Genetic Heterogeneity in Uveal Melanoma Assessed by Multiplex Ligation-Dependent Probe Amplification

Transcription

1 Anatomy and Pathology Genetic Heterogeneity in Uveal Melanoma Assessed by Multiplex Ligation-Dependent Probe Amplification Justyna Dopierala, 1 Bertil E. Damato, 2 Sarah L. Lake, 1 Azzam F. G. Taktak, 3 and Sarah E. Coupland 1 From the 1 Department of Pathology, School of Cancer Studies, and the 3 Department of Medical Physics and Clinical Engineering, University of Liverpool, Liverpool, United Kingdom; and 2 St. Paul s Eye Hospital, Royal University Liverpool Hospital, Liverpool, United Kingdom. Supported by the Eye Tumor Research Fund (Royal Liverpool University Hospital), which provides a PhD stipend to JD; Eye Tumor Research Fund (RLBUHT) Grant CRR10416; and Fight-for-Sight UK Grant CRR Submitted for publication December 1, 2009; revised March 28 and April 18, 2010; accepted April 25, Disclosure: J. Dopierala, None; B.E. Damato, None; S.L. Lake, None; A.F.G. Taktak, None; S. E Coupland, None Corresponding author: Sarah E. Coupland, Department of Pathology, University of Liverpool, Liverpool, UK, s.e.coupland@liverpool.ac.uk. PURPOSE. To determine intratumor genetic heterogeneity in uveal melanoma (UM) by multiplex ligation dependent probe amplification (MLPA) in formalin-fixed, paraffin-embedded (FFPE) tumor tissues. METHODS. DNA was extracted from whole tumor sections and from two to nine different areas microdissected from 32 FFPE UMs. Thirty-one loci on chromosomes 1, 3, 6, and 8 were tested with MLPA for copy number changes. The tumor was considered heterogeneous at a locus if (1) the difference in dosage quotients (DQs) of any two areas was 0.2 or more, and (2) the DQs of the areas belonged to different ranges. RESULTS. Comparison of MLPA data obtained from microdissected areas of the UMs showed heterogeneity in 1 to 26 examined loci in 24 (75%) tumors, with only 25% of the tumors being homogeneous. Intratumor heterogeneity of 3p12.2, 6p21.2, and 8q11.23 was most common, occurring in 30% of the UMs. Gains of chromosome 3 were observed in four UMs, with three of these tumors showing the highest degree of heterogeneity. Copy number variation was associated with differences in tumor cell type, but not with differences in tumor pigmentation or reactive inflammation. UMs with genetic heterogeneity across multiple sample sites showed equivocal MLPA results when the whole tumor section was examined. These results suggest that different clones dilute MLPA results. CONCLUSIONS. Heterogeneity of chromosomal abnormalities of chromosomes 1, 3, 6, and 8 is present in most UMs. This heterogeneity causes equivocal MLPA results. One random tumor sample may not be representative of the whole tumor and, therefore, may be insufficient for prognostic testing. (Invest Ophthalmol Vis Sci. 2010;51: ) DOI: / iovs In almost 50% of all patients with uveal melanoma (UM), metastatic disease develops that usually involves the liver and is almost inevitably fatal. 1 Such metastatic disease occurs almost exclusively in patients with tumors that show partial or complete deletion of chromosome Tumor dimensions at the time of initial ocular treatment and mitotic count give an indication of the likely survival time in the presence of monosomy 3. 5 Epithelioid melanoma cells, closed connective tissue loops, and a high mitotic count also suggest that a biopsy result indicating disomy 3 may be erroneous. 2,6 12 We have developed online tools for performing multivariate analyses of UM, to estimate, with a reasonable degree of reliability, the survival probability of individual patients. 13 Such personalized prognostication enables reassurance of UM patients with good prognoses while indicating more intensive care for those having a high risk of metastasis. Genetic typing of UMs should also facilitate studies of systemic adjuvant therapy by excluding patients with a low risk of metastasis. We and others have been typing UMs by using a variety of methods, such as cytogenetics and gene expression profiling. 14 For several years, we relied on fluorescence in situ hybridization (FISH), but this method required larger, fresh tumor samples and tested only one centromeric locus on chromosome 3, so that partial deletions were missed. 15 In late 2006, we replaced FISH with multiplex ligation-dependent probe amplification (MLPA), 16 which simultaneously tests 31 genomic sequences on chromosomes 1, 3, 6, and 8, requiring smaller tumor samples that can be either fresh or formalin-fixed and paraffin-embedded (FFPE). 17 In 2009, we validated this method in 73 UMs from patients treated between 1998 and This evaluation showed that equivocal (borderline) MLPA results for chromosome 3 loci indicate a high risk of metastasis, suggesting that this phenomenon could occur as a result of melanoma cell heterogeneity, with disomy 3 cells diluting monosomy 3 cell clones. Several research groups have reported on histologic and genetic intratumoral heterogeneity of UM Such heterogeneity gives rise to a risk of sampling error when performing microbiopsy. In these studies of genetic heterogeneity of UM, only chromosome 3 was tested, and usually only one locus on this chromosome was assessed. We thought it would be useful to study multiple loci, not only on chromosome 3 but also on chromosomes 1, 6, and 8, which are known to develop abnormalities of prognostic significance. The goals of this study, therefore, were to gain more knowledge on intratumoral heterogeneity in UM by assessing copy number variations within different areas of UM using MLPA on FFPE tumor tissue. If we could confirm such heterogeneity, we then proceeded to attempt to determine whether there was an association between the heterogeneous loci and equivocal results of the same loci detected in the whole tumor sections (i.e., a dilution effect). We hope that our findings will facilitate the interpretation of UM microbiopsy results of these tumors. Investigative Ophthalmology & Visual Science, October 2010, Vol. 51, No Copyright Association for Research in Vision and Ophthalmology

2 IOVS, October 2010, Vol. 51, No. 10 Tumor Heterogeneity in Uveal Melanoma 4899 METHODS Patients The present study was performed on archival UM tissue originating from 38 patients who underwent enucleation at the Liverpool Ocular Oncology Centre (LOOC) in 2007 and None of these patients underwent eye-saving treatment (e.g., radiotherapy) before enucleation. Selected UMs were large enough to allow sampling of several distinct areas, each measuring 0.6 mm in diameter for genetic analysis (i.e., a basal diameter exceeding 11.9 mm and a thickness exceeding 5 mm). Informed consent was obtained, and the study was conducted in accordance with the Declaration of Helsinki and with institutional review board approval. Histomorphologic Assessment Four-micrometer sections were cut from each FFPE UM block and stained with hematoxylin and eosin (H&E) and periodic acid-schiff (PAS), with and without hematoxylin counterstain. Histopathologic examination was performed as described elsewhere 15 for assessment of cell type 2 ; presence of PAS closed connective tissue loops; degree of pigmentation; mitotic count (number of mitoses per 40 high-power fields; one HPF 40 objective); and the presence of admixed reactive cells (macrophages and/or lymphocytes). Degrees of reactive inflammation and pigmentation were graded as follows: none, 0; mild, 1; moderate, 2; and strong/numerous, 3 (Fig. 1). This assessment was performed for the whole tumor as well as for each of the microdissected areas (described later). The H&E section was also used to guide microdissection for DNA extraction. Sampling of FFPE Tissue and DNA Extraction For DNA extraction, 20- m-thick whole tumor sections were cut, and two to nine tumor areas were either microdissected using a scalpel or a 0.6-mm-diameter donor punch (manual arrayer; Beecher Instruments, Sun Prairie, WI). When possible, samples were obtained from the tumor apex and base as well as from anterior and posterior portions of the UM. Areas within or at the edge of a UM consisting purely of blood vessels, necrotic tumor, or dense macrophage/lymphocyte infiltrates were intentionally avoided in obtaining the samples. Extraocular melanoma was not sampled, as only two UMs demonstrated extraocular extension (Table 1). Our procedure resulted in a total of 187 UM samples. Nontumor controls comprised 18 FFPE normal choroid and 6 FFPE reactive tonsils. Tissue lysis and protein digestion were performed in 125 L of lysis buffer (50 mm Tris-HCl [ph 8.2], 1 mm EDTA, 100 mm NaCl, 0.5% Tween-20, 0.5% NP40, and 20 mm DTT). Proteinase K (Qiagen, Crawley, UK) was added to the final concentration of 0.8 mg/ml, and after 36 hours of incubation (24 hours at 56 C and 12 hours at 37 C), RNA was cleaved by the addition of RNase A (Sigma-Aldrich Co., Ltd., Gilingham, UK) to a final concentration of 20 g/ L. DNA was extracted (DNeasy Blood and Tissue protocol; Qiagen) and was then eluted in 40 to 50 L of AE buffer. DNA Quantification and Quality Assessment The DNA concentration and absorbance were measured with a spectrophotometer (NanoDrop; Thermo Scientific, Wilmington, DE) at 280 and 260 nm. Multiplex PCR was adapted from van Dongen et al., 23 using the RAG1, PLZF, AF4 exon 3, and AF4 exon 11 primers (Eurofins MWG Operon, London, UK), and was performed on samples with a concentration exceeding 40 ng/ L, to assess DNA quality with a thermal cycler (model TC-412; Techne, Staffordshire, UK). The reaction volume was 25 L and contained 100 ng of DNA and 1 highperformance buffer; 2 mm MgCl 2 ; 0.8 mm dntp mix; units of polymerase (ThermoStart; ABgene-Thermo Fisher Scientific); 0.5% BSA (Sigma-Aldrich Co., Ltd.); 0.1 M forward and reverse primers for RAG1, PLZF, and AF4 exon 11; and 0.2 M forward and reverse primers for AF4 exon 3. PCR products were visualized on 2% agarose gels (150 ma for 30 minutes) stained with 1 SYBR DNA gel stain (SYBR Safe; Invitrogen, Paisley, UK), with a gel-imaging system (Bio Doc-It; Ultra-Violet Products, Ltd., Cambridge, UK). MLPA Procedure and Sequencing The MLPA procedure and sequencing were performed in 32 tumors, as previously reported 14 (Salsa P027.B1 Uveal Melanoma kit; MRC-Holland, Amsterdam, The Netherlands). In brief, six nontumor control samples were used in each MLPA assay, with 200 ng of DNA being analyzed for both tumor and nontumor samples. MLPA reactions were performed on a thermal cycler (G-Storm GS1; Gene Technologies Ltd, Essex, UK) with fragment detection being performed on a genetic analyzer (ABI-3130XL and GeneMapper software; Applied Biosystems [ABI], Foster City, CA). Raw data were received as peak heights, as a measure of peak intensity for each of the 43 probes (31 test probes and 12 control probes). MLPA was performed in triplicate for each UM sample. FIGURE 1. The grading system used for degree of pigmentation and density of inflammatory cells within the UM: (a) UM with no pigmentation and no inflammatory infiltrate (grade 0 for both categories); UM with (b) mild inflammatory infiltrate and (c) mild pigmentation (grade 1); UM with (d) moderate reactive cell infiltrate and (e) moderate pigmentation (grade 2); and; UM with (f) numerous reactive lymphocytes and (g) heavy pigmentation (grade 3). MLPA Data Analysis Analysis was performed using an adapted version of a spreadsheet (Excel; Microsoft, Redmond, WA) designed by the National Genetics Reference Laboratory (NGRL), Manchester, UK ( ngrl.org.uk/manchester/), as described previously. 18 We modified this method to exclude UM control loci that seemed abnormal. Such abnormality was confirmed by manual analysis: abnormal probes (a dosage quotient [DQ] outside the range) that were not detected by the program were excluded, and those falsely excluded were retrieved. The MLPA data were considered reliable if the number of control probes within the normal range was 7. The DQ was categorized as suggested by the NGRL as a deletion (D), 0.65; equivocal deletion (E), ; normal diploid (N), ; equivocal amplification (Q), ; and amplification (A) 1.35.

3 4900 Dopierala et al. IOVS, October 2010, Vol. 51, No. 10 TABLE 1. Clinical and Histological Characteristics of 32 Uveal Melanomas Tested for Genetic Heterogeneity UM Age Sex FU (y) LBD EOM Closed Loops MC EC Cell Het Chr 3 Status* (Overall) Chr 3 Status (Microdissection) H02 79 F L 2 L H04 81 F N 3 G,1 N H05 68 M L 2 L H06 56 F L 4 L H07 49 F L 2 L H08 85 F G/N 2 G/N,2 L H09 69 M L 2 L H10 60 F L 3 G/N,2 L H12 63 F L 7 L,1 N H14 56 M N 4 N H15 71 M L 4 L H16 77 F L 3 L H17 60 F N 5 N H18 67 M N 3 N H20 59 M U 9 N H21 68 M N 4 N H23 67 M L 4 L H24 63 F L 5 L H25 81 F U 5 U H26 63 F L 3 L H27 74 F EL 4 EL H28 54 M L 8 L H29 94 F G 4 N/G H30 61 M L 4 L H41 72 F N 4 N M08 88 M L 3 L M12 72 M L 2 L M13 41 F N 2 N M16 65 F L 2 L M17 63 M L 2 L M18 58 M N 2 N M27 58 M L 2 L FU, follow-up; LBD, largest basal tumor diameter (measured in mm); EOM, extraocular melanoma growth; MC, mitotic count (number per 40/HPF); EC, epithelioid cells; Cell het, cellular heterogeneity. * Chromosome 3 status as determined with routine diagnostic MLPA analysis on a single tumor sample. Chromosome 3 status as determined with MLPA on multiple microdissected sites within the UM. The numbers indicate how many samples were microdissected in each tumor. The data highlighted in bold indicate a change in the interpretation of chromosome 3 status. L, loss; N, normal (diploid); G, gain; EL, equivocal loss; U, uninterpretable. Establishment by MLPA of the Criteria for Heterogeneity Samples taken from each UM were assessed for heterogeneity at each locus by calculating a difference between the maximum and minimum values of DQs for a locus. This method identified the tumor samples that showed the greatest intratumoral difference at each locus. Therefore, calculation of all the differences of all sample combinations was unnecessary. DQ heterogeneity was significant if the difference between two loci was 0.2. This value was the difference between the upper and lower limits of the ranges for deletion, equivocal deletion, equivocal amplification, and amplification. The difference between the upper and lower limit of the diploid range was 0.3. Consequently, a second condition for the definition of a heterogeneous locus was that the DQ difference of 0.2 or more had to belong to different DQ ranges. Correlation between Genetic and Histologic Heterogeneity Chromosomal and histologic heterogeneity findings were coded 0 for no heterogeneity and 1 for heterogeneity. Fisher s exact test was used to correlate chromosomal heterogeneity with histologic heterogeneity with respect to cell type, closed connective tissue loops, degree of pigmentation, and reactive inflammation (SPSS ver. 16.0; SPSS, Chicago, IL). Correlation of DQ Heterogeneity with MLPA Result from Whole Tumor Sections To determine whether equivocal MLPA results obtained from whole UM sections could be explained by intratumoral variation in copy number, we scored the DQs obtained from the whole tumor sections as follows: 1, for equivocal loss or gain; and 0, for unequivocal outcome (i.e., normal diploid and amplification or deletion). For the microdissected tumor area, we used the binary coding specified earlier. The 2 test was used to test for any association after dichotomization of the cohort for heterogeneity and equivocality. RESULTS Patients and Samples A total of 160 samples from 32 UMs were studied after 27 samples from 6 tumors were excluded because of inadequate DNA quality or poor MLPA reproducibility. The 32 UMs included in the study were from 15 men and 17 women who had a mean age of 67 years (range, 41 94). The largest basal tumor diameter (LBD) averaged 16.0 mm (range, ). Two UMs showed extraocular growth. Histologic examination showed epithelioid cells in 19 tumors, PAS closed connective tissue loops in 18 tumors, and a high mitotic count (i.e., exceeding 5/40 HPF) in 16 UMs (Table 1). Two tumors showed partial loss of chromosome 3, and

4 IOVS, October 2010, Vol. 51, No. 10 Tumor Heterogeneity in Uveal Melanoma 4901 LOCI TESTED BY MLPA 60% 50% UVEAL MELANOMAS EXAMINED FIGURE 2. Distribution of heterogeneity of 31 loci across chromosomes 1, 3, 6, and 8, tested by MLPA (P027.B1). Purple: heterogeneous loci; blue: loss; light blue: equivocal loss; green: normal; orange bars: equivocal gain; red bars: gain; x: unreliable results; FR, frequency of the heterogeneous loci; INF, number of heterogeneous loci per tumor/across tumors; HET LOCI, number of heterogeneous loci per tumor. 21 showed complete monosomy 3 in at least one microdissected area. The median postoperative follow-up time was 1.2 years (range, ). Five patients (M8, H12, H23, H24, and H28) developed metastatic disease by the time of study closure, and one of these patients (M8) died during the study period. Intratumoral Chromosomal Heterogeneity According to the criteria for the study, genetic heterogeneity was detected in 24 (75%) of the 32 UMs in between 1 and 26 of the tested 31 loci across chromosomes 1, 3, 6, and 8 (Fig. 2). The percentage of loci that were heterogeneous varied among the UMs (Fig. 3). The most heterogeneous loci, in decreasing frequency, were CDKN1A (6p21.2) in 11 (35%) of 31 (95% confidence interval [CI], 19% 52%); RP1 (8q11.23) in 11 (34%) of 32 (95% CI, 18% 51%); and ROBO1 (3p12.2) in 10 (31%) of 32 (95% CI, 15% 47%) of tumors. (Fig. 4). Percentage of tumors 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% -5% -10% 0% 1-10% 11-20% 21-30% 31-40% 41-60% 61-70% 71-80% 81-90% Percentage of heterogenous loci per tumor FIGURE 3. The percentage of tumors according to the percentage of heterogeneous loci per tumor. INCREASING FREQUENCY OF HETEROGENOUS LOCI Percent 40% 30% 20% 10% 0% -10% Gene Locus FIGURE 4. Frequency of the heterogeneous loci on chromosomes 1, 3, 6, and 8. Eight (25%) of the evaluated 32 melanomas were homogeneous for all 31 loci tested by MLPA (Fig. 2). Three of these were of disomy 3 type. They had a mean LBD of 16.3 mm, with a median mitotic count of 3 of 40 HPFs, and epithelioid cells in one tumor only. Five UMs showed monosomy 3, and these had a median LBD of 15.2 mm, a median mitotic count of 9 of 40 HPFs, and epithelioid cells in two tumors. There were no tumors with partial deletion of chromosome 3 in this group. There was no significant difference in the mean LBD (t-test P 0.533), mean mitotic count (t-test P 0.54), prevalence of epithelioid cellularity (Fisher s exact test: P 0.219), and the prevalence of monosomy 3 (Fisher s exact test: P 1.00), when homogeneous and heterogeneous UMs were compared. Heterogeneity of Chromosome 3 Fifteen (47%) of the examined 32 UMs showed intratumoral variation of at least one locus on chromosome 3 (Fig. 2). Of these, 13 also demonstrated heterogeneity for at least one locus on chromosomes 1, 6, and/or 8. The most common heterogeneous loci on chromosome 3 were ROBO1 (3p12.2), VHL (3p25.3), and FANCD2 (3p25.3), which occurred in 31% (95%, CI 15% 47%), 28% (95% CI, 0.16% 44%), and 22% (95% CI, 14% 36%) of all tumors, respectively (Fig. 2). In 11 (73%) of the 15 UMs with heterogeneity of chromosome 3, the variation involved one to six scattered loci (i.e., they were not clustered in similar chromosomal regions). In the remaining four UMs, 11 of the 13 tested loci on chromosome 3 showed heterogeneity (Fig. 2). When UMs heterogeneous for chromosome 3 loci were compared with those tumors with homogeneous chromosome 3 loci, there was no significant difference between the mean LBD, the mean mitotic rate, the presence of epithelioid cells, and the presence of partial or total chromosome 3 loss for any of the chromosome 3 loci (Table 2). Of note, three of the four UMs with chromosome 3 heterogeneity of 11 loci showed gain of one or both arms of chromosome 3 in at least one part of the tumor (Fig. 5). Of the remaining 17 UMs that were homogenous for all 13 loci on chromosome 3, 5 were of the disomy 3 type, whereas 12 showed monosomy 3 and none showed partial chromosome 3 loss. Correlation between Genetic and Histomorphologic Heterogeneity Intratumoral histomorphologic heterogeneity for melanoma cell type was detected in 10 of 32 UMs. Such cellular heterogeneity correlated with intratumoral heterogeneity of VHL (3p25.3), CDKN1A, and RP1 (Table 3), as well as heterogeneity of connective tissue loops (P 0.001). Heterogeneity of the

5 4902 Dopierala et al. IOVS, October 2010, Vol. 51, No. 10 TABLE 2. A Comparison of UMs with Heterogeneous Chromosome 3 Loci with Those Tumors with Homogeneous Chromosome 3 Loci FANCD2 (i) (3p25.3) FANCD2 (ii) (3p25.3) VHL (3p25.3) MLH1 (3p22.1) CTNNB1 (3p22) SEMA3B (3p21.3) FHIT(i) (3p14.2) FHIT(ii) (3p14.2) ROBO1 (3p12.2) CPO (3q12) RHO (3q21.3) MME (3q25.1) OPA1 (3q29) LBD Mitotic rate Epithelioid cells Monosomy 3 (MLPA) P-values were obtained from the Fisher s exact test and the t-test. There was no significant difference between the mean LBD, the mean mitotic rate, the presence of epithelioid cells, and the presence of partial or total chromosome 3 loss for any of the chromosome 3 loci. latter also correlated with an intratumoral copy number difference in CDKN1A. There was no correlation between genetic heterogeneity and variation in reactive inflammation or tumor pigmentation (Table 3). Distribution of Chromosomal Changes across Uveal Melanomas Not Showing Heterogeneity at a Particular Locus The most common abnormalities in our cohort of UMs were deletions of chromosome 3 and gains of the long arm of chromosome 8. Loss of loci located on chromosome 3 was present in between 25% (for VHL, 3p25.3) and 58% (for FHIT(ii), 3p14.2) of the UMs. Gains of loci on 8q were present in between 41% (for RP1, 8q11.23) and 66% (DDEF1, 8q24.2) of the tumors. Depending on which locus was considered, amplifications on 6p occurred in 19% to 26% of the UMs. Deletions of loci were seen in 13% to 28% and 19% of the tumors on 1p and 6q, respectively. Duplications on 1p, 3, and 6q and deletions on 6p were rare (Fig. 6). Correlation between Genetic Heterogeneity and Equivocal MLPA Results Equivocal results obtained from the whole tumor sections correlated with the underlying intratumoral copy number differences ( 2 test, P 0.001). DISCUSSION The main finding of our investigation using MLPA on FFPE material is that 75% of the analyzed UMs showed intratumoral heterogeneity of chromosomes 1, 3, 6, or 8. Almost 50% of tumors showed intratumoral heterogeneity of at least one locus of chromosome 3. The loci showing heterogeneity most commonly were ROBO1, CDKN1A, and RP1. Most of the UMs showing heterogeneity of chromosome 3 also showed heterogeneity of loci on the other examined chromosomes. Intratumoral heterogeneity of some loci correlated with variation in cell type but not with reactive inflammation or degree of pigmentation. Genetic heterogeneity correlated with equivocal MLPA results obtained from whole tumor sections. To our knowledge, no studies have been undertaken to investigate the intratumoral heterogeneity of multiple gene loci in UM. In this cohort of UMs, 24 (75%) of 32 tumors showed intratumoral heterogeneity of 1 to 26 loci across chromosomes 1, 3, 6, and 8. Only one quarter of the studied UMs were homogeneous for all 31 loci tested by MLPA. Small tumors (LBD 11.9 mm; thickness 5 mm) were excluded from the study because it was not possible to sample multiple sites from such cases. We plan, however, to examine increasingly smaller UMs and to compare the MLPA results with those in the present study. The degree of genetic heterogeneity varied between the examined UMs (Fig. 3). Slight variation involving up to six scattered loci was seen in seven UMs, whereas five tumors showed heterogeneity of 40% of loci on at least three of the four chromosomes examined. The heterogeneity of single isolated loci should be interpreted with caution, as these differences could be an artifact, as a result of DNA degradation after formalin fixation. However, copy number differences involving several loci across a large chromosomal area are associated with true heterogeneity between these regions. The most heterogeneous loci detected using MLPA in decreasing frequency were CDKN1A (35% of UMs), RP1 (34%), and ROBO1 (31%). The least heterogeneous locus was MYCBP, showing a copy number variation in only three tumors. Because of the short follow-up period and the small number of patients with metastatic disease, the clinical relevance of these findings has yet to be determined. With respect to chromosome 3 only, approximately half of the examined UMs showed heterogeneity of at least one locus, with 11 of the tumors demonstrating copy number variation in up to 6 loci. When compared with baseline MLPA data used for classifying the UM for chromosome 3 status, this heterogeneity in the loci between areas did not result in a change in interpretation (Table 1). Four of these UMs, however, showed marked heterogeneity involving 11 of the 13 loci examined on chromosome 3. The intratumoral heterogeneity in these four UMs was enough to result in contradictory interpretation of the chromosomal status in different parts of the same tumor (Table 1). For example, one microdissected area of the UMs indicated monosomy 3, and another area showed amplification of several loci on chromosome 3 (Figs. 5F J). Other groups have reported chromosome 3 heterogeneity in 6.6% 20 to 14% 21 of tumors; however, they applied only one chromosome 3 probe, using FISH or chromogenic (C)ISH, and so the prevalence of this phenomenon is probably underreported. The advantage of MLPA over these cytogenetic methods is that it allows for the simultaneous examination of 13 chromosome 3 loci. Some researchers have suggested that chromosome 3 abnormalities are more common at the tumor base 22 and are associated with melanoma cell type 19 and presence of PAS connective tissue loops. 21 In the present study, we demonstrated a strong correlation between genetic heterogeneity of some chromosomal loci and differences in cell type, as well as the presence or absence of PAS connective tissue loops. Because the copy number variation affected numerous different loci on chromosome 3 and the number of UMs examined in this cohort was relatively small, the statistical analysis for a potential association between the geographic location of a microdissected sample (i.e., apex versus base or anterior versus posterior) and copy number variation was not possible. An interesting finding was that genetic heterogeneity of chromosomes 1, 3, 6, and 8 did not correlate with intratumoral variation in reactive inflammation. This result suggests that the copy number variation observed using MLPA was most likely to be caused by the presence of melanoma cell clones differing in chromosome 3 copy number. This impression is supported by our finding that equivocal MLPA data from whole tumor sec-

6 Tumor Heterogeneity in Uveal Melanoma IOVS, October 2010, Vol. 51, No. 10 A B C 4903 D ii I iii area (i) E area (ii) area (iii) 1.8 GAIN CHROMOSOME 1p CHROMOSOME 3 CHROMOSOME 6 CHROMOSOME LOSS NORMAL area (i) F area (ii) G area (iii) H I ii I area (ii) 2.5 GAIN area (i) J 2.0 CHROMOSOME 1p CHROMOSOME 3 CHROMOSOME 6 CHROMOSOME 8 NORMAL 1.5 LOSS FIGURE 5. MLPA results for three different areas of UM H04, and for two different areas of UM H08. Images and MLPA data for similar areas within these two tumors are not depicted. (A E) UM H04. Area (ii) showed two copies of 3; normal 1, and 6; and loss of 8p and gains on 8q. Areas (i) and (iii) showed gains of multiple loci on both arms of 3 and differences in the status of 1p, 6p, and 8. Area (iv) (not shown) was similar to (i). There were no significant differences in histomorphology in the areas examined. (F J) UM H08. Area (i) showed monosomy 3 and loss of chromosome 1. Area (iii) was similar (not shown). Area (ii) showed a partial gain of 3p25.3-3p22 and two copies of 3q, a partial loss of 1p, and gains of 6p and 8q. Area (iv) (not shown) was similar to (ii). Histologically, there are differences only in the degree of pigmentation of the spindle cells tions were strongly associated with copy number variation in the microdissected areas (P 0.001). However, to support the hypothesis that admixed reactive cells in UMs play a minor role in producing heterogeneous MLPA data, we are currently investigating macrophage- and lymphocyte-rich UMs by using fluorescence-activated cell sorting methods to separate neoplastic from nonneoplastic cell populations. Also of particular interest was that four of the UMs analyzed demonstrated gains on chromosome 3, which consisted of partial or total gain of one or both arms of chromosome 3. To our knowledge, this finding has not been reported. Although MLPA cannot detect polyploidy, we believe that the gains observed on chromosome 3 were aneuploid aberrations, since area (i) area (ii) loss of 1p was also present in three of these UMs, and loss of 6q and 8p was found in one other UMs. Three of these four UMs showed marked intratumor heterogeneity of chromosome 3. Monosomy 3 is considered an early event in UM24; therefore, UMs showing monosomy 3 in at least one area of the tumor and two or more copies of chromosome 3 in other regions of the same tumor are particularly interesting, suggesting that additional copies of chromosome 3 were gained during tumor progression. In summary, we have demonstrated that (1) genetic heterogeneity of chromosomes 1, 3, 6, and 8 is present in most FFPE UMs, so that a single random sample may not be representative of the whole tumor, with the result that false

7 4904 Dopierala et al. IOVS, October 2010, Vol. 51, No. 10 TABLE 3. Probabilities Obtained by Fisher s Exact Test Cell Type Closed Connective Tissue Loops Pigmentation Admixed Reactive Cells Loops Pigmentation Reactive cells MFN2 (1p36.22) NBL1 (1p36.13) PTAFR (1p34) MYCBP (1p34) MUTYH (1p33) RPE65 (1p31) NOTCH2 (1p11.2) FANCD2 (i) (3p25.3) FANCD2 (ii) (3p25.3) VHL (3p25.3) MLH1 (3p22.1) CTNNB1 (3p22) SEMA3B (3p21.3) FHIT (i) (3p14.2) FHIT (ii) (3p14.2) ROBO1 (3p12.2) CPO (3q12) RHO (3q21.3) MME (3q25.1) OPA1 (3q29) PEC1 (6p25.2) DCDC2 (6p22.2) CDKN1A (6p21.2) RUNX2 (6p21) CTGF (6q23.1) IGF2R (6q26) NRG1 (8p12) RP1 (8q11.23) MYC (i) (8q24.12) MYC (ii) (8q24.12) DDEF1 (8q24.2) The null hypothesis was that there is no association between intratumoral heterogeneity of histomorphologic variables (cell type, closed connective tissue loops, pigmentation of tumor cells, and admixed reactive cells) and heterogeneity of 31 loci tested by MLPA. Bold values indicate strong evidence against the null hypothesis (i.e., an association between a variable in a column with a variable in a row). reassurance regarding the survival prognosis is provided; (2) although genetic heterogeneity for chromosome 3 occurred in 47% of the UMs, it involved only a few scattered loci in most tumors and did not change the interpretation of chromosome 3 status between the microdissected areas; (3) UMs containing clones with gains in chromosome 3 occur and tend to show a high degree of heterogeneity; and (4) equivocal MLPA results are likely to be caused by genetically Percent 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% HET GAIN EQ GAIN NORMAL EQ LOSS LOSS Gene locus FIGURE 6. Distribution of chromosomal abnormalities. Het, heterogeneous loci; gain, gain in copy number; eq gain, equivocal gain; normal, normal copy number; eq loss, equivocal loss; loss, loss in copy number.

8 IOVS, October 2010, Vol. 51, No. 10 Tumor Heterogeneity in Uveal Melanoma 4905 different melanoma cell clones and not by reactive cells, although further investigation using fresh tumor material is needed to confirm this impression. Acknowledgments The authors thank Helen Kalirai for fruitful discussion of the results and Roger Mountford (Molecular Genetics Department, Liverpool Women s Hospital) for the use of the sequencing facilities. References 1. Kujala E, Makitie T, Kivela T. Very long-term prognosis of patients with malignant uveal melanoma. Invest Ophthalmol Vis Sci. 2003; 44: Prescher G, Bornfeld N, Hirche H, Horsthemke B, Jockel KH, Becher R. Prognostic implications of monosomy 3 in uveal melanoma. Lancet. 1996;347: Kilic E, van Gils W, Lodder E, et al. Clinical and cytogenetic analyses in uveal melanoma. Invest Ophthalmol Vis Sci. 2006;47: Scholes AG, Damato BE, Nunn J, Hiscott P, Grierson I, Field JK. Monosomy 3 in uveal melanoma: correlation with clinical and histologic predictors of survival. Invest Ophthalmol Vis Sci. 2003; 44: Damato B, Coupland SE. A reappraisal of the significance of largest basal diameter of posterior uveal melanoma. Eye. 2009;23: White VA, Chambers JD, Courtright PD, Chang WY, Horsman DE. Correlation of cytogenetic abnormalities with the outcome of patients with uveal melanoma. Cancer. 1998;83: McLean IW, Foster WD, Zimmerman LE, Gamel JW. Modifications of Callender s classification of uveal melanoma at the Armed Forces Institute of Pathology. Am J Ophthalmol. 1983;96: Coupland SE, Campbell I, Damato B. Routes of extraocular extension of uveal melanoma: risk factors and influence on survival probability. Ophthalmology. 2008;115: Folberg R, Pe er J, Gruman LM, et al. The morphologic characteristics of tumor blood vessels as a marker of tumor progression in primary human uveal melanoma: a matched case-control study. Hum Pathol. 1992;23: Kivela T, Makitie T, Al-Jamal RT, Toivonen P. Microvascular loops and networks in uveal melanoma. Can J Ophthalmol. 2004;39: Aalto Y, Eriksson L, Seregard S, Larsson O, Knuutila S. Concomitant loss of chromosome 3 and whole arm losses and gains of chromosome 1, 6, or 8 in metastasizing primary uveal melanoma. Invest Ophthalmol Vis Sci. 2001;42: Parrella P, Sidransky D, Merbs SL. Allelotype of posterior uveal melanoma: implications for a bifurcated tumor progression pathway. Cancer Res. 1999;59: Damato B, Eleuteri A, Fisher AC, Coupland SE, Taktak AF. Artificial neural networks estimating survival probability after treatment of choroidal melanoma. Ophthalmology. 2008;115: Onken MD, Worley LA, Person E, Char DH, Bowcock AM, Harbour JW. Loss of heterozygosity of chromosome 3 detected with single nucleotide polymorphisms is superior to monosomy 3 for predicting metastasis in uveal melanoma. Clin Cancer Res. 2007;13: Damato B, Duke C, Coupland SE, et al. Cytogenetics of uveal melanoma: a 7-year clinical experience. Ophthalmology. 2007; 114: Damato B, Coupland SE. Translating uveal melanoma cytogenetics into clinical care. Arch Ophthalmol. 2009;127: Schouten JP, McElgunn CJ, Waaijer R, Zwijnenburg D, Diepvens F, Pals G. Relative quantification of 40 nucleic acid sequences by multiplex ligation-dependent probe amplification. Nucleic Acids Res. 2002;30:e Damato B, Dopierala J, Klaasen A, van Dijk M, Sibbring J, Coupland S. Multiplex ligation-dependent probe amplification of uveal melanoma: correlation with metastatic death. Invest Ophthalmol Vis Sci. 2009;50: Sandinha T, Farquharson M, McKay I, Roberts F. Correlation of heterogeneity for chromosome 3 copy number with cell type in choroidal melanoma of mixed-cell type. Invest Ophthalmol Vis Sci. 2006;47: Mensink HW, Vaarwater J, Kilic E, et al. Chromosome 3 intratumor heterogeneity in uveal melanoma. Invest Ophthalmol Vis Sci. 2009;50: Maat W, Jordanova ES, van Zelderen-Bhola SL, et al. The heterogeneous distribution of monosomy 3 in uveal melanomas: implications for prognostication based on fine-needle aspiration biopsies. Arch Pathol Lab Med. 2007;131: Schoenfield L, Pettay J, Tubbs RR, Singh AD. Variation of monosomy 3 status within uveal melanoma. Arch Pathol Lab Med. 2009;133: van Dongen JJ, Langerak AW, Bruggemann M, et al. Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT Leukemia. 2003;17: Prescher G, Bornfeld N, Friedrichs W, Seeber S, Becher R. Cytogenetics of twelve cases of uveal melanoma and patterns of nonrandom anomalies and isochromosome formation. Cancer Genet Cytogenet. 1995;80:40 46.