Abstract
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.
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 3.
2–4 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 2000.
18 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.
19–22 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,
19–22 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.
For DNA extraction, 20-μm-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.
Table 1. Clinical and Histological Characteristics of 32 Uveal Melanomas Tested for Genetic Heterogeneity
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 | 1.49 | 15 | − | − | 3 | − | − | L | 2×L |
| H04 | 81 | F | 1.40 | 15 | − | − | 9 | + | − | N | 3×G,1×N |
| H05 | 68 | M | 1.35 | 15 | − | + | 52 | + | − | L | 2×L |
| H06 | 56 | F | 1.33 | 16 | − | − | 8 | + | + | L | 4×L |
| H07 | 49 | F | 1.31 | 16 | − | + | 9 | − | + | L | 2×L |
| H08 | 85 | F | 1.26 | 18 | + | + | 8 | − | − | G/N | 2×G/N,2×L |
| H09 | 69 | M | 1.25 | 13 | − | − | 6 | − | − | L | 2×L |
| H10 | 60 | F | 1.23 | 14 | − | + | 2 | − | + | L | 3×G/N,2×L |
| H12 | 63 | F | 1.34 | 18 | − | − | 13 | + | + | L | 7×L,1×N |
| H14 | 56 | M | 1.14 | 12 | − | + | 2 | − | − | N | 4×N |
| H15 | 71 | M | 1.14 | 12 | − | − | 3 | − | − | L | 4×L |
| H16 | 77 | F | 1.14 | 15 | − | − | 4 | + | − | L | 3×L |
| H17 | 60 | F | 1.21 | 13 | − | − | 3 | + | − | N | 5×N |
| H18 | 67 | M | 1.21 | 17 | − | + | 2 | + | − | N | 3×N |
| H20 | 59 | M | 0.94 | 19 | − | + | 15 | + | + | U | 9×N |
| H21 | 68 | M | 0.93 | 15 | − | − | 6 | − | − | N | 4×N |
| H23 | 67 | M | 0.89 | 19 | − | + | 2 | − | − | L | 4×L |
| H24 | 63 | F | 0.88 | 17 | − | − | 2 | + | + | L | 5×L |
| H25 | 81 | F | 0.88 | 16 | − | − | 4 | − | − | U | 5×U |
| H26 | 63 | F | 0.87 | 14 | − | − | 2 | − | − | L | 3×L |
| H27 | 74 | F | 0.81 | 15 | − | + | 3 | + | + | EL | 4×EL |
| H28 | 54 | M | 0.96 | 20 | − | + | 12 | + | + | L | 8×L |
| H29 | 94 | F | 0.94 | 19 | − | + | 18 | + | − | G | 4×N/G |
| H30 | 61 | M | 0.94 | 18 | − | + | 8 | + | + | L | 4×L |
| H41 | 72 | F | 2.04 | 12 | − | − | 14 | − | − | N | 4×N |
| M08 | 88 | M | 0.36 | 19 | − | + | 2 | + | + | L | 3×L |
| M12 | 72 | M | 2.03 | 14 | + | + | 15 | + | − | L | 2×L |
| M13 | 41 | F | 1.96 | 22 | − | + | 1 | + | − | N | 2×N |
| M16 | 65 | F | 1.72 | 18 | − | + | 19 | + | − | L | 2×L |
| M17 | 63 | M | 1.79 | 18 | − | − | 2 | + | − | L | 2×L |
| M18 | 58 | M | 1.70 | 16 | − | + | 2 | − | − | N | 2×N |
| M27 | 58 | M | 1.67 | 13 | − | + | 17 | + | − | L | 2×L |
Tissue lysis and protein digestion were performed in 125 μL 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 μg/μL. DNA was extracted (DNeasy Blood and Tissue protocol; Qiagen) and was then eluted in 40 to 50 μL of AE buffer.
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.
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).
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 χ2 test was used to test for any association after dichotomization of the cohort for heterogeneity and equivocality.
Distribution of Chromosomal Changes across Uveal Melanomas Not Showing Heterogeneity at a Particular Locus
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 sections 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 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 UM
24 ; 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 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 different melanoma cell clones and not by reactive cells, although further investigation using fresh tumor material is needed to confirm this impression.
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 CRR10082.
Disclosure:
J. Dopierala, None;
B.E. Damato, None;
S.L. Lake, None;
A.F.G. Taktak, None;
S. E Coupland, None
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.