Multiplex Ligation-Dependent Probe Amplification Analysis of Uveal Melanoma with Extraocular Extension Demonstrates Heterogeneity of Gross Chromosomal Abnormalities

Abstract

Purpose.: To determine whether biopsy of extraocular extension of uveal melanoma (UM) is representative of the intraocular tumor with respect to copy number of chromosomes 1p, 3, 6, and 8.

Methods.: Multiplex ligation-dependent probe amplification (MLPA) using the P027 assay was performed on formalin-fixed, paraffin-embedded sections from 10 UMs. The intraocular and extraocular parts of the tumor were microdissected and analyzed separately.

Results.: Of the 10 UMs analyzed, seven showed heterogeneity for at least one chromosome arm; the most frequently heterogeneous chromosome arm was 6p. No heterogeneity of 8p was observed between the intraocular and extraocular areas of the tumor. One tumor showed monosomy 3 in the intraocular area of the tumor but loss of the 3q arm only for the extraocular area.

Conclusions.: Biopsy of an extraocular tumor extension may not be representative of the underlying UM with respect to chromosome 1p, 3, 6, and 8q abnormalities detectable by MLPA. These results suggest that for UM with extraocular extension, both the intraocular and the extraocular parts of the tumor should be sampled for accurate genetic prognostic testing.

Uveal melanoma (UM) is an unusual tumor in that nonrandom gross chromosomal changes are strongly associated with prognosis, in particular monosomy 3 and polysomy 8q. ^{1 –4} Clinical and histomorphologic characteristics of the tumor are also predictive, with poor prognosis correlating with features such as large tumor size, ciliary body involvement, epithelioid cytomorphology, extracellular matrix periodic acid-Schiff-positive (PAS+) loops, high mitotic count, and extraocular extension. ^1,5,6  

Previous studies have shown the prevalence of extraocular extension to range from <1% to 14.6%. ^{7 –10} Extraocular extension correlates with metastatic death; this is thought to be due to the association with features of increased malignancy, such as epithelioid cell morphology, and of advanced disease, such as tumor size. ^{8,10 –13} However, the route of extraocular extension appears not to be important. ⁸  

Monosomy 3 was first observed in UM in the early 1990s ^14,15 and remains one of the strongest genetic prognostic factors for UM metastasis >10 years after the first study identified its importance. ^2,4 The incorporation of molecular genetic testing, particularly for the presence or absence of monosomy 3, into routine care has proved successful for UM patients, ^1,16,17 allowing patients with a poor prognosis to be more rigorously followed, with the aim of identifying and treating metastases earlier. However, whether heterogeneous distribution of monosomy 3 cell populations exists within a single tumor remains controversial. ^{18 –22} Studies have investigated whether histopathologic features indicating poor prognosis occur focally with monosomy 3 cell populations in UM. In mixed cell tumors, areas of epithelioid cells were found to be monosomy 3 ²² ; however, vasculogenic mimicry patterns were not limited to areas of monosomy 3 cells. ²⁰  

To our knowledge, no previous studies have investigated whether extraocular UM shows the same chromosomal changes as the intraocular component of the tumor. The aim of this study was, therefore, to use multiplex ligation-dependent probe amplification (MLPA) to determine whether UM with a poor prognostic feature, extraocular spread, were heterogeneous for the most common gross chromosomal abnormalities occurring in UM of chromosomes 1p, 3, 6, and 8. 

Materials and Methods

Tumor Samples

Formalin-fixed, paraffin-embedded (FFPE) enucleated eyes of patients with extraocular tumor extension were available from the archives of the Department of Pathology, Royal Liverpool University Hospital. Of these, 10 patient samples, diagnosed between 2002 and 2009, were selected for the study, because both intraocular and extraocular areas of the UM were sufficiently large to microdissect and yield a minimum of 700 ng DNA for quality control PCR and for analysis by MLPA in triplicate. 

None of the tumors used in this MLPA analysis had developed from a preexisting naevus or had been previously treated by radiotherapy or chemotherapy before enucleation. Informed consent was obtained from each patient, and research was performed according to the tenets of the Declaration of Helsinki. Ethical approval was obtained for this study from the Local Research Ethics Committee (LREC number 014/103). 

DNA Extraction

After examination of hematoxylin and eosin- and PAS-stained 20-μm FFPE sections, areas with greater than 90% UM cells were microdissected from the intraocular and the extraocular tumor components and dispensed into individual microfuge tubes. DNA extraction was performed as described in Lake et al. ²³ using a modified DNA purification kit (DNeasy Blood & Tissue Kit; Qiagen, Crawley, UK). In brief, tissues were lysed in proteinase K buffer (1.6–0.8 mg/mL proteinase K, 50 mM Tris, pH 8.5, 0.1 M NaCl, 1 mM EDTA pH 8.0, 0.5% Tween-20, 0.5% NP40, 20 mM dithiothreitol) for 16 hours at 56 °C followed by a further 24 hours at 37°C. The DNeasy protocol was modified to include two washes with AW1 buffer, and DNA was eluted in 50 μL AE buffer. DNA quantity and A260/280 ratio were assessed with a spectrophotometer (NanoDrop; Thermo-Fisher Scientific, Cambridge, UK). 

Quality Control PCR

A multiplex PCR, adapted from the technique of van Dongen et al., ²⁴ was performed to ensure sufficient DNA quality for analyses. Twenty-five-microliter reactions contained 1× high-performance buffer, 2 mM MgCl₂, 0.5% BSA (Sigma-Aldrich Company Ltd, Gillingham, UK), 0.8 mM dNTP mix (ABgene, UK), 0.625 U polymerase (ThermoStart; ABgene, Epsom, UK), 0.1 μM forward and reverse primers for RAG1, PLZF, and AF4 exon 11, 0.2 μM forward and reverse primers for AF4 exon 3 (Eurofins; MWG Operon, London, UK), and 100 ng DNA. Reactions were performed using a thermal cycler (TC-412; Techne, Staffordshire, UK). PCR products were visualized on 2% agarose gels stained with 1× DNA gel stain (SYBR Safe; Invitrogen, Paisley, UK) using an imaging system (Bio Doc-It; Ultra-Violet Products Ltd, Cambridge, UK). 

Multiplex Ligation-Dependent Probe Amplification

As described in Lake et al., ²³ the MLPA assay (SALSA P027; MRC-Holland, Amsterdam, The Netherlands) was performed using 200 ng DNA from the test UMs and three to six nontumor control DNAs from normal tonsil or choroid in each assay. MLPA reactions were performed using a thermal cycler (G-Storm GS1; Gene Technologies Ltd, Essex, UK). Amplified fragments were detected (3130 Genetic Analyzer; Applied Biosystems, Foster City, CA), and software tools (GeneMarker SoftGenetics, State College, PA) were used to determine peak heights as a measure of intensity. Intraocular and extraocular tumor areas were tested independently three times. 

During the course of this study, the P027 assay was revised by the manufacturer (MRC-Holland). In the new assay, test probes FOXC1 (6p25) and SERPINB9 (6p25.2) were replaced with PECI (6p25.2) and DCDC2 (6p22.2). As a result of this modification, tumors UM1–6 were tested with the P027 assay and UM7–10 with the P027-B1 assay. The new assay, however, was validated in routine prognostic testing. ²⁵ In addition, the intraocular and extraocular melanoma components of UM1 and UM4 were analyzed using both assays. No discrepancy was observed between the results for chromosome 6p from the P027 and P027B1 assays. 

Amplification or deletion of individual loci was determined from the peak heights based on the guidelines set out by the National Genetics Reference Laboratory (NGRL) (http:/www.ngrl.org.uk/Manchester/mlpapubs.html) in which dosage quotients (DQs) are determined for each locus. A DQ value of 0.85 to 1.15 indicates a diploid signal for that particular locus; DQs of <0.85 to 0.65 and >1.15 to 1.35 were designated by the NGRL to be equivocal loss or equivocal gain of the locus, respectively. Increased prognostic accuracy was demonstrated by Damato et al. ²⁵ when such equivocal (borderline) changes were considered abnormal. As a result of these data, we designated any DQ below 0.85 as loss and any DQ above 1.15 as gain. 

Criteria for calling copy number of chromosome arms were modified from those described by Lake et al. ²³ to allow for the observation of no disomy copy number signal on a chromosome arm. Consequently, where a minimum of one probe gave a signal indicating no loss or gain, the chromosome arm was considered disomic (see Table 2, abbreviation “Di”). Loss (L) or gain (G) of a chromosome arm was determined only if all MLPA probes on that arm showed loss (DQ <0.85) or gain (DQ >1.15). Where probes on the same arm gave differing aberrant copy number signals (e.g., deletion and amplification) but no diploid signal, the copy number was considered undeterminable and designated unclassified (U). 

Results

Patients

The 10 UMs analyzed were from patients whose median age was 68.5 years (range, 53–82 years); five were men and five were women. Five UMs arose in the ciliary body. Clinical and histopathologic features of all UMs examined in this study (including at the seventh TNM stage) are summarized in Table 1. ²⁶ These UMs had a median LBD of 15.5 mm (range, 11–21 mm) and were classified histologically as of mixed cell type in five cases, spindle in two, and epithelioid in three, according to the modified Callendar classification. ²⁷ Closed extracellular matrix PAS+ loops ¹ were detected in five of the UMs. The mitotic count, determined on hematoxylin and eosin sections, as described previously, ¹ ranged from 2 to 61 mitotic figures per 40 high-power fields (median, 4). 

Table 1.

 

Clinicopathological Features and Tumor Classifications

Tumor	Sex	Age at Primary Management (y)	Ciliary Body Involvement	Histopathologic Features				TNM Stage†	Fatal UM Metastasis
				Cell Type	Closed Loops	Mitotic Rate*	LBD (mm)
UM1	M	59	N	Mix	Y	2	16	3c IIIA	N
UM2	M	68	N	Mix	N	2	12	3c IIIA	N
UM3	M	68	Y	E	N	61	15	3d IIIB	Y
UM4	F	71	Y	Mix	Y	4	16	3d IIIB	Y
UM5	F	69	Y	E	N	15	11	1d IIA	Y
UM6	F	76	N	S	N	2	11	2c IIIA	Y
UM7	F	53	N	S	N	2	18	4c IIIB	N
UM8	F	76	Y	Mix	Y	19	12	2d IIIA	N
UM9	M	82	Y	E	Y	18	18	3d IIIB	Y
UM10	M	67	N	Mix	Y	4	21	4e IIIC	N

 LBD, largest basal tumor diameter; M, male; F, female; E, epithelioid cells; S, spindle cells; Mix, mixed cell type; N, no; Y, yes.

*  Mitotic rate assessed as number of mitotic figures per 40 hpf.

†  Based on the AJCC Cancer Staging Manual, seventh edition. ²⁶

The National Cancer Registry informed us when any of our patients died if they were resident in mainland Britain. Patient mortality resulting from metastatic UM was confirmed for five patients in the study (UM3-UM6 and UM9). All remaining patients were UK residents with no known metastases and were alive at the close of the study. 

Multiplex Ligation-Dependent Probe Amplification

Chromosome arm copy number detected by MLPA is given for each UM in Table 2, with the dosage quotients for each locus tested by MLPA detailed in Supplementary Table S1. Of the 10 UMs analyzed, only three did not show any gross chromosomal differences between the intraocular and extraocular areas of the tumor. These were UM6, UM9, and UM10. There was no heterogeneity of chromosome 8p in any of the examined UMs. 

Table 2.

 

Chromosome Arm Copy Number Detected by MLPA

Tumor	Sample	Chromosome Arm
		1p	3p	3q	6p	6q	8p	8q
UM1	Intraocular	Di	Di	Di	Di	Di	L	U
	Extraocular	Di	Di	L	Di	Di	L	U
UM2	Intraocular	Di	Di	Di	G	Di	L	Di
	Extraocular	Di	Di	Di	Di	Di	L	Di
UM3	Intraocular	Di	Di	Di	G	Di	L	U
	Extraocular	Di	Di	Di	Di	L	L	U
UM4	Intraocular	Di	L	L	Di	Di	L	G
	Extraocular	Di	Di	L	Di	Di	L	G
UM5	Intraocular	Di	Di	Di	G	Di	Di	G
	Extraocular	Di	Di	Di	Di	Di	Di	G
UM6	Intraocular	L	Di	Di	G	L	Di	U
	Extraocular	L	Di	Di	G	L	Di	U
UM7	Intraocular	Di	Di	Di	G	L	Di	U
	Extraocular	Di	Di	Di	Di	Di	Di	Di
UM8	Intraocular	U	Di	U	Di	L	G	Di
	Extraocular	Di	Di	Di	Di	Di	G	Di
UM9	Intraocular	Di	L	L	Di	L	L	G
	Extraocular	Di	L	L	Di	L	L	G
UM10	Intraocular	Di	L	L	Di	Di	L	G
	Extraocular	Di	L	L	Di	Di	L	G

 D, disomy; L, loss of all loci on the chromosome arm; G, gain of all loci on the chromosome arm; U, unclassified.

The chromosome arm demonstrating the most frequent heterogeneity was 6p, showing gains in the intraocular component and disomy in the extraocular area in four UMs (UM2, UM3, UM5, and UM7). UM2 and UM5 showed heterogeneity only of 6p. The extraocular tumor of UM7 was disomy for chromosome 6, but the intraocular area showed gain of 6p and loss of 6q, suggesting the presence of isochromosome 6p. 

Three tumors were heterogeneous for chromosome 3 loss: UM1 had loss of 3q only in the extraocular tumor; UM4 had loss of one copy of chromosome 3 in the intraocular area of the tumor and of 3q in the extraocular tumor component; and UM8 had unclassified 3q in the intraocular tumor. 

With regard to chromosome 8q, classification of chromosome arm loss or gain was not achieved in 4 of 10 UMs. The most heterogeneous tumors were UM7 and UM8; three chromosome arms were heterogeneous for copy number between the intraocular and extraocular tumor samples. 

Discussion

In the present study, we demonstrated that chromosomal abnormalities in extraocular UM differ significantly from those detected in intraocular UM. To the best of our knowledge, this finding has not been previously reported in the literature. The heterogeneity we observed is in agreement with several studies indicating that UM is genetically heterogeneous. ^{18 –22,28} Our study is of clinical relevance because, although the extraocular growth may be more accessible for tumor sampling, it may not be representative of the usually larger, intraocular portion of the UM. 

The strengths of our study lie in the different approach we took in assessing heterogeneity compared with that of previous studies. We used MLPA rather than fluorescence in situ hybridization (FISH) because our experience demonstrated that MLPA has greater sensitivity than FISH for UM prognostic testing ² ; analyzed heterogeneity of chromosomes 1p, 3, 6, and 8, not just of chromosome 3; and used the MLPA data to determine copy number of each chromosome arm tested in the intraocular and extraocular regions (i.e., not of a whole chromosome or, conversely, of individual genes). ^{18 –22} This study applied the same experimental methods used by Dopierala et al., ¹⁸ in their study of heterogeneity by MLPA in UM. However, the analysis of the data in the present study differs from used by Dopierala et al., ¹⁸ in that less emphasis was placed on heterogeneity in the loss or gain of individual loci across the tumor; instead, the present investigation concentrated on copy number changes for individual chromosome arms between different tumor areas. 

A major weakness of this study is the small cohort analyzed, preventing meaningful statistical analysis of the MLPA data to determine whether heterogeneity influenced prognostic testing results. Given that only 14.6% of UM patients treated in Liverpool have extraocular extension, ⁸ only a small cohort of patients was available from whom we could select the samples for this study. Long-term prospective studies would allow a larger cohort to be examined and potentially appropriate statistical analysis to be performed. Seven hundred nanograms of DNA was needed for MLPA analyses from the FFPE material; therefore, larger tumors were selected, and this might have introduced some bias into the study by possibly favoring older or more aggressive tumors. 

Although monosomy 3 remains a strong cytogenetic indicator of the metastatic potential of UM, the inclusion of chromosome 8q gains in prognostic testing greatly improves the ability to predict metastasis. ^2,25 Five of the 10 UMs (UM3-UM6 and UM9) analyzed were from patients confirmed to have died of metastatic melanoma. Of these, UM4 and UM9 showed evidence of monosomy 3 cell populations. Interestingly, in the case of UM4, the intraocular component showed monosomy 3 whereas the extraocular tumor showed loss of 3q only. Polysomy 8q was present in both the intraocular and the extraocular areas of 3 of the 5 UMs (UM4, UM5, and UM9), which ultimately produced metastases. The remaining two metastatic UMs were unclassified for 8q copy number in both the intraocular and the extraocular tumor areas. The unclassified calls for chromosome 8q were the result of gain of the MYC and DDEF1 loci but loss of the remaining gene probe, RP1, on 8q. Lack of agreement between the copy numbers observed for RP1, MYC, and DDEF1 occurred in 4 of 10 UMs and resulted in 8q copy number being unclassified for these samples. This could be caused by the specific amplification and loss of these genes; however, without more detailed analysis of 8q copy number by other techniques (e.g., comparative genomic hybridization or genomic microarray), it is not possible to determine the background in which these gene aberrations have taken place. The specific overrepresentation of MYC and DDEF1 is in agreement with observations from previous studies. ^29,30 Incorporation of additional probes on chromosome 8q in future versions of the P027 assay would improve the confidence with which chromosome arm copy number could be detected. In addition, this would allow delineation of the importance of MYC and DDEF1 in UM progression. 

Chromosome 6p was the most heterogeneous of the chromosome arms tested in this study; four tumors, (UM2, UM3, UM5, and UM7) were polysomic for 6p in the intraocular sample but disomic in the extraocular area. Several studies have identified polysomy 6p as a frequent aberration in UM. ^{31 –33} From the limited cohort of 10 UMs analyzed in this study, a trend in the presence of heterogeneity of chromosome 6p and patient survival was not observed. The lack of association between the presence of polysomy 6p and monosomy 3, indicating the proposed bifurcate pathway for UM pathogenesis, is upheld in the samples studied. However, the trend of polysomy 6p to be associated with a good prognosis is not supported here, as demonstrated by 2 of the 5 UM patients with fatal metastases having gain of 6p. ^{34 –36} These three tumors also showed gain of the MYC and DDEF1 loci on 8q in both tumor areas; association of 6p gains and 8q gains has previously been observed by our research group in routine samples. ²⁵ These observations, coupled with the high level of heterogeneity of the 6p arm, may suggest that development of 6p gain in UM may not have a strong influence on patient survival or may be the result of genomic instability (i.e., a passenger aberration). ²⁵ The lack of detectable monosomy 3 in these samples could be explained by either the presence of small deletions of chromosome 3 below the threshold of detection for MLPA, as seen in SNP microarray studies, ²¹ or the presence of isodisomy of chromosome 3, as observed by Onken et al. ³⁷ in 6% of UMs. 

Chromosome 8p showed no evidence of intratumor heterogeneity. However, a single locus was tested in the P027 assay on chromosome 8p; the sparse coverage of this region of the genome may contribute to the lack of heterogeneity observed. 

Monosomy 3 was detected in three tumors: UM4, UM9, and UM10. Extraocular and intraocular samples from UM9 and UM10 were monosomy 3, but only the intraocular sample was monosomy 3 for UM4, with the intraocular sample showing loss of 3q. In addition to the strong association of monosomy 3 with poor prognosis in UM patients, studies have found that monosomy 3 cells dominate the cell population in liver metastases. ^30,38 Hence, a monosomy 3 cell population, present solely in the intraocular area of a UM, could be sufficient for metastasis, as seen in UM4. Previous studies have reported small numbers of UM with partial loss of chromosome 3; this loss ranged from small deletions to loss of a whole chromosome arm, though the latter event was less frequent. ^{39 38,40 –42} Minimal regions of loss of 3p and 3q have been defined; however, no strong association between the loss of a single arm, or a smaller region of chromosome 3, and metastasis development has been determined. In the present study, two tumors show loss of the 3q arm in the extraocular area; UM1 has not metastasized, UM4 has metastasized. From this limited data set, no further conclusions can currently be drawn as to the significance of this partial chromosome loss. Chromosome 3 copy number data suggest that UM3, UM5, and UM6 may belong to the rare subset of disomy 3 metastasizing UM that harbor many small deletions of chromosome 3 while retaining two copies of the chromosome. ^1,23 Conversely, UM10 showed loss of chromosome 3 and gain of 8q in both the intraocular and the extraocular areas of the tumor, but the patient was free of metastasis at the completion of the study, 26 months after enucleation. This is perhaps not surprising because, despite the association of poor prognosis with chromosomal abnormalities of 3 and 8q in UM, disease-specific mortality has been determined as 66% at 5 years after diagnosis. ¹  

UM4 is the only tumor in which genetic testing of the extraocular tumor alone could have given a false indication of the chromosome 3 status. However, several features of the tumor would have indicated that the patient belonged to the “high-risk” group for the development of metastases and should be closely monitored. These included 3q loss in the extraocular area, polysomy 8q in both tumor areas, ciliary body involvement, and epithelioid cells and closed extracellular matrix PAS+ loops. In our center, clinical, histomorphologic, and genetic data are used in an accelerated failure time model to generate personalized prognostic curves (www.ocularmelanomaonline.com) ⁴³ ; that is, we do not rely completely on only one particular feature of the tumor. Therefore, prognostic curves for UM4 were generated on the basis of all the clinical and histomorphologic parameters described here. Figure 1A is based on UM4 being monosomy 3, and Figure 1B is based on UM4 being disomy 3. The inclusion of multiple tumor features compensates for the extraocular disomy 3 signal, and poor prognosis is indicated in both prognostic curves. Although UM is undoubtedly genetically heterogeneous and multiple samples should be taken for UM prognostic genetic testing, this study emphasizes the importance of considering multiple prognostic features to accurately predict patient prognosis. 

Figure 1.

 

Personalized prognostic curves for UM4 generated using www.ocularmelanomaonline.com. All clinical (age, sex, largest ultrasound diameter, ciliary body involvement, presence of extraocular extension) and histomorphologic (presence of epithelioid cells, PAS+ loops, mitotic rate) features used to generate both curves were identical. (A) Curve generated for the intraocular tumor with monosomy of chromosome 3. (B) Curve for the disomy 3 extraocular tumor area. Dashed lines: 95% confidence intervals for the survival of the patient over an 8-year period after diagnosis of UM. The line between them is the predicted survival curve for the person. Solid line: survival of an age- and sex-matched control population. The y-axis scale is 0 to 1, with 1 representing 100%. Both curves indicate poor patient prognosis irrespective of chromosome 3 status.

Further investigation of genetic heterogeneity in a larger cohort of UMs with extraocular extension and known clinical course is necessary to confirm which abnormalities, when present in discrete areas of the UM, are associated with poor patient prognosis. This would further improve prognostic testing for UM patients, allowing those with a poor prognosis to be enrolled in the appropriate clinical trials and to facilitate the use of genetic markers of prognosis, in conjunction with clinical and histomorphologic features, as robust surrogate end points for such trials. 

Supplementary Materials

Table st1, XLS - Table st1, XLS 

Footnotes

 Supported by Fight for Sight Grant 685 (SL); Eye Tumor Research Fund, Royal Liverpool Hospital (JD); National Health Service (AT); National Commissioning Group, National Health Service (BD, MB, SC).

Footnotes

 Disclosure: S.L. Lake, None; B.E. Damato, None; J. Dopierala, None; M.M. Baudo, None; A.F.G. Taktak, None; S.E. Coupland, None