, we propose that 1p31 harbors genes involved in the progression of uveal melanoma with monosomy 3. © 2005 Wiley-Liss, Inc.
Uveal melanoma is the most common intraocular malignancy of the eye, with an incidence of about 6 per million people per year. Long-term studies have shown that metastases are the cause of death in approximately 60% of patients.1 Loss of an entire chromosome 3 (monosomy 3) is the most frequent genetic alteration. It occurs in about 50% of tumors and is a strong predictor for metastatic disease.2, 3, 4, 5, 6 Based on these findings and on the results of global gene expression analysis in primary uveal melanomas, we have proposed that tumors with monosomy 3 and tumors with disomy 3 represent 2 distinct entities and that chromosome 3 harbors genes involved in the development or progression of 1 class of uveal melanomas.7 The analysis of rare cases with partial chromosome 3 deletions has pointed to 3p25 and 3q24–26 as critical regions, but genes that are targeted by these deletions have not yet been identified.8, 9
Melanoma-related mortality is high in patients with monosomy 3, but the onset of metastatic disease is highly variable. Some patients with uveal melanoma die of metastases within the 1st year after diagnosis of the primary tumor, whereas in others metastases occur more than 30 years later.1, 10 This suggests that tumor progression might be modulated by chromosome 3 independent factors, e.g., gains and losses on other chromosomes, which are subjected to nonrandom changes in uveal melanoma. Candidate regions include 1p, 6p, 6q and 8q.2
Alterations of the short arm of chromosome 1 (1p) are frequently found in tumors of neural crest origin, such as neuroblastoma, meningioma and cutaneous melanoma, and are associated with advanced or metastatic disease.11, 12, 13 Chromosome 1 aberrations also occur in uveal melanoma, but at a lesser frequency, and their role in tumor formation or progression is unknown.2, 14, 15, 16, 17 Using cytogenetic techniques, Aalto et al.17 have shown that 1p31 was deleted in all primary uveal melanomas and metastases with losses of 1p. Interestingly, by global expression analysis, we have identified genes that map within a 4-Mb region within 1p31 and show reduced or absent expression in tumors with monosomy 3 (PDE4B, IL12RB2 and ITGB3BP),7 an observation that can be explained by the loss of 1p31 material.
To investigate whether alterations of chromosome 1, in particular 1p31, might contribute to the progression of uveal melanoma, we have studied a large series of uveal melanomas by microsatellite analysis (MSA) and have correlated the findings to the metastases-free survival of the patients.
Material and methods
Patients and tumor specimens
All 70 patients were treated at the Ophthalmology Department of the Universitätsklinikum Essen by primary enucleation without prior radiation or chemotherapy. A total of 51 tumors were obtained from patients with clinical follow-up data up to 11 years (Table I), 19 tumors were from recent patients without follow-up but they had been subjected to a global expression analysis in a previous study.7 Tumor and blood samples were stored at –80 and –20°C, respectively. DNA extraction and purification has been described elsewhere.18, 19 Survival time was defined as the time from enucleation to metastasis (event), or to the last time that the patient was known to be alive and free of metastasis (right-censored) or to the date of death for patients whose death was considered to be unrelated to uveal melanoma (right-censored). We assessed the influence of 1p alterations on metastasis-free survival by Kaplan-Meier estimates and log-rank test using SAS PROF LIFETEST (SAS Institute, Inc., Cary, NC). We tested associations between categorical variables by Fisher's exact test.
|Tumor ID||Chromosome 3 status (CGH)1||Chromosome 1p status (MSA)2||Cell type||Ciliar body involvement||Protuberance [mm]||Extra ocular growth||Patient survival||Survival time [month]|
|G 1305||Normal||Part. LOH||Mixed||No||3,8||No||Alive||106|
|M 2254||Normal||Part. LOH||Spindle||No||8,5||No||DNT||7|
Microsatellite analysis (MSA)
MSA with fluorescence-labeled primers was performed as described elsewhere.18 Briefly: 50 ng of tumor or blood DNA were added to a 20 μl reaction mixture containing 1× GeneAmp PCR buffer, 1.25 mM of each deoxynucleotide triphosphate, 0.2 units Taq polymerase (Ampi Taq, Applied Biosystems, Foster City, CA) and 8 pmol of each primer pair. Cycling was performed in a GeneAmp PCR system 9700 thermocycler (PE, Applied Biosystems, Foster City, CA) with an initial denaturation step of 2 min at 95°C, 35 cycles of denaturation (15 sec at 95°C), annealing (30 sec at primer specific temperature), extension (30 sec at 72°C) followed by a final extension step at 72°C for 7 minutes. Amplification products were analyzed using an ABI Prism 3100 and the GeneScan™ and Genotyper® software (Applied Biosystems). To correct for differences in amplification efficiency of distinct marker alleles, the ratio of allele peak areas (allele ratio, AR) in the tumor as given by the GeneScan and Genotyper software was normalized against the peak ratio obtained in corresponding constitutional DNA.
Based on a previous study comparing MSA and comparative genomic hybridization data obtained from the same tumor and blood samples, we defined the ARs as follows: retention of heterozygosity (ROH), AR ≤ 1.2; allelic imbalance (AI), 1.2 ≤ AR ≤ 2.5; loss of heterozygosity (LOH), AR > 2.5.18 The allele with the largest peak area in the tumor was always defined as allele 1 to obtain values ≥ 1. The following loci were tested by MSA (distance from 1pter): D1S214 (6 Mb), D1S507 (14 Mb), D1S233 (30 Mb), D1S190 (44 Mb), D1S1661 (50 Mb), D1S438 (63 Mb), D1S198 (66 Mb), D1S219 (69 Mb), D1S248 (106 Mb), D1S533 (191 Mb) (GesGen, Invitrogen Corp., Carlsbad, CA). The approximate map location and primer sequences are available from the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/mapview/map_search.cgi?taxid=9606).
Quantitative real-time PCR
To determine the relative DNA copy number of D1S190, D1S233 and D10S558, we performed real-time Taqman®-PCR assays on an ABI Prism® 7000 sequence detection system as described elsewhere.20 For all loci tested, standard curves were generated. All reactions were performed in duplicate in a total reaction volume of 25 μl including 120 nM of each primer, 250 nM MGB-Probe (VIC-5′GTGTGTGTGTGTGTGTG 3′-DQ; Applied Biosystems), 12,5 μl of 2× Universal PCR MasterMix (no uracil-DNA glycosylase UNG) (Applied Biosystems) and 5 U AmpliTaq Gold™ (Applied Biosystems). The following cycle parameters were used: 1 step at 95°C for 10 min, followed by 40 cycles at 95°C for 15 sec and 60°C for 1 min.
Based on global gene expression analysis of 20 primary uveal melanomas, we previously identified 2 classes of uveal melanomas.7 All of the tumors in 1 class had disomy 3, whereas all but 1 of the tumors in the other class had monosomy 3 (Fig. 1). Although the analysis suggested the existence of subclasses, there was no obvious correlation with any of the clinicopathological features studied or with alterations of chromosomes 6 and 8. In the present study, we typed the tumors and corresponding blood samples at 7 1p microsatellite loci to analyze a possible correlation between 1p alterations and gene expression–based tumor classification. Among 19 tumors studied, we found 3 tumors with LOH at all informative markers. These tumors are part of a subclass of 4 tumors with monosomy 3 (Fig. 1). None of the tumors with disomy 3 showed LOH of 1p.
To extend this analysis and to identify common regions of 1p losses, we studied tumor and blood samples from another 51 patients. These patients were selected because we had long-term follow-up data on them. A total of 27 tumors had monosomy 3, 23 tumors had disomy 3, and 1 tumor had partial monosomy 3p (Table I; Fig. 2). We analyzed these samples with 10 chromosome 1 markers including the markers described above, 2 additional markers on band 1p31, and 1 marker on 1q.
In 8 (30%) of 27 tumors with monosomy 3, all informative 1p markers showed ROH (ARs < 1.2). In 7 tumors (26%) at least 1 1p marker showed an AR in the range of 1.3 to 2.5, which is designated as AI.18 One of these tumors, G1405, showed AI (AR = 1.4–1.9) of all informative 1p markers, suggesting an additional copy of the short arm of chromosome 1. In 7 other tumors, marker D1S248 showed AI, suggesting that gain of material in this centromeric region might be a frequent alteration in tumors with monosomy 3. In 4 tumors (14%) with monosomy 3, a heterogeneous pattern was obtained. In 1 of these tumors (G1497) we found ROH at D1S190, but LOH at the flanking loci D1S198 and D1S233. As this result could indicate the presence of a homozygous deletion of D1S190, we compared the relative allele doses at D1S190, D1S233 and D10S558 in tumor G1497 by quantitative real-time PCR.20 No significant reduction of D1S190 was seen (data not shown). These results are compatible with a complex chromosome 1 rearrangement or may be related to technical problems. Discounting this tumor, we can define a smallest region of overlap (SRO) (Fig. 2a) of approximately 55 Mb flanked by the markers D1S198 and D1S507.
In tumors with disomy 3, 20 tumors (87%) did not have any LOH on 1p. Of these, 9 tumors had AI at 1 or a few loci. LOH of at least 1 marker was present in only 3 tumors (13%), resulting in an SRO defined by the loci D1S507 and D1S214, which map to 1p36 (Fig. 2b). Of note, LOH of the entire short arm of chromosome 1 was not detected in any tumor with disomy 3. This confirms our findings in the 19 tumors used for cluster analysis (see above).
We also analyzed a single tumor with partial monosomy 3. In this tumor, all but 1 informative 1p marker showed LOH.
To evaluate if genes affected by chromosome 1 alterations might modulate disease progression, we associated the metastases-free survival of patients with the chromosome 1 status by Kaplan-Meier analysis. As tumors with disomy 3 rarely give rise to metastatic disease we included only tumors with monosomy 3. Kaplan-Meier analysis did not show a correlation between 1p alterations and metastases-free survival of patients with monosomy 3 (data not shown).
We have used MSA to determine chromosome 1p alterations in 70 primary uveal melanomas. Among 37 tumors with monosomy 3, 11 tumors (30%) have LOH for all of 1p. In contrast, none of 32 tumors with disomy 3 had LOH for all of 1p. This difference is highly significant (p = 0.0001). Based on a temporal analysis of cytogenetic imbalances, Hoglund et al.16 have shown that loss of 1p is secondary to loss of chromosome 3 and that these aberrations define one of several cytogenetic pathways in melanoma formation. The chromosome 3- 1p- pathway appears to result in tumors with a specific gene expression profile, as suggested by our cluster analysis (Fig. 1). Therefore, it is reasonable to assume that this subclass of tumors has distinct biological properties. In this regard, it is of interest to note that 1p LOH is correlated with tumor size, stage and ciliary body involvement.14 A correlation of 1p alterations and tumor-related survival of patients can be deduced from the results of a CGH based study on 29 uveal melanomas by Alto et al.17 They found complete or partial loss of 1p in 5 tumors. In 3 of these tumors, loss of chromosome 3 was also observed. All 5 patients died of their tumor disease. These data suggest, that loss of 1p in patients with monosomy 3 might correlate with poor prognosis and support the hypothesis that genes located on 1p might be relevant for uveal melanoma progression. In view of our studies and that of others, it is surprising that 2 of the 5 patients with loss of 1p that died of metastatic disease, had disomy 3.3, 4, 6 It is possible that these 2 patients had uniparental disomy of chromosome 3, a state that is not detectable by chromosome analysis or CGH. In the present study, we could not find a correlation of 1p alterations with metastases-free survival in patients with monosomy 3. However, this might be due to the relatively small study size.
To define the region on chromosome 1p that is targeted by allelic losses, we compared the extent of LOH and found an SRO of approximately 55 Mb, which extends from 1p31 to 1p35. This SRO might overlap with the SRO defined by Aalto et al.17 (Fig. 3). However, the alignment is hampered by the fact that the SROs were determined with different techniques (MSA vs. CGH). This region at 1p31 does not overlap with SROs that have been described for other tumors of neural crest origin.11, 12, 13, 21
By global gene expression analysis of 20 uveal melanomas, we had identified 3 genes on 1p31 that showed reduced (ITGB3BP) or absent expression (PDE4B and IL12RB2) in almost all tumors with monosomy 3.7 Two of these genes (ITGB3BP and PDE4B) map within the SRO as defined in the present study (Fig. 3), indicating that loss of genetic material of 1p31 might contribute to the downregulation of these genes. Compared to tumors with disomy 3, the steady state level of ITGB3BP, which encodes the integrin binding protein beta-3, was reduced by a factor of 2 in tumors with monosomy 3, which is consistent with the loss of 1 1p allele in these tumors. However, as we detected 1p LOH in only 3 of these tumors (Fig. 1), we would have to assume that in the other tumors ITGB3BP expression is lost as the result of a local deletion or mutation.
PDE4B, which encodes the cAMP-specific phosphodiesterase 4B, is a candidate tumor suppressor gene, because it maps within the common region of LOH and shows loss of expression in almost all uveal melanomas with monosomy 3. To account for the expression findings, we would have to assume a local deletion or mutation of the remaining allele in the 3 tumors with 1p LOH and of both alleles in the other tumors. A similar reasoning would have to apply to IL12RB2, which encodes the interleukin 12 receptor beta-2 and maps just outside the SRO. Interestingly, epigenetic silencing, i.e., promoter methylation, of IL12RB2 has been observed in malignant B cells.22 While these 3 genes are candidate genes for the progression of uveal melanoma with monosomy 3, they are unlikely to play a specific role in the subclass of tumors with monosomy 3 and 1p LOH. To further evaluate the role of 1p alterations in uveal melanoma progression it will be necessary to narrow the SRO by the analysis of additional tumors and to perform mutation and expression studies of candidate genes mapping within the critical region.
D.L. is financed through KFO109, Klinische Forschergruppe Ophthalmologische Onkologie und Genetik, Teilprojekt I-1.