SEARCH

SEARCH BY CITATION

Keywords:

  • cytogenetics

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

Blokx W A M, van Dijk M C R F & Ruiter D J (2010) Histopathology56, 121–132 Molecular cytogenetics of cutaneous melanocytic lesions – diagnostic, prognostic and therapeutic aspects

This review intends to update current knowledge regarding molecular cytogenetics in melanocytic tumours with a focus on cutaneous melanocytic lesions. Advantages and limitations of diverse, already established methods, such as (fluorescence) in situ hybridization and mutation analysis, to detect these cytogenetic alterations in melanocytic tumours are described. In addition, the potential value of more novel techniques such as multiplex ligation-dependent probe amplification is pointed out. This review demonstrates that at present cytogenetics has mainly increased our understanding of the pathogenesis of melanocytic tumours, with an important role for activation of the mitogen-activated protein kinase (MAPK) signalling pathway in the initiation of melanocytic tumours. Mutations in BRAF (in common naevocellular naevi), NRAS (congenital naevi), HRAS (Spitz naevi) and GNAQ (blue naevi) can all cause MAPK activation. All these mutations seem early events in the development of melanocytic tumours, but by themselves are insufficient to cause progression towards melanoma. Additional molecular alterations are implicated in progression towards melanoma, with different genetic alterations in melanomas at different sites and with varying levels of sun exposure. This genetic heterogeneity in distinct types of naevi and melanomas can be used for the development of molecular tests for diagnostic purposes. However, at the moment only few molecular tests have become of diagnostic value and are performed in daily routine practice. This is caused by lack of large prospective studies on the diagnostic value of molecular tests including follow-up, and by the low prevalence of certain molecular alterations. For the future we foresee an increasing role for cytogenetics in the treatment of melanoma patients with the increasing availability of targeted therapy. Potential targets for metastatic melanoma include genes involved in the MAPK pathway, such as BRAF and RAS. More recently, KIT has emerged as a potential target in melanoma patients. These targeted treatments all need careful evaluation, but might be a promising adjunct for treatment of metastatic melanoma patients, in which other therapies have not brought important survival advantages yet.


Abbreviations:
AI

allelic imbalance

CGH

comparative genomic hybridization

(F) ISH

(fluorescence) in situ hybridization

FISH

fluorescence in situ hybridization

GIST

gastrointestinal stromal tumour

HRMA

high-resolution melting analysis

ISH

in situ hybridization

LOH

loss of heterozygosity

MAPK

mitogen-activated protein kinase

MLPA

multiplex ligation-dependent probe amplification

PCR

polymerase chain reaction

RT

reverse transcriptase

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

Melanocytes, pigment-producing cells in the skin, eye and mucosal surfaces, can give rise to a large spectrum of proliferative lesions varying from naevocellular naevi to melanomas. Both are frequently encountered in daily pathology practice. The histological diagnosis of certain melanocytic tumours is, even among experts, notoriously difficult and hampered by lack of objective criteria and consensus on criteria.1 Since melanomas can be highly aggressive neoplasms, proper diagnosis is of great importance in determining optimal treatment. Integration of genetic data in histopathological diagnosis can result in better and more reproducible classification.

This review intends to update current knowledge and insights regarding cytogenetics in the pathogenesis of melanocytic tumours, and how this at the moment influences or in future might influence diagnosis, prognosis and treatment of melanocytic lesions. Focus will be on cutaneous melanocytic lesions. Cytogenetics includes all studies of the structure and function of chromosomes; however, our accent will be on molecular cytogenetics such as fluorescence in situ hybridization (FISH), comparative genomic hybridization (CGH) and mutation analysis.

Detection methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

Several techniques can be used to detect cytogenetic alterations in melanocytic lesions for research and diagnostic purposes. We will summarize the most important methods to detect DNA gains and deletions and DNA mutations, with emphasis on advantages and limitations of these techniques.

Detection of loss of heterozygosity or allelic imbalance

Studies such as allelic imbalance (AI) analysis [also called detection of loss of heterozygosity (LOH)] have shown important differences in DNA copy number in naevi and melanoma.2,3 A polymerase chain reaction (PCR)-based technique is used that can detect copy number alterations of microsatellites (short tandem repeats of 1–4 nucleotides). A major obstacle for a more general application of AI analysis in routine diagnostic practice concerning melanocytic lesions is the need for DNA from non-tumour cells from the same patient. Normal DNA is often present in limited or insufficient amounts in perilesional tissue in a received specimen. Furthermore, only limited numbers of chromosomal locations can be analysed in one experiment and LOH analysis is prone to PCR artefacts, especially if limited amounts of DNA are used. Therefore, in our opinion LOH has no direct diagnostic application, especially considering that there are other techniques such as in situ hybridization (ISH) and multiplex ligation-dependent probe amplification (MLPA) that are either more reliable or can analyse more loci, respectively.3

CGH and array CGH

CGH is based on the comparison of genomic tumour and normal reference DNA (not necessarily from the same patient), each labelled with different fluorochromes and subsequently hybridized on normal metaphase chromosomes or an array of small spots of genomic DNA. During this co-hybridization, tumour and reference DNA compete for the same target on the metaphase chromosomes. The relative amount of tumour and reference DNA bound to a given chromosomal region is dependent on the relative abundance of these sequences in both DNA samples. Therefore, intensity difference in the fluorescent hybridization pattern of tumour and normal DNA can be interpreted as copy number differences between the two genomes.4

For diagnostic application it is an advantage that (array) CGH can be performed on paraffin-embedded tissue and that a large amount of targets can be analysed simultaneously without prior knowledge of the expected alterations. However, it is a laborious technique that requires relatively large amounts of good-quality DNA, highly specialized personnel and equipment, limiting its applicability to highly qualified molecular core-labs only. Therefore, CGH is mainly used in a research setting to reveal genetic differences between lesions.5 Possible diagnostic applications of CGH could be atypical Spitz tumours, and proliferative nodules in congenital naevocellular naevi.5 However, for diagnostic purposes other techniques are favoured for practical and financial reasons.

(Fluorescence) in situ hybridization

(Fluorescence) in situ hybridization [(F) ISH] has been used in the study of tumours since the 1980s to visualize copy number changes as well as translocations. A small DNA fragment of known origin (probe) is (fluorescently) labelled and hybridized to metaphase chromosomes, nuclear suspensions or interphase nuclei in tissue slides. The number of spots per nucleus is indicative of the copy number of the chromosome locus analysed. Sequences of whole genomes, centromeres, telomeres or specific regions on genes can be used as probes. A translocation can be detected by a spot that splits into two separate spots, or by red- and green-labelled probes that hybridize on each of the translocated genes and result in a yellow spot. (F) ISH can be applied on paraffin-embedded tissue slides and therefore provides a link between cytogenetics, histology and, if necessary, immunohistochemistry. In melanomas, for example, cells with chromosomal aberrations that are present in the primary tumour can be detected in the deep margin of a cutaneous melanoma or in adjacent normal-appearing skin.6

Disadvantages of the use of (F) ISH on tissue sections is the presence of incomplete nuclei in the section causing artificial loss of chromosome parts, which makes the technique difficult to evaluate: about 100–200 nuclei need to be analysed for a reliable copy number of the probe used.7 Only a limited number of locations can be tested in one experiment even with the application of two- or three-colour FISH. Because the representative area of a naevus is frequently too small, there are insufficient numbers of nuclei for reliable evaluation.8 To overcome this problem, (F) ISH can be performed on whole nuclei, nuclear suspensions or touch preparations. However, it remains difficult to obtain sufficient nuclei from small melanocytic lesions, and in this way the link with the histological information is lost.

Multiplex ligation-dependent probe amplification

MLPA is a method that was first published in 2002 and is based on the annealing of up to 45 probes each consisting of two oligonucleotides that hybridize adjacently on a specific chromosome region.9 Besides a target-specific sequence, both oligonucleotides also contain a universal PCR primer, whereas one oligonucleotide additionally contains a so-called stuffer sequence that has a probe-specific length. After ligation of the hybridized adjacent oligonucleotides the probe is amplified using the universal PCR primers. As each probe has a unique length due to the stuffer sequence, electrophoresis can be used to separate and quantify the amount of PCR product that is indicative of the DNA copy number (Figure 1). We have shown that MLPA is a sensitive technique, but false-positive results may occur.10 The applicability of MLPA in routine diagnostics needs to be confirmed, but this method seems to be a good alternative for AI analysis and ISH because no normal patient DNA is needed and up to 45 probes can be tested in a single experiment.

image

Figure 1.  Figure demonstrating the basic principle of the multiplex ligation-dependent probe amplification (MLPA) technique. The MLPA reaction can be divided into five major steps: (i) DNA denaturation and hybridization of MLPA probes; (ii) ligation reaction; (iii) polymerase chain reaction (PCR) reaction; (iv) separation of amplification products by electrophoresis; and (v) data analysis. During the first step, the DNA is denatured and incubated overnight with a mixture of MLPA probes. MLPA probes consist of two separate oligonucleotides, each containing one of the PCR primer sequences. The two probe oligonucleotides hybridize to immediately adjacent target sequences. Only when the two probe oligonucleotides are both hybridized to their adjacent targets can they be ligated during the ligation reaction. Because only ligated probes will be exponentially amplified during the subsequent PCR reaction, the number of probe ligation products is a measure for the number of target sequences in the sample. The amplification products are separated using capillary electrophoresis. Probe oligonucleotides that are not ligated contain only one primer sequence. As a consequence, they cannot be amplified exponentially and will not generate a signal. The removal of unbound probes is therefore unnecessary in MLPA and makes the MLPA method easy to perform. Adapted from Schouten et al.12 (http://www.mlpa.com).

Download figure to PowerPoint

Mutation analysis

The recent discovery of several specific mutations in melanocytic lesions has rapidly increased the understanding of genetic changes that may underlie the development of benign and malignant melanocytic lesions. A major advantage of this technique is that it is sensitive and needs only a limited amount of tumour DNA. Mutations that alter the protein due to amino acid substitution must be discriminated from mutations that result in the same amino acid (silent mutations or polymorphisms). The conventional nucleotide sequence in an amplified DNA fragment of interest is analysed after a sequence reaction by electrophoresis. The disadvantage of this screening method is that for reliable mutation detection a tumour percentage of 70% tumour cells is needed.

High-resolution melting analysis (HRMA) and mutation-specific MLPA are two recently developed methods that overcome this problem. HRMA is a relatively new method based on the dissociative behaviour of DNA when heated in such a high resolution that it can detect single base pair sequence variations.11 It is a rapid method to detect hot-spot mutations and can identify at least 5% of mutated cells in a background of wild-type DNA.12 Mutation-specific MLPA can combine copy number detection and the presence of hot-spot mutations in a single experiment.13 Mutation-specific MLPA, like HRMA, can also detect a specific mutation in a sample with a considerably lower percentage of tumour cells when compared with conventional PCR and sequencing (only 10% tumour cells are needed with MLPA, compared with 70% for conventional PCR and sequencing, manuscript in preparation). Both HRMA and MLPA are rapid, sensitive and cost-limited approaches that can be used in a diagnostic setting.

Cytogenetics in the pathogenesis of melanocytic lesions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

The development of a normal melanocyte into a melanoma is characterized by certain histological, and so far only partially unravelled genetic alterations (Figure 2). At a molecular level, the mitogen-activated protein kinase (MAPK) signalling pathway and PTEN/AKT pathway are both involved in the growth control of melanocytic cells.14 Activation of these pathways via somatic mutations in the RAS and RAF genes is thought to be one of the first steps in the development of common naevi.15

image

Figure 2.  Schematic events in the development of melanocytic tumours. RAF, RAS and GNAQ mutations lead to the development of naevi out ofmelanocytes. These mutations occur exclusive of each other. HRAS mutations are confined to Spitz naevi, and GNAQ mutations are found in blue naevi. The next step is thought to be the development of a premalignant lesion from a pre-existent naevus or de novo. This stage is caused by progressive accumulation of genetic alterations with additional mutations in genes like CDKN2A or PTEN mutations and/or gains or losses of chromosome parts.

Download figure to PowerPoint

Data thus far point towards at least a triple concept in the development of melanocytic lesions. A point mutation (V600E) in the BRAF gene (chromosome 7q34) is thought to be an early event in the genesis of naevocellular naevi and cutaneous melanomas derived from intermittent sun-exposed skin, and this mutation is thought to be induced by this intermittent type of sun exposure. Melanomas derived from other sites (chronic sun-exposed or sun-protected sites) less often contain these BRAF mutations. In addition, certain benign naevi such as blue naevi, Spitz naevi and congenital naevocellular naevi do not or rarely contain BRAF mutations, and instead in part contain other mutations, e.g. in the NRAS or HRAS genes.16 Very recently somatic mutations of GNAQ, occurring exclusively in codon 209 in the RAS-like domain, were found in uveal melanomas and blue naevi, both lacking BRAF or NRAS mutations.17 Like mutations in BRAF, these RAS mutations and GNAQ mutations can cause MAPK activation and form an alternative route for melanocytic neoplasia.17 All these mutations seem to be an early event in the development of melanocytic tumours and by themselves are insufficient to cause progression towards melanoma.

In further development from melanocytic naevus towards dysplastic naevus additional molecular alterations, like loss of wild-type INK4A and loss of PTEN, are implicated. In progression towards melanoma additional genetic alterations such as loss of chromosome 6q, del11q, and gain of 7q have been described.18 An extensive overview of all pathways involved in melanoma genesis is beyond the scope of this review. Excellent updates are given by Miller et al.15, Fecher et al.19, Dahl et al.14 and Carlson et al.20

Cytogenetic alterations in naevocellular naevi and melanomas

Current knowledge on chromosomal gains and losses in diverse subtypes of naevi, and reported frequencies of mutated genes in these lesions, are both summarized in Table 1. This table illustrates that both different subtypes of benign naevi as well as subtypes of melanoma are cytogenetically heterogeneous, and these distinct genetic changes in diverse melanocytic lesions could be of use in a diagnostic setting (see below). However, most molecular studies thus far have not been designed for diagnostic purposes, offering only potential diagnostic benefit.

Table 1.   Summary of reported chromosomal aberrations in benign and dysplastic naevi in literature, and frequencies of mutations in genes implicated in initiating development of melanocytic tumours
Lesion typeChromosomal gains or losses (CGH)BRAF mutationv-raf murine sarcoma viral oncogene homologue B1, 7q34NRAS mutationNeuroblastoma RAS viral oncogene homologue, 1p13.2HRAS mutationHarvey rat sarcoma viral oncogene homologue, 11p15.5GNAQ mutation
  1. ND, not determined.

Common acquired naevusNDUp to 87.5%16Not extensively studied, NRAS mutation present in two naevi associated with SSM with NRAS mutation490ND
Dysplastic or atypical naevusND52–62%50,51Up to 71% (only seven cases tested in patients with germ-line CDKN2A mutations)52053ND
Blue naevusNone26,540–12%55,56056ND83%17
Spitz naevus+11p in 20%028,55028Up to 29%28ND
Congenital naevusNone26.Numerical chromosomal aberration in atypical nodule (86%)2730% medium-sized lesions88% small size1664%160ND

Molecular techniques such as array-based CGH have revealed that there are distinct pathways involved in the development of melanoma,19,21 with different genetic alterations in melanomas at different sites and with varying levels of sun exposure (Table 2). Interestingly, mucosal melanomas were found to contain an amplification of the 4q12 chromosome. In humans, the KIT gene maps to 4q12-13, in the neighbourhood of the gene encoding PDGFRA (platelet-derived growth factor receptor-alpha) receptor kinase. KIT and PDGRFA are both class III receptor tyrosine kinases.22 Activating KIT mutations has been identified as a key factor in the pathogenesis of gastrointestinal stromal tumours (GISTs), and this discovery had a great impact on the diagnosis and treatment of these GISTs.23 More recently, activating KIT mutations have also been implicated in the pathogenesis of certain melanoma subtypes,22,24,25 and first single case reports of successful KIT targeting in melanoma are emerging.

Table 2.   Summary of the findings in four subtypes of melanoma as identified by Curtin et al.21 based on genetic findings
Melanoma subtypeBRAF mutationNRAS mutationChromosomal aberrations
  1. NCSD, non-chronic sun damage; CSD, chronic sun damage.

Skin NCSD59%22%Increased copy no: 6p, 7,8q, 17q, 20q
Reduced copy no: 9p, 10, 21q
Skin CSD11%15%Increased copy no: 6p, 7,8q, 17q, 20q
Reduced copy no: 9p, 10, 21q
Mucosal11%5%Increased copy no: 1q, 6p, 7, 8q, 11q13, 17q, 20q
Reduced copy no: 3q, 4q, 6q, 8p, 9p, 10,11p, 11q, 21q
Amplification: 1q13, 4q12, 12q14
Acral23%10%Increased copy no: 6p, 7,8q, 17q, 20q
Reduced copy no: 6q, 9p, 10,11q, 21q
Amplification: 5p15, 5p13, 11q13, 12q14

Clinical applications of cytogenetics in melanocytic tumours

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

In diagnosis

Copy number changes for discriminating between benign and malignant melanocytic lesions

CGH has demonstrated that most melanomas (96%) have multiple genetic aberrations (mean number of aberrations 7.5), while these are absent in common naevi see (Table 1). This could potentially allow differentiation of more than 96% of melanomas from naevi.26 Only a specific subtype of benign naevus, so-called Spitz naevus, contains aberrations, which are virtually only single 11p gains.18,26 Also in so-called proliferative nodules in congenital naevi in children, CGH could be helpful in assessing the diagnosis. True melanomas arising in congenital naevi show CGH patterns comparable to typical melanoma, while atypical nodules contain numerical aberrations of entire chromosomes, which are seen only in a minority (5%) of melanomas.27

Mutation analysis to differentiate between Spitz naevus and spitzoid melanoma

A minor part of the Spitz naevi is unique among melanocytic tumours for having a single gain of 11p (26%) and/or HRAS mutations, the latter occurring in up to 29% of cases.26,28 Thus far no HRAS mutations have been reported in spitzoid melanomas. Demonstration of either a HRAS mutation or a single 11p gain therefore seems indicative of benign behaviour. In difficult to diagnose spitzoid lesions, so-called Spitzoid tumours of uncertain malignant potential, assessment of one of these cytogenetic aberrations could be of help in making a more accurate diagnosis, as illustrated in Figure 3. For HRAS mutation analysis, PCR and sequencing is at present the most appropriate technique. For determination of DNA gains or losses, FISH and CGH have proved diagnostically applicable.26,29 In future, other techniques such as MLPA and HRMA are of potential value.

image

Figure 3.  A case of a Spitzoid melanocytic tumour in a 23-year-old woman located on the right arm. The overview histological picture shows a wedge-shaped rather symmetrical tumour. There is a small epidermal component. The detail shows clearly spitzoid melanocytes with large epithelioid cells; several mitoses are seen (arrows), up to 6/1 mm2, also near the base. The relatively high mitotic number at this age with deep mitoses was considered unusual for ordinary Spitz naevus, and based on histology alone we preferred a diagnosis of Spitzoid tumour of uncertain malignant potential (STUMP), probably low risk). We performed HRAS mutation analysis, which showed a HRAS codon 61 mutation: c.182A[RIGHTWARDS ARROW]T (p. GLN61Leu), supporting a final diagnosis of Spitz naevus. HRAS mutations have been described in Spitz naevi, but thus far never in melanoma. The presented case illustrates the potential value of HRAS mutation analysis in case of STUMP lesions.

Download figure to PowerPoint

Cytogenetics to distinguish a second new primary melanoma from a cutaneous melanoma metastasis

The histological differential diagnosis between a second primary cutaneous melanoma and cutaneous melanoma metastasis in a patient with a previous history of melanoma can be very difficult, especially when metastatic melanoma manifests tropism to the epidermis.30 The differentiation, however, is of great prognostic and clinical importance, since a skin metastasis implies systemic spread, which in contrast to a second primary tumour is incurable. In sporadic melanoma somatic CDKN2A mutations have been described.31,32 In a case report we have described useful application of CDKN2A mutation analysis for discriminating between a cutaneous melanoma metastasis and a new primary melanoma of the skin, in a patient with a history of cutaneous melanoma.33

The exact prevalence of CDKN2A mutations in (metastasizing) melanoma is not known. CDKN2A point mutations are detected in approximately 10% of sporadic melanomas, although considerable discrepancies are reported between different studies with mutations ranging between 0 and 25%. One study analysed CDKN2A mutations in 12 primary sporadic melanomas and nine corresponding metastases.31 In four of the primary tumours and in none of the metastases mutational inactivation of the p16INK4A protein was detected. Future studies on the prevalence of somatic CDKN2A mutations in primary and metastatic melanoma are necessary to determine the value and possible broader application of testing this mutation in histological problematic cases in a diagnostic setting.

In addition to CDKN2A testing, in some cases RAF or RAS mutation analysis could be performed, possibly together with KIT mutation analysis in case of a clinical problem in differentiating visceral metastasis of cutaneous melanoma from a new primary non-cutaneous melanoma (as we described in a case in 2008).34 Cutaneous melanomas frequently harbour either BRAF or NRAS mutations and these are uncommon in non-cutaneous types. In addition, these NRAS and BRAF mutations are preserved in the progression of cutaneous melanomas.35 This is an infrequent problem, since as a rule of thumb a melanoma metastasis is far more likely than a new primary non-cutaneous melanoma in a patient with a previous history of cutaneous melanoma,36 even in case of a long interval or a peculiar metastatic pattern.34 Similarly, this BRAF and RAS mutation analysis might be of use in case of visceral melanoma localization in patients with a blank history for cutaneous melanoma to support a diagnosis of primary mucosal melanoma.

Cytogenetics in familial melanoma

Melanoma may occur as part of a family cancer syndrome such as familial retinoblastoma and Li Fraumeni syndrome. However, the majority of families with a history of melanoma will predominantly have only melanoma.37 In most families with more than four melanoma cases the genetic base is formed by a germ-line mutation at the CDKN2A locus of chromosome 9 (45%). The CDKN2A (INK4A-ARF) locus, which maps to chromosome 9p21, encodes for two proteins p16INK4A (exons 1α, 2 and 3) and p14ARF (exon 1β and 2). Both proteins are tumour suppressors and therefore mutations in the CDKN2A locus can contribute to tumorigenesis.38 Most families with at last three melanoma patients studied within the GenoMEL project have p16 mutations (41%). Mutations in p14 are less frequent. Small numbers of patients have hereditary mutations in the CDK4 gene.37,39 GenoMEL is an international research consortium with special focus on identifying melanoma susceptibility genes (http://www.genomel.org).

In some healthcare systems CDKN2A gene testing for familial melanoma is already practised. CDKN2A testing at present if performed is probably best restricted to certain clinical settings, but is important to detect new kindreds with hereditary melanoma predisposition and to offer preventive measures to relatives of identified germ-line mutation carriers.39

Cytogenetics in the differential diagnosis of clear cell sarcoma (melanoma of the soft parts) and primary cutaneous or metastatic melanoma

Clear cell sarcoma, in contrast to melanoma, has been found to contain a balanced t(12;22)(q13;q13) aberration in 50–75% of cases.40 In diagnostically challenging cases, reverse transcriptase (RT)-PCR and FISH to detect this translocation represent valuable tools for the clinicopathological differentiation between these two entities, which may demonstrate significant overlap at the microscopic level, with an important impact on patient management and prognosis.

In prognostication

RT-PCR of sentinel lymph node biopsies

RT-PCR is said to detect one tumour cell out of 106–107 non-tumour cells, while immunohistochemistry can detect one melanoma cell in a background of 104–105 non-tumour cells.18

While immunohistochemistry improves the sensitivity of metastasis detection with 10–45% compared with histology alone, the use of RT-PCR technique on frozen tissue increases the detection of suspected occult metastases by up to 70% (RT-PCR for melanoma-related marker gene expression, like tyrosinase and Mart-1).20,41 However, pathologists should not only identify the presence of melanoma metastasis within a sentinel lymph node, but also measure the size and location of the metastasis within the lymph node.42 At present the role of non-histopathological methods to investigate the sentinal lymph node, like RT-PCR, remains unclear. These methods have the problem that a histopathological correlate of the suspected tumour cells is lacking and misinterpretation of nodal naevus cells as metastasis may occur. This may be overcome by using more specific methods that detect only melanoma cells and not benign melanocytes,41 but even then the diameter and location of the deposits can no longer be assessed.

Determination of prognosis in an individual melanoma patient

Gene expression profiling in a research setting has demonstrated that 254 out of 11 043 genes were differentially expressed between patients with disease-free survival below or greater than 4 years.43 Some of these genes also on the protein level were associated with survival. Although to date several protein and genetic markers have been found to be related to survival, thus far they have been applied only in a research setting, because none proved solid enough to guide melanoma patient management.20 At present, tests for melanoma tumour markers or for circulating melanoma cells are not performed in routine staging or monitoring of patients because of insufficient sensitivity and specificity to predict recurrence or progression of melanoma accurately. For further reading, one is referred to Carlson et al.18 It can be expected that novel, more accurate techniques will become important in prognostication of melanoma patients, while early detection of systemic disease makes treatment possible before bulky tumour has developed.

In treatment

Targeted therapy for melanoma is a promising adjunct for treatment, especially while other therapies have not yet brought important survival advantages.

In melanoma, several key signalling pathways are deregulated. Of these the (RAF/RAS/) MAPK pathway involved in control of cell growth signals, cell survival and invasion seems to play the most important role in (early) melanoma development.

Important potential targets for metastatic melanoma therefore include BRAF, RAS and MAPK. In addition, targets like VEGF, PTEN, CDK4 and more recently KIT have emerged (for more extensive review see Carlson et al.18).

In this review, we will highlight only the potential role of KIT targeting in melanoma, which seems very promising.

KIT-targeted treatment

Recent published studies and findings with respect to KIT aberrations in human melanoma tissue (excluding results in melanoma cell lines) are summarized in Table 3. The most recent and largest series thus far describes 189 melanomas tested for KIT mutations in exons 11, 13 and 17.44 This study demonstrates that KIT mutations are most common in acral and mucosal subtypes, respectively 23% and 15.6%. PDGFRA mutations (also frequently mutated in GIST) were not tested. The study included 45 mucosal melanomas (36 head and neck, nine anorectal/vaginal/vulvar); which types contained the mutations was not further specified. Almost all KIT mutations found were to be of an imatinib-sensitive type. Imatinib (Gleevec; Novartis Pharmaceuticals, Camberley, UK) inhibits the enzymatic activity of tyrosinase kinases including KIT and PDGFRA. Imatinib in GISTs has become an effective and important therapeutic option.

Table 3.   Overview of recent published studies on frequency of KIT and PDGFRA mutations within tissue sections of human melanomas
Author, yearMelanoma subtypeGenes and exons testedFrequency KIT mutationFrequency PDGRFA mutation
  1. ND, not determined; NCSD, non-chronic sun damage; CSD, chronic sun damage.

Pache, 2003Ten uveal melanomasKIT exon 2, 8, 9,11, 13, 170ND
All-Ericsson, 2004Fifteen uveal melanomas and three uveal melanoma cell linesKIT exon 110ND
Curtin, 200638 mucosal nos28 acra18 skin NCSD18 skin CSDKIT exon 11,13, 17, 18PDGRFA 10, 12, 14, 188/38 (21%) mucosal3/28 (11%) acral0/18 (0%) skin NCSD3/18 (16%) skin CSD5 mutations exon 13, 5 mutations exon 11, 4 otherOnly cases with K642E mut in exon 13 also BRAF mutation0, only cases with KIT amplification (n = 7) tested
Hodi, 2007One rectal melanomaKIT exon 11,13, 17100% (1/1)exon 11 seven codon duplicationND
Antonescu, 200720 anal melanomasKIT exon 11,13,17PDGRFA exon 12,1815% (3/20)all KIT mutations were a heterozygous exon 11 L576P substitution0
Rivera, 200818 oral melanomasKIT exon 11, 1322% (4/18)2 mutations in exon 11 and 2 in exon 13 (K642E)ND
Lutzky, 2008One metastatic anal melanomaKIT exon 11, 13, 17, 18K642E mutation exon 13ND
Quintas-Cardama, 2008One stage IV anal mucosal melanomaKIT exon 1100% (1/1)exon 11 V560DND
Beadling, 2008189 melanomas13 acral, 45 mucosal, 13 conjunctival, 58 cutaneous, 60 choroidalKIT exon 11, 13, 17. Only in subset exon 8 and 923% acral (3/13)15.6% mucosal (7/45)7.7% conjunctival(1/13)1.7% cutaneous (1/58)0% choroidalND

At present there are only two cases reported on successful imatinib treatment with major objective responses in two patients with metastatic mucosal anorectal melanoma. One case of complete response of an anal melanoma to sorafenib has been reported. In all cases the melanoma harboured a KIT mutation.45–47 There was one abstract (presented as a poster in the European Organization for Research and Treatment of Cancer melanoma congress, Scheveningen, the Netherlands, 2–4 October 2008) on a single Chinese case of acral melanoma with KIT mutation (exon 11 L576P) by Si et al., treated with imatinib 400 mg/day, with 80% shrinkage of lesions.

In a multicentre Phase II trial in metastatic melanoma patients (n = 26, 23 cutaneous, two ocular and one mucosal melanoma) treated with high-dose imatinib, no responses were reported; however, these patients were not tested for the presence of a KIT or PDGRFA mutation.48 Only KIT and PDGRFA immunohistochemistry was performed.

This illustrates the importance of proper selection of patients prior to treatment with imatinib, including mutation analysis of the KIT and PDGFRA genes. From studies on GISTs it is known that KIT immunohistochemistry cannot reliably predict cases with KIT or PDGRFA mutations that are responsive to imatinib.22

In our opinion, more extensive research on larger series of diverse types of metastatic melanoma patients on the frequency of KIT and possibly PDGFRA mutations and effect of KIT-targeted therapies has to be done, before this laborious mutation analysis can become a more standard procedure in melanoma patient care.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

Cytogenetics could become of importance in diagnosis and treatment of melanocytic tumours. However, studies published thus far have been mostly retrospective and descriptive. This has increased our understanding of the pathogenesis of melanocytic tumours, and might lead to clues for future successful therapeutic intervention.

In diagnosis, large prospective studies on the diagnostic value of molecular tests including follow-up to establish the value in daily practice are needed. As the prevalence of certain molecular alterations is low, most probably a multimarker approach will be needed.

An important future role for cytogenetics in the treatment of melanoma patients is emerging. It is to be hoped that targeted treatment will bring the long expected and hoped for survival increase in metastatic melanoma patients, for whom at present no curative treatment is available. New techniques with higher sensitivity and specificity are promising. Good correlation with histopathology, clinical data and follow-up remain important, however. Only by integrating all these data can we achieve a future, more reproducible classification of melanocytic tumours that includes more accurate prognostic features and can make tailor-made therapy possible.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References

This review, and part of the research referred to in this review, was supported by the Dutch Cancer Society (grant no. KUN 2006-3700). We thank Mrs. A. Klaasen for technical assistance.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Detection methods
  5. Cytogenetics in the pathogenesis of melanocytic lesions
  6. Clinical applications of cytogenetics in melanocytic tumours
  7. Conclusions
  8. Acknowledgements
  9. References