PDF hosted at the Radboud Repository of the Radboud University Nijmegen

Transcription

1 PDF hosted at the Radboud Repository of the Radboud University Nijmegen The following full text is a publisher's version. For additional information about this publication click this link. Please be advised that this information was generated on and may be subject to change.

2 Genetic alterations in rare subtypes of melanocytic tumors; towards improved diagnosis and therapy Ilse van Engen- van Grunsven

3 Publication of this thesis was financially supported by Radboud Universiteit Nijmegen and Servier Nederland Farma b.v.. ISBN Design/lay-out Promotie In Zicht, Arnhem Print Ipskamp Printing, Enschede A.C.H. van Engen-van Grunsven, 2017 All rights are reserved. No part of this book may be reproduced, distributed, stored in a retrieval system, or transmitted in any form or by any means, without prior written permission of the author.

4 Genetic alterations in rare subtypes of melanocytic tumors; towards improved diagnosis and therapy Proefschrift ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen op gezag van de rector magnificus prof. dr. J.H.J.M. van Krieken, volgens besluit van het college van decanen in het openbaar te verdedigen op vrijdag 24 maart 2017 om uur precies door Adriana Carolina Hendrika van Engen-van Grunsven geboren op 22 maart 1977 te Rosmalen

5 Promotoren Prof. dr. P. Wesseling Copromotoren Dr. W.A.M. Blokx Dr. P.J.T.A. Groenen Manuscriptcommissie Prof. dr. ir. J.J.M. van der Hoeven Prof. dr. M.A.A.P. Willemsen Prof. dr. J.J. van den Oord (UZ Leuven, België)

6 CONTENTS Part I 7 Chapter 1 Aim and general outline 9 Chapter 2 Update on Molecular Pathology of Cutaneous Melanocytic Lesions: What is New in Diagnosis and Molecular Testing for Treatment? Front Med (Lausanne) 2014;1:39: Part II 43 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features. Am J Surg Pathol 2010;34: Whole-genome copy-number analysis identifies new leads for chromosomal aberrations involved in the oncogenesis and metastastic behavior of uveal melanomas. Melanoma Res 2015;25: Mutations in G protein encoding genes and chromosomal alterations in primary leptomeningeal melanocytic neoplasms. Pathol Oncol Res 2015;21: Improved discrimination of melanotic schwannoma from melanocytic lesions by combined morphological and GNAQ mutational analysis. Acta Neuropathol 2010;120: Genomic analysis of leptomeningeal and skin melanocytic tumors in neurocutaneous melanosis; potential clues for diagnosis and prognosis. Revised manuscript published in Acta Neuropathol Feb;133(2): NRAS mutations are more prevalent than KIT mutations in melanoma of the female urogenital tract - a study of 24 cases from the Netherlands. Gynecol Oncol 2014;134: Part III 155 Chapter 9 Summarizing discussion and future perspectives 157 Part IV 183 Nederlandse samenvatting Dankwoord Curriculum vitae List of publications

7

8 Part I

9

10 1 Aim and general outline

11

12 Aim and general outline Melanin, a pigment that colors the skin, hair, and iris, is produced by specialized cells called melanocytes. These melanocytes are derived from the neural crest (1). During embryonic development they colonize the skin, the uveal tract of the eye and, to a lesser extent, multiple other tissues throughout the body such as the meninges and the anogenital tract (2). Melanocytes at these different sites can give rise to different types of neoplasms, which can be benign (melanocytic nevi, melanocytomas) or malignant (melanomas) (3-7). Cutaneous melanoma is the most common malignant melanocytic tumor (~95%), while mucosal melanoma, uveal melanoma (UM), and primary melanoma of the central nervous system account for approximately 1.3%, 3% and 1% of all melanoma cases respectively (8-10). 1 While the molecular underpinnings of the most frequent cutaneous melanocytic neoplasms (melanomas, melanocytic nevi) have been extensively studied, elucidation of the molecular pathogenesis of the more rare melanocytic tumors is lagging behind. The aim of the research described in this thesis was to reduce this knowledge gap, in order to facilitate the diagnosis and to discover potential therapeutic targets in these less frequent melanocytic neoplasms. The review on the molecular pathology in melanocytic tumors we present in chapter 2 serves as an introduction to the rest of the thesis. The first part of this review focuses on the three subgroups of spitzoid melanocytic tumors and their specific phenotypical and genetic characteristics: the HRAS-mutated spitzoid tumors, the melanocytic BAP1-associated intradermal tumors (MBAITs), and the spitzoid lesions with genomic rearrangements involving kinases. Subsequently, the molecular background of congenital melanocytic nevi and neurocutaneous melanosis is described. Melanomas from different origin have been shown to contain different mutation types and mutation frequencies, which can be of aid in the differential diagnosis of a melanoma or a melanoma metastasis. The last part of the review focuses on targeted therapy for metastatic melanoma. The HRAS-mutated spitzoid tumors are described in chapter 3. This type of spitzoid tumors is characterized by a typical phenotype. They are predominantly located in the dermis, have a wedged shape, relatively low cellularity, show desmoplasia and have an infiltrating base. Recognition of this subgroup is important since they have a favorable prognosis and can be easily overdiagnosed as a melanoma. The results of whole genome copy-number analysis of twenty-five UMs as well as mutation and MLPA analysis of these tumors are described in chapter 4. Our findings point to both genetic changes that may play a role in the initiation of UM, as well as to chromosomal areas that may indicate metastatic behavior. 11

13 Chapter 1 Mutation analysis of twenty primary leptomeningeal melanocytic neoplasms (LMNs) showed that these harbor oncogenic mutations in GNAQ and GNA11 as is presented in chapter 5. This knowledge can be of help in the differential diagnosis between a primary LMN and a metastasis to the central nervous system of a cutaneous melanoma, which has implications for treatment and prognosis. Based on the presence of GNAQ and GNA11 mutations in both LMNs as well as in uveal melanomas (UMs), we also hypothesized that they might share gross chromosomal aberrations. The LMNs were tested for copy number aberrations frequently present in UMs by multiplex ligation-dependent probe amplification (MLPA) and OncoScan analysis. We demonstrate how GNAQ mutation analysis can also be of aid in the histological challenging differential diagnosis of melanotic schwannoma with LMNs in chapter 6. GNAQ mutations were highly specific for LMNs compared to melanotic schwannoma. Correct diagnosis of the latter is important to identify those patients who need further evaluation for possible underlying Carney Complex (an autosomal dominant syndrome associated with spotty pigmentation of the skin, endocrinopathy, and endocrine and non endocrine tumors (for example myxomas of the heart)). Chapter 7 describes whole genome copy-number variation analysis as well as mutation analysis in five central nervous system melanocytic tumors and three congenital melanocytic nevi of five patients with the rare syndrome of neurocutaneous melanosis. The findings might have implications for diagnosis, prognosis, and therapy. In chapter 8, mutation analysis in a series of twenty-four female urogenital tract melanomas is presented. Our finding that NRAS mutations were more prevalent than KIT mutations has potential implications for therapy in the advanced setting. A summary, general conclusions, and future perspectives related to the work described in this thesis are provided in chapter 9. 12

14 Aim and general outline REFERENCES 1. Le Douarin NMK. The neural crest. 2nd ed. Cambridge: Cambridge University Press; Mort RL, Jackson IJ, Patton EE. The melanocyte lineage in development and disease. Development. 2015;142(4): Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, et al. Distinct sets of genetic alterations in melanoma. The New England journal of medicine. 2005;353(20): Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT in distinct subtypes of melanoma. Journal of clinical oncology: official journal of the American Society of Clinical Oncology. 2006;24(26): Bauer J, Curtin JA, Pinkel D, Bastian BC. Congenital melanocytic nevi frequently harbor NRAS mutations but no BRAF mutations. The Journal of investigative dermatology. 2007;127(1): Bastian BC. The molecular pathology of melanoma: an integrated taxonomy of melanocytic neoplasia. Annual review of pathology. 2014;9: Murali R, Wiesner T, Rosenblum MK, Bastian BC. GNAQ and GNA11 mutations in melanocytomas of the central nervous system. Acta neuropathologica. 2012;123(3): Chang AE, Karnell LH, Menck HR. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer. 1998;83(8): Seddon JM, Gragoudas ES, Glynn RJ, Egan KM, Albert DM, Blitzer PH. Host factors, UV radiation, and risk of uveal melanoma. A case-control study. Archives of ophthalmology. 1990;108(9): Seetharamu N, Ott PA, Pavlick AC. Mucosal melanomas: a case-based review of the literature. The oncologist. 2010;15(7):

15

16 2 Update on molecular pathology of cutaneous melanocytic lesions; what is new in diagnosis and molecular testing for treatment? Adriana C.H. van Engen-van Grunsven*, Heidi Kusters-Vandevelde^, Patricia J.T.A. Groenen*, Willeke A.M.Blokx* * Department of Pathology, Radboud university medical center, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands; ^ Department of Pathology, Canisius Wilhelmina Hospital, P.O. Box 9015, 6500 GS Nijmegen, The Netherlands Front Med (Lausanne) 2014;1:39:1-13.

17 Chapter 2 ABSTRACT In this article we give an update on recent findings regarding molecular pathology in cutaneous melanocytic tumours. The focus lies on use of genetics in the diagnosis of distinct subtypes of spitzoid tumours that are often characterized by specific phenotypic-genotypic alterations that can frequently be recognized by adequate histological examination. Typical illustrating cases are given in order to increase recognition of these lesions in daily dermatopathology practice. New molecular findings in the pathogenesis of congenital melanocytic tumours and neurocutaneous melanosis are reviewed. In addition use of mutation analysis in the differential diagnosis of melanoma metastasis is discussed. Finally application of mutation analysis in targeted therapy in advanced melanoma with advantages of new techniques such as next generation sequencing is described. 16

18 Update on molecular pathology of cutaneous melanocytic lesions INTRODUCTION During the past 20 years, there has been a rapid development of molecular techniques increasing the possibilities of genetic testing in all kinds of tumors including melanocytic tumors. This has led to a rapid gain in our knowledge on the development of melanocytic tumors, which can help in diagnosis, prognosis, and treatment of melanocytic tumors. In 2010, we wrote a review on the molecular pathology of these tumors (1). Since developments in the field are fast, we will address in this paper relevant new findings from the recent years within the field of molecular pathology in melanocytic lesions. The first part has a focus on new findings with respect to diagnosis and pathogenesis. The focus will be on new molecular findings in spitzoid tumors, in congenital melanocytic nevi (CMN) and neurocutaneous melanocytosis, and the use of molecular tests in the (differential) diagnosis of melanoma. The second part will will be devoted to application of molecular pathology in the treatment of melanoma, and will briefly address new technological developments such as next generation sequencing (NGS) techniques in this setting. 2 PART 1. MOLECULAR PATHOLOGY IN THE DIAGNOSIS AND PATHOGENESIS OF MELANOCYTIC TUMOURS Early events in the development of melanocytic tumours are often hotspot mutations in genes involved in the MAPK pathway, which is one of the most important pathways involved in melanocytic tumor development (Figure 1). Important oncogenes in this pathway are BRAF (7q34), NRAS (1p13), HRAS (11p15), GNAQ (9p21), GNA11 (19p13), and KIT (4q12) (2-5). Mutations in these genes are mostly mutually exclusive, by themselves do not cause malignant progression, stay present with malignant progression, and activate the MAPK pathway. Different subtypes of benign and malignant melanocytic tumours are characterized by different mutations in these genes of the MAPK pathway. In common nevi for instance, BRAF and NRAS mutations are present in % (6, 7) and 20%, respectively. In large congenital nevi upto 80% NRAS mutations are reported (7, 8). In blue nevi, mainly GNAQ (83%) and GNA11 (7%) mutations are found (9), and in Spitz nevi HRAS mutations are reported in 20-29% (7, 10). Especially in Spitz tumors several new data indicate that these tumors are genetically more diverse than was previously thought. We will discuss these new findings below in part 1, together with new insights in the pathogenesis of CMN and the rare disease of neurocutaneous melanocytosis. We will also address the role of molecular pathology in the differential diagnosis of (metastatic) melanoma. The distinct mutations in different melanoma types will be discussed later in part 2 (see also Table 1). 17

19 Chapter 2 GNAQ GNA11 RAS KIT Tyrosinase Kinase inhibitor BRAFi BRAF P13K/AKT MEKi MEK ERK PROLIFERATION Figure 1 This figure depicts two important pathways involved in the development of melanocytic tumours and melanoma: the MAPK-pathway and the AKT/PI3K-pathway. Activation of both routes leads to proliferation. NRAS plays an important role in both pathways. Mutations in GNAQ, GNA11, and BRAF all lead to activation of the MAPK-pathway only, while mutations in KIT and NRAS can activate both pathways. Different inhibitors can be applied in targeted therapy of advanced melanoma patients, and their points of action are indicated also. This figure shows that MEK-inhibitors can be effective in case of several different gene mutations in the MAPKpathway, because they exert their effect in the distal part of the pathway What is new in spitzoid melanocytic tumors? At present roughly three subgroups of spitzoid melanocytic tumors can be identified based on distinct genetic alterations. The first one is the group of the HRAS-mutated spitzoid tumors (11). The second group is the one of the BAP1-mutated spitzoid lesions (12, 13), and the third group consists of spitzoid tumors with kinase fusion (14). The first two groups seem to be characterized mostly by a typical phenotype that can be recognized or at least suspected upon histological evaluation. Most of the HRAS-mutated spitzoid tumors are typically wedge shaped, dermal-based lesions, with an infiltrative margin, consisting of mostly spindle shaped cells, and showing marked desmoplasia (11, 15). This group is relevant to discriminate because of the favorable prognosis and to prevent melanoma overdiagnosis. 18

20 Update on molecular pathology of cutaneous melanocytic lesions Table 1 Overview of frequencies of gene mutations in different melanoma subtypes derived from different locations GNA11 19p13 GNAQ 9p21 KIT 4q12 NRAS 1p13.2 BRAF 7q34 localisation primary melanoma melanoma from CSDS/LMM 8% 15% 28% 1.4% 0 melanoma from NCSD skin 60% 22% 0%-very low 0 0 ALM 22% 10% 23-36% 0 0 mucosal melanoma 3-11% 5-24% 16-39% 0 0 uvea melanoma 0 % 0% 0% 45-50% 32% melanoma from the CNS 0% 0-low in adults. 0% 30% (adults) 30% (adults) Frequently mutated in melanoma in context of NCM in children. (pre-clinical) MEK inhibitors (pre-clinical) MEK inhibitors Imatinib, Nilotinib, Sunitinib, Dasatinib sensitive to treatment with BRAF inhibitors MEK inhibitors. Resistant to BRAFi In different studies there is some variation in reported frequencies (5, 67, 68). Abbreviations: CSDS= chronic sun damaged skin; LLM= lentigo malignant melanoma; ALM= acrolentiginous melanoma; CNS= central nervous system. 2 19

21 Chapter 2 Several studies have reported the presence of HRAS mutations in spitzoid tumors with benign behavior, and absence in clear-cut spitzoid melanomas (10, 15, 16). There is only one recent paper mentioning the occurrence of HRAS mutations in upto 10% (2/20 cases examined) of primary cutaneous melanomas (17). In this paper no histology of the lesions is shown or described, therefore whether these lesions were spitzoid or not remains unclear, and no follow-up data of the patients are included to confirm the proposed diagnosis of melanoma by the authors. Furthermore, this paper also gives mutation frequencies of BRAF (25%) and NRAS (10%) that are quite different from most studies in the melanoma field. In 2010, we described a series of 24 HRAS mutated spitzoid melanocytic tumors (11). In 7/24 (29%) of these lesions the initial diagnosis or important differential diagnosis had been melanoma based upon histological examination alone. These were mainly cases in adults that displayed rather frequent or deep mitotic activity. In five cases, more than 2 mitoses/1mm2 or deep mitoses were present. In this series with a mean and median follow-up of 10.5 years, no recurrences or metastases occurred. An example of a HRAS-mutated spitzoid lesion is depicted in Figure 2. The second group, which we preferentially call MBAITs (melanocytic BAP1- associated intradermal tumors, but are also called Wiesner tumor or BAPoma) are often polypoid, dermal based, consisting of large epitheloid spitzoid-looking cells, that can have a small common nevus component at the margin (especially in BAP1-germline mutated lesions), and in one third of cases there are prominent tumour-infiltrating lymphocytes (TILs). These BAP1-mutated melanocytic lesions were first described by Wiesner et al.. They described two families, one in Australia and one in Germany, in which a total of sixteen individuals were affected by atypical cutaneous melanocytic tumours, in association with cutaneous and uveal melanomas (12). The affected family members were found to have a BAP1 germline mutation. Subsequently, it was found that these spitzoid looking MBAITs besides a BAP1 mutation also contained a BRAF mutation. Later, these lesions were also described in a sporadic setting in so-called atypical spitzoid tumors (ASTs), without having an underlying BAP1 germline mutation (13, 18). In 2012, Wiesner et al. described a series of 32 ASTs (18). Nine cases (28%) showed BAP 1 protein expression loss while BRAF protein expression was present. In 8/9 cases (89%) a BRAF V600E mutation was found, and in 5/9 cases (55%) a somatic BAP1 mutation was present. No HRAS mutations were found. Histology was comparable to the BAP1-germline mutation associated MBAITs, demonstrating a dermal based lesion, plump epithelioid cells, giant cells, with in 1/3 cases prominent TILs, and absence of prominent fibrosis. These cases were not described to contain a small nevoid component in contrast to the germline BAP1-associated cases. 20

22 Update on molecular pathology of cutaneous melanocytic lesions A 2 C HRAS exon 3 B Figure 2 HRAS-mutated spitzoid tumour in a female patient of 47 years of age. The lesion was located on the right lower arm and measured 0.8 cm in diameter. The lesion was diagnosed elsewhere as STUMP (spitzoid tumor of uncertain malignant potential). A, B: The lesion was symmetrical, wedge shaped, mainly dermal located, with epithelioid spitzoid looking cells in an eosinophilic hyalinised stroma ( A HE 25x, B HE 400x ). There were 2 mitoses present per 1mm 2. Deep mitoses were absent. C: Sequence plot of the mutation HRAS c.182a>t(p.(gln61leu)) with 44% of mutant alleles. The genomic sequence of the gene investigated is marked in green bars and the protein sequence information is on top of the gene sequence. The combined (forward and reverse) gene sequence information is highlighted in light green and the expected protein sequence is placed on top. The forward sequence information is indicated as light blue bars and reverse sequence information as purple bars. The square box shows the number and percentage of the different nucleotides at the variant position; the red vertical bar indicates the HRAS hotspot mutation position. The sequence plots are generated in SeqNext (from JSI medical systems GmbH). Note; the HRAS RefSeq is NM_ Yeh et al. recently described genomic loss, determined with array CGH, of >1Mbp of chromosome 3 in a region containing the BAP1 locus in a series of 29 cases out of 436 ambiguous melanocytic tumours (6.7%) (19): 22 cases showed partial loss of chromosome 3, while 7 cases demonstrated monosomy of chromosome 3. In 11 21

23 Chapter 2 cases BAP1 mutation analysis was performed with in 10 cases a loss of function mutation of BAP1, and in the remaining single case with wild-type BAP1, BAP1 protein expression was lost on the immunohistochemical level. In the cases with loss of 1 copy of BAP1, the BAP1 protein expression was always lost. So immunohistochemistry for the BAP1 protein seems to correlate well with the genetic findings. Reported follow-up in their series was good, although short (median 17 months), with no recurrences, and in one patient a negative sentinel lymph node. Morphology of 17 lesions with biallelic loss of BAP1 that looked spitzoid was as described previously by others, with TILs being present in 50% of cases, and 31% having a small nevoid component at the margin. In these cases, a junctional component was present composed of the common nevus component. In one case without a common nevus component, a junctional spitzoid component was present. In 12/17 cases a BRAF V600E mutation was present, 4 cases were wildtype BRAF, and 1 case showed an NRAS Q61R mutation. The latter is rarely reported at present, but we have also encountered such a case ourselves recently (paper submitted). In our patient no underlying BAP1 mutation was found. The other reported NRAS-mutated case (19) in literature was not tested for a germline mutation in BAP1. Most of the MBAIT cases reported thusfar do not seem to be tested for the presence of NRAS mutations, so the number of NRAS mutated cases may be larger, and the genetic make-up of these BAP1 associated melanocytic lesions could be broader than currently thought. The fact that the combined MBAIT lesions show only BAP1 loss in the epithelioid component suggests that they probably develop from a common nevus (that is mostly BRAF- and seldom NRAS-mutated) (7). MBAITs probably have a low risk of developing into melanoma, but at present, data about behaviour are insufficient to draw definite conclusions. In the two largest series thusfar by Pouryazdanparast et al. (20) and Yeh et al. (19), reporting 28 and 29 cases respectively, follow-up was favourable without recurrences. Follow-up was relatively short with a mean of 21 months and median of 17 months respectively. Pouryazdanparast performed FISH (using probes targeting chromosome 6p25 (RREB1), chromosome 6q23 (MYB), chromosome 11q13 (CCND1), and the centromeric portion of chromosome 6 (CEP 6)) on these lesions, which was negative in all cases. The difference in outcome between the uveal lesions and the skin lesions with a BAP1 mutation may be related to the presence of different oncogenic driver mutations in uveal lesions, which harbor GNAQ or GNA11 mutations (21) instead of BRAF or NRAS mutations. The suggested progression-promoting effect of mutated BAP1 is in line with the tumor suppressive function of intact BAP1 as a deubiquitylase required for efficient assembly of the homologous recombination (HR) factors BRCA1 and RAD51 after DNA double strand breaks (DSBs) (22) (23). BAP1 is recruited to DNA damage sites 22

24 Update on molecular pathology of cutaneous melanocytic lesions together with ASXL1 and deubiquitilates Ub-H2AK119 at sites of DNA damage (22, 23). In this way, it promotes error-free repair of these lesions. Defective HR and increased sensitivity to radiation due to BAP1 deficiency may, therefore, lead to genomic instability, a hallmark of cancer (22, 23). Moreover, BAP1 prevents proteasomal degradation of the conserved epigenetic regulator host cell factor-1 (HCF-1) and, consistent with this, Dey et al. showed that BAP1 KO splenocytes contained far less HCF-1 than their wild type counterparts (24). It is thought that BAP1 regulates gene expression via stabilization of HCF-1(24). These two examined functions of BAP1 could explain tumor-progression due to altered BAP1 expression. The most important reason for recognition of MBAITs at present is that they can be a marker of an underlying BAP1-associated germline mutation/cancer syndrome. Individuals with a BAP1 germline mutation have an increased risk to develop cutaneous and ocular melanoma and mesothelioma, apart from the risk to develop multiple MBAITs (25). Management of MBAIT lesions should consist of complete removal of the lesion and advise for genetic counseling to exclude a potential underlying cancer syndrome. Typical BRAF-mutated MBAIT cases from our own practice are shown in Figures 3 and 4 (Courtesy Dr. R. Kornegoor, Department of Pathology, Gelre Hospitals, Apeldoorn, The Netherlands). In case of suspicion of a MBAIT lesion, we recommend at least a BRAF and BAP1 immunostaining, but preferentially BRAF, NRAS, HRAS, and BAP1 mutation analysis is performed. In case a laboratory cannot perform these, we recommend consultation because of the important implications of a correct diagnosis. Mutation analysis of the BAP1 gene is difficult, since it is a complex gene and mutations can be present along all of the 17 exons of the gene. A low tumour percentage due to the small size of a lesion or small size of the spitzoid component, or the presence of a lot of TILs, can all hamper BAP1 mutation analysis in this setting. 2 The third group of spitzoid lesions, those with kinase fusions has only recently been described. Wiesner et al. described the presence of alterations in ROS1, NTRK1, ALK, BRAF, and RET in respectively 17%, 16%, 10%, 5% and 3% of spitzoid tumours. These alterations were present along the entire spectrum of the spitzoid tumours (55% in Spitz nevi, 56% in AST, and 39% in spitzoid melanoma) and these alterations therefore seem early events, and are not useful (yet) for differentiating benign from malignant spitzoid lesions. No clear distinct phenotypes were described at this time. Recently Busam et al. described 17 cases of Spitz tumors with ALK fusion, including 5 Spitz nevi and 12 AST (26). Clinically, these lesions were often polypoid. Histology showed a compound, mostly dermal located lesions with a plexiform growth, and consisting of fusiform melanocytes. In only 2 cases the lesional cells contained pigment. 23

25 Chapter 2 B B C C AA B B µm µm C C D D BRAFexon exon15 15 BRAF Figure 3 BRAF-mutated MBAIT in a 31-years old female located on the lower back (Courtesy Dr. R. Kornegoor, Gelre Hospitals, Apeldoorn, The Netherlands). A: overview (HE 25x), showing a polypoid lesion above the skin surface. The central part, B (HE 400x) consists of large epithelioid cells with intermingled lymphocytes. The margin, C (HE 100x), shows a common compound nevus. D: Molecular analysis showed a BRAF V600E mutation. In D Sequence plot of the mutation BRAF c.1799_1800tg>aa (p.(val600glu) alias p.v600e) with 7% mutant allel (very low load). Note; the BRAF RefSeq is NM_ The sequence plots are visualized as described under Fig. 2. Immunohistochemistry demonstrated positive BRAF protein expression, and loss of BAP1 protein expression in the large epithelioid cells (not shown). 24

26 Update on molecular pathology of cutaneous melanocytic lesions 1mm B 2 A B C D E Figure 4 BRAF-mutated MBAIT-lesion located on the posterior side of the ear in a 55-year-old female (Courtesy Dr. R. Kornegoor, Gelre Hospitals, Apeldoorn, The Netherlands). A: overview showing a nodular dermal based lesion with lymphocytes at the lateral margin. In this case no common nevus component was present (HE, 25x). B: a higher magnification demonstrates the large epithelioid aspect of the cells (HE, 200x). In C absence of BAP1 protein expression is seen in the large epithelioid melanocytes. Nuclei in the keratinocytes in the epidermis stain positive, serving as a positive internal control (400x). In D, BRAF protein is expressed by the large epithelioid melanocytes (400x). E: Sequence plots of the mutation BRAF c.1799t>a (p.(val600glu) alias p.v600e) with 23% mutant alleles and of the BAP1 (RefSeq NM_ ) c.581-1g>a splice site mutation with 34% mutant alleles. The sequence plots are visualized as described under Fig. 2. In the BAP1 plot the intron-exon boundary is represented as grey-white parts. 25

27 Chapter 2 All cases showed ALK protein expression. ALK FISH was positive in all cases (using a commercially available break-apart probe, Abbott Molecular, Des Plaines, IL): in 11/17 cases the fusion partner was tropomyosin 3 (TPM-3), in 6/17 cases the fusion partner was dynactin 1 (DCTN-1). FISH for copy number alterations did not met the criteria for melanoma diagnosis in any case. Array-CGH revealed no chromosomal gains or losses. In 2 cases a sentinel node procedure was performed and in both cases showed small nests in the subcapsular sinus. Both patients were alive and well after <1 year and after 4 years of follow-up. At present the follow-up time and number of cases is too limited to draw definite conclusions about the prognosis of this group of Spitz tumors with ALK-fusions. An ALK-positive Spitz tumor is presented in Figure 5. (Courtesy Prof. Dr. J. van den Oord, KU Leuven, Belgium). The anaplastic lymphoma kinase (ALK) gene, located on chromosome 2p23, is a receptor tyrosinase kinase protein and is capable of causing diverse tumour types of different lineages through a variety of molecular mechanisms (27). The most common mechanism of ALK activation is a genomic rearrangement involving the ALK locus, with a breakpoint in the 3 end of the ALK gene, with the other breakpoint involving a diverse group of genes leading to formation of a fusion oncogene that encodes for a fusion oncoprotein that is able to self-associate (27). Another way of ALK activation is by gain-of-function point mutations, but these are less frequent than ALK rearrangements. Also, ALK amplifications have been described in several tumor types, leading to the presence of multiple copies of the wild-type full-lenght ALK gene. But the way this contributes to tumor development is poorly understood (27). As mentioned above, in spitzoid tumours at present, only ALK rearrangements have been tested and described, and present in upto 19% of cases (14), with TPM-3 and DCTN-1 as fusion partners (14, 26). ALK rearrangements leading to TPM3-ALK fusion have also been reported in anaplastic large cell lymphoma and papillary thyroid carcinoma, and result in a constitutive tyrosinase kinase activity, in this way causing tumor development (27). Relevance of diagnosis of Spitz tumors with kinase fusions at the moment is mainly for treatment in malignant lesions (eg ALK, RET, ROS 1 alterations can be targeted with kinase inhibitors, such as crizotinib, cabozantinib, vandetanib). Table 2 summarises the characteristic features of the three above-described subtypes of spitzoid melanocytic lesions, including the most common histological findings, and genetic and prognostic features, with clinical relevance of recognition of these subtypes. 26

28 Update on molecular pathology of cutaneous melanocytic lesions A B 2 C D E F Figure 5 ALK-positive spitzoid tumour diagnosed as spitzoid tumour of uncertain malignant potential (STUMP). This lesion was seen in consultation elsewere (Courtesy Prof. Dr. J. van den Oord, KU Leuven, Belgium). The lesion was clinically polypoid and presented on the arm of a 4-year old girl. There were 11 mitoses/1mm2 present. Follow-up data are not available yet. In A HE overview (25x); in B and C larger maginification (HE x resp.) showing spindle shaped cells and some lymphocytes intermingled; in D PHH3 staining demonstrating mitoses (400x); in E mitf (250x); and in F ALK (250x). 1.2 Molecular background of congenital melanocytic nevi (CMN) and neurocutaneous melanosis (NCM) Congenital melanocytic nevi (CMN) are pigmented moles present at birth or shortly thereafter (28). They vary in size from small to very large or giant and hundreds of CMN can be present in one patient (29). CMN are considered to be a sporadic event with the exception of a few familial cases (30, 31). Especially giant CMN (> 40cm in size) are associated with an increased risk for cutaneous melanoma (up to 10 15% life-time risk) (32). 27

29 Chapter 2 Table 2 Overview of frequencies of gene mutations in different melanoma subtypes derived from different locations Subtypes spitzoid melanocytic tumours Histological characteristics (phenotype) Genetic alterations HRAS-mutated spitzoid tumours - dermal based lesions -wedge shaped -desmoplasia - often spindle shaped cells - HRAS mutations or amplification BAP 1-mutated spitzoid tumours (MBAIT) - mostly dermal based -often polypoid -often TILs -epithelioid cells, a common nevus component can be present - BAP1 mutation - often with combined BRAF V600E mutation. Rare cases with combined NRAS mutation reported Spitzoid tumours with kinase fusions -no distinct phenotypes identified yet. -ST with ALK fusions recently reported as often polypoid with plexiform growth of fusiform melanocytes (26) - specific fusion of certain genes (ROS1 17%, NTRK1 16%, ALK 10%, BRAF 5%, RET 3%) Relevance of diagnosis - single 11p gain can be present -favourable prognosis (prevention of overdiagnosis of melanoma). - loss of (a part of ) chromosome 3 can be present -can be a marker of a BAP1 cancer syndrome. -genetic counceling should be advised. -behaviour not clear yet. -in ALK fused STs no chromosomal gains or losses reported thusfar (26) -in case of malignant/ metastatic cases targeted treatment is an option (ALK, ROS, and RET alterations can be targeted by kinase inhibitors like crizotinib, cabozantinib, vandetanib). -behaviour not clear yet. Abbreviations: TILs=tumour-infiltrating lymphocytes; ST=Spitz tumour; MBAIT=melanocytic BAP1 associated intradermal tumour. Both NRAS and BRAF mutations have been detected in CMN in a mutually exclusive pattern, and there is a genotype-phenotype correlation between the size of CMN and type of mutation. Especially, large and giant CMN harbor somatic NRAS mutations (> 75%) in contrast to medium-sized CMN that are less frequently NRAS-mutated and especially small CMN and acquired melanocytic nevi that frequently carry BRAF V600E mutations (up to 80% in acquired nevi) (33). Activating BRAF V600E mutations 28

30 Update on molecular pathology of cutaneous melanocytic lesions in large CMN are rare (approximately 15 %) (6). A few cases of large CMN were shown to harbor a chromosomal translocation involving BRAF resulting in gain-of-function, and this could represent an alternative mechanism of BRAF activation in BRAF- or NRAS-wild-type CMN (34). Recently, using several highly sensitive techniques, Charbel et al. showed that large and giant CMN contain NRAS Q61 mutations (Q61R or Q61K) in up to 94.7% of cases (35). In addition, using whole-exome sequencing they found no other coding mutations in five large/giant CMN implying that at present NRAS mutations are the sole recurrent mutation in these lesions. Somatic NRAS mutations, therefore, seem to be sufficient to drive melanocytic proliferations in utero. In addition, identical NRAS Q61 mutations were recently demonstrated in multiple CMN samples from the same patient (36). This finding suggests that multiple CMN are clonal proliferations caused by a single, postzygotic NRAS mutation in a neuro-ectodermal precursor cell rather than de novo proliferations arising from different mutations. Kinsler et al. recently proposed the term CMN syndrome as they observed that patients with CMN often have characteristic facial features (such as wide or prominent forehead and apparent hypertelorism) (37). The osteocartilageneous structures of the face are neuro-ectodermal in origin and can be affected by mutations in components of the RAS/RAF/MEK/ERK pathway in patients with germ-line RASopathies who have characteristic facial features (38). Kinsler et al. hypothesized that the occurrence of a postzygotic, somatic NRAS mutation in early neuro-ectodermal precursors might be responsible for the characteristic facial features in patients with CMN as well (36). In addition, as they observed a high prevalence of red hair in families of children with CMN, a germline predisposition for the development of the CMN syndrome was hypothesized. Indeed, they showed that certain germ-line allele variants of the melanocortin-1-receptor (MC1R), known to be responsible for the red hair/fair skin/freckling phenotype, are associated with the presence of CMN (as well as with more extensive CMN) (39). CMN can be associated with a spectrum of neurological abnormalities, including malformations (for instance Dandy-Walker malformation), communicating hydrocephalus, arachnoid cysts, CNS tumors (for instance astrocytoma and choroid plexus papilloma) but also melanotic abnormalities (36). The latter is called neurocutaneous melanosis (NCM) and refers to the presence of large of multiple CMN in association with melanin depositions in the brain parenchym (visible on T1-weighted MRI) or melanocytic tumors like leptomeningeal melanocytosis or melanoma (40). Leptomeningeal melanocytosis consists of a diffuse proliferaton of histological benign appearing melanocytes in the leptomeninges, without CNS invasion, and carries a poor prognosis once symptomatic (40). 2 29

31 Chapter 2 Recently, it was shown that postzygotic, somatic NRAS mutations contribute to the pathogenesis of NCM. Identical, somatic NRAS Q61 mutations were detected in multiple CMN as well as in the CNS melanocytic tumor in the same NCM patients (41). Kinsler et al. also detected a somatic NRAS Q61 mutation in three non melanocytic CNS tumors occuring in patients with CMN (including choroid plexus papilloma, neurocristic hamartoma, meningioma) (36). The presence of identical, somatic NRASQ61 mutations in both CMN as well as in CNS melanocytic (and non-melanocytic) neoplasms in the same patients suggests that these mutations occur in the developing neuro-ectoderm early during embryogenesis. In fact, this pathogenetic mechanism fits the spectrum of mosaic RASopathies that are characterized by postzygotic mutations resulting in the presence of at least two genetically distinct cell populations in the same organism (42). For instance, postzygotic, somatic RAS mutations (HRAS, KRAS) were recently shown to be present in distinct lesions of the keratinocytic epidermal nevus syndrome, a mosaic RASopathy characterized by the presence of epidermal nevi in association with extra-cutaneous abnormalities (CNS, ocular, skeletal, cardiovascular, and genitourinary system) (42). The phenotype in the spectrum of mosaic RASopathies is most likely determined by type of mutation, the timing of the mutation, and affected cell type (42, 43). A mouse model has demonstrated a role for postzygotic, early embryonic NRAS mutations in the pathogenesis of NCM. By using the Cre-loxP technology, Pedersen et al. showed that expression of oncogenic NRAS G12D in melanocytes early during embryogenesis resulted in a mouse phenotype strongly resembling NCM in human beings (41). As NRAS mutations occur in benign lesions such as CMN, they are in itself insufficient for malignant transformation of melanocytes, and better insight in the genetic aberrations eventually leading to melanoma is needed. NRAS mutations are a therapeutic target for treatment with MEK inhibitors. In a Phase II trial, treatment with MEK162 was shown to have effect in some patients with advanced NRAS-mutated melanoma and a Phase III trial is ongoing (44). For patients with NRAS-mutated CNS melanocytic tumors treatment with MEK162 might be of benefit as well (45) Molecular diagnostics in the differential diagnosis of melanoma (metastasis) Melanomas are known for their wide range of cytomorphologic features and architectural patterns and may mimic various non-melanocytic tumors (reviewed by Banerjee et al. (46, 47)). In most cases a diagnosis can be rendered by careful examination of the histomorphology and by sometimes adding immunohistochemical stains. In some cases, however, especially in recurrences and metastases, melanomas can show an aberrant immunophenotype with loss of lineage-specific markers. In part of these cases, molecular analysis of genes in the MAPK pathway and CDKN2A (7) mutation analysis can be useful in the differential diagnosis. In addition, molecular 30

32 Update on molecular pathology of cutaneous melanocytic lesions analysis can be of help in determining the site of the primary melanoma in case of metastasis with an unknown primary since melanomas from different locations have different mutation types (see Table 1). Furthermore, molecular analysis can help to discriminate between a melanoma metastasis and a second primary melanoma. One of the lesions that can be hard to distinguish from melanoma is clear cell sarcoma (CCS) as they share histopathological features and cannot be distinguished by immunohistochemistry (48). Classically, CCS is a deep soft tissue tumor associated with tendons and aponeuroses (49), but it can also present as a cutaneous lesion (50), and then has to be differentiated from metastatic or primary nodular melanoma. While most cutaneous melanomas harbor a BRAF or NRAS mutation (see Table 1), CCS in approximately 75% have a t(12;22)(a13;q12) or less commonly a t(2;22) (q34;q12) translocation leading to the EWSR1/ATF1 or EWSR1/CREB1 fusion transcripts (51, 52). Yang et al. (53) performed BRAF and NRAS mutation analysis, as well as FISH analysis for the EWST1/ATF1 fusion gene in 31 melanoma cases and 16 CCSs. They found the translocation in 78.6 % of the CCSs but in none of the melanomas, whereas BRAF and NRAS mutations were present in, respectively, 51.6% and 12.9% of the melanomas and not in any of the CCSs. Hantschke et al. (50) describe twelve cases of cutaneous CCS in which FISH analysis for the t(12;22)(a13;q12) translocation contributed in the differential diagnosis with melanoma. This implicates that this type of analysis can be of great aid in the differential diagnosis between CCS and melanoma. Recently, we described two cases in which mutation analysis lead to the correct diagnosis of (dedifferentiated) metastatic melanoma (54). Both patients had a history of melanoma and presented several years after their primary diagnosis with a lesion histologically and immunohistochemically different from the primary melanoma and mimicking a solitary fibrous tumor (SFT). Using BRAF and NRAS mutation analysis, it could be proven that both lesions were melanoma metastases instead of SFT. Many other soft tissue tumors can mimick melanoma or melanoma metastasis like dermatofibrosarcoma protuberans, malignant peripheral nerve sheath tumors (MPNST) (55) and sarcomas (46, 56-58). Although CDKN2A mutation frequency in sporadic melanomas as well as consistency in melanoma metastasis are reported to be low (reported CDKN2A mutation frequency in upto 25% of primary melanomas, 0-14% in melanoma metasases, with a 31% consistency being reported between a primary and the metastasis (59, 60)), CDKN2A mutation analysis can be an alternative way to confirm a diagnosis of metastatic melanoma by showing a clonal relationship between a primary melanoma and a metastasis (after exclusion of a CDKN2A germline mutation (61)), as is illustrated by a recent case from our own practice. A 36-year-old woman with a history of a superficial spreading melanoma of the back 16 years earlier, presented with a large lung mass which was thought to be a non-small cell lesion, probably squamous cell carcinoma based on fine needle aspiration (EUS-FNA): melanoma markers HMB

33 Chapter 2 and MART were negative while the squamous cell marker - p63- was weakly positive. Evaluation of the pneumonectomy specimen, however, led to a differential diagnosis with metastatic melanoma, since immunohistochemical stainings on the complete tumor now available for histological evaluation showed in part features consistent with melanocytic differentiation. Vimentin, tyrosinase and mitf were weakly positive, however S100 and SOX10 were both negative. Therefore additional BRAF and CDKN2A mutation analysis was performed. Identical mutations in both genes were present in the primary cutaneous melanoma and the lung mass -in absence of a germline CDKN2A mutation- confirming that the lung mass was a late metastasis of the cutaneous melanoma. This case is illustrated in Figure 6. In a patient with a history of melanoma, it can occasionally be difficult to differentiate between a second primary cutaneous melanoma or a melanoma metastasis based on histomorphology, especially when there is epidermotropism (62). If the first melanoma was not a cutaneous melanoma, the problem can often be solved by mutation analysis since different types of melanomas have mutations in different genes of the MAPK pathway as mentioned in Table 1. If the primary melanoma was a cutaneous melanoma, mutation analysis of NRAS and BRAF can be of help, but since these mutations are hotspot mutations, it s of limited use. Usage is additionally hampered by frequent occurrence of heterogeneity in melanoma with respect to NRAS and BRAF status, leading to a substantial discordance in mutation status in these genes in a primary compared to the metastases (59, 63). CDKN2A mutation analysis can sometimes be helpful as we have published before (64), and is illustrated by the case above, to differentiate between a new primary and a melanoma metastassis. The advantage of using CDKN2A mutations for clonality is that CDKN2A mutations are unique mutations instead of hot spot mutations. Although CDKN2A mutation frequency in sporadic melanomas is low, CDKN2A mutation analysis to our opinion is worth trying, as there is a substantial difference in the prognosis between a second primary melanoma and a melanoma metastasis. PART 2. MOLECULAR PATHOLOGY IN MELANOMA TREATMENT Metastatic melanoma treatment got a great impulse after the discovery that the (hot spot) mutations in genes involved in the MAPK pathway in melanoma, as well as in GNAQ and GNA11, proved targetable on the protein level by specific inhibitors. GNAQ and GNA11 mutations are mainly present in uveal and primary brain melanomas, and combinations of MEK inhibitors with either PI3K or mtor inhibitors have show efficacy in GNAQ- and GNA11-mutant melanomas (65, 66). A simplified scheme of the most important pathways involved in melanoma pathogenesis, and points of action of specific inhibitors are depicted in Figure 1. 32

34 Update on molecular pathology of cutaneous melanocytic lesions 2 A B D C E F G Figure 6 Case of a superficial spreading melanoma of the back with lung metastasis sixteen years later in a female patient. In A an overview of the primary cutaneous melanoma showing centrally an asymmetrical melanocytic lesion (HE, 25x). In B, a more detailed picture showing clear atypia with loss of maturation and ascension in the epidermis (HE, 100x). In C, D, and E, the lung lesion at different magnifications (HE C 25x, D and E 100x) showing a nodular proliferation of atypical partly plasmacytoid cells resembling the primary cutaneous melanoma. Immunohistochemical S100 stain however was negative (F S100 immunostaining, 100x). Internal and external controles were positive. G: Sequence plots demonstrated an identical c.168_170delinstg(p.(arg58fs)) in the CDKN2A gene in both the skin melanoma and the lung tumour, which was absent in normal tissue of the patient. This confirmed that the lung tumour was a metastasis of the previous melanoma. The sequence plots are visualized as described under Fig

35 Chapter 2 Melanomas from distinct locations have been shown to contain different mutation types and frequencies (5, 67, 68). An overview of the most frequent mutations in the distinct melanoma subtypes and their frequencies, as well as the different therapeutic options, are depicted in Table 1. The most frequently tested genes at present in melanoma are BRAF and NRAS, because these mutations are most prevalent in melanoma, especially in those arising from the skin. Testing for targeted treatment in melanoma should be guided by the localization of the primary tumor. Nowadays this issue becomes less important because several molecular labs use platforms that detect various regions from multiple genes within one test. In case of molecular testing for targeted therapy in metastatic melanoma there are several important issues. Molecular testing for therapy requires a close collaboration between clinician, pathologist, and molecular biologist. Testing for treatment should only be performed when targeted therapy is considered as a therapeutic option. This can be judged best by the clinician. In most cases, testing is only indicated in inoperable stage 3 or stage 4 melanoma patients. Tests need to be performed in an accredited -molecular laboratory that guarantees that the laboratory techniques and processes are performed standardized and yield high quality results which implies that only validated tests are used. The role of the pathologist is to confirm the diagnosis of melanoma metastasis, to check representativity of the tumor tissue to be examined, to indicate the area for macrodissection (if needed) by the technician for DNA extraction, to indicate possible tumor heterogeneity and to estimate the neoplastic cell percentage within the tested sample. The knowledge and the expertise of trained molecular biologists is used to come to optimal test-results and adequate interpretation and reporting of molecular test results. The molecular report should contain information on the specific tissue block tested, the percentage of neoplastic cells, the type of molecular test used and the sensitivity of the test, the type of genes and exons or mutations thereof which are tested, and an accurate description of the mutation present according to the Human Genome Variation Society nomenclature. The molecular biologist, together with the pathologist, is responsible for proper integration and interpretation of the molecular results in the pathology report. The molecular test is preferentially performed on a recent metastasis. Reason for this is firstly confirmation of the diagnosis of melanoma metastasis. Secondly, there is considerable tumour heterogeneity with a reported discrepancy by Colombino et al. of % in the BRAF and NRAS mutation status of the metastatic melanoma when compared to the primary (59, 63). Concordance seems dependent on the location of the metastases and is highest for visceral (92.5%) and nodal metastases (91%), but 34

36 Update on molecular pathology of cutaneous melanocytic lesions relatively low for brain (79%) and skin metastases (71%). Especially in the latter therefore testing should be performed on the metastasis. Saint-Jean et al. (69) also reported discordant BRAF results. They performed multiple tests in a subgroup of 74 patients: in 10/74 (13.5%) of these patients BRAF status was discordant between distinct samples of a patient. In two patients, the discordance was present between the primary and a metastasis. In six patients the discordance was present between two distinct metastases. The authors state that without repeated testing, five patients would unjustly have been excluded from treatment with the BRAF-inhibitor, vemurafenib. They advocate to test other samples in case no BRAF mutation has been detected. We can agree with this especially in selected cases, when the primary location is likely to be associated with a BRAF mutation and when no mutation in other targetable genes (like NRAS or KIT) has been identified. A molecular test should be able to detect all relevant and targetable mutations in a gene. The most frequently tested gene for melanoma treatment at present is BRAF. The most frequent BRAF mutation is a mononucleotide point mutation in codon 600 (CTG) of exon 15, c.1799t>a (p.(val600glu), in which the valine (V) of codon 600 is replaced by glutamine (E). This mutation is therefore known as the BRAF V600E mutation and is present in about 70-75% of all BRAF-mutated melanomas (67, 70, 71). In a recent study in which 1112 primary and metastatic melanomas from different locations were analysed for BRAF mutations (774 skin melanomas, 111 acral melanomas, 26 mucosal melanomas, 23 uveal melanomas, 1 leptomeningeal melanoma, and 177 metastases), 44.9% of the cases harbored a BRAF mutation: 75.4% of the cases, mutations were BRAF V600E either deriving from the c.1799t>a or from a c delinsAA mutation. Of the remaining BRAF non-v600e cases (24.6%), the most frequently seen mutation was the BRAF V600K (17.2%); BRAF V600R or BRAF V600D mutated cases were found in low percentages (4.6% ). BRAF exon 11 mutations were also observed in a low percentage (0.4%). There are several studies that have reported that all melanomas with BRAF codon 600 mutations are sensitive to BRAF inhibitors such as vemurafenib (Zelboraf, Roche Molecular Systems Inc.) and dabrafenib (GSK )(72, 73). This implicates that a BRAF test needs to be able to detect not only the BRAFc.1799T>A(p. (Val600Glu)/ BRAF V600E mutation, but the other codon 600 mutations as well. A test must be able to detect BRAF mutations that have been reported to be insensitive to BRAF inhibitor treatment, like the kinase dead mutation BRAF D594 (74, 75). The molecular test for therapy should be performed within a short turn-around time since mostly this kind of targeted therapy will be given in rapidly progressive metastatic melanoma patients. A turn around time of 5 working days is feasible within our hands. At present we perform NGS-based mutation testing using Ion Torrent Personal Genome Machine (IT-PGM) from life Technologies for analysis of gene-panels for diagnostic purposes. 2 35

37 Chapter 2 At the moment there is a tendency towards testing with NGS methods for targeted treatment in diverse cancer types. The sensitivity of NGS is higher than Sanger sequencing (detection of 2-10% versus 15-25% allele frequency). Moreover, the amount of DNA that is needed for the analysis of gene panels is very low, only 10 ng for all amplicons for instance when using the Ion Torrent Personal Genome Machine (IT-PGM) from life Technologies versus 10ng needed per amplicon for Sanger sequencing. The turnaround time and costs of NGS methods can be competitive with respect to low throughput technologies in centers that have sufficient numbers of samples. For diagnostic requests in melanoma in our department a custom-designed gene panel for amplicon-sequencing of BRAF (NM_ ) exon 15, NRAS (NM_ ) exon 2 and 3, HRAS (NM_ ) exon 2 and 3, AKT1 (NM_ ) exon 3, GNAQ (NM_ ) exon 4 and 5, GNA11 (NM_ ) exon 4 and 5, KIT NM_ ) exon 8, 9, 11, 13, 14 and PDGFRA (NM_ ) exon 12, 14, 18 is performed. The use of small, dedicated gene panels and efficient loading of the chips for IT-PGM also significantly reduce costs per case. The major benefit of (targeted) NGS is that it uncovers all kinds of mutations in selected genomic regions instead of only mutations at predefined positions. 36

38 Update on molecular pathology of cutaneous melanocytic lesions REFERENCES 1. Blokx WA, van Dijk MC, Ruiter DJ. Molecular cytogenetics of cutaneous melanocytic lesions - diagnostic, prognostic and therapeutic aspects. Histopathology. 2010;56(1): Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, et al. Distinct sets of genetic alterations in melanoma. N Engl J Med. 2005;353(20): Solus JF, Kraft S. Ras, Raf, and MAP kinase in melanoma. Adv Anat Pathol. 2013;20(4): Takata M, Murata H, Saida T. Molecular pathogenesis of malignant melanoma: a different perspective from the studies of melanocytic nevus and acral melanoma. Pigment Cell Melanoma Res. 2010;23(1): Whiteman DC, Pavan WJ, Bastian BC. The melanomas: a synthesis of epidemiological, clinical, histopathological, genetic, and biological aspects, supporting distinct subtypes, causal pathways, and cells of origin. Pigment Cell Melanoma Res. 2011;24(5): Ichii-Nakato N, Takata M, Takayanagi S, Takashima S, Lin J, Murata H, et al. High frequency of BRAFV600E mutation in acquired nevi and small congenital nevi, but low frequency of mutation in medium-sized congenital nevi. J Invest Dermatol. 2006;126(9): Bastian BC. The molecular pathology of melanoma: an integrated taxonomy of melanocytic neoplasia. Annu Rev Pathol. 2014;9: Bauer J, Curtin JA, Pinkel D, Bastian BC. Congenital melanocytic nevi frequently harbor NRAS mutations but no BRAF mutations. J Invest Dermatol. 2007;127(1): Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, O Brien JM, et al. Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature. 2009;457(7229): van Dijk MC, Bernsen MR, Ruiter DJ. Analysis of mutations in B-RAF, N-RAS, and H-RAS genes in the differential diagnosis of Spitz nevus and spitzoid melanoma. The American journal of surgical pathology. 2005;29(9): van Engen-van Grunsven AC, van Dijk MC, Ruiter DJ, Klaasen A, Mooi WJ, Blokx WA. HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features. Am J Surg Pathol. 2010;34(10): Wiesner T, Obenauf AC, Murali R, Fried I, Griewank KG, Ulz P, et al. Germline mutations in BAP1 predispose to melanocytic tumors. Nat Genet. 2011;43(10): Busam KJ, Wanna M, Wiesner T. Multiple epithelioid Spitz nevi or tumors with loss of BAP1 expression: a clue to a hereditary tumor syndrome. JAMA dermatology. 2013;149(3): Wiesner T, He J, Yelensky R, Esteve-Puig R, Botton T, Yeh I, et al. Kinase fusions are frequent in Spitz tumours and spitzoid melanomas. Nature communications. 2014;5: Bastian BC, LeBoit PE, Pinkel D. Mutations and copy number increase of HRAS in Spitz nevi with distinctive histopathological features. The American journal of pathology. 2000;157(3): Da Forno PD, Pringle JH, Fletcher A, Bamford M, Su L, Potter L, et al. BRAF, NRAS and HRAS mutations in spitzoid tumours and their possible pathogenetic significance. The British journal of dermatology. 2009;161(2): Richetta AG, Silvestri V, Giancristoforo S, Rizzolo P, D Epiro S, Graziano V, et al. Mutational profiling in melanocytic tumors: multiple somatic mutations and clinical implications. Oncology. 2014;86(2): Wiesner T, Murali R, Fried I, Cerroni L, Busam K, Kutzner H, et al. A distinct subset of atypical Spitz tumors is characterized by BRAF mutation and loss of BAP1 expression. Am J Surg Pathol. 2012;36(6): Yeh I, Mully TW, Wiesner T, Vemula SS, Mirza SA, Sparatta AJ, et al. Ambiguous Melanocytic Tumors With Loss of 3p21. Am J Surg Pathol Pouryazdanparast P, Haghighat Z, Beilfuss BA, Guitart J, Gerami P. Melanocytic nevi with an atypical epithelioid cell component: clinical, histopathologic, and fluorescence in situ hybridization findings. Am J Surg Pathol. 2011;35(9): Field MG, Harbour JW. Recent developments in prognostic and predictive testing in uveal melanoma. Curr Opin Ophthalmol. 2014;25(3): Yu H, Pak H, Hammond-Martel I, Ghram M, Rodrigue A, Daou S, et al. Tumor suppressor and deubiquitinase BAP1 promotes DNA double-strand break repair. Proc Natl Acad Sci U S A. 2014;111(1):

39 Chapter Ismail IH, Davidson R, Gagne JP, Xu ZZ, Poirier G, Hendzel MJ. Germ-line Mutations in BAP1 Impair its Function in DNA Double-Strand break Repair. Cancer Res Dey A, Seshasayee D, Noubade R, French DM, Liu J, Chaurushiya MS, et al. Loss of the tumor suppressor BAP1 causes myeloid transformation. Science. 2012;337(6101): Carbone M, Ferris LK, Baumann F, Napolitano A, Lum CA, Flores EG, et al. BAP1 cancer syndrome: malignant mesothelioma, uveal and cutaneous melanoma, and MBAITs. Journal of translational medicine. 2012;10: Busam KJ, Kutzner H, Cerroni L, Wiesner T. Clinical and Pathologic Findings of Spitz Nevi and Atypical Spitz Tumors With ALK Fusions. Am J Surg Pathol. 2014;38(7): Marino-Enriquez A, Dal Cin P. ALK as a paradigm of oncogenic promiscuity: different mechanisms of activation and different fusion partners drive tumors of different lineages. Cancer genetics. 2013;206(11): Krengel S. Nevogenesis--new thoughts regarding a classical problem. Am J Dermatopathol. 2005;27(5): Ruiz-Maldonado R. Measuring congenital melanocytic nevi. Pediatric dermatology. 2004;21(2): Danarti R, Konig A, Happle R. Large congenital melanocytic nevi may reflect paradominant inheritance implying allelic loss. European journal of dermatology : EJD. 2003;13(5): de Wijn RS, Zaal LH, Hennekam RC, van der Horst CM. Familial clustering of giant congenital melanocytic nevi. J Plast Reconstr Aesthet Surg. 2010;63(6): Krengel S, Hauschild A, Schafer T. Melanoma risk in congenital melanocytic naevi: a systematic review. Br J Dermatol. 2006;155(1): Dessars B, De Raeve LE, Morandini R, Lefort A, El Housni H, Ghanem GE, et al. Genotypic and gene expression studies in congenital melanocytic nevi: insight into initial steps of melanotumorigenesis. J Invest Dermatol. 2009;129(1): Dessars B, De Raeve LE, El Housni H, Debouck CJ, Sidon PJ, Morandini R, et al. Chromosomal translocations as a mechanism of BRAF activation in two cases of large congenital melanocytic nevi. J Invest Dermatol. 2007;127(6): Charbel C, Fontaine RH, Malouf GG, Picard A, Kadlub N, El-Murr N, et al. NRAS mutation is the sole recurrent somatic mutation in large congenital melanocytic nevi. J Invest Dermatol. 2014;134(4): Kinsler VA, Thomas AC, Ishida M, Bulstrode NW, Loughlin S, Hing S, et al. Multiple Congenital Melanocytic Nevi and Neurocutaneous Melanosis Are Caused by Postzygotic Mutations in Codon 61 of NRAS. J Invest Dermatol. 2013: Kinsler V, Shaw AC, Merks JH, Hennekam RC. The face in congenital melanocytic nevus syndrome. American journal of medical genetics Part A. 2012;158A(5): Zenker M. Clinical manifestations of mutations in RAS and related intracellular signal transduction factors. Current opinion in pediatrics. 2011;23(4): Kinsler VA, Abu-Amero S, Budd P, Jackson IJ, Ring SM, Northstone K, et al. Germline melanocortin-1-receptor genotype is associated with severity of cutaneous phenotype in congenital melanocytic nevi: a role for MC1R in human fetal development. J Invest Dermatol. 2012;132(8): Kadonaga JN, Frieden IJ. Neurocutaneous melanosis: definition and review of the literature. J Am Acad Dermatol. 1991;24(5 Pt 1): Pedersen M, Kusters-Vandevelde HV, Viros A, Groenen PJ, Sanchez-Laorden B, Gilhuis JH, et al. Primary melanoma of the CNS in children is driven by congenital expression of oncogenic NRAS in melanocytes. Cancer discovery. 2013;3(4): Hafner C, Groesser L. Mosaic RASopathies. Cell cycle. 2013;12(1): Avitan-Hersh E, Tatur S, Indelman M, Gepstein V, Shreter R, Hershkovitz D, et al. Postzygotic HRAS mutation causing both keratinocytic epidermal nevus and thymoma and associated with bone dysplasia and hypophosphatemia due to elevated FGF23. The Journal of clinical endocrinology and metabolism. 2014;99(1):E Ascierto PA, Schadendorf D, Berking C, Agarwala SS, van Herpen CM, Queirolo P, et al. MEK162 for patients with advanced melanoma harbouring NRAS or Val600 BRAF mutations: a non-randomised, open-label phase 2 study. Lancet Oncol. 2013;14(3):

40 Update on molecular pathology of cutaneous melanocytic lesions 45. Kusters-Vandevelde HV, Willemsen AE, Groenen PJ, Kusters B, Lammens M, Wesseling P, et al. Experimental treatment of NRAS-mutated neurocutaneous melanocytosis with MEK162, a MEK-inhibitor. Acta neuropathologica communications. 2014;2(1): Banerjee SS, Harris M. Morphological and immunophenotypic variations in malignant melanoma. Histopathology. 2000;36(5): Banerjee SS, Eyden B. Divergent differentiation in malignant melanomas: a review. Histopathology. 2008;52(2): Bearman RM, Noe J, Kempson RL. Clear cell sarcoma with melanin pigment. Cancer. 1975;36(3): Enzinger FM. Clear-Cell Sarcoma of Tendons and Aponeuroses. An Analysis of 21 Cases. Cancer. 1965;18: Hantschke M, Mentzel T, Rutten A, Palmedo G, Calonje E, Lazar AJ, et al. Cutaneous clear cell sarcoma: a clinicopathologic, immunohistochemical, and molecular analysis of 12 cases emphasizing its distinction from dermal melanoma. Am J Surg Pathol. 2010;34(2): Mrozek K, Karakousis CP, Perez-Mesa C, Bloomfield CD. Translocation t(12;22)(q13;q ) in a clear cell sarcoma of tendons and aponeuroses. Genes Chromosomes Cancer. 1993;6(4): Antonescu CR, Nafa K, Segal NH, Dal Cin P, Ladanyi M. EWS-CREB1: a recurrent variant fusion in clear cell sarcoma--association with gastrointestinal location and absence of melanocytic differentiation. Clin Cancer Res. 2006;12(18): Yang L, Chen Y, Cui T, Knosel T, Zhang Q, Geier C, et al. Identification of biomarkers to distinguish clear cell sarcoma from malignant melanoma. Hum Pathol. 2012;43(9): Bekers EM, van Engen-van Grunsven AC, Groenen PJ, Westdorp H, Koornstra RH, Bonenkamp JJ, et al. Metastatic melanoma mimicking solitary fibrous tumor: report of two cases. Virchows Arch. 2014;464(2): King R, Busam K, Rosai J. Metastatic malignant melanoma resembling malignant peripheral nerve sheath tumor: report of 16 cases. Am J Surg Pathol. 1999;23(12): Lodding P, Kindblom LG, Angervall L. Metastases of malignant melanoma simulating soft tissue sarcoma. A clinico-pathological, light- and electron microscopic and immunohistochemical study of 21 cases. Virchows Arch A Pathol Anat Histopathol. 1990;417(5): Zelger BG, Steiner H, Wambacher B, Zelger B. Malignant melanomas simulating various types of soft tissue tumors. Dermatol Surg. 1997;23(11): Magro CM, Crowson AN, Mihm MC. Unusual variants of malignant melanoma. Mod Pathol. 2006;19 Suppl 2:S Colombino M, Capone M, Lissia A, Cossu A, Rubino C, De Giorgi V, et al. BRAF/NRAS mutation frequencies among primary tumors and metastases in patients with melanoma. J Clin Oncol. 2012;30(20): Piccinin S, Doglioni C, Maestro R, Vukosavljevic T, Gasparotto D, D Orazi C, et al. p16/cdkn2 and CDK4 gene mutations in sporadic melanoma development and progression. International journal of cancer. 1997;74(1): Monzon J, Liu L, Brill H, Goldstein AM, Tucker MA, From L, et al. CDKN2A mutations in multiple primary melanomas. N Engl J Med. 1998;338(13): Crowson AN, Magro CM, Mihm MC. The melanocytic proliferations. New York: Wiley-Liss; Colombino M, Lissia A, Capone M, De Giorgi V, Massi D, Stanganelli I, et al. Heterogeneous distribution of BRAF/NRAS mutations among Italian patients with advanced melanoma. Journal of translational medicine. 2013;11: Blokx WA, Lesterhuis WJ, Andriessen MP, Verdijk MA, Punt CJ, Ligtenberg MJ. CDKN2A (INK4A-ARF) mutation analysis to distinguish cutaneous melanoma metastasis from a second primary melanoma. Am J Surg Pathol. 2007;31(4): Khalili JS, Yu X, Wang J, Hayes BC, Davies MA, Lizee G, et al. Combination small molecule MEK and PI3K inhibition enhances uveal melanoma cell death in a mutant GNAQ- and GNA11-dependent manner. Clin Cancer Res. 2012;18(16): Ho AL, Musi E, Ambrosini G, Nair JS, Deraje Vasudeva S, de Stanchina E, et al. Impact of combined mtor and MEK inhibition in uveal melanoma is driven by tumor genotype. PLoS One. 2012;7(7):e

41 Chapter Greaves WO, Verma S, Patel KP, Davies MA, Barkoh BA, Galbincea JM, et al. Frequency and Spectrum of BRAF Mutations in a Retrospective, Single-Institution Study of 1112 Cases of Melanoma. J Mol Diagn. 2013;15(2): Lovly CM, Dahlman KB, Fohn LE, Su Z, Dias-Santagata D, Hicks DJ, et al. Routine multiplex mutational profiling of melanomas enables enrollment in genotype-driven therapeutic trials. PLoS One. 2012;7(4):e Saint-Jean M, Quereux G, Nguyen JM, Peuvrel L, Brocard A, Vallee A, et al. Is a single BRAF wild-type test sufficient to exclude melanoma patients from vemurafenib therapy? J Invest Dermatol. 2014;134(5): Rubinstein JC, Sznol M, Pavlick AC, Ariyan S, Cheng E, Bacchiocchi A, et al. Incidence of the V600K mutation among melanoma patients with BRAF mutations, and potential therapeutic response to the specific BRAF inhibitor PLX4032. Journal of translational medicine. 2010;8: Amanuel B, Grieu F, Kular J, Millward M, Iacopetta B. Incidence of BRAF p.val600glu and p.val600lys mutations in a consecutive series of 183 metastatic melanoma patients from a high incidence region. Pathology. 2012;44(4): Klein O, Clements A, Menzies AM, O Toole S, Kefford RF, Long GV. BRAF inhibitor activity in V600R metastatic melanoma. Eur J Cancer Menzies AM, Long GV, Murali R. Dabrafenib and its potential for the treatment of metastatic melanoma. Drug design, development and therapy. 2012;6: Heidorn SJ, Milagre C, Whittaker S, Nourry A, Niculescu-Duvas I, Dhomen N, et al. Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through CRAF. Cell. 2010;140(2): Stones CJ, Kim JE, Joseph WR, Leung E, Marshall ES, Finlay GJ, et al. Comparison of responses of human melanoma cell lines to MEK and BRAF inhibitors. Frontiers in genetics. 2013;4:66. 40

42 Update on molecular pathology of cutaneous melanocytic lesions 2 41

43

44 Part II

45

46 3 HRAS mutated Spitz tumors: a subtype of Spitz tumors with distinct features Adriana C.H. van Engen- van Grunsven, MD 1, Marcory C.R.F. van Dijk, MD PhD 2, Dirk J. Ruiter, MD PhD 1, 3, Annelies Klaasen 1, Wolter J. Mooi MD PhD 4, Willeke A.M. Blokx MD PhD 1 1 Department of Pathology, Radboud University Nijmegen Medical Center, Nijmegen, 2 Department of Pathology and Medical Biology, University Medical Center Groningen, Groningen, 3 Department of Anatomy, Radboud University Nijmegen Medical Center, Nijmegen and 4 Department of Pathology, VU University Medical Center, Amsterdam, The Netherlands. Am J Surg Pathol 2010;34:

47 Chapter 3 ABSTRACT It is often very difficult to confidently distinguish benign and malignant Spitz lesions, and a diagnosis of Spitz tumor of unknown malignant potential (STUMP) is rendered. To address this problem, we performed molecular genetic analysis in a large group of Spitz tumors (93 Spitz nevi and 77 STUMPs) and identified a subgroup of 24 lesions harboring a HRAS mutation. This subgroup lay predominantly in the dermis, had a relatively low cellularity, showed desmoplasia (with single cells interspersed between the collagen bundles) and had an infiltrating base. In 7 of these 24 cases (29%) melanoma had been the initial diagnosis, or an important differential diagnostic consideration, mainly based on the presence of multiple or deeply located mitotic figures, especially in adult patients. In our series none of the patients with the HRAS mutated lesions developed recurrences or metastases (mean and median follow-up: 10.5 years). This was in accordance with the literature: review showed that no HRAS mutations had so far been reported in Spitzoid melanomas. We therefore conclude that HRAS mutation analysis may be a useful diagnostic tool to help differentiate between Spitz nevus and Spitzoid melanoma, thereby reducing the frequency of overdiagnosis of melanoma, and to help predict the biological behavior of a STUMP. Moreover, this might be a first step towards a more reproducible classification of Spitz tumors combining histological and genetic data. 46

48 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features INTRODUCTION Spitz tumors are notoriously difficult to classify. They form a diagnostic challenge for pathologists and are frequently submitted for consultation to expert dermatopathologists. Because of similarities with melanoma, Spitz nevus is not uncommonly misdiagnosed as melanoma, and vice versa (1). There have been many attempts to define diagnostic criteria to correctly classify Spitz tumors, but this has failed to lead to consensus on diagnostic criteria even among expert dermatopathologists (2-9). Moreover, even consensus on the diagnosis does not guarantee a correct prediction of outcome (4, 9). The lack of objective histological criteria in Spitz tumors illustrates the need for additional parameters to aid in this distinction. There are other tumor types (lymphoma, breast cancer) where integration of genetic and clinical data besides morphology and immunophenotype has resulted in an improved and more reproducible tumor classification and has even led to the discovery of new entities within the spectrum, and to new therapeutical approaches. For Spitz tumors the demonstration of HRAS mutations in benign lesions seems a first break-through in an improved classification combining histopathology and mutation analysis (10). About 15% of Spitz nevi harbor an activating mutation (1, 10-12) of this oncogene which is part of the MAPK signaling transduction pathway (13, 14). Until now, no HRAS mutations have been found in spitzoid melanomas (1, 10-12) and so far no malignant progression of a Spitz tumor with a HRAS mutation has been described. Presence of a HRAS mutation in a Spitz tumor therefore seems to point to a good prognosis and/or a good clinical outcome (10, 11). HRAS mutation analysis therefore seems a prime candidate as an adjunct in the differential diagnosis of benign versus malignant Spitz tumor. The use of molecular techniques for the diagnosis of Spitz tumors is not yet widely used in common daily practice. In the past years, we have performed HRAS, BRAF, and NRAS mutation analysis on a large group of Spitz lesions (93 Spitz nevi and 77 STUMPs). After integration of histological and molecular data and review of all HRAS mutated cases in our series, we noticed remarkable histomorphological similarities between all HRAS mutated Spitz tumors. Although this was previously briefly described by Bastian et al. (10), in our opinion most pathologists are unfamiliar with these rather distinct features in HRAS mutated Spitz tumors. Improved recognition and increased knowledge of these features may significantly improve the accuracy of the diagnosis. Therefore, in this article we will first highlight these features in a group of 24 HRAS mutated Spitz tumors. Secondly, we hope to point to the role direct molecular genetic testing could play in reaching a correct diagnosis and a sound prediction of outcome for the individual patient. 3 47

49 Chapter 3 MATERIALS AND METHODS Patients and histopathology For this retrospective study, formalin fixed and paraffin embedded (FFPE) tissues of 24 HRAS mutated Spitz tumors were reviewed. The larger part of the material from our consultation practice had come from several pathology departments in The Netherlands. In the majority of cases the initial diagnosis had been made by a general pathologist, but in a few cases it had been reported by another expert dermatopathologist (cases 4, 11 and 15). The final histological diagnosis was made by a panel of three expert dermatopathologists (WB, MvD, DR), blinded to the initial diagnosis and to the results of mutation analysis. Molecular analysis DNA extraction About three manually dissected sections of 10 µm FFPE tissue with an estimated tumor cell percentage of at least 70% were used for DNA extraction. After deparaffinization and rehydration, the tissue sections were incubated in proteinase K, followed by subsequent affinity-purification of the DNA (QIAGEN GmbH, Germany). DNA sample concentration and quality were assessed spectrophotometrically (260/280 nm using a NanoDrop spectrophotometer, Peqlab Biotechnologies, Erlangen, Germany) and by PCR-amplification using the BIOMED-2 control gene primer set (15). All extracted DNAs allowed amplification of at least the 200 bp amplicon of the control gene PCR. Mutation analysis Direct sequence analysis of the HRAS gene was performed on all samples. For the BRAF and NRAS genes this was performed on 22 of the samples. For 2 lesions this could not be performed, due to lack of tissue. Exon 3 of HRAS and NRAS and exon 15 of BRAF were sequenced. These are known hotspot regions for oncogenic mutations in melanocytic lesions of the skin (1, 16-19). The sequences of the primers Table 1 Primers used for mutation analyses Gene Exon Forward (Fw) Reverse (Rv) Primer sequence 5-3 BRAF 15 Fw CCTTTACTTACTACACCTCAG Rv AAAAATAGCCTCAATTCTTAC NRAS 3 Fw GATTCTTACAGAAAACAAGTGG Rv TAATGCTCCTAGTACCTGTAGAG HRAS 3 Fw CTGCAGGATTCCTACCGGA Rv ACTTGGTGTTGTTGATGGCA 48

50 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features used are listed in Table 1. All primers contained a M13 forward or reverse consensus sequence for sequencing the different exons. PCR amplification of exon 15 of BRAF and exon 3 of NRAS and HRAS was performed in a total volume of 20 µl. PCR mix containing 50 ng DNA, buffer IV (Integro), 3 mm MgCl 2, 200 µm of each deoxynucleotide triphospate, 30 µg bovine serum albumin (Sigma), 10 pmol of each primer and 0.25 units of thermostable DNA polymerase (Sigma). DNA amplification was performed in a PTC 200 Thermal Cycler (MJ Research). The PCR was started with 5 min at 94 C and followed with 30 cycles of denaturation 45 seconds at 94 C, annealing at 60 C for 45 seconds and extension at 72 C for 45 seconds, with a final extension at 72 C for 5 minutes. All PCR products were purified with MinElute plates (Qiagen). One µl of the PCR product was used for the sequence reaction on an ABI PRISM 3700 DNA analyzer (Applied Biosystems). Both strands were sequenced using the M13 primers. 3 RESULTS Clinical features Of the 24 patients, 14 were females and 10 were males. There was no predilection site (6 lesions were localized in the head and neck region, 2 on the trunk, 1 on the buttock, 5 on the upper leg, 5 on the lower leg/foot and 3 on the distal forearm/hand. From two patients we could not retrieve information regarding the exact location). The ages ranged from 8 to 59 years (mean age: 25 years, median age: 22.5 years). At our referral center we classified 17 cases based on histopathological features alone as a Spitz nevus, and 7 cases as a STUMP. The initial diagnosis made by the original pathologist however was in 3 cases melanoma (cases 4, 11, and 15), and in 4 cases melanoma was included in the differential diagnosis (cases 6, 14, 16, and 23). In the majority of cases these initial diagnoses were made by general pathologists, but a few cases (cases 4, 11 and 15) had already been sent for consultation to an expert dermatopathologist. Patient characteristics are given in Table 2. Histological features All lesions had both architectural and cellular features consistent with a diagnosis of a Spitz tumor. The lesions were symmetrical and sharply demarcated at the margins and all consisted largely of spindle and/or epithelioid cells with large vesicular nuclei containing prominent eosinophilic nucleoli and abundant compact, eosinophilic cytoplasm. The most important criterion for diagnosing STUMP (i.e., refraining from making a confident diagnosis of Spitz nevus) or for diagnosing melanoma was the presence of more than 2 mitoses /1 mm 2 and the presence of deep mitoses, especially in patients 49

51 Chapter 3 Table 1 Patient characteristics Pat Sex Age Location Size (mm) Clark level Mitoses Mitoses/ 2 mm 2 1 F 19 head/ neck < 10* F 18 buttock < 12* M 17 trunk NA F 12 lower leg/ foot F 42 lower leg/ foot < 14* F 22 head/ neck NA NA 1-7 M 10 head/ neck M 13 NA M 11 NA F 8 head/ neck NA M 39 upper leg F 22 distal forearm/ hand NA F 16 upper leg NA M 34 lower leg/ foot NA F 26 lower leg/ foot M 59 upper arm NA M 28 upper leg M 8 head/ neck < 10* F 27 distal forearm/ hand NA F 47 upper leg NA M 29 head/ neck F 32 trunk F 32 upper leg F 23 lower leg NA= not available; Sex: F= female, M= male; : 1= no mitoses, 2= only superficial and not atypical, 3= atypical and/or deep; : dermal number of mitotic figures per 2 mm 2 : - = no mitoses, += < 2, ++= 2-5, +++= > 5; * size of the excision; : biopsy; dd= differential diagnosis; Spitz N= Spitz nevus, STUMP= Spitz tumor of unknown malignant potential; HRAS mutation: 1= c.181a>t (p.gln61leu), 2= c.181a>g (p.gln61arg), 3= c.182_183 delins TA CAG>CTA, 4= Gln61His c.183 G>T CAG>CAT, 5= c.181_182 delins AG CAG>AGG, 6= c.182_183 delins GA CAG>CGA 50

52 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features Initial diagnosis Final Diagnosis Follow-up (years) Spitz N Spitz N 11 1 Spitz N Spitz N 11 3 Spitz N STUMP 10 1 Spitzoid melanoma STUMP 9 1 Spitz N STUMP 8 2 STUMP dd Spitz N and melanoma Spitz N 7 2 Spitz N Spitz N 18 4 Spitz N Spitz N 13 1 Spitz N Spitz N 11 2 Spitz N Spitz N 12 5 Spitzoid melanoma STUMP 9 2 Spitz N dd STUMP STUMP NA 1 No diagnosis Spitz N 4 1 Spitz N dd melanoma Spitz N 4 2 Spitzoid melanoma Spitz N 8 2 Spitz N dd nevoid melanoma STUMP 3 2 Spitz N STUMP 20 6 Spitz N Spitz N 18 1 Spitz N Spitz N 16 1 Spitz N Spitz N NA 1 Spitz N Spitz N 2 1 Spitz N Spitz N 16 1 Spitz N dd melanoma Spitz N 9 1 Spitz N Spitz N 13 1 HRAS mutation (type) 3 51

53 Chapter 3 Table 3 Histologic features of HRAS mutated Spitz tumors Histologic feature Number of lesions Predominantly intradermal localization n= 24 (24/24= 100%) Relatively low cellularity n= 10 (10/24= 42%) Single cells between collagen bundles n= 23 (23/24= 96%) Desmoplasia n= 21 (21/24= 88%) Infiltrative base n= 22 (22/23*= 96%) * in one case the base could not be examined due to extensive positive resection margins over 17 years of age. Deep mitoses were defined as mitoses in the deep half and/or at the base of the lesion. Mitoses were quantified by searching for a hotspot in the dermis and counting 1 mm 2. The tumors were all predominantly intradermal with no or very little epidermal involvement, had a relatively low cellularity and showed desmoplasia with single melanocytic cells interspersed between the dermal collagen bundles. Where this growth pattern continued into the base of the lesion, the term infiltrative growth pattern was applied. An overview of the distinctive histologic features is given in Table 3. It is illustrated in Figure 1. Finally, the lesions contained very little or no melanin pigment and cells had vesicular nuclei: these features were not considered distinctive since they are quite common in Spitz tumors of all types. Molecular findings Six different mutations were found in exon 3 of HRAS. There was no correlation between the type of mutation and the phenotype of the lesion; the different types of mutation were associated with the same histomorphologic changes. In 21 of the lesions BRAF exon 15 and NRAS exon 3 were also tested, but no mutations were found. Follow-up Follow-up data was retrieved from the Dutch nationwide histopathology and cytopathology data network and archive (PALGA) (20). Follow-up data was available in 22 of the 24 cases; median and mean follow-up were both 10.5 years (range 2 to 20 years). All patients had remained free of disease. 52

54 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features A B 3 C D E F G H Figure 1 Four HRAS mutated Spitz tumors. Note the architectural similarities with the mainly intradermal localisation and the relatively low cellularity. A, B: case 2; Spitz nevus. B shows the desmoplasia with single cells between the collagen bundles. C, D: case 14; Spitz nevus, initial differential diagnosis melanoma. D shows the infiltrative base of the lesion. E, F: case 5; spitzoid tumor of unknown malignant potential (STUMP). F shows the desmoplasia. G, H: case 11; STUMP, the initial diagnosis was spitzoid melanoma. There are multiple dermal mitoses (H, arrows). 53

55 Chapter 3 DISCUSSION In this article we describe the distinct histological features of HRAS exon 3 mutated Spitz tumors. Meticulous evaluation of the histomorphological features of these lesions remains of great importance to help avoid misdiagnosis. Awareness of the characteristic features described in this article and mutation analysis may contribute to this. This is especially important since it is in general hard to predict the outcome of an atypical Spitz tumor in an individual patient (4). However, the presence of a HRAS mutation seems to be associated with a favorable prognosis: so far, none of the patients with HRAS mutated Spitz tumors developed recurrent disease, and no HRAS mutations have been found, or previously reported, in spitzoid melanomas with distant metastases or fatal outcome. Follow-up over 6-8 years of the 12 cases with 11p copy number increase that Bastian et al. (10) described, showed no evidence of metastases. Da Forno et al. (11) performed HRAS mutation analysis in 21 Spitzoid melanomas and no mutations were found. In addition, these workers did not find HRAS mutations in 22 non-spitzoid melanomas. They stated that, based on the known HRAS mutation frequency of approximately 15% in Spitz nevi and the absence of HRAS mutations in spitzoid melanomas, it would seem unlikely that HRAS mutated Spitz nevi should progress into spitzoid melanomas. In a series of 36 Spitzoid melanomas Van Dijk et al. (1) found no HRAS mutations. They identified HRAS mutations in 4/14 (29%) cases of Spitz nevi, and in 3/22 (14%) cases of atypical Spitz nevi. Follow-up of 7.8 years and 6 years, respectively, showed no evidence of recurrence or metastases in lesions harboring a HRAS mutation. Takata et al. (21) found no HRAS mutations in 22 non-spitzoid melanomas. However, they did not find any HRAS mutations in 11 typical Spitz nevi and 16 atypical Spitz tumors, either, which is in contrast to the other studies referred to here. All the studies that have determined the HRAS mutation status are consistent that no HRAS mutations are found in spitzoid melanomas and so far no metastases have been reported for Spitz tumors with a HRAS mutation. This was also the case in our series, in which none of the patients developed metastases, even after many years of follow-up. Detection of a HRAS mutation therefore seems to correlate with benign biological behavior. An overview of the mentioned studies is provided in Table 4. In older patients especially, where the a priori chance of a melanoma is higher, Spitz nevi are at risk of being misdiagnosed as a melanoma. In the present series of 24 cases of HRAS mutated Spitz tumors, in 7 cases (29%) the initial diagnosis was melanoma or melanoma was an important differential diagnostic consideration. We ve illustrated that in the HRAS mutated Spitz tumors, even in higher age groups, mitoses are frequently seen; in 5 cases (1, 11, 12, 16, 17) more than 2 dermal mitoses/ mm 2 and/or deep mitoses were seen (21% of cases). Deep mitoses were defined as 54

56 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features Table 4 Comparison of HRAS mutation data and follow-up in published studies of Spitz tumors Study Spitz N Atypical Spitz N STUMP Spitzoid melanoma Bastian et al 2000 (10) 8/ Van Dijk et al 2005 (15) /36-4/ / Takata et al 2007 (21) 0/11 0/16-0/22 - Da Forno et al / (11) Present study Follow-up (years)* 3 Spitz N= Spitz nevus; STUMP= Spitz tumor of unknown malignant potential. * Follow up in years. No recurrences or metastases were reported. mitoses present in the lower half and/or at the base of a lesion. If one is unfamiliar with the distinct histopathological features of HRAS mutated Spitz tumors, characterized by a predominantly intradermal localization, a relatively low cellularity, desmoplasia and single cells between the collagen bundles, an infiltrative base and no or very little melanin pigment, the risk of misdiagnosis is probably very significant. Bastian et al. (10) already described the characteristic histopathological appearance in 12 Spitz nevi with 11p copy number increase involving the HRAS gene. In 8 of these 12 lesions a HRAS mutation was found. Consistent with their data we found that these lesions were mainly intradermal, showed single cells, often with vesicular nuclei, lying between the collagen bundles and desmoplasia. Bastian and colleagues described eosinophilic membranes and marked nuclear pleomorphism. Although we could not identify such eosinophilic membranes, some cells were surrounded by eosinophilic material, which we interpreted as the presence of a small amount of collagen around the cells. We also did not consider the nuclear pleomorphism to exceed that normally seen in Spitz nevi. The reported mutation frequencies in Spitz tumors vary considerably between different studies, not only for HRAS, but also for BRAF and NRAS as reviewed by Da Forno et al. (17). Possible explanations are lack of consistency of the diagnostic criteria and variations in age distribution. Moreover, there might have been a difference in the percentage of tumor cells in the lesions used for DNA isolation. In most of the studies this is not specified, while a reliable PCR and mutation analysis requires a tumor cell percentage of at least 50%; a lower percentage might lead to false negative 55

57 Chapter 3 results. This requirement of a high tumor percentage is at present an important limitation for the diagnostic use of HRAS and BRAF mutation analysis by traditional PCR and sequence analysis in melanocytic tumors. A practical consequence for the general pathologist is that there should be enough tumor material left when a case is sent for consultation. Another limitation is the relative low frequency of HRAS mutations in Spitz tumors. Therefore only a small proportion of the lesions can be classified based on the HRAS mutation status. Perhaps more sensitive techniques, such as high resolution melting assay (HRMA), may allow an increase in the number of Spitz tumors that can be analyzed and reveal that more Spitz tumors contain mutations than presently thought. We expect that, in the near future, classification of melanocytic tumors and more specifically of Spitz tumors will combine morphology and genetic data, and that this integrated classification will allow for a more objective approach to predict the biological behavior of individual tumors. Despite the limitations of the present technique of HRAS mutation analysis by PCR and sequencing, it does seem a most rewarding step towards a more reproducible and objective classification in a subset of Spitz tumors. We therefore advocate that mutation analysis for HRAS be performed, preferentially together with BRAF in case a diagnosis of STUMP is considered since this is more common in spitzoid melanomas and very rarely found in Spitz nevi (22). Acknowledgement This article was supported by the Dutch Cancer Society (grant no. KUN ). The authors thank Dr L.M.Th. Sterk (Diaconessenziekenhuis Leiden), Dr P.Th.G.A. Nooijen (Jeroen Bosch Medical Center Den Bosch), Dr J.C. den Hollander (Erasmus Medical Center Rotterdam) and Dr A.M.W. van Marion (Vie Curie Medical Center Venlo) for kindly providing the Spitz tumors. We thank Dr S. Dubois (Meander Medical Center Amersfoort) for critically reviewing the manuscript and Dr J. van der Laak (Radboud University Nijmegen Medical Center) for improving the figure. 56

58 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features REFERENCES 1. van Dijk MC, Aben KK, van Hees F, Klaasen A, Blokx WA, Kiemeney LA, et al. Expert review remains important in the histopathological diagnosis of cutaneous melanocytic lesions. Histopathology. 2008; 52(2): Barnhill RL. The Spitzoid lesion: rethinking Spitz tumors, atypical variants, Spitzoid melanoma and risk assessment. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2006;19 Suppl 2:S Barnhill RL. The spitzoid lesion: the importance of atypical variants and risk assessment. The American Journal of dermatopathology. 2006;28(1): Barnhill RL, Argenyi ZB, From L, Glass LF, Maize JC, Mihm MC, Jr., et al. Atypical Spitz nevi/tumors: lack of consensus for diagnosis, discrimination from melanoma, and prediction of outcome. Human pathology. 1999;30(5): Evans AT. Spitz tumours. Diagn Histopathol. 2008;14: Mooi WJ, Krausz T. Spitz nevus versus spitzoid melanoma: diagnostic difficulties, conceptual controversies. Advances in anatomic pathology. 2006;13(4): Spatz A, Calonje E, Handfield-Jones S, Barnhill RL. Spitz tumors in children: a grading system for risk stratification. Archives of dermatology. 1999;135(3): van Dongen JJ, Langerak AW, Bruggemann M, Evans PA, Hummel M, Lavender FL, 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(12): Cerroni L, Barnhill R, Elder D, Gottlieb G, Heenan P, Kutzner H, et al. Melanocytic tumors of uncertain malignant potential: results of a tutorial held at the XXIX Symposium of the International Society of Dermatopathology in Graz, October The American journal of surgical pathology. 2010;34(3): Bastian BC, LeBoit PE, Pinkel D. Mutations and copy number increase of HRAS in Spitz nevi with distinctive histopathological features. The American journal of pathology. 2000;157(3): Da Forno PD, Pringle JH, Fletcher A, Bamford M, Su L, Potter L, et al. BRAF, NRAS and HRAS mutations in spitzoid tumours and their possible pathogenetic significance. The British journal of dermatology. 2009;161(2): Gill M, Cohen J, Renwick N, Mones JM, Silvers DN, Celebi JT. Genetic similarities between Spitz nevus and Spitzoid melanoma in children. Cancer. 2004;101(11): McCubrey JA, Steelman LS, Chappell WH, Abrams SL, Wong EW, Chang F, et al. Roles of the Raf/MEK/ ERK pathway in cell growth, malignant transformation and drug resistance. Biochimica et biophysica acta. 2007;1773(8): Russo AE, Torrisi E, Bevelacqua Y, Perrotta R, Libra M, McCubrey JA, et al. Melanoma: molecular pathogenesis and emerging target therapies (Review). International journal of oncology. 2009;34(6): van Dijk MC, Bernsen MR, Ruiter DJ. Analysis of mutations in B-RAF, N-RAS, and H-RAS genes in the differential diagnosis of Spitz nevus and spitzoid melanoma. The American journal of surgical pathology. 2005;29(9): Barbacid M. ras genes. Annual review of biochemistry. 1987;56: Da Forno PD, Fletcher A, Pringle JH, Saldanha GS. Understanding spitzoid tumours: new insights from molecular pathology. The British journal of dermatology. 2008;158(1): Hocker T, Tsao H. Ultraviolet radiation and melanoma: a systematic review and analysis of reported sequence variants. Human mutation. 2007;28(6): Omholt K, Karsberg S, Platz A, Kanter L, Ringborg U, Hansson J. Screening of N-ras codon 61 mutations in paired primary and metastatic cutaneous melanomas: mutations occur early and persist throughout tumor progression. Clinical cancer research : an official journal of the American Association for Cancer Research. 2002;8(11):

59 Chapter Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, et al. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cellular oncology : the official journal of the International Society for Cellular Oncology. 2007;29(1): Takata M, Lin J, Takayanagi S, Suzuki T, Ansai S, Kimura T, et al. Genetic and epigenetic alterations in the differential diagnosis of malignant melanoma and spitzoid lesion. The British journal of dermatology. 2007;156(6): Blokx WA, van Dijk MC, Ruiter DJ. Molecular cytogenetics of cutaneous melanocytic lesions - diagnostic, prognostic and therapeutic aspects. Histopathology. 2010;56(1):

60 HRAS-mutated Spitz tumors: A subtype of Spitz tumors with distinct features 3 59

61

62 4 Whole-genome copy-number analysis identifies new leads for chromosomal aberrations involved in the oncogenesis and metastastic behavior of uveal melanomas Adriana C. van Engen-van Grunsven a, *, Marjolein P. Baar a, *, Rolph Pfundt b, Jos Rijntjes a, Heidi V. Küsters-Vandevelde c, Ann-Laure Delbecq d, Jan E. Keunen d, Jeroen B. Klevering d, Pieter Wesseling a,c,e, Willeke A. Blokx a and Patricia J. Groenen a *Both authors contributed equally to this work; shared first authorship. a Department of Pathology, Radboud University Medical Center, Nijmegen, b Radboud University Medical Center, Nijmegen, c Department of Pathology, Canisius Wilhelmina Hospital, Nijmegen, d Department of Ophthalmology, Radboud University Medical Center, Nijmegen, e Department of Pathology, VU University Medical Center, Amsterdam. Melanoma Res 2015;25:

63 Chapter 4 ABSTRACT To further elucidate the genetic underpinnings of uveal melanoma (UM) and identify new markers that correlate with disease outcome, archival formalin-fixed paraffin embedded enucleation specimens from 25 patients with UM and a mean follow-up of 14 years were analyzed for whole genome copy number alterations using OncoScan analysis. Copy-number alterations of chromosomes 1, 3, 6 and 8 were also analyzed in these tumors using multiplex ligation-dependent probe amplification, and mutations in GNAQ, GNA11 and BAP1 were searched for by Sanger sequencing. Our study confirms the previously reported GNAQ and GNA11 mutation frequencies in UMs as well as the presence of monosomy 3 as a factor strongly indicating poor prognosis. Two cases with metastatic disease, but without monosomy of chromosome 3, showed loss of a small region in the distal part of chromosome 2p. Also, UMs leading to metastatic disease had more chromosomal aberrations than those without metastases. Three UMs lacking a GNAQ or GNA11 mutation showed a gain of chromosome 8q; one of these cases demonstrated extensive chromothripsis. Another case (with suspect lung metastasis) showed focal chromothripsis. Our whole genome copy-number analysis shows that focal loss of chromosome 2p may be involved in metastatic spread of UMs without monosomy 3; metastatic UMs carry more chromosomal aberrations than those without metastases; and chromothripsis may play a role in the oncogenesis of Ums, but does not necessarily indicate a poor prognosis. The clinical and particularly diagnostic utility of these findings needs to be corroborated in a larger set of patients with UM. 62

64 Whole-genome copy-number analysis of uveal melanomas INTRODUCTION Uveal melanomas (UMs) are the most common primary intraocular malignancy in adults, with an incidence of 6 per million population in the USA (1). Approximately 50% are fatal, usually as a result of metastatic disease (2) for which, even nowadays, there is no effective treatment. In the past, prognostic factors employed to indicate patients at risk of metastatic disease included clinical and histological features, such as tumor diameter and location (2,3), cell type, closed connective tissue loops, and mitotic rate (3,4). Like in other neoplasms, genetic alterations have been proven to play an important role in the development and progression of UM. Over the past 15 years, it has become clear that these genetic alterations are better predictors of prognosis than the traditional clinicopathological factors (5,6). Recently, gene expression profiling has shown that UMs can be divided into two distinct molecular subgroups with different prognoses: class 1 lesions, with a low risk of metastasis, and class 2 lesions, with a high risk of metastatis (5,7). In most of the cases with metastasis (84%), there is a monosomy of chromosome 3, which is often associated with an inactivating mutation in the tumor-suppressor gene BRCA1- associated protein-1 (BAP1, a ubiquitin carboxyterminal hydrolase) (8) located on the remaining chromosome 3 (9). BAP1 mutations are considered to be late events that coincide with the onset of metastatic behavior (9). In contrast, GNAQ and GNA11 mutations at codon 209 (Q209) or 183 (R183) are thought to occur early in the pathogenesis of UMs (10). GNAQ and GNA11 mutations have been described in 45% and 32% of UMs, respectively (10-14). Both GNAQ and GNA11 encode for heterotrimeric guanine nucleotide-binding protein (G protein) αq subunits that mediate signals between G-protein-coupled receptors and their downstream pathways (15) and are involved in determining skin and hair pigmentation (16,17). Mutations result in constitutive mitogen-activated protein (MAP) kinase activation, turning them into oncogenes. The GNAQ and GNA11 mutational status is not significantly correlated with disease-free or overall survival (11,14). However, not all UMs harbor a GNAQ or GNA11 mutation. In earlier studies, in 17% (14) and 9% (18) of UMs, respectively, no GNAQ or GNA11 mutation was found. In addition, not all cases of metastatic UM carry mutations in BAP1 (9) or show monosomy 3, suggesting that additional oncogenic events are involved in UM initiation and progression. In the present study, we performed a whole-genome screen of copy-number alterations using OncoScan analysis. This platform allows for analysis of over 335,000 probes with a 50- to 100-kb copy-number resolution in ~900 cancer genes and a 300-kb genome-wide copy-number resolution outside these cancer genes. In addition, we applied multiplex ligation-dependent probe amplification (MLPA) analysis. MLPA 4 63

65 Chapter 4 is a more targeted approach and routinely used in many pathology laboratories. In the present study, in particular, chromosomes 1p, 3, 6, and 8 were analyzed by MLPA. Both OncoScan and MLPA analysis can theoretically be exploited for assessment of DNA obtained from formalin-fixed paraffin-embedded (FFPE) tissue blocks. Finally, we also performed mutation analysis of GNAQ, GNA11, and BAP1 by Sanger sequencing in 25 primary UM cases. Results of the genomic analysis were correlated with clinicopathological characteristics and follow-up. We confirmed existing prognostic chromosomal aberrations and identified potential new mechanisms for UM initiation and chromosomal areas that may correlate with metastatic behavior. MATERIALS AND METHODS Patients and material For this retrospective study, FFPE enucleation specimens from 25 patients treated for primary UM between 1992 and 2007 at the Department of Ophthalmology of the Radboud University Medical Center in Nijmegen, the Netherlands, were included. Mean follow-up was 14 years. Follow-up data were available for 19 of the patients, as shown in Table 1. For four patients, follow-up was unknown, and for two patients, follow-up was uncertain (possible metastasis but not proven). As far as we know, none of the patients had a family history of UM. The FFPE samples were retrieved from our archives and the tissue blocks stored at room temperature. Hematoxylin and eosin-stained sections of all cases were reviewed (ACvEvG, PW) and all cases were scored with respect to important pathological factors, such as tumor location and size and cell type. The diagnosis of melanoma was made on hematoxylin and eosin-stained sections. Anonymized patient information and follow-up data until 2011 were obtained, and data assessment was performed according to the Human tissue and medical research: code of conduct for responsible use ( sites/default/files/digital_version_first_part_code_of_conduct_in_uk_2011_ pdf). DNA extraction and sequence analysis DNA extraction and sequence analysis were performed as previously described (19). The DNA quality of the samples was tested using the BIOMED-2 gene control polymerase chain reaction (PCR) (see Table 1). Mutation detection was performed for hotspot mutations in exon 5 (codon Q209) and exon 4 (codon R183) in GNAQ and GNA11 (accession number (no.) NT_ and NG_ , respectively) and for all exons of BAP1 (accession no.: NC_ ) for all cases. In selected cases, exon 15 of BRAF (accession no.: NG_ ) and exons 2 and 3 of NRAS (accession 64

66 Whole-genome copy-number analysis of uveal melanomas no.: NG_ ) were analyzed for hotspot mutations. Primer sequences are listed in supplementary Table 1. In one of the cases with a BAP1 mutation (case UM 23), the presence of a germline mutation was excluded by testing normal DNA. In the other BAP1-mutated cases this could not be performed as the patients did not give permission for testing for germline mutations and/or there was a lack of normal DNA. GNAQ and GNA11 mutation analysis has only been performed in tumor material since no germline mutations have been described in these genes in patients with UM (20). Analysis of copy-number variations OncoScan analysis Copy-number alterations were determined by OncoScan array analysis (molecular inversion probe (MIP) technology) (essentially as described by Wang et al. (21)). This array consists of more than 335,000 markers, which cover the entire genome with high resolution of cancer genes. DNA extracted from the lesions was processed by the Affymetrix Research Services Laboratory (ARSL, Santa Clara, CA, USA). The normalized OncoScan data were analyzed with the Nexus copy-number software version 5.1 (Biodiscovery, Inc, El Segundo, CA, USA) for detection and analysis of copy-number variation. 4 MLPA procedure and data analysis To evaluate copy-number changes, we used MLPA kit P027-B1 (MRC-Holland, Amsterdam, The Netherlands) containing probes on chromosomes 1p (seven probes), 3p (nine probes) and 3q (four probes), 6p (four probes) and 6q (two probes), and 8p (one probe) and 8q (four probes). Twelve other probes were used as reference probes ( Immortalized cell lines of healthy controls were used as normal control samples. MLPA was performed according to the manufacturer s instructions. RESULTS Patients and material Patient and tumor characteristics are listed in Table 1. Of the 25 patients, 12 were female and 13 male. Mean age of the patients was 59 years. Twenty tumors were located at the choroid, four at the ciliary body, and one was ciliochoroidal. The mean diameter of the tumors was 1.3 cm. Cell type was either epithelioid (four cases), spindle (nine cases) or mixed (twelve cases). 65

67 Chapter 4 Table 1 Overview of the different cases Case nr gender age tumor location tumor diameter (cm) cell type follow up neoplastic cell load (%) DNA quality (bp) GNAQ/ GNA11 hotspot mutation Chrom. 3 loss UM1 f 66 choroid 1,5 epithelioid liver metastasis 50 <100 GNA11 a + + d BAP1 mutation / % of gene evaluated UM2 m 54 choroid 1,0 epithelioid liver metastasis GNA11 a + + d UM3 f 62 choroid 1,5 epithelioid unknown GNAQ b + + d UM4 m 56 choroid 1,5 mixed liver metastasis GNA11 a + not tested e UM5 f 38 ciliochoroidal 1,6 spindle liver metastasis 70 <100 GNA11 a + not tested e UM6 m 62 choroid 1,6 mixed pancreas carcinoma, metastasis unknown GNAQ c + -/ 45% UM7 m 64 choroid 1,7 mixed liver metastasis 70 <100 GNA11 a + -/ 90% UM8 m 61 ciliary body 1 mixed liver metastasis GNA11 a + -/ 50% UM9 m 66 choroid 0,9 epithelioid liver metastasis GNA11 a + -/ 20% UM10 f 70 choroid 1,3 spindle liver metastasis <100 GNAQ c + -/ 45% UM11 f 58 ciliary body 1,5 mixed lung metastasis? 70 <100 GNAQ c + -/ 50% UM12 f 60 choroid 1,2 mixed unknown GNAQ c + -/ 85% UM13 m 25 choroid 2,0 spindle no metastasis GNA11 a - -/ 20% UM14 f 62 choroid 0,8 spindle no metastasis GNAQ b - -/ 30% UM15 m 52 ciliary body 1,0 spindle liver metastasis, GNA11 a - -/ 90% prostate carcinoma UM16 m 54 choroid 1,0 spindle prostate carcinoma, GNAQ c - -/ 85% no metastasis UM17 m 60 choroid 1,0 spindle unknown GNAQ b - -/ 90% 66

68 Whole-genome copy-number analysis of uveal melanomas Table 1 Continued BAP1 mutation / % of gene evaluated Chrom. 3 loss GNAQ/ GNA11 hotspot mutation DNA quality (bp) cell type follow up neoplastic cell load (%) tumor diameter (cm) gender age tumor location Case nr GNA11 a - -/ 85% UM18 f 68 choroid 0,9 mixed melanoma left shoulder; metastasis?, carcinoma breast and lung. UM19 m 28 choroid 1,6 mixed liver metastasis 90 <100 GNAQ c - -/ 85% UM20 f 70 choroid 1,5 mixed pancreas carcinoma GNAQ c - -/ 90% UM21 m 26 choroid 1,8 spinde no metastasis GNAQ b - -/ 50% UM22 f 55 choroid 1,3 spindle no metastasis GNA11 a - -/ 50% UM23 f 70 choroid 1,1 mixed liver metastasis d UM24 m 57 ciliary body 1,5 mixed liver metastasis / 5% e UM25 f 68 choroid 1,2 mixed no metastasis not tested e Gender: m= male; f= female; age in years. The tissues that were used for DNA-extraction were histologically scored to determine the percentage of neoplastic cells. The quality of the DNA that was extracted was determined by control gene PCR that amplifies fragments of 100, 200, 300 and 400 base pairs (bp) in one PCR. The number of the bp presented represents the largest fragment that could be amplified and is thus a measure of the DNA quality. The? at case UM11 indicates that the probable lung metastases were not histologically proven. + indicates there is a loss or a mutation while indicates there is no loss or mutation. The mutational status of GNAQ and GNA11 codons 209 and 183 were tested. Mutations were only found at codon 209. Mutation types: a GNA11 c.626a>t (p.(gln209leu)). b GNAQ c.626a>t(p. (Gln209Leu)). c GNAQ c.626a>c(p.(gln209pro)). In the three cases in which no GNAQ or GNA11 mutation was found, BRAF exon 15 and NRAS exon 2 and 3 mutations were also tested, but no mutations were found. The chromosome 3 data of the MLPA were consistent with those found by the OncoScan analysis. All seventeen exons of BAP1 were analyzed by Sanger sequencing. Evaluation of the sequence was dependent on the DNA quality of the sample and the amount of DNA available. The percentage of exons that could be evaluated is given in the table. e indicates insufficient amount of DNA. d BAP1 mutations: UM1: exon 6 c.422a>g (p.(his141arg)); UM2: exon 12 c.1177c>t (p.(gln393x)); UM3: exon 15 c.1924delg (p.(glu642argfs*13)); UM23: c.117_122+6del (p.(gln40profs*27)) 4 67

69 Chapter 4 Mutation analysis All 25 primary UMs were tested for hotspot mutations in GNAQ and GNA11. The results are summarized in Table 1. Twenty-two cases (88%) had either a GNAQ or GNA11 mutation (n = 11 for GNAQ as well as for GNA11). These mutations were mutually exclusive and included only exon 5 (Q209) mutations. The three UMs without GNAQ or GNA11 hotspot mutations were analyzed for mutations in the entire exons 4 and 5 of GNAQ and GNA11 by Sanger sequencing, as well as for mutations in exon 15 of BRAF, and exons 2 and 3 of NRAS. However, no mutations in these exons were observed. BAP1-mutation analysis was performed on all samples. Unfortunately, the DNA extracted was suboptimal for proper Sanger sequencing of BAP1: appropriate evaluation of all 17 exons of BAP1 in these FFPE-derived DNA specimens would require design and analysis of small (<100 bp) amplicons for sequence analysis and a substantial amount of DNA per sample, which was not feasible. We therefore proceeded with our routine BAP1-Sanger sequencing protocol, which uses amplicons of bp, to analyze the BAP1 mutational status. The percentage of the exons that could be properly evaluated is given in Table 1. The majority of the BAP1 coding sequence could be evaluated by sequence analysis in the DNA specimens that could amplify 200 or 300 bp of the control gene PCR, while part of the BAP1 coding sequence could be evaluated in highly degraded cases that could only amplify 100 bp or less of the control gene PCR (see Table 1). We identified BAP1 mutations in four of the 15 UMs with monosomy 3 (26%). Copy-number variations by OncoScan analysis OncoScan analysis was successfully performed on all cases. Due to the technology of the molecular inversion probes that hybridize to a DNA target in two locations spanning a 40-bp footprint (21), even the suboptimal DNAs extracted from the relatively old FFPE samples were suitable for this analysis. The analysis gave a complete overview of copy-number variations of the entire genome. We found a difference in the number of copy-number variations between patients with and without metastatic disease. As illustrated in Figure 1, the 12 patients with known metastases (cases 1, 2, 4, 5, 7, 8, 9, 10, 15, 19, 23, and 24) had more chromosomal aberrations than the six cases without metastases (cases 13, 14, 16, 20, 21, and 22). The occurrence of aberrations of chromosome 1, 3, 6, and 8 were separately analyzed in relation to metastases. Seven cases were excluded from this analysis: five because of unknown or uncertain follow-up data (cases 3, 6, 11, 12, 17, and 18) (see Table 1) and one case (case 25) because a typical pattern of chromosomal imbalance was shown (chromothripsis; to be discussed later), which precludes appropriate analysis of the chromosomal copy-number alterations. Ten of the 12 patients with a metastasis showed monosomy of chromosome 3 (83%). In all of these cases, the entire 68

70 Whole-genome copy-number analysis of uveal melanomas chromosome 3 was lost. Furthermore, this loss was correlated with a loss of chromosome 1p and a gain of chromosome 8q, as is shown in Figure 1B. The six patients without metastases did not show monosomy 3, but all of these patients did have a gain of chromosome 6p (such a gain being present in four of the 12 patients with metastastic disease). Furthermore, loss of chromosome 6q was quite common in both groups (half of the cases in both the non-metastastic and the metastatic tumors), as well as a chromosome 8q gain (three and 11 cases in non-metastastic metastatic tumors, respectively). The minimal common overlapping regions in the non-metastatic and the metastatic tumors are indicated in Figure 1C. Two patients (cases 15 and 19) had metastatic disease, but no monosomy of chromosome 3. An overview of the OncoScan data of these two patients is shown in Figure 2. The tumors of both these patients showed loss of a small region on the distal part of chromosome 2p (2Mb, chr2: 1-2,356,641), which is only found in these two cases. Furthermore, in both UMs, a high copy-number gain of chromosome 8q, loss of chromosome 1p, gain of chromosome 6p, and loss of chromosome 6q were found, but this was also seen in other cases. As can be seen in Table 1, one of these cases had a GNAQ mutation and the other had a GNA11 mutation. In addition, these two patients differed substantially in age, the tumors had different locations and cell types, but otherwise these two tumors did not show any specific histological characteristics. The three cases in our series without a GNAQ or GNA11 mutation (and without hotspot BRAF or NRAS mutations) all showed a gain of chromosome 8q. Interestingly, one of these cases (case 25) showed alternating small gains and deletions that is typical of chromothripsis (21) in multiple chromosomes, notably chromosomes 1, 3, 6, and 8. The chromothripsis of chromosome 3 had been interpreted as a partial loss on the MLPA. Another case (case 11, with a GNAQ mutation) also showed chromothripsis, but only of chromosome 13 (see Figure 3). Besides the chromosomal changes described above, the OncoScan analysis showed that no other chromosomal abnormalities were shared in a large percentage of the patients. Our MLPA analysis revealed monosomy of chromosome 3 in 15 of the UMs, which is completely concordant with the results obtained by OncoScan analysis. Of note, in the MLPA platform we used, chromosome 3 is represented by 13 probes, allowing for relatively robust testing of chromosome 3 status. In contrast, MLPA analysis of copy-number variations in chromosomes 1p (seven probes), 6p (four probes), 6q (two probes), 8p (one probe) and 8q (four probes) in old archival FFPE material, as used in our study, often showed discordant results in flanking probes and was therefore considered to be insufficiently robust. 4 69

71 Chapter 4 A B 70

72 Whole-genome copy-number analysis of uveal melanomas C Gain/loss UM A, without or UM B, with metastasis Chromosomal region (hg19) Chromosome arm 6p 8,107,512-26,128,335 Gain UM A 6q 89,797, ,192,220 Loss UM A 8q 115,540, ,274,826 Gain UM A 1p 0-32,069,468 Loss UM B 3p 0-80,757,601 Loss UM B 6p 3,031,935-3,656,577 Gain UM B 6p 4,052,997-4,661,624 Gain UM B 6p 32,874,356-34,655,361 Gain UM B 6p 43,689,975-44,313,688 Gain UM B 6q 128,347, ,899,992 Loss UM B 8q 95,637, ,274,826 Gain UM B Figure 1 Comparison of the number of copy-number variations in uveal melanomas without (A) and with (B) metastasis and minimal common overlapping regions (C). OncoScan data of the six cases without metastasis (A) and of the 12 cases with metastasis (B). The upper parts of A and B show the aggregated gains and losses of the different cases, the height of the bars indicating the number of cases with the gain or loss. In the lower parts, the rows represent the different samples. Gains and losses of the different chromosomes are represented by blue and red lines, respectively, under the different chromosomes (depicted in columns) and the length of these lines indicates the size of the gain or loss. The yellow and purple lines indicate the loss of heterozygosity and genomic imbalance, respectively. Comparison of A and B shows that the cases with metastasis have more chromosomal aberrations than those without. The minimal common overlapping regions of gain or loss that are detected in at least three out of six UM cases without metastasis and in five out of 12 UM cases with metastasis are provided in C. 4 71

73 Chapter 4 A B Figure 2 OncoScan data of the two cases (UM15 (A) and UM19 (B)) with metastasis, but without monosomy 3, and the 10 cases with metastasis and monosomy 3 (C). The different chromosomes are shown with the location of the gains (in blue) and losses (in red). The two cases without monosomy 3 share chromosome 6 aberrations, loss of chromosome 1, and a high copy-number gain of chromosome 8q, aberrations also seen in cases with monosomy 3. Loss of a small region at the distal part of chromosome 2p (indicated by an arrow) is seen in both uveal melanomas without monosomy 3 (A and B) but not in those with monosomy 3 (C). The shorter region of loss of chromosome 2 in case UM19 (B) may narrow down the region of interest with regard to the location of a putative tumor-suppressor gene. 72

74 Whole-genome copy-number analysis of uveal melanomas C Figure 2 Continued. DISCUSSION 4 The present study was performed to further unravel genetic alterations involved in the initiation and progression of UM by performing a whole-genome screen of copy-number alterations by using OncoScan analysis, MLPA, and mutation analysis. Of the 25 UM cases included in this study, 11 (44%) had a GNA11 Q209 hotspot mutation and another 11 (44%) a GNAQ Q209 hotspot mutation. The three cases (12%) without GNAQ or GNA11 mutations (and without hotspot BRAF and NRAS mutations), all showed a gain of 8q. In addition, one of these cases (case 25) showed a very typical pattern of alternating small gains and deletions in multiple chromosomes that is typical of chromothripsis (22) (see Figure 3). To the best of our knowledge, chromothripsis has not been described in UM before. In chromothripsis, a single catastrophic event leads to fragmentation of specific regions of the genome, after which DNA repair mechanisms piece the fragments together in the wrong order. This results in a genomic landscape with multiple chromosomal rearrangements, which is thought to promote cancer development (23). It has been observed in various cancer types (22) and has been associated with a poor prognosis in multiple myeloma (24), acute myeloid leukemia (25), medulloblastoma (25,26), and recently in cutaneous melanoma (27). Interestingly, the chromosomes affected in this tumor are chromosomes 1, 3, 6, and 8, which also frequently show other aberrations in UMs. Furthermore, a GNAQ or GNA11 mutation was lacking in case 25 and the patient had not developed metastases after 17 years of follow-up. Less extensive chromothripsis involving only chromosome 13 (a chromosome not known to play a part in UM development or progression) was found in another case. We did not find any copy-number variations in chromosome 13 in the other cases. The role of chromo- 73

75 Chapter 4 A B Figure 3 OncoScan data of the two cases (UM11 (A) and UM25 (B)) with chromothripsis. The graphs show the typical pattern of many small deletions and gains. A shows chromothripsis of chromosome 3 in case UM25 (which showed chromothripsis of chromosomes 1 and 6 as well). B shows chromothripsis of chromosome 13 in case UM11 (i.e. the only chromosome affected by chromothripsis in this case). 74

76 Whole-genome copy-number analysis of uveal melanomas thripsis in UMs is intruiging, but not yet understood, and (as demonstrated by one of our cases) its presence does not necessarily signify a grim prognosis. The GNAQ and GNA11 mutation frequencies we found are in accordance with those reported in the literature. Van Raamsdonk et al. (14) reported GNA11 Q209 hotspot mutations in 32% of UMs and GNAQ Q209 hotspot mutations in 45%, whereas no GNAQ or GNA11 hotspot mutations were found in 17%. Daniels et al. (18) found GNA11 Q209 hotspot mutations in 43% of UMs and GNAQ Q209 hotspot mutations in 44%, and no GNAQ or GNA11 hotspot mutations in 9%. In these studies, R183 exon 4 hotspot mutations were reported in 6% and 4% of the cases, respectively. In our relatively small series, we did not find exon 4 hotspot mutations. Metastasis occurred in 27% of the patients with a GNAQ mutation and in 73% of the patients with a GNA11 mutation, corroborating the suggestion of previous studies that GNA11 mutations are correlated with a higher chance of metastatic disease (17,27,28,29). UMs with metastases showed more chromosomal aberrations, as assessed by OncoScan analysis, than those without (Figure 1). An association between the degree of chromosomal instability and disease outcome has also recently been described in cutaneous melanoma (30). However, in our series, there were, individual metastatic cases with relatively few aberrations and cases without metastases with relatively many aberrations, indicating that the number of aberrations cannot be used for determination of the prognosis in an individual case. Monosomy 3 was seen in 15 of the cases, which is in concordance with the literature. In all cases with monosomy 3 in our series, the whole chromosome 3 was lost. Loss of the 1p chromosomal region was seen in 65% of our cases with metastatic disease, but not in those without. Many genes are located on this chromosome arm, which makes it difficult to determine if a specific gene located here might be important for metastatic behavior. A gain of chromosome 8q was previously reported to be correlated with metastatic disease (31), but this aberration also occurred in patients without metastatic disease (see Figure 1). A gain of chromosome 6p has been associated with a good prognosis (32,33), and this is confirmed by our results. All patients without metastasis showed a gain of chromosome 6, while such a gain was only found in 33% (four out of 12) of the patients with metastatic disease. Two patients with metastatic disease, but without monosomy 3, showed, besides aberrations on chromosomes 1, 6, and 8, loss of an overlapping small region on the distal part of chromosome arm 2p. Interestingly, these are the only tumors with this loss. To our knowledge, no (tumor-suppressor) genes located on this specific part of chromosome 2 have yet been reported that could be involved in the development or progression of UMs. Recently, the presence of somatic mutations in SF3B1 (34) and EIF1AX (35) has been described in UM with disomy 3 in particular (48% and 29%, respectively, in a series of 31 UM cases with disomy 3 vs 5.7% in 35 UM cases with monosomy 3 (35)). Mutations in these genes are associated with a good prognosis, and are mutually exclusive with BAP1 4 75

77 Chapter 4 mutations. The SF3B1 gene is located at chromosome 2q33.1. EIF1AX is located at Xp22.1 (35). The knowledge of mutations in SF3B1 and EIF1AX can enhance patient prognosis beyond that determined by tumor characteristics, chromosome 3 status, and BAP1 mutation status (36). As discussed above, the combination of archival FFPE material (that had been stored for at least 14 years and of which often only PCR-amplicons of less than 100 bp could be retrieved) and the fact that for adequate assessment of BAP1 mutations, the 17 exons comprising of 2190 nt need to be interrogated allowed us to identify BAP1 mutations in only three of our cases with monosomy 3 (25%). Of note, it has been reported that 84% of UMs carry a BAP1 mutation (9). It should also be mentioned that, in this study, only enucleation specimens were used, which implies we have studied a subgroup of relatively large UMs since large UMs, in particular, are treated with enucleation. There might be tumor heterogeneity. Maat et al. (37) have studied the heterogeneity of monosomy 3 status in UMs and have shown that large tumors can have intratumor heterogeneity. Onken et al. (38), however, performed transcriptomic profiling of fine-needle biopsies from UMs and showed that there was little or no intratumor heterogeneity. In the present study, tumor heterogeneity was not addressed due to the lack of availability of multiple samples from the tumors. In conclusion, our mutation analysis of FFPE material from 25 primary UM cases confirmed previously reported results on GNAQ and GNA11 mutation frequencies in UM as well as on the presence of monosomy 3 as a factor strongly indicating poor prognosis. Additionally, whole-genome (OncoScan) copy-number analysis revealed that cases with metastases generally have more chromosomal aberrations than those without, and identified chromothripsis and chromosome 8q gain as potential mechanisms for initiation of UM. Furthermore, our results indicate that loss of a small region on chromosome 2p may be involved in the development of metastatic disease. The exact role of these genetic aberrations in the oncogenesis and/or metastatic spread of UMs needs to be further investigated in a larger study. Acknowledgements The authors thank S. Hendriks-Cornelissen, Department of Pathology, Radboud University Medical Center for technical assistance. 76

78 Whole-genome copy-number analysis of uveal melanomas REFERENCES 1 Shields CL, Shields JA. Ocular melanoma: relatively rare but requiring respect. Clin Dermatol 2009; 27: Kujala E, Makitie T, Kivela T. Very long-term prognosis of patients with malignant uveal melanoma. Invest Ophthalmol Vis Sci 2003;44: 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, Woolson RF, Jeng G, Montague PR 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: Onken MD, Worley LA, Char DH, Augsburger JJ, Correa ZM, Nudleman E et al. Collaborative Ocular Oncology Group Report Number 1: Prospective Validation of a Multi-Gene Prognostic Assay in Uveal Melanoma. Ophthalmology 2012;119: Gill HS, Char DH. Uveal melanoma prognostication: from lesion size and cell type to molecular class. Can J Ophthalmol 2012;47: Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res 2004;64: Jensen DE, Proctor M, Marquis ST, Gardner HP, Ha SI, Chodosh LA et al. BAP1: a novel ubiquitin hydrolase which binds to the BRCA1 RING finger and enhances BRCA1-mediated cell growth suppression. Oncogene 1998;16: Harbour JW, Onken MD, Roberson ED, Duan S, Cao L, Worley LA et al. Frequent Mutation of BAP1 in Metastasizing Uveal Melanomas. Science 2010;330: Onken MD, Worley LA, Long MD, Duan S, Council ML, Bowcock AM et al. Oncogenic mutations in GNAQ occur early in uveal melanoma. Invest Ophthalmol Vis Sci 2008;49: Bauer J, Kilic E, Vaarwater J, Bastian BC, Garbe C, de KA. Oncogenic GNAQ mutations are not correlated with disease-free survival in uveal melanoma. Br J Cancer 2009;101: Lamba S, Felicioni L, Buttitta F, Bleeker FE, Malatesta S, Corbo V et al. Mutational profile of GNAQQ209 in human tumors. PLoS One 2009;4:e Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, O Brien JM et al. Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature 2009;457: Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, Vemula S, Wiesner T et al. Mutations in GNA11 in Uveal Melanoma. N Engl J Med 2010;363: Neves SR, Ram PT, Iyengar R. G protein pathways. Science 2002;296: Van Raamsdonk CD, Barsh GS, Wakamatsu K, Ito S. Independent regulation of hair and skin color by two G protein-coupled pathways. Pigment Cell Melanoma Res 2009;22: Van Raamsdonk CD, Fitch KR, Fuchs H, de Angelis MH, Barsh GS. Effects of G-protein mutations on skin color. Nat Genet 2004;36: Daniels AB, Lee JE, Macconaill LE, Palescandolo E, Van Hummelen P, Adams SM et al. High throughput mass spectrometry-based mutation profiling of primary uveal melanoma. Invest Ophthalmol Vis Sci 2012;53: Kusters-Vandevelde HV, Klaasen A, Kusters B, Groenen PJ, van Engen-van Grunsven IA, van Dijk MR et al. Activating mutations of the GNAQ gene: a frequent event in primary melanocytic neoplasms of the central nervous system. Acta Neuropathol 2010;119: Hawkes JE, Campbell J, Garvin D, Cannon-Albright L, Cassidy P, Leachman SA. Lack of GNAQ and GNA11 Germ-Line Mutations in Familial Melanoma Pedigrees with Uveal Melanoma or Blue Nevi. Front Oncol Jun;3: Wang Y, Cottman M, Schiffman JD. Molecular inversion probes: a novel microarray technology and its application in cancer research. Cancer Genet. 2012;205: Stephens PJ, Greenman CD, Fu B, Yang F, Bignell GR, Mudie LJ et al. Massive genomic rearrangement acquired in a single catastrophic event during cancer development. Cell 2011;144:

79 Chapter 4 23 Forment JV, Kaidi A, Jackson SP. Chromothripsis and cancer: causes and consequences of chromosome shattering. Nat Rev Cancer 2012;12: Magrangeas F, vet-loiseau H, Munshi NC, Minvielle S. Chromothripsis identifies a rare and aggressive entity among newly diagnosed multiple myeloma patients. Blood 2011;118: Rausch T, Jones DT, Zapatka M, Stütz AM, Zichner T, Weischenfeldt J et al. Genome sequencing of pediatric medulloblastoma links catastrophic DNA rearrangements with TP53 mutations. Cell 2012;148: Northcott PA, Shih DJ, Peacock J, Garzia L, Morrissy AS, Zichner et al. Subgroup-specific structural variation across 1,000 medulloblastoma genomes. Nature 2012;488: Patel M, Smyth E, Chapman PB, Wolchok JD, Schwartz GK, Abramson DH et al. Therapeutic implications of the emerging molecular biology of uveal melanoma. Clin Cancer Res 2011;17: Ewens KG, Kanetsky PA, Richards-Yutz J, Purrazzella J, Shields CL, Ganguly TI et al. Chromosome 3 status combined with BAP1 and EIF1AX mutation profiles are associated with metastasis in uveal melanoma. Invest Ophthalmol Vis Sci Jun 26;55(8): Griewank KG, van de Nes J, Schilling B, Moll I, Sucker A, Kakavand H et al. Genetic and clinico-pathologic analysis of metastatic uveal melanoma. Mod Pathol Feb;27(2): Hirsch D, Kemmerling R, Davis S, Camps J, Meltzer PS, Ried T et al. Chromothripsis and focal copy number alterations determine poor outcome in malignant melanoma. Cancer Res 2013;73: Sisley K, Rennie IG, Parsons MA, Jacques R, Hammond DW, Bell SM et al. Abnormalities of chromosomes 3 and 8 in posterior uveal melanoma correlate with prognosis. Genes Chromosomes Cancer 1997;19: Damato B, Dopierala J, Klaasen A, van Dijk M, Sibbring J, Coupland SE. Multiplex ligation-dependent probe amplification of uveal melanoma: correlation with metastatic death. Invest Ophthalmol Vis Sci 2009;50: 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: Harbour JW, Roberson ED, Anbunathan H, Onken MD, Worley LA, Bowcock AM. Recurrent mutations at codon 625 of the splicing factor SF3B1 in uveal melanoma. Nat Genet 2013;45: Martin M, Maßhöfer L, Temming P, Rahmann S, Metz C, Bornfeld N et al. Exome sequencing identifies recurrent somatic mutations in EIF1AX and SF3B1 in uveal melanoma with disomy 3. Nat Genet. 2013;45(8): Field MG, Harbour JW. Recent developments in prognostic and predictive testing in uveal melanoma. Curr Opin Ophthalmol May;25(3): Review. 37 Maat W, Jordanova ES, van Zelderen-Bhola SL, Barthen ER, Wessels HW, Schalij-Delfos NE et al. The heterogeneous distribution of monosomy 3 in uveal melanomas: implications for prognostication based on fine-needle aspiration biopsies. Arch Pathol Lab Med Jan;131(1): Onken MD, Worley LA, Dávila RM, Char DH, Harbour JW. Prognostic testing in uveal melanoma melanoma by transcriptomic profiling of fine needle biopsy specimens. J Mol Diagn Nov;8(5):

80 Whole-genome copy-number analysis of uveal melanomas Supplementary Table 1 Primers used for mutational analysis Gene Exon Forward (Fw) Primer sequence 5-3 Reverse (Rv) BAP1 1 Fw TGTAAAACGACGGCCAGT CTGGACATGGCGCTGAG Rv CAGGAAACAGCTATGACC GGTGATGGTCAGGCAGG BAP1 2 Fw TGTAAAACGACGGCCAGT AGAGCGACCCAGGTGAG Rv CAGGAAACAGCTATGACC GCAGTAGGGAAGGACAGC BAP1 3 Fw TGTAAAACGACGGCCAGT CTCACTCATCAGGGGCTG Rv CAGGAAACAGCTATGACC CTGTTCTCTGGGACCTTCC BAP1 4 Fw TGTAAAACGACGGCCAGT TGATTGTCTTCTCCCCTTTG Rv CAGGAAACAGCTATGACC CAAAAACATGGCAGCATCC BAP1 5 Fw TGTAAAACGACGGCCAGT GCTTGCAGTGAGGGGTG Rv CAGGAAACAGCTATGACC GGTCAGATCTGCCCAGTTG BAP1 6 Fw TGTAAAACGACGGCCAGT CATAGTCCTACCTGAGGAG Rv CAGGAAACAGCTATGACC ATGATACTCCCCCTACTCC BAP1 7 Fw TGTAAAACGACGGCCAGT TGGGAGTAGGGGGAGTATC Rv CAGGAAACAGCTATGACC CCTAGGAGGTAGGCAGAG BAP1 8 Fw TGTAAAACGACGGCCAGT CCAAACTCAGGGTTTCCTTC Rv CAGGAAACAGCTATGACC CTCTCTGTCCCTCCCAAAG BAP1 9 Fw TGTAAAACGACGGCCAGT TATCTGCCTCAACCTGATG Rv CAGGAAACAGCTATGACC GAGGAGGAATGCAGGGAG BAP1 10 Fw TGTAAAACGACGGCCAGT AGCCCGGGTCTACCCTTTC Rv CAGGAAACAGCTATGACC TGTCAGACATTAGCGGGTG BAP1 11a Fw TGTAAAACGACGGCCAGT GGGAGACTGTGAGCTTTTC Rv CAGGAAACAGCTATGACC ATTGTCTAGAAAGGCCGGC BAP1 11b Fw TGTAAAACGACGGCCAGT GGTTCATGCGCACAAGC Rv CAGGAAACAGCTATGACC TTGCCTGTTGCAGCCTC BAP1 12 Fw TGTAAAACGACGGCCAGT CCGAGCAGCACTTGTTTG Rv CAGGAAACAGCTATGACC CAGGATGGGATCCGAAG BAP1 13a Fw TGTAAAACGACGGCCAGT ATGGTCACCTGGCCCGTTC Rv CAGGAAACAGCTATGACC TGAGGGCTGCGAGTGTG BAP1 13b Fw TGTAAAACGACGGCCAGT ACCTCTCAATTCCTCTGTCC Rv CAGGAAACAGCTATGACC CAGGCTGTCATCCTCTCC BAP1 13c Fw TGTAAAACGACGGCCAGT CTCGCCTATCCGCTCAG Rv CAGGAAACAGCTATGACC CTCCTGGGTGCACCAAG BAP1 14 Fw TGTAAAACGACGGCCAGT CAAAGTGTCCTGCACTCTG Rv CAGGAAACAGCTATGACC CCCTCTGCCAGGATTAAAG 4 79

81 Chapter 4 Supplementary Table 1 Continued Gene Exon Forward (Fw) Primer sequence 5-3 Reverse (Rv) BAP1 15 Fw TGTAAAACGACGGCCAGT TAGCTGCCTATTGCTCGTG Rv CAGGAAACAGCTATGACC CAGTGGACCTCGGGAGAG BAP1 16 Fw TGTAAAACGACGGCCAGT CTGCTCTGGCAAGATTGG Rv CAGGAAACAGCTATGACC AGGGGAGCTGAAGGACAC BAP1 17 Fw TGTAAAACGACGGCCAGT CTCAGCTCCTGGCCTGAG Rv CAGGAAACAGCTATGACC GGCCACACGGCAAGAGTG GNAQ 4 Fw TGTAAAACGACGGCCAGT TTCTCATTGTGTCTTCCCTCCT Rv CAGGAAACAGCTATGACC AAGGCATAAAAGCTGG- GAAAT GNAQ 5 Fw TTCCCTAAGTTTGTAAGTAGTGC Rv TCCATTTTCTTCTCTCTGACC GNA11 4 Fw TGTAAAACGACGGCCAGT GTGCTGTGTCCCTGTCCTG Rv CAGGAAACAGCTATGACC GGCAAATGAGCCTCTCAGTG GNA11 5 Fw GGTGGGAGCCGTCCTGGGATTGC Rv CCCACCTCGTTGTCCGAC GNAQ 4a Fw TGTAAAACGACGGCCAGT GTCACTGACATTCTCATTGT- GTCTT Rv CAGGAAACAGCTATGACC GTGGTGGGGACTCGAACTCT GNAQ 4b Fw TGTAAAACGACGGCCAGT ACCTGCCTACGCAACAAGAT Rv CAGGAAACAGCTATGACC AAGGCATAAAAGCTGG- GAAAT GNAQ 5a Fw TGTAAAACGACGGCCAGT TTTCCCTAAGTTTGTAAG- TAGTGCT Rv CAGGAAACAGCTATGACC GGTGACATTTTCAAAG- CAGTG GNAQ 5b Fw TGTAAAACGACGGCCAGT AAGGTCAGA- GAGAAGAAAATGGA Rv CAGGAAACAGCTATGACC AGTTCACTCCATTCCCCACA GNA11 4a Fw TGTAAAACGACGGCCAGT GTGCTGTGTCCCTGTCCTG Rv CAGGAAACAGCTATGACC GTGGGCAGGTAGCCCAAG GNA11 4b Fw TGTAAAACGACGGCCAGT ACCGCTGTGTTGCAGCTA Rv CAGGAAACAGCTATGACC TGATGTTCTCCAGGTCGAAA GNA11 4c Fw TGTAAAACGACGGCCAGT ACCGGCATCATCGAGTACC Rv CAGGAAACAGCTATGACC GGCAAATGAGCCTCTCAGTG GNA11 5a Fw TGTAAAACGACGGCCAGT CCAGGTGGCTGAGTCCTG Rv CAGGAAACAGCTATGACC CGACGAGAAACATGATGGAT 80

82 Whole-genome copy-number analysis of uveal melanomas Supplementary Table 1 Continued Gene Exon Forward (Fw) Primer sequence 5-3 Reverse (Rv) GNA11 5b Fw TGTAAAACGACGGCCAGT CCACTGCTTTGAGAACGTGA Rv CAGGAAACAGCTATGACC CACTGCACACAGCCCAAG BRAF 15 Fw TGTAAAACGACGGCCAGT TTCATGAAGACCTCACAG- TAAAAA Rv CAGGAAACAGCTATGACC AGCCTCAATTCTTACCATCCA NRAS 2 Fw TGTAAAACGACGGCCAGT ATTACTGGTTTCCAACAGG Rv CAGGAAACAGCTATGACC CTGGATTGTCAGTGCGCTTT NRAS 3 Fw TGTAAAACGACGGCCAGT CCCCAGGATTCTTA- CAGAAAA Rv CAGGAAACAGCTATGACC TCGCCTGTCCTCATGTATTG BRAF 15 Fw TGTAAAACGACGGCCAGT ACTTACTACACCTCAGA- TATATTTCTTCAT Rv CAGGAAACAGCTATGACC GATGGGACCCACTCCATC NRAS 2 Fw TGTAAAACGACGGCCAGT ATTACTGGTTTCCAACAGG Rv CAGGAAACAGCTATGACC TGGGTAAAGATGATCCGACA NRAS 3 Fw TGTAAAACGACGGCCAGT GGTTATAGATGGTGAAACC Rv CAGGAAACAGCTATGACC TAATGCTCCTAGTACCTGTA- GAG 4 For some genes different primers with amplicons of different sizes were used. 81

83

84 5 Mutations in G protein encoding genes and chromosomal alterations in primary leptomeningeal melanocytic neoplasms Heidi V.N. Küsters-Vandevelde, MD 1, Ilse A.C.H. van Engen-van Grunsven, MD 2, Sarah E. Coupland, MBBS, PhD 3, Sarah L. Lake PhD 3, Jos Rijntjes 2, Rolph Pfundt PhD 4, Benno Küsters MD, PhD 2,5, Pieter Wesseling MD, PhD 1,2,6, Willeke A.M. Blokx MD, PhD 2, Patricia J.T.A.Groenen PhD 2 1 Department of Pathology, Canisius Wilhelmina Hospital, Nijmegen, 2 Department of Pathology, Radboud University Medical Center, Nijmegen, 3 Pathology, Department of Molecular and Clinical Cancer Medicine, University of Liverpool, 4 Department of Human Genetics, Radboud University Medical Center, Nijmegen, 5 Department of Pathology, Maastricht University Medical Center, Maastricht, 6 Department of Pathology, VU University Medical Center, Amsterdam Pathol Oncol Res 2015;21:

85 Chapter 5 ABSTRACT Limited data is available on the genetic features of primary leptomeningeal melanocytic neoplasms (LMNs). Similarities with uveal melanoma were recently suggested as both entities harbor oncogenic mutations in GNAQ and GNA11. Whether primary LMNs share additional genetic alterations with uveal melanoma including copy number variations is unknown. Twenty primary LMNs ranging from benign and intermediate-grade melanocytomas to melanomas were tested by direct sequencing for hotspot mutations in the genes GNA11, GNAQ, BRAF, NRAS and HRAS. Furthermore, the lesions were tested for copy number variations of chromosomes frequently present in uveal melanoma (1p, 3, 6 and 8q) by multiplex ligation-dependent probe amplification (MLPA). Genomewide analyses of copy number alterations of two leptomeningeal melanocytic neoplasms were performed using the OncoScan SNP-array. GNAQ Q209 mutations were present in eleven LMNs, while two of 20 cases carried a GNA11 Q209 mutation. No BRAF, HRAS or NRAS hotspot mutations were detected. Monosomy 3 and gain of 8q were present in one leptomeningeal melanoma, and one intermediate-grade melanocytoma harbored a gain of chromosome 6. With MLPA, the melanocytomas did not show any further gross chromosomal variations. Our data shows that primary LMNs, like uveal melanoma, harbor oncogenic mutations in GNAQ and GNA11 but lack mutations in BRAF, NRAS and HRAS. This finding may help in the differential diagnosis between a primary LMN and a metastasis from a cutaneous melanoma to the central nervous system. Copy number variations in some aggressive LMNs resemble those present in uveal melanoma but their prognostic significance is unclear. 84

86 Molecular alterations in primary leptomeningeal melanocytic neoplasms INTRODUCTION Primary leptomeningeal melanocytic neoplasms (LMNs) comprise a spectrum of neoplasms ranging from benign melanocytoma to melanoma (1). Melanocytomas (MCs) are slow-growing neoplasms generally characterized by a benign clinical course, although local recurrence and malignant transformation can occur (2-5). Intermediate-grade melanocytomas (IMCs) show histological features that are suggestive of more aggressive behavior, and their clinical behavior is unpredictable (1, 6). Primary leptomeningeal melanomas (LMMs) are histological similar to melanomas arising in other sites but rarely metastasize to remote organs (7, 8). Systematic study of LMNs is hampered by their very low incidence, and the genetic events underlying the oncogenesis of these tumors are largely unknown. Several recent studies, including our previous study, reported the presence of activating GNAQ Q209 mutations in a substantial proportion of LMNs (37%) (9-11). Identical GNAQ Q209 mutations are present in uveal melanomas (UMs) and blue nevi, but they are very rare in cutaneous melanomas (CMs) (12-15). The GNAQ gene encodes the α-subunit of G-proteins (Gα q ) that are associated with G protein coupled receptors on the cell surface (16). Substitution of the critical glutamine (Q at codon 209) or arginine (R at codon 183) of Gα q leads to decreased GTPase activity, preventing hydrolysis of GTP and locking GNAQ in its active, GTP-bound state. This results eventually in constitutive activation of the MAPK pathway (12, 14, 16-18). Activating mutations in R183 of GNAQ are very rare in UMs and blue nevi, and little is known about its incidence in LMNs (19). Activating mutations in GNA11, a homologue of GNAQ located on chromosome 19p13, are present in UMs and blue nevi, and were more recently demonstrated in some cases of LMNs as well (10, 14, 19, 20). The gross chromosomal alterations present in LMNs are yet to be investigated. In UMs, monosomy 3 is a frequent event and strongly associated with metastatic disease (50%) (21, 22). Other frequent chromosomal aberrations in UMs include gain of chromosome 8q that is also associated with metastasis, and loss of chromosome 6p, that is associated with good prognosis (21). Based on the presence of GNAQ and GNA11 mutations in both LMNs and UMs, we hypothesized that LMNs might also share some gross chromosomal alterations with UMs as well. In this study, we investigated a series of twenty LMNs for mutations in Q209 and R183 of both GNA11 and GNAQ, as well as for other well-known hotspot mutations in BRAF, NRAS and HRAS. Additionally, lesions were tested for copy number variations of chromosomes frequently present in UMs (1p, 3, 6, 8q) using MLPA and two LMN cases were selected for in depth investigation of copy number alterations across the genome using the OncoScan SNP-array. Genetic changes were correlated with clinicopathological characteristics and patient follow-up data. 5 85

87 Chapter 5 MATERIALS AND METHODS Patients and histopathology For this retrospective study, we collected formalin-fixed, paraffin-embedded tissues of 20 LMNs, including 14 cases from our previously published series (9). The diagnosis of melanocytoma (MC) (n=11), intermediate-grade melanocytoma (IMC) (n=5) or leptomeningeal melanoma (LMM) (n=4) was based on the criteria previously described (1, 6, 9). None of these patients had clinical evidence of primary melanoma outside the central nervous system (CNS). DNA extraction and mutational analysis DNA extraction and assessment of DNA quality by the BIOMED-2 control gene PCR were performed as previously described (9). All but seven DNA samples allowed amplification of the 200bp amplicon, minimally (23). The remaining cases (n=7) allowed amplification of the 100bp-amplicon only, and still could be evaluated by mutational analysis (Table 1). Twenty LMNs were investigated for hotspot mutations: Q209 mutations in exon 5 and R183 mutations of exon 4 in GNAQ and GNA11; V600 mutations in BRAF; G12, G13, G15 mutations in exon 2 and Q61 mutations in exon 3 of the NRAS gene; and Q61 mutations in exon 3 of HRAS. Primer sequences are listed in supplementary table 1. PCR conditions as well as assessment on the ABI PRISM 3730 DNA analyzer were standardized and described previously (9). PCR for direct sequence analysis of exon 4 of GNAQ and GNA11 were performed using the Amplitaq Gold 360 Master Mix (Applied Biosystems) with standardized cycling conditions and duplicate PCRs were submitted for Sanger sequencing on an ABI PRISM 3730 DNA analyzer (Life Technologies, Foster City, USA). Determination of copy number variations 1. MLPA procedure We used the MLPA P027-B1 assay prepared by MRC-Holland (Amsterdam, the Netherlands) containing probes on chromosomes 1p, 3, 6, and 8q (as the MLPA assay contains only one probe for chromosome arm 8p we did not include 8p analysis in our study). A further twelve probes were used as reference probes (details available at; Immortalized cell lines of healthy controls were used as baseline control samples. MLPA was performed according to the manufacturer s instructions, and as previously described (21). Samples were analyzed on an ABI 3730 DNA analyzer using the Genescan software (Life Technologies, Foster City, USA). Details on data analysis can be found in Online Resource 2. 86

88 Molecular alterations in primary leptomeningeal melanocytic neoplasms 2. OncoScan analysis One LMM and one IMC (Table 1), selected due to the presence of copy number variations (CNVs) detected by MLPA, were further investigated for CNVs across the whole genome by OncoScan analysis (molecular inversion probe technology), as described by Wang et al. (24). This array consists of ~335,000 probes of which the majority (~283,000) are SNP based. DNA was processed by the Affymetrix Research Services Laboratory (ARSL, Santa Clara, California, USA). The normalized OncoScan data ( 2 Log(test/reference)-ratios and B-allele frequency plots) were analyzed with the Nexus copy number software version 5.1 (Biodiscovery, Inc, El Segundo, California, USA) for copy number calling. Statistical analysis Fisher s exact test was used to test whether significant associations were present between mutational status and cell type; location of the lesion (spinal versus intracranial); disease-related death; and clinical behavior of the lesion (stable versus progressive lesions, the latter including local recurrence, leptomeningeal seeding and/or distant metastasis). Statistical analyses were performed using SPSS 16.0 for Windows (IBM Statistics 20, IBM Corporation, Armonk, New York, United States). RESULTS 5 Patients and histopathological characteristics Our study group consisted of 11 MCs, five IMCs and four LMMs (Table 1). Most lesions were located around the spinal cord (13/20). MC recurred in four patients and aggressive behavior with leptomeningeal seeding or locoregional disease progression was present in three MC patients (#3, #5, #7). One IMC recurred with leptomeningeal seeding (patient #14), while two IMCs had stable disease (patients #12 and #13). Follow-up data was available for three of four LMM patients and all had succumbed to their disease. One LMM patient presented with a melanoma in the sacral region and developed distant metastases in the bone and lungs two years after initial diagnosis (patient #18). The late manifestation of other melanoma locations in this patient is supportive of the sacral lesion being the primary tumor. The other two LMM patients demonstrated subsequent leptomeningeal seeding but did not develop distant metastases (patients #19 and #20). Mutational analysis A summary of the mutational analyses is provided in Table 1. Briefly, mutations in codon 209 of GNAQ in were detected in 11 neoplasms (55%), while in two other LMNs a mutation in codon 209 of GNA11 was observed (10%). In this series, GNA11 Q209 and 87

89 Chapter 5 Table 1 Patient characteristics, histopathology and results of mutational analysis and copy number changes detected by MLPA Patient Sex Age Diagnosis Location Cell Pigmentation Follow-up type 1* F 27 MC right CPA E no Local recurrence. No distant metastases. 2 *, # M 41 MC C0-C3 E 3+ Local recurrence. Stable since. Alive. 3 *, #, + M 27 MC cerebellar tentorium Mx 2+ LM seeding. Stable since. Alive. 4* M 55 MC C3-6 Mx 2+ Local recurrence. Stable since. Alive. 5 *, #, + na na MC C5-6 E 3+ Tumor spread in neck and vertebra. Deceased. 6 *, #, + M 41 MC Th6 S 3+ Local recurrence. 7 *, #, + F 45 MC L3-4 Mx 3+ LM seeding. No distant metastases. Alive. 8 * na na MC Th11 S 1+ na 9 *, + na na MC cerebellar S 2+ na 10 *, #, + na na MC na S 3+ na 11 na na MC EM, ID E 3+ No recurrence. Stable. Alive. 12 na na IMC IM Mx 3+ No recurrence. Stable. Alive. 13 na na IMC cauda S 2+ No recurrence. Stable. Alive. 14 & M na IMC IM E 1+ Local recurrence. LM seeding. Alive. 15 *, + na na IMC LM E 3+ na 16 * F 68 IMC cerebellar tentorium E 1+ Deceased (not disease related) 17 *, # na na LMM LM E no na 88

90 Molecular alterations in primary leptomeningeal melanocytic neoplasms GNA11 ex 5 GNA11 ex 4 GNAQ ex 5 GNAQ ex 4 Chrom 1p Chrom 3p Chrom 3q Chrom 6p Chrom 6q Chrom 8q wt wt wt wt wt wt c.626a>c (p.(gln209pro)) wt wt wt c.626a>c (p.(gln209pro)) wt na na na na na na wt wt c.626a>c (p.(gln209pro)) wt wt wt wt wt na na na na na na na na c.626a>t (p.(gln209leu)) wt wt c.626a>t (p.(gln209leu)) na na na na na na na wt na na na na na na 5 wt wt c.626a>t (p.(gln209leu)) wt wt wt wt wt na na na na na na wt wt wt wt na na na na na na wt wt c.626a>t (p.(gln209leu)) wt wt wt wt wt na na wt wt c.626a>t (p.(gln209leu)) wt wt wt c.626a>t (p.(gln209leu)) wt gain & gain & na c.626a>t (p.(gln209leu)) na na wt na na na na na na na wt wt wt wt na wt wt - na na na na - 89

91 Chapter 5 Table 1 Continued Patient Sex Age Diagnosis Location Cell Pigmentation Follow-up type 18 *, & F 59 LMM S2 Mx 1+ Distant metastases (bone, lungs). Deceased. 19 F 32 LMM Th10-11 Mx 1+ LM seeding. No extradural disease. Deceased. 20 M 55 LMM L1-L2 Mx 1+ LM seeding. No distant metastases. Deceased. F female, M male, MC melanocytoma, IMC intermediate-grade melanocytoma, LMM leptomeningeal melanoma, CPA cerebello-pontine angle, EM extramedullary, ID intradural, LM leptomeningeal, S spindle, E epithelioid, Mx mixed; * cases of which results of mutation analysis were in part described previously (9), - no copy number changes with MLPA, na not available (sex, age, location, follow-up) or not assessable (MLPA and/or mutation analysis), # cases with 100bp-amplicon in the BIOMED-2 gene control PCR (n=7). GNAQ Q209 mutations were mutually exclusive. Both GNA11 Q209 mutations resulted in substitution of glutamine at position 209 by leucine c.626a>t (p.(gln209leu)) alias Q209L (Figure 1a), and were present in an IMC and LMM (patients #16 and #19) (Figure 1b, c). GNAQ Q209 mutations were present in seven of 11 MCs (64%), two of five IMCs (40%) and two of four LMMs (50%). These GNAQ Q209 mutations resulted in substitution of glutamine to leucine in 64% (7/11) and to proline in 36% of cases (4/11). In all samples, hotspot mutations in exon 4 of GNAQ and GNA11, and in BRAF, HRAS or NRAS were absent. There was no significant association between GNAQ/GNA11 mutational status and cell type, location of the lesion, disease-related death, or clinical behavior (Fisher s exact, P>0.05). Copy number variations MLPA results for eleven cases were of sufficient quality for further analysis (Table 1). Five MC cases did not harbor gross CNVs for the chromosome regions investigated (Figure 2a, patients #1, #2, #4, #8 and #11). Three of four IMCs did not show chromosomal CNVs (patient #12, #13 and #16; patient #12 and #13 had stable disease and patient #16 died of non-disease related illness. Of note, patient #12 could only be evaluated for 1p, 3p, 3q and 6p). One IMC showed gain of chromosome 6 confirmed by OncoScan analysis (Figure 3a, Online Resource 3, patient #14). This was a GNAQ-mutated IMC showing local recurrence and leptomeningeal seeding at follow-up. With the OncoScan analysis, no other gross chromosomal CNVs were detected in this sample. Of the two LMMs that were evaluated using MLPA, one 90

92 Molecular alterations in primary leptomeningeal melanocytic neoplasms GNA11 ex 5 GNA11 ex 4 GNAQ ex 5 GNAQ ex 4 Chrom 1p Chrom 3p Chrom 3q Chrom 6p Chrom 6q Chrom 8q wt wt c.626a>t wt - loss & loss & - - gain & (p.(gln209leu)) c.626a>t wt wt wt gain loss - (p.(gln209leu)) wt wt c.626a>c (p.(gln209pro)) wt na na na na na na The remaining cases (n=13) allowed amplification of the 200bp-amplicon, minimally, & copy number alterations of two cases confirmed by use of OncoScan analysis (Affymetrix); Cases not assessable in the mutational analysis (data not further shown): case 6 for GNA11 and NRAS, case 7 for HRAS, case 10 and 18 for NRAS ex 2, case 15 for GNA11, NRAS and HRAS, case 17 for GNA11 exon 4; Nine cases not assessable with MLPA (na), possibly due to the combination of low quality DNA (100bpamplicon in the BIOMED-2 gene control PCR, marked with # ) and/or heavily pigmentation, marked with +. GNA11-mutated LMM showed gain of 6p, and loss of 6q (Figure 2b, patient #19). The patient with this LMM suffered from leptomeningeal seeding at presentation but had no extradural melanoma dissemination. The second case was a GNAQ-mutated LMM that demonstrated monosomy 3 and gain of 8q, confirmed by OncoScan analysis (patient #18) (Figure 3b). In addition, the OncoScan data revealed the following CNVs for this LMM: loss of chromosome 1p, loss of chromosome 2p, loss of chromosome 8p, loss of chromosomes 10, 11 and 18, and gains of chromosomes 4p, 17 and 22 (Figure 3b, Online Resource 4). 5 91

93 Chapter 5 a b c Figure 1 The GNA11Q209L mutation was present in patient #16 (IMC) and patient #19 (LMM) a. Forward sequence tracing for GNA11 surrounding codon 209 of exon 5, showing the mutation c.626 A>T (p.(gln209leu)) for patient #16 (IMC). b. GNA11-mutated IMC in the cerebellar tentorium (magnification 400x) showing an epithelioid cell type and increased mitotic activity (arrow) (patient #16). c. GNA11-mutated melanoma in a 32-year-old female (magnification 200x) located at the spinal region Th10-11 (patient #19), showing a mixed cell type, nuclear pleomorphism and mitotic activity (arrow). 92

94 Molecular alterations in primary leptomeningeal melanocytic neoplasms 5 Figure 2 MLPA results for patient #8 (GNAQ-mutated MC) and patient #19 (GNA11-mutated LMM) a. MLPA results of a GNAQ-mutated MC located in the spinal canal at level of the 11 th thoracic vertebrae (patient #8), without evident copy number changes. The ratio of 1.3 at 8p12 (probe 27) and the ratio of at 8q24.12 (probes 29 and 30) were considered within the normal range. b. GNA11-mutated primary leptomeningeal melanoma showing gain of 6p and loss of 6q by MLPA (patient #19). The ratio of 1.3 at 8q24.12 (probe 30) was considered within the normal range. 93

95 Chapter 5 Figure 3 Copy number variation plots obtained by OncoScan SNP-analysis for patient #14 (a) (IMC) and patient #18 (b) (LMM), using the NEXUS 5.1 software package. The normalized and log2-transformed patient-over-reference intensity ratios (2Log(T/R)) for each array probe are displayed (blue dots) according to their chromosomal position (from 1pter Yqter), with each chromosome separated with a dotted vertical line. The results of a statistical copy number algorithm are shown by the black horizontal lines, with a normal copy number (n=2) at the 0-line, a loss (deletion) going down (negative probe values) and a gain (duplication/ amplification) going up (positive probe values). The corresponding allelic frequency plots, which are also obtained by the OncoScan SNP-analysis, are shown in the Supplementary Figures. The information regarding the CNVs and the corresponding allele frequency plots is needed for a correct interpretation of the copy number results (see Online Resources 3 and 4). The genome profile of the IMC (patient #14) is depicted in panel (a) and demonstrates gain of chromosome 6 and (normal) male hemizygosity for chromosome X. The genome profile of the LMM in patient #18 is depicted in panel (b), and demonstrates losses of chromosome 1p, 2p, 3, 8p, 10, 11 and 18. Clear gains are present of chromosome 4p, 8q, chromosome 17 (loss of 17q21.22q25.3) and 22. DISCUSSION We previously demonstrated the presence of oncogenic mutations in Q209 of GNAQ in a series of LMNs (37%) (9). In the present study, we broadened our sequencing analysis and tested a series of 20 LMNs for the presence of hotspot mutations in 94

96 Molecular alterations in primary leptomeningeal melanocytic neoplasms Q209 and R183 of GNA11 and GNAQ, as these hotspots were more recently shown to be mutated in cases of LMNs as well (10, 19, 20). We identified two LMNs harboring a GNA11 mutation at codon 209 (Q209L) (10%). This result is very similar to the study of Gessi et al. who detected a GNA11 Q209L mutation in two out of 16 LMNs (13%) (10). Our study thus confirms that GNA11 Q209 mutations are a rare event in LMNs (~10%). A similarly low GNA11 Q209 mutational frequency has been reported in blue nevi (7%) while in UMs this frequency is higher (32%) (14). Furthermore, the MC cases in our series did not show any GNA11 Q209 mutations (n=11) and this was also the case in the six MCs in the series of Gessi et al. (10). However, GNA11 Q209 mutations do occur in MCs, as two MC cases have been reported to harbor such mutation (Q209P and Q209L, respectively) (19, 20). In 55% of LMNs, we detected mutations in codon 209 of GNAQ, which is in the range of Gessi et al. (43%) and similar to the GNAQ Q209 mutational frequency in UMs (45%) and blue nevi (55%) (9, 10, 14). In UMs and blue nevi, mutations at codon 183 of both GNAQ and GNA11 are rare (1-3% in either gene, tested in >200 cases) (14). In contrast, little is known about the GNAQ R183 and GNA11 R183 mutational frequency in LMNs. Up to now, only one smaller case series searched for mutations in this codon of both genes and detected one activating GNAQ R183 mutation in five LMN cases (19). In our study, however, mutations in codon 183 of GNAQ and GNA11 were absent, suggesting that these mutations are a rare event in LMNs. At the moment, it is unclear whether GNAQ or GNA11 mutations in LMNs have any association with histological characteristics or clinical behavior. In UM, such an association has not been found (14, 25, 26). Furthermore, although inhibition of the GNAQ/GNA11-dependent MAPK signaling pathway is now under investigation for UMs, it remains to be determined whether a similar approach would be effective in LMNs (27-33). In our present series of adult LMN cases, we did not find any BRAF, NRAS or HRAS mutations. Similar results were obtained by Wang et al. (34). Of note, in contrast to the adult setting, we and others recently showed that mutations at Q61of NRAS rather than GNAQ or GNA11 are involved in the development of LMNs in children (10, 35, 36). The incidence of CNS metastasis in patients with melanoma ranges from 10% to 40% in clinical studies (37). This high frequency of CNS metastasis in melanoma patients is in contrast to the very low incidence of primary melanoma of the CNS (1). Still, the discrimination between a primary versus metastatic melanocytic tumor in or around the CNS is relevant because of different patient work-up and the substantially better prognosis of patients with a primary neoplasm, even when the tumor is malignant (38-41). In contrast to primary LMNs, mutations in GNAQ and GNA11 are absent or very rare in cutaneous melanoma (CM) (1.4%, tested in > 150 cases), acral melanoma (0%) and mucosal melanoma (0%) (14, 32). As a consequence, molecular testing of these genes could be helpful for elucidating the origin of a melanocytic tumor in or 5 95

97 Chapter 5 around the CNS. For instance, demonstration of a GNAQ or GNA11 mutation in a CNS melanocytic tumor favors a primary LMN, although a metastasis from a UM or malignant blue nevus still needs to be excluded. Conversely, the demonstration of a BRAF mutation favors a metastasis from a CM as these melanomas show a high frequency of BRAF mutations (50%) (42, 43) (Figure 4). Figure 4 Frequency of GNAQ, GNA11 and BRAF hotspot mutations for LMNs, UMs and CMs. Demonstration of a GNAQ or GNA11 mutation in a CNS melanocytic tumor may be helpful in discriminating a primary LMN from a CM metastasis in/around the CNS, as CMs only very rarely harbor GNAQ or GNA11 mutations. Vice versa, the demonstration of a BRAF mutation strongly favors a metastasis from a CM, as CMs show a high frequency of BRAF mutations. Note: the mutation frequencies include the hotspot mutations at codon 209 of GNAQ and GNA11, and codon 600 of BRAF, and were based on the following series, including the present study: (9, 10, 13, 14, 34, 46-49). Based on the presence of GNAQ and GNA11 mutations in both LMNs and UMs, we postulated that UMs and especially aggressive LMNs might share chromosomal copy number variations. The strong association of monosomy 3 and gain of 8q with poor UM patient prognosis is well documented (21, 22, 44). Gain of 6p in UMs is often mutually exclusive with monosomy 3 and correlates with a better prognosis (21, 22, 45). However, in our study, only two lesions carried the common gross chromosomal abnormalities seen in UM: one GNAQ-mutated LMM showed monosomy 3 and gain of 8q; and one GNA11-mutated LMM showed gain of 6p and loss of 6q. The former patient developed metastases in the bone and lungs two years after initial diagnosis, 96

98 Molecular alterations in primary leptomeningeal melanocytic neoplasms but no liver metastases were demonstrated. The latter patient showed leptomeningeal seeding but no distant metastases. In addition, one GNAQ-mutated IMC, presenting with local recurrence and leptomeningeal seeding, also showed gain of chromosome 6. Due to the limited number of patients that could be tested, however, it is still unclear if the gross chromosomal changes in LMNs have any prognostic implications. We conclude that LMNs share some genetic aberrations with UMs, especially GNAQ Q209 and GNA11 Q209 mutations, although the latter are, in contrast to UMs, an infrequent event in LMNs. Mutations in GNAQ R183 and GNA11 R183 were absent in our series, suggesting that these mutations are rare in LMNs, as in UMs. The demonstration of a GNAQ or GNA11 mutation in a LMN may have diagnostic value, allowing the discrimination of a primary CNS melanocytic tumor from a cutaneous melanoma metastasis. Copy number variations in some aggressive LMNs resemble those present in UMs including monosomy 3 and gain of 8q but their exact prognostic significance is so far unclear. Acknowledgments We thank Sandra Hendrikx-Cornelissen and Monique Goossens for their skillful assistance with MLPA and data analysis; Roel Besseling for technical assistance with the layout of the figures, and Dr. Eugene Verwiel for general discussions and technical evaluation of the OncoScan results. We have no funding resources to declare. 5 97

99 Chapter 5 REFERENCES 1. Brat DJ, Perry A. Melanocytic lesions. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, editors. WHO classification of tumours of the central nervous system. Lyon: IARC; p Bydon A, Gutierrez JA, Mahmood A. Meningeal melanocytoma: an aggressive course for a benign tumor. J Neurooncol. 2003;64(3): Liubinas SV, Maartens N, Drummond KJ. Primary melanocytic neoplasms of the central nervous system. J Clin Neurosci. 2010;17(10): Perrini P, Caniglia M, Pieroni M, Castagna M, Parenti GF. Malignant transformation of intramedullary melanocytoma: case report. Neurosurgery. 2010;67(3):E867-E9. 5. Guo J, Si L, Kong Y, Flaherty KT, Xu X, Zhu Y, et al. Phase II, open-label, single-arm trial of imatinib mesylate in patients with metastatic melanoma harboring c-kit mutation or amplification. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(21): Brat DJ, Giannini C, Scheithauer BW, Burger PC. Primary melanocytic neoplasms of the central nervous systems. Am J Surg Pathol. 1999;23(7): Arunkumar MJ, Ranjan A, Jacob M, Rajshekhar V. Neurocutaneous melanosis: a case of primary intracranial melanoma with metastasis. Clin Oncol (R Coll Radiol ). 2001;13(1): Koenigsmann M, Jautzke G, Unger M, Theallier-Janko A, Wiegel T, Stoltenburg-Didinger G. June 2002: 57-year-old male with leptomeningeal and liver tumors. Brain Pathol. 2002;12(4): Kusters-Vandevelde HV, Klaasen A, Kusters B, Groenen PJ, van Engen-van Grunsven IA, van Dijk MR, et al. Activating mutations of the GNAQ gene: a frequent event in primary melanocytic neoplasms of the central nervous system. Acta Neuropathol Gessi M, Hammes J, Lauriola L, Dorner E, Kirfel J, Kristiansen G, et al. GNA11 and N-RAS mutations: alternatives for MAPK pathway activating GNAQ mutations in primary melanocytic tumours of the central nervous system. Neuropathology and applied neurobiology Fuld AD, Speck ME, Harris BT, Simmons NE, Corless CL, Tsongalis GJ, et al. Primary melanoma of the spinal cord: a case report, molecular footprint, and review of the literature. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(17):e Onken MD, Worley LA, Long MD, Duan S, Council ML, Bowcock AM, et al. Oncogenic mutations in GNAQ occur early in uveal melanoma. Invest Ophthalmol Vis Sci. 2008;49(12): Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, O Brien JM, et al. Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature. 2009;457(7229): Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, Vemula S, Wiesner T, et al. Mutations in GNA11 in Uveal Melanoma. N Engl J Med Lamba S, Felicioni L, Buttitta F, Bleeker FE, Malatesta S, Corbo V, et al. Mutational profile of GNAQQ209 in human tumors. PLoS One. 2009;4(8):e Hurowitz EH, Melnyk JM, Chen YJ, Kouros-Mehr H, Simon MI, Shizuya H. Genomic characterization of the human heterotrimeric G protein alpha, beta, and gamma subunit genes. DNA Res. 2000;7(2): Markby DW, Onrust R, Bourne HR. Separate GTP binding and GTPase activating domains of a G alpha subunit. Science. 1993;262(5141): Radhika V, Dhanasekaran N. Transforming G proteins. Oncogene. 2001;20(13): Murali R, Wiesner T, Rosenblum MK, Bastian BC. GNAQ and GNA11 mutations in melanocytomas of the central nervous system. Acta Neuropathol Strom RG, Shvartsbeyn M, Rosenblum MK, Hameed MR, Nafa K, Mikolaenko I, et al. Melanocytic tumor with GNA11 p.q209l mutation mimicking a foramen magnum meningioma. Clin Neurol Neurosurg Damato B, Dopierala JA, Coupland SE. Genotypic profiling of 452 choroidal melanomas with multiplex ligation-dependent probe amplification. Clin Cancer Res. 2010;16(24): Prescher G, Bornfeld N, Hirche H, Horsthemke B, Jockel KH, Becher R. Prognostic implications of monosomy 3 in uveal melanoma. Lancet. 1996;347(9010):

100 Molecular alterations in primary leptomeningeal melanocytic neoplasms 23. van Dongen JJ, Langerak AW, Bruggemann M, Evans PA, Hummel M, Lavender FL, 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(12): Wang Y, Carlton VE, Karlin-Neumann G, Sapolsky R, Zhang L, Moorhead M, et al. High quality copy number and genotype data from FFPE samples using Molecular Inversion Probe (MIP) microarrays. BMC Med Genomics. 2009;2: Bauer J, Kilic E, Vaarwater J, Bastian BC, Garbe C, de KA. Oncogenic GNAQ mutations are not correlated with disease-free survival in uveal melanoma. British journal of cancer. 2009;101(5): Koopmans AE, Vaarwater J, Paridaens D, Naus NC, Kilic E, de Klein A. Patient survival in uveal melanoma is not affected by oncogenic mutations in GNAQ and GNA11. Br J Cancer. 2013;109(2): Mitsiades N, Chew SA, He B, Riechardt AI, Karadedou T, Kotoula V, et al. Genotype-dependent sensitivity of uveal melanoma cell lines to inhibition of B-Raf, MEK, and Akt kinases: rationale for personalized therapy. Invest Ophthalmol Vis Sci. 2011;52(10): Wu X, Zhu M, Fletcher JA, Giobbie-Hurder A, Hodi FS. The Protein Kinase C Inhibitor Enzastaurin Exhibits Antitumor Activity against Uveal Melanoma. PLoS One. 2012;7(1):e von Euw E, Atefi M, Attar N, Chu C, Zachariah S, Burgess BL, et al. Antitumor effects of the investigational selective MEK inhibitor TAK733 against cutaneous and uveal melanoma cell lines. Molecular cancer. 2012;11: Khalili JS, Yu X, Wang J, Hayes BC, Davies MA, Lizee G, et al. Combination small molecule MEK and PI3K inhibition enhances uveal melanoma cell death in a mutant GNAQ- and GNA11-dependent manner. Clinical cancer research: an official journal of the American Association for Cancer Research. 2012;18(16): Ho AL, Musi E, Ambrosini G, Nair JS, Deraje Vasudeva S, de Stanchina E, et al. Impact of combined mtor and MEK inhibition in uveal melanoma is driven by tumor genotype. PLoS One. 2012;7(7):e Carjaval RD SJ, Quevedo F, et al., editor Phase II study of selumetinib versus temozolomide in gnaq/ gna11 mutation uveal melanoma ASCO Annual Meeting; June Chen X, Wu Q, Tan L, Porter D, Jager MJ, Emery C, et al. Combined PKC and MEK inhibition in uveal melanoma with GNAQ and GNA11 mutations. Oncogene Wang H, Zhang S, Wu C, Zhang Z, Qin T. Melanocytomas of the Central Nervous System: a clinicopathological and molecular study. European journal of clinical investigation Pedersen M, Kusters-Vandevelde HV, Viros A, Groenen PJ, Sanchez-Laorden B, Gilhuis JH, et al. Primary melanoma of the CNS in children is driven by congenital expression of oncogenic NRAS in melanocytes. Cancer discovery. 2013;3(4): Kinsler VA, Thomas AC, Ishida M, Bulstrode NW, Loughlin S, Hing S, et al. Multiple Congenital Melanocytic Nevi and Neurocutaneous Melanosis Are Caused by Postzygotic Mutations in Codon 61 of NRAS. The Journal of investigative dermatology Bafaloukos D, Gogas H. The treatment of brain metastases in melanoma patients. Cancer treatment reviews. 2004;30(6): Greco Crasto S, Soffietti R, Bradac GB, Boldorini R. Primitive cerebral melanoma: case report and review of the literature. Surgical neurology. 2001;55(3):163-8; discussion Theunissen P, Spinc le G, Pannebakker M, Lambers J. Meningeal melanoma associated with nevus of Ota: case report and review. Clin Neuropathol. 1993;12(3): Rodriguez y Baena R, Gaetani P, Danova M, Bosi F, Zappoli F. Primary solitary intracranial melanoma: case report and review of the literature. Surgical neurology. 1992;38(1): Skarli SO, Wolf AL, Kristt DA, Numaguchi Y. Melanoma arising in a cervical spinal nerve root: report of a case with a benign course and malignant features. Neurosurgery. 1994;34(3):533-7; discussion Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, et al. Mutations of the BRAF gene in human cancer. Nature. 2002;417(6892): Broekaert SM, Roy R, Okamoto I, van den Oord J, Bauer J, Garbe C, et al. Genetic and morphologic features for melanoma classification. Pigment cell & melanoma research. 2010;23(6):

101 Chapter van den BT, Kilic E, Paridaens D, de KA. Genetics of uveal melanoma and cutaneous melanoma: two of a kind? Dermatol Res Pract. 2010;2010: Parrella P, Sidransky D, Merbs SL. Allelotype of posterior uveal melanoma: implications for a bifurcated tumor progression pathway. Cancer research. 1999;59(13): Cruz F, 3rd, Rubin BP, Wilson D, Town A, Schroeder A, Haley A, et al. Absence of BRAF and NRAS mutations in uveal melanoma. Cancer research. 2003;63(18): Kilic E, Bruggenwirth HT, Verbiest MM, Zwarthoff EC, Mooy NM, Luyten GP, et al. The RAS-BRAF kinase pathway is not involved in uveal melanoma. Melanoma research. 2004;14(3): Weber A, Hengge UR, Urbanik D, Markwart A, Mirmohammadsaegh A, Reichel MB, et al. Absence of mutations of the BRAF gene and constitutive activation of extracellular-regulated kinase in malignant melanomas of the uvea. Laboratory investigation; a journal of technical methods and pathology. 2003;83(12): Janssen CS, Sibbett R, Henriquez FL, McKay IC, Kemp EG, Roberts F. The T1799A point mutation is present in posterior uveal melanoma. British journal of cancer. 2008;99(10):

102 Molecular alterations in primary leptomeningeal melanocytic neoplasms ELECTRONIC SUPPLEMENTARY MATERIAL Supplementary Table 1 Primers used for mutational analysis Gene Exon Forward (Fw) Primer sequence 5-3 Reverse (Rv) GNA11 5 Fw GGTGGGAGCCGTCCTGGGATTGC Rv CCCACCTCGTTGTCCGAC GNA11 4 Fw GTGCTGTGTCCCTGTCCTG Rv GGCAAATGAGCCTCTCAGTG GNAQ 5 Fw TTCCCTAAGTTTGTAAGTAGTGC Rv TCCATTTTCTTCTCTCTGACC GNAQ 4 Fw TTCTCATTGTGTCTTCCCTCCT Rv AAGGCATAAAAGCTGGGAAAT BRAF 15 Fw CCTTTACTTACTACACCTCAG Rv AAAAATAGCCTCAATTCTTAC NRAS 2 Fw GGTTTCCAACAGGTTCTTGC Rv CCACTGGGCCTCACCTCTATG NRAS 3 Fw GATTCTTACAGAAAACAAGTGG Rv TAATGCTCCTAGTACCTGTAGAG HRAS 3 Fw CTGCAGGATTCCTACCGGA Rv ACTTGGTGTTGTTGATGGCA 5 All primers except for the GNAQ reverse primer contained a M13 forward or reverse consensus sequence enabling standardized sequencing of different exons Online Resource 2. MLPA data analysis Data analysis was performed in Microsoft Excel as follows: 1) the average control probe fraction of the control samples is calculated by dividing the peak value of each probe amplification product by the combined value of all the probes of the control samples (n=3), and subsequently to calculate the average of the obtained values, 2) the average patient probe fraction is calculated by dividing the peak value of the patient probe amplification by the combined value of all patient probes, and, subsequently to divide this value by the average control probe fraction, 3) next, the obtained value is divided by the average of the reference probes of the patient sample, and 4) is corrected for tumor load. A normal or disomy copy number was determined from a signal of ; above 1.2 was indicated an amplification and below 0.8 indicated a deletion. 101

103 Chapter 5 Online Resource 3 CNV plot and allelic frequency plot obtained by OncoScan SNP-analysis for patient #14 (IMC) Online Resource 4 CNV plot and allelic frequency plot obtained by OncoScan SNP-analysis for patient #18 (LMM) 102

104 Molecular alterations in primary leptomeningeal melanocytic neoplasms 5 103

105

106 6 Improved discrimination of melanotic schwannoma from melanocytic lesions by combined morphological and GNAQ mutational analysis Heidi V.N. Küsters-Vandevelde, MD 1, *, Ilse A.C.H. van Engen-van Grunsven, MD 2, *, Benno Küsters MD, PhD 2, Marcory RCF van Dijk MD, PhD 3, Patricia J.T.A.Groenen PhD 2, Pieter Wesseling MD, PhD 1,2,6, Willeke A.M. Blokx MD, PhD 2, 1 Department of Pathology, Canisius Wilhelmina Hospital, Nijmegen, 2 Department of Pathology, Radboud University Medical Center, Nijmegen, 3 Department of Pathology and Medical Biology, University Medical Centre Groningen, 6 Department of Pathology, VU University Medical Center, Amsterdam *Both authors contributed equally to this work; shared first authorship. Acta Neuropathol 2010;120:

107 Chapter 6 ABSTRACT The histological differential diagnosis between melanotic schwannoma, primary leptomeningeal melanocytic lesions and cellular blue nevus can be challenging. Correct diagnosis of melanotic schwannoma is important to select patients who need clinical evaluation for possible association with Carney Complex. Recently, we described the presence of activating codon 209 mutations in the GNAQ gene in primary leptomeningeal melanocytic lesions. Identical codon 209 mutations have been described in blue nevi. The aim of the present study was to (1) perform a histological review of a series of lesions (initially) diagnosed as melanotic schwannoma and analyze them for GNAQ mutations, and (2) test the diagnostic value of GNAQ mutational analysis in the differential diagnosis with leptomeningeal melanocytic lesions. We retrieved 25 cases that were initially diagnosed as melanotic schwannoma. All cases were reviewed using established criteria and analyzed for GNAQ codon 209 mutations. After review, nine cases were classified as melanotic schwannoma. GNAQ mutations were absent in these nine cases. The remaining cases were reclassified as conventional schwannoma (n=9), melanocytoma (n=4), blue nevus (n=1) and lesions that could not be classified with certainty as melanotic schwannoma or melanocytoma (n=2). GNAQ codon 209 mutations were present in 3/4 melanocytomas and the blue nevus. Including results from our previous study in leptomeningeal melanocytic lesions, GNAQ mutations were highly specific (100%) for leptomeningeal melanocytic lesions compared to melanotic schwannoma (sensitivity 43%). We conclude that a detailed analysis of morphology combined with GNAQ mutational analysis can aid in the differential diagnosis of melanotic schwannoma with leptomeningeal melanocytic lesions. 106

108 Improved discrimination of melanotic schwannoma from melanocytic lesions INTRODUCTION Melanotic schwannoma is a rare variant of schwannoma composed of cells having the ultrastructure and immunophenotype of Schwann cells but containing melanosomes in varying stages of maturation (24,25). They show a peak incidence in the fourth decade, and affect males and females equally. About half of these tumors arise in close proximity to the neural axis. These neoplasms are often associated with spinal nerve roots, cranial nerves or arising from the paraspinal sympathetic chain (5,15). The other half arises in a wide variety of sites including the gastrointestinal tract, soft tissues, skin, heart and liver (3,8,10,13,25). The biological behavior of melanotic schwannomas can be unpredictable with metastasis reported in up to 10% of cases, sometimes even in absence of histological features suggestive of aggressive behavior (5,27). A specific subgroup of melanotic schwannomas contain psammoma calcifications. About half of these so-called psammomatous melanotic schwannomas occur in the setting of the autosomal dominant Carney complex, a multiple neoplasia syndrome characterized by a markedly increased risk to develop myxomas, mucocutaneous lentigines and blue nevi, endocrine hyperactivity and/or psammomatous melanotic schwannomas (5). Isolated cases of non-psammomatous melanotic schwannomas have been associated with Carney complex as well (6). Nearly two thirds of patients with Carney complex have mutations spread over the PRKAR1A gene, which encodes the R1α regulatory subunit of cyclic-amp-dependent protein kinase A (14,29). The differential diagnosis between melanotic schwannomas and pigmented melanocytic lesions can be difficult, especially in small biopsy specimens. It includes primary or metastatic melanoma, a diagnosis with obvious prognostic and therapeutic relevance. In addition, melanotic schwannomas may histologically closely resemble leptomeningeal melanocytomas and cellular blue nevi (20). Immunohistochemical stains are not always useful in sorting out this differential diagnosis. All these lesions generally express S-100 (which can be explained by their common neural crest origin) and one or more melanocytic markers. Although stains for components of the basement membrane can be useful in discriminating tumors with schwannian differentiation from tumors with melanocytic differentiation, overlap in staining patterns has been noted and can impair interpretation, especially in small biopsy specimens (4,12,21,26). Recently, we reported the presence of activating mutations in the GNAQ gene at codon 209 in primary leptomeningeal melanocytomas (50%) and melanomas (25%) (16). The GNAQ gene maps on chromosome 9q21, and encodes the α-subunit of a heterotrimeric GTP-binding protein that couples G protein coupled receptor signalling to the MAP kinase pathway (22). GNAQ codon 209 mutations form an alternative route to MAP kinase activation (28). Identical oncogenic mutations in the GNAQ gene have been reported in blue nevi (83%) and uveal melanomas (UMs) (46%) (17,28)

109 Chapter 6 The aims of the present study were (1) to perform a thorough histological review of a series of lesions that were (initially) diagnosed as melanotic schwannoma and to analyze these cases for mutations in the GNAQ gene, and (2) to assess the diagnostic value of GNAQ mutational analysis in the differential diagnosis of melanotic schwannoma with primary leptomeningeal melanocytic lesions by including data from our previous study on GNAQ mutations in leptomeningeal melanocytic lesions (16). MATERIALS AND METHODS Patients and histopathology For this retrospective study, formalin-fixed and paraffin-embedded (FFPE) tissues of melanotic schwannomas were retrieved from archives of various Departments of Pathology in The Netherlands. The cases were obtained through the Dutch nationwide histopathology and cytopathology data network and archive (PALGA) (7). All cases (n=25) were initially diagnosed as melanotic schwannomas. FFPE tissues were sectioned at 4 µm, deparaffinized, and subjected to antigen retrieval by limited protein digestion or citrate (CC1) and microwave treatment. Slides were then incubated at room temperature with antibodies to S-100 protein (polyclonal, 1:800, Dako Co., Lot ), Melan-A (Mart-1, 1:800, Neomarkers, Lot 799P807C), HMB-45 (1:50, Neomarkers, Lot 23068), MIB-1 (Ki-67, 1:200, Dako Co., Lot ), EMA (E-29, 1:50, Dako Co., Lot ), Collagen type IV (monoclonal, 1: 1250, Sigma, Lot 062K4819) and Laminin (polyclonal, 1:100, Thermo, Lot 82H907B). Aminoethylcarbazole served as chromogen. Reticulin stains (Laguesse) were performed on each case. Schmorl and Perl s (iron) stains were performed to confirm melanin pigmentation. Histology was revised independently by two pathologists (HK, BK) without knowledge of molecular results. Revision of the slides included scoring of histological criteria and immunohistochemical stains. The deposition of basement membrane material in the reticulin, Collagen type IV and Laminin stains was categorized into three staining patterns. The pericellular pattern was represented by an extensive network of basement membrane material around individual tumor cells, consistent with schwannian differentiation (12,24). The nested/ perilobular pattern consisted of deposition of basement membrane material around nests of tumor cells (rather than around individual tumor cells), and is typically seen in melanocytic lesions (26). The biphasic staining pattern was characterized by areas with nested staining as well as areas with pericellular staining, and has been described in melanotic schwannomas (12,24). After revision, the cases were reclassified as: (1) melanotic schwannoma, with or without psammoma bodies and/or fat, (2) conventional schwannoma, (3) melanocytoma, (4) 108

110 Improved discrimination of melanotic schwannoma from melanocytic lesions lesion not further classified with certainty as melanotic schwannoma or melanocytoma, or (5) blue nevus. Criteria used for classification are presented in Table 1. The diagnosis of `melanotic schwannoma` and `conventional schwannoma` was based on WHO criteria (24). Additional detailed histological criteria of melanotic schwannoma were retrieved from Mooi (20), and included circumscribed, often encapsulated tumors consisting of spindle to epithelioid cells, with or without psammoma bodies and/or fat, variable melanin pigmentation, S-100 positivity and expression of Melan-A and/or HMB-45. Schwannomas lacking positivity for both HMB-45 and Melan-A were classified as conventional schwannomas. A pericellular staining pattern in the basement membrane stains served as an additional criterion for conventional schwannoma, while in melanotic schwannomas also a biphasic staining pattern can be present (12,24). The diagnosis of `melanocytoma` was based on histological criteria as described by Brat et al. (1,2). This included circumscribed, non-encapsulated lesions consisting of nests or fascicles of spindle and/or epithelioid cells with little nuclear atypia and variable melanin pigmentation. Immunohistochemical criteria included S-100 positivity, expression of Melan-A and/or HMB-45, and a nested staining pattern in the basement membrane stains. Lesions with increased mitotic activity or invasion of neuroglial tissue were classified as intermediate-grade melanocytomas. The diagnosis of cellular blue nevus was based on histological criteria described by Mooi (19). This included a primary location of the lesion within the dermis with a biphasic architecture, especially at the base of the lesion, consisting of nodules and sheets of pale, oval melanocytes intermingled with pigmented dendritic melanocytes diffusely arranged in sclerotic stroma. For statistical evaluation, 19 leptomeningeal melanocytic lesions from our previous study were included (16). 6 DNA extraction About five manually dissected sections of 10 µm FFPE tissue with an estimated tumor cell percentage of at least 60% were used for DNA extraction. After heating in ATL buffer, the tissue sections were incubated in proteinase K for one hour, followed by subsequent purification of the DNA according to the manufacturer (QIAamp DNA Mini Kit, QIAGEN GmbH, Germany). DNA sample concentration was assessed spectrophotometrically (Varian Cary 50 spectrophotometer, Agilent Technologies). Mutational analysis Direct sequence analysis of GNAQ exon 5 was performed on DNA-specimens from 25 l pigmented lesions of schwannian or melanocytic origin. PCR amplification was performed in a total volume of 25 µl, containing 50 ng DNA, PCR-buffer IV (Integro), 37 mm MgCl 2, 250 µm of dntps, 37.5 µg bovine serum albumin (Sigma), 10 pmol of 109

111 Chapter 6 Table 1 Morphological, esp. histopathological criteria used for classification Conventional Melanotic Leptomeningeal schwannoma 1 schwannoma 1,3 melanocytoma (MC) 2 Cellular blue nevus 3 Localization Majority outside central nervous system, most often skin and subcutaneous tissue Near the neural axis (esp. spinal nerve roots, cranial nerves) or peripherally located in a wide variety of sites (soft tissue, skin etc.) Mostly at the cervical and thoracic spinal level, posterior fossa and sometimes supratentorial compartment Skin (mainly dermis); frequently buttocks or sacrococcygeal region Origin Nerve sheath (schwann cell) Nerve sheath (schwann cell) Leptomeningeal melanocyte Dermal melanocyte Growth pattern, Circumscribed, margin encapsulated Melanin pigmentation Psammoma bodies and/or fat Schwannian Circumscribed, often encapsulated Circumscribed, nonencapsulated; MC with invasion of CNS is classified as intermediate-grade MC Absent Variably, often heavily pigmented Variably, often heavily pigmented; can rarely be absent Absent Present in the psammomatous form Absent Absent features 4 conventional schwannoma Present Often less pronounced than in Non-encapsulated; both infiltrative and pushing border Variably, can rarely be absent Absent Degenerative changes reminiscent of ancient schwannoma can be present Cell phenotype Spindle Spindle, epithelioid Spindle, epithelioid Spindle, ovoid, dendritic (biphasic architecture) Variable; absence of necrosis, mitotic activity < 2 per 2mm 2 Nuclear atypia Generally absent; ancient changes ( degenerative atypia`) can be present Usually mild; prominent nuclear atypia with increased mitotic activity and necrosis is indicative of aggressive behavior Generally absent; MC with increased mitotic activity (2-5 per 10 HPF) is classified as intermediate-grade MC S-100 Positive Positive Positive Positive HMB-45 and/ or Negative Positive Positive Positive Melan-A Basement Pericellular pattern Pericellular or biphasic (pericellular and Nested pattern Predominantly nested membrane staining nested) pattern pattern 1 Based on WHO criteria (24) 2 Based on criteria described by Brat et al. (1,2) 3 Based on criteria described by Mooi (19,20) 4 Esp. Antoni A and B pattern, Verocay bodies and lipid-laden macrophages 110

112 Improved discrimination of melanotic schwannoma from melanocytic lesions forward primer (p740, sequence: 5 TGTAAAACGACGGCCAGTTTCCCTAAGTTTG- TAAGTAGTGC 3 ) and reverse primer (p 705, sequence: 5 CAGGAAACAGCTAT- GACCCTTACCTCATTGTCTGAC 3, or p695, sequence: 5 TCCATTTTCTTCTCTCT- GACC 3 ) and 0.05 units of thermostable DNA polymerase (Sigma). Primers p740 and p705 contained a M13 forward and a M13 reverse consensus sequence, respectively, enabling standardized sequencing of different exons. DNA amplification was performed in duplicate in a PTC 200 Thermal Cycler (MJ Research). The PCR was started with 5 min at 92 C and followed with 35 cycles of denaturation 45 s at 94 C, annealing at 62 C for 45 s and extension at 72 C for 45 s, followed by a final extension at 72 C for 20 min and cooling down for 5 min at 20 C. PCR products were purified with MinElute plates (Qiagen). After analysis of the PCR products by agarose gel electrophoresis (2%; Invitrogen), 1-4 µl of the PCR products were sent in for sequence analysis using an ABI PRISM 3700 DNA analyzer (Applied Biosystems). Because the DNA-quality of some specimens was suboptimal, different primer combinations were used in the PCR. The DNA-samples of cases 1, 2, 3, 5, 6, 13, 14, 22 and 23 were of sufficient quality and the primer combination p740 and p705, yielding a PCR product of 224 basepairs (bp), was used for DNA-amplification. Sequence analysis was performed using the M13 forward primer and reverse primer on duplicate DNA-amplifications of the same specimen. For the DNA-samples that were strongly degraded (cases 4, 7, 8, 9, 10, 11, 12, 15, 16, 17, 18, 19, 20, 21, 24, 25) the primer combination p740 and p695, yielding a PCR product of 134 bp, was used. Sequence analysis (in duplicate) of these samples was performed using the forward primer only, since the reverse primer p695 is closely flanking the mutation hotspot. By sequence analysis of the entire PCR products, it was shown that the primers used specifically amplified the GNAQ gene (accession nr: NM_ ) and not its pseudogene (GNAQP1; accession nr: NG_ v). The amplified gene sequence was specific for the GNAQ gene, and nucleotides that are present in the pseudogene were not detected (9). 6 Statistical analysis The diagnostic value of GNAQ mutational analysis in the differential diagnosis of melanotic schwannoma with primary leptomeningeal melanocytic lesions was tested using additional data from our previous study on GNAQ mutations in primary leptomeningeal melanocytic lesions (16). Sensitivity and specificity of GNAQ mutational analysis was calculated. To assess whether the difference in GNAQ mutational status between melanotic schwannoma and leptomeningeal melanocytic lesions was statistically significant, we performed Fisher s exact test (SPSS 16.0 for Windows). 111

113 Chapter 6 RESULTS Patient and histopathological characteristics We retrieved 25 lesions initially diagnosed as melanotic schwannomas from the archives of various Departments of Pathology in The Netherlands. After revision, the lesions were reclassified using the above-mentioned criteria into nine melanotic schwannomas, nine conventional schwannomas, four melanocytomas, one blue nevus and finally two lesions that could not be further classified with certainty as melanotic schwannoma or melanocytoma. Patient and histopathological characteristics are summarized in Table 2. As far as data were available, the melanotic schwannomas had an equal sex distribution and the mean age was 37.2 years. Most melanotic schwannomas were located around the spinal cord (7/9), while 2 lesions were located in the extremities (lower arm and lower leg). As far as was assessable, melanotic schwannomas were circumscribed, encapsulated tumors (6/9) (Figure 1a) with variably but mostly intense melanin pigmentation (6/9) (Figure 1b). Eight of nine melanotic schwannomas contained psammoma bodies (Figure 1a, b) and, in addition, one melanotic schwannoma contained large, empty vacuoles resembling fat (case 8) (Figure 1c). Tumor cells were arranged in sheets and fascicles and varied from spindle to epithelioid, often with nuclear pseudoinclusions (Figure 1c) and conspicuous nucleoli. In some cases there was nuclear palissading and one case showed Antoni A and B areas together with Verocay bodies (case 4). Marked nuclear pleomorphism, necrosis, and invasion in surrounding structures was absent. These tumors were all negative in the EMA stains (excluding meningeoma). MIB-1 staining was less than 1 percent. S-100 stains showed a diffuse and strong expression in all melanotic schwannomas (Figure 1d). In addition, all lesions showed positivity for Melan-A and HMB-45 (Figure 1e). Reticulin, Collagen type IV and Laminin stains showed in four of nine cases an extensive network of basement membrane material around individual tumor cells (pericellular pattern), consistent with schwannian differentiation (Figure 1f). In five melanotic schwannomas, a biphasic staining pattern was present consisting of areas with pericellular deposition of basement membrane material next to areas with deposition around nests of tumor cells (nested/perilobular pattern). One lesion initially diagnosed elsewhere as a `peripherally located melanotic schwannoma` was reclassified as a cellular blue nevus (case 19) (19). Nine lesions initially diagnosed elsewhere as melanotic schwannoma were reclassified as conventional schwannomas. These lesions showed the typical histological features of schwannomas with frequent Antoni A and B areas, nuclear palissading, blood vessel hyalinization and wavy nuclei. We did not classify them as melanotic schwannoma because positivity for both Melan-A and HMB-45 was lacking. They all contained focally some fine intracytoplasmic grayish pigment in tumor cells (Schmorl 112

114 Improved discrimination of melanotic schwannoma from melanocytic lesions positive). Furthermore, extensive areas with deposition of basement membrane material around individual tumor cells in the basement membrane stains was present. Based on the criteria described by Brat et al. (1,2), four lesions located around the spinal cord were reclassified as melanocytomas. Three of them showed an infiltrative growth pattern in the neuroglial tissues and were classified as intermediate-grade melanocytomas (Figure 2a). All lesions consisted of fascicles and nests of spindle and/or epithelioid cells with variable melanin pigment. Nuclei were round to oval and uniform in size, with mostly small nucleoli (Figure 2b). Prominent nuclear atypia was absent. In contrast to the melanotic and conventional schwannomas, S-100 was only focally positive (Figure 2c), while Melan-A and HMB-45 were more diffusely positive (Figure 2d, e). MIB-1 staining was less than 1 percent in all cases. Reticulin, Collagen type IV and Laminin stains showed deposition of basement membrane material around vascular structures and nests of tumor cells (nested/perilobular), but not around individual tumor cells (Figure 2f). Finally, two lesions could not be classified with certainty into melanotic schwannoma or melanocytoma (case 24 and 25). Case 24 was a variably, partly heavily pigmented lesion consisting of spindle to pleiomorphic cells (Figure 3a, c). The lesion showed infiltrative growth in or from a ganglion at the lumbar spinal region (Figure 3b). The nuclei showed moderate to strong atypia with conspicuous nucleoli (Figure 3c). MIB-1 staining was less than 10%. Psammoma bodies were absent. S-100, Melan-A and HMB-45 stains were diffusely positive. The reticulin, Collagen type IV and Laminin stains showed a biphasic pattern (Figure 3d,e). Based on these histological features we favored the diagnosis of melanotic schwannoma with atypical features suggestive of aggressive behavior. For case 25, only limited biopsy material was available which impaired unequivocal histological assessment. It concerned a heavily pigmented lesion located at the cerebellopontine angle consisting of spindle cells with mild cytonuclear atypia and small nucleoli. S-100 was focally positive while the Melan-A and HMB-45 stains were diffusely positive. A nested/perilobulated staining pattern was present in the reticulin, Collagen type IV and Laminin stains. In this case, we favored the diagnosis of melanocytoma. 6 Results of GNAQ mutational analysis All lesions were tested for mutations in exon 5 of the GNAQ gene. The results are presented in Table 2. We detected mutations in exon 5 of GNAQ in 3 of 4 melanocytomas and the blue nevus (c.626a>t (p.(gln209leu))). No mutations were detected in melanotic or conventional schwannomas. In the two lesions which we could not classify with certainty as melanotic schwannoma or melanocytoma (case 24 and 25), only case 25 (the probable melanocytoma case) had a mutation in codon 209 of exon 5 (c.626a>t (p.(gln209leu))). 113

115 Chapter 6 Table 2 Patient and histopathological characteristics and results of mutational analysis Patient Sex Age Diagnosis Localization Origin from nerve / ganglion Circumscribed, capsule Pigmentation Psammoma bodies 1 F 75 MS L1-L2 na F 35 MS L M 42 MS L2 na na na MS C na na MS C1 na na na na MS C4 na na M 35 MS Lumbal + na F 23 MS Lower arm na + Focal 1+ +, fat 9 M 13 MS Lower leg + + Focal M 72 Sw Vestibular nerve na na F 60 Sw Th11 na na Focal F 32 Sw Retroperitoneum na + Focal M 42 Sw Sacrum na + Focal na na Sw Intercostal nerve na M 49 Sw S1 na na Focal F 74 Sw S2 + na Focal F 74 Sw Groin na na Focal M 44 Sw Sacral na + Focal na na blue nevus Foot na na na IMC Cauda na No, invasion of 1+ - CNS 21 na na IMC IM na No, invasion of 3+ - CNS 22 na na IMC IM na No, invasion of 1+ - CNS 23 na na MC EM na na F 41 NOS L5 + na na na NOS CPA na na 3+ - na data not available, MS melanotic schwannoma, Sw conventional schwannoma with limited pigment (neuromelanin, lipofuscin), IMC intermediate-grade melanocytoma, MC melanocytoma, NOS lesion not further classified with certainty as melanotic schwannoma or melanocytoma, IM intramedullary, EM extramedullary, CPA cerebellopontine angle, S spindle cell phenotype, E epithelioid cell phenotype, O ovoid, Pa palissading of nuclei, V Verocay bodies, A Antoni A and B pattern, mi mild, mo moderate, sev severe, B biphasic, P pericellular, N nested. 114

116 Improved discrimination of melanotic schwannoma from melanocytic lesions Cell type Schwannian features Nuclear pleomorphism S-100 HMB-45 Melan- A Basement membrane staining GNAQ exon 5 mutation S Pa mi B - E,S Pa mi B - S - mi B - E,S A, V mi P - S - mi P - E, S - mi B - S Pa mi B - E - mi P - E, S - mi P - S Pa mi 2+ neg neg P - S Pa, V mi 2+ neg neg P - S Pa, A mi 2+ neg neg P - S Pa mi 2+ neg neg P - S Pa, A mi 2+ neg neg P - S Pa, A mi 2+ neg neg P - S Pa mi 2+ neg neg P - S Pa, A mi 2+ neg neg P - S Pa, A mi 2+ neg neg P - S,O - mi Focal Foc. 1+ N c.626a>t (p.(gln209leu)) S - mi Focal N c.626a>t (p.(gln209leu)) E,S - mi Focal N - 6 E - mi Focal N c.626a>t (p.(gln209leu)) E - mi Focal N c.626a>t (p.(gln209leu)) E,S - mo, sev B - S - mi Focal N c.626a>t (p.(gln209leu)) 115

117 Chapter 6 Figure 1 Melanotic schwannoma (case 4) showing a circumscribed tumor, covered by a fibrous capsule and containing several psammoma bodies (a) (x50). Detail of the same lesion showing spindle cell morphology, an ample amount of melanin pigment and a psammoma body (case 4) (b) (x200). Melanotic schwannoma (case 8) with more epithelioid cell morphology, nuclear pseudoinclusions (arrow) and large vacuoles resembling fat (c) (x400). Strong and diffuse positivity for respectively S-100 and HMB-45 in a melanotic schwanomma (case 4) (d, e) (x25, x100). Same lesion with pericellular deposition of basement membrane material (case 4) (f) (Laguesse stain, x400). Diagnostic value of GNAQ mutational analysis In our previous study on GNAQ mutations in primary leptomeningeal melanocytic lesions (16), codon 209 mutations of GNAQ were present in seven of 19 lesions. Taking these previous results into account, a total GNAQ mutation frequency of 43% (10/23) in primary leptomeningeal melanocytic lesions was detected. No GNAQ 116

118 Improved discrimination of melanotic schwannoma from melanocytic lesions 6 Figure 2 Intermediate-grade melanocytoma (case 20) with invasion of neuroglial tissue; note the Rosenthal fibers (arrow) in the surrounding neuropil (a) (x200). Epithelioid cell morphology with bland nuclei and small nucleoli in a melanocytoma (case 22) (b) (x400). Focal positivity for S-100 in the same lesion (case 22) (c) (x200). Diffuse positivity for HMB-45 and Melan-A, respectively, in the same lesion (d, e) and a nested staining pattern in the reticulin stain (f) (case 22) (Laguesse stain, x200). mutations were detected in the melanotic schwannomas. The mutational status between melanotic schwannoma and leptomeningeal melanocytic lesions was significantly different (P = 0.030). GNAQ mutations were highly specific for leptomeningeal melano cytic lesions compared to melanotic schwannoma (100%), however, sensitivity for leptomeningeal melanocytic lesions was low (43%). 117

119 Chapter 6 Figure 3 Case 24 was a variably, partly heavily pigmented lesion consisting of spindle to pleomorphic cells (a, c) (x200), showing infiltrative growth in a ganglion at the lumbar spinal region with dispersed incorporated ganglion cells (arrow) (b) (x400). The nuclei showed marked atypia with conspicuous nucleoli (c) (x200). In the reticulin stain, both a pericellular (d) and nested staining pattern was seen (e) (Laguesse, x100). We, therefore, favored the diagnosis of a melanotic schwannoma with atypical features suggestive of aggressive behavior. Molecular analysis of this case did not reveal mutations in the GNAQ gene. DISCUSSION Especially on biopsy material, the differential diagnosis of melanotic schwannoma with primary leptomeningeal melanocytic lesions and, in case of peripherally located lesions, cellular blue nevi can pose a problem to pathologists since these lesions have overlapping histological features and immunohistochemical stains are not always useful. All these lesions are variably pigmented, spindle and/or epithelioid cell tumors, expressing S-100 (owing to their common neural crest origin) and one or more melanocytic markers. However, some histological features favor the diagnosis of melanotic schwannoma, such as circumscribed, often encapsulated growth, and/ or the presence of psammoma bodies. Psammoma bodies are generally absent in melanocytic tumors, although one case of a psammomatous melanoma arising in an intradermal nevus has been described (18). Stains for components of the basement 118

120 Improved discrimination of melanotic schwannoma from melanocytic lesions membrane can be useful in discriminating tumors with schwannian differentiation from tumors with melanocytic differentiation (12). Deposition of basement membrane material around individual tumor cells (`pericellular` pattern) is a feature of tumors with schwannian differentiation, and is generally not pronounced in melanocytic tumors (23,26). However, stains for basement membrane in melanin producing tumors should be interpreted with caution. First, in melanotic schwannomas a biphasic staining pattern has been described with deposition of basement membrane material around both individual tumor cells as well as around nests of tumor cells (12), as was the case in five of nine melanotic schwannomas in our study. On biopsy material, with only limited tissue to evaluate, this might result in misinterpreting the lesion as a melanocytic tumor. Secondly, most melanomas show absence of basement membrane material around individual tumor cells (26), except for spindle cell melanomas and desmoplastic neurotropic melanomas in which also a biphasic staining pattern has been described (4,21,26). Recently, we reported the presence of activating mutations in codon 209 of the GNAQ gene in primary leptomeningeal melanocytic lesions (16). This gene encodes the α-subunit of q class heterotrimeric GTP-binding proteins which couple G protein coupled receptors to the MAP kinase pathway. Identical codon 209 mutations have been reported in UMs (46%) and blue nevi (83%) (28). These codon 209 mutations occur in the ras-like domain turning GNAQ into a dominant acting oncogene (28). The aim of the present study was to perform a thorough histological review of a series of lesions that were initially diagnosed as melanotic schwannoma and analyze these cases for mutations in the GNAQ gene. In addition, we evaluated the diagnostic value of GNAQ mutational analysis in the sometimes difficult differential diagnosis with primary leptomeningeal melanocytic lesions. Twenty five cases initially diagnosed as melanotic schwannoma were reviewed and reclassified using established criteria into nine melanotic schwannomas, nine conventional schwannomas, four melanocytomas, one blue nevus and two lesions that could not be further classified with certainty as melanotic schwannoma or melanocytoma. The fact that in only nine of 25 cases the initial diagnosis melanotic schwannoma remained unchanged after review illustrates the difficult differential diagnosis with esp. conventional schwannoma and leptomeningeal melanocytic lesions. Including results from our previous study (16), a total GNAQ mutation frequency of 43% was present in leptomeningeal melanocytic lesions (10/23). No GNAQ mutations were detected in the melanotic schwannomas (0/9). These results show that GNAQ mutations are highly specific for leptomeningeal melanocytic lesions compared to melanotic schwannoma (100%). Thus, GNAQ mutational analysis can have diagnostic value: in case a GNAQ mutation is present this supports the diagnosis of a melanocytic lesion. It is important to note here that lack of a GNAQ mutation does not rule out the diagnosis of a melanocytic lesion. Correct diagnosis of melanotic schwannoma has clinical relevance as melanotic 6 119

121 Chapter 6 schwannoma can occur in the setting of the autosomal dominant Carney complex and may require additional clinical investigations. GNAQ codon 209 mutations in blue nevi have been reported in up to 83% (28). This high mutation frequency suggests that GNAQ mutational analysis can also be useful in the differential diagnosis of peripherally located melanotic schwannoma with blue nevus. Noteworthy, we reclassified a substantial number of lesions initially diagnosed as melanotic schwannoma into conventional schwannoma (n=9), based on absence of staining for Melan-A and HMB-45. These lesions all contained some areas with fine intracytoplasmic grayish Schmorl positive pigment in tumor cells. However, the absence of staining for Melan-A and HMB-45 contradicts the presence of true melanosomes in these cells and suggests that the pigment is rather autophagosomerelated (e.g. neuromelanin, lipofuscin) (11). Care should thus be taken not to over-diagnose melanotic schwannoma in order to prevent unnecessary clinical work-up for Carney Complex. At the moment, it is unclear why mutations in the GNAQ gene are especially found in a subgroup of melanocytic lesions, including leptomeningeal melanocytic lesions, UMs and blue nevi but not in CM (16,17,28). The fact that these GNAQ-mutated melanocytic lesions seem to be derived from non-epithelium related melanocytes may indicate that GNAQ mutations preferentially occur in melanocytes present in extra-epithelial structures such as leptomeninges, dermis and uvea. However, whether the site-specific occurrence of GNAQ mutations is a consequence of the mutations themselves or represents different target cell populations needs to be further determined. GNAQ codon 209 mutations seem to be an early event in the pathogenesis since they occur in benign as well as malignant melanocytic lesions. These activating codon 209 mutations form an alternative route for MAP kinase activation and probably represent an initial stimulus for melanocytes to proliferate (22,28). Inhibition of GNAQ-dependent MAP kinase signalling may form a novel option for targeted therapy of these tumors. In conclusion, this study illustrates the difficult differential diagnosis of melanotic schwannoma with leptomeningeal melanocytoma, and shows that a detailed analysis of morphology combined with molecular analysis of the GNAQ hotspot mutation can aid herein. Correct diagnosis of melanotic schwannoma is important to select those patients who need further clinical evaluation for possible association with Carney Complex. Acknowledgments We thank the PALGA foundation, the national histopathology and cytopathology database in The Netherlands, for providing us with anonymized patient information. We are grateful to Clemens Prinsen, Mandy Wouters and Inge Arbeider for performing 120

122 Improved discrimination of melanotic schwannoma from melanocytic lesions the DNA isolations, and Susan Van de Meer, Susan Noorlander, Renske Zenhorst, Coos Diepenbroeck and Monique Link for supervising and/or performing the histochemical and immunohistochemical stains. We thank Paul Rombout for his technical assistance with the molecular analyses and Peter de Wilde for assistance with the statistical analysis

123 Chapter 6 REFERENCES 1. Brat DJ, Perry A (2007) Melanocytic lesions. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) WHO classification of tumours of the central nervous system, 4 th edn. IARC, Lyon, p Brat DJ, Giannini C, Scheithauer BW, Burger PC (1999) Primary melanocytic neoplasms of the central nervous systems. Am J Surg Pathol 23: Burns DK, Silva FG, Forde KA, Mount PM, Clark HB (1983) Primary melanocytic schwannoma of the stomach. Evidence of dual melanocytic and schwannian differentiation in an extra-axial site in a patient without neurofibromatosis. Cancer 52: Carlson JA, Dickersin GR, Sober AJ, Barnhill RL (1995) Desmoplastic neurotropic melanoma. A clinicopathologic analysis of 28 cases. Cancer 75: Carney JA (1990) Psammomatous melanotic schwannoma. A distinctive, heritable tumor with special associations, including cardiac myxoma and the Cushing syndrome. Am J Surg Pathol 14: Carrasco CA, Rojas-Salazar D, Chiorino R, Venega JC, Wohllk N (2006) Melanotic nonpsammomatous trigeminal schwannoma as the first manifestation of Carney complex: case report. Neurosurgery 59:E1334-E Casparie M, Tiebosch AT, Burger G et al (2007) Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol 29: Di Bella C, Declich P, Assi A et al (1997) Melanotic schwannoma of the sympathetic ganglia: a histologic, immunohistochemical and ultrastructural study. J Neurooncol 35: Dong Q, Shenker A, Way J et al (1995) Molecular cloning of human G alpha q cdna and chromosomal localization of the G alpha q gene (GNAQ) and a processed pseudogene. Genomics 30: Font RL, Truong LD (1984) Melanotic schwannoma of soft tissues. Electron-microscopic observations and review of literature. Am J Surg Pathol 8: Fukuda T, Igarashi T, Hiraki H, Yamaki T, Baba K, Suzuki T (2000) Abnormal pigmentation of schwannoma attributed to excess production of neuromelanin-like pigment. Pathol Int 50: Huang HY, Park N, Erlandson RA, Antonescu CR (2004) Immunohistochemical and ultrastructural comparative study of external lamina structure in 31 cases of cellular, classical, and melanotic schwannomas. Appl Immunohistochem Mol Morphol 12: Kaehler KC, Russo PA, Katenkamp D et al (2008) Melanocytic schwannoma of the cutaneous and subcutaneous tissues: three cases and a review of the literature. Melanoma Res 18: Kirschner LS, Sandrini F, Monbo J, Lin JP, Carney JA, Stratakis CA (2000) Genetic heterogeneity and spectrum of mutations of the PRKAR1A gene in patients with the carney complex. Hum Mol Genet 9: Krausz T, Azzopardi JG, Pearse E (1984) Malignant melanoma of the sympathetic chain: with a consideration of pigmented nerve sheath tumours. Histopathology 8: Küsters-Vandevelde HV, Klaasen A, Küsters B et al (2009) Activating mutations of the GNAQ gene: a frequent event in primary melanocytic neoplasms of the central nervous system. Acta Neuropathol 119: Lamba S, Felicioni L, Buttitta F et al (2009) Mutational profile of GNAQQ209 in human tumors. PLoS One 4:e Monteagudo C, Ferrandez A, Gonzalez-Devesa M, Llombart-Bosch A (2001) Psammomatous malignant melanoma arising in an intradermal naevus. Histopathology 39: Mooi WJ, Krausz T (2007) Blue nevus and related lesions. In: Mooi WJ, Krausz T (eds) Pathology of Melanocytic Disorders, 2 nd ed. Hodder Arnold, London, p Mooi WJ, Krausz T (2007) Other extracutaneous melanotic tumors. In: Mooi WJ, Krausz T (eds) Pathology of Melanocytic Disorders, 2 nd ed. Hodder Arnold, London, p Prieto VG, Woodruff JM (1998) Expression of basement membrane antigens in spindle cell melanoma. J Cutan Pathol 25: Ross EM, Wilkie TM (2000) GTPase-activating proteins for heterotrimeric G proteins: regulators of G protein signaling (RGS) and RGS-like proteins. Annu Rev Biochem 69:

124 Improved discrimination of melanotic schwannoma from melanocytic lesions 23. Schaumburg-Lever G, Lever I, Fehrenbacher B et al (2000) Melanocytes in nevi and melanomas synthesize basement membrane and basement membrane-like material. An immunohistochemical and electron microscopic study including immunoelectron microscopy. J Cutan Pathol 27: Scheithauer BW, Louis DN, Hunter S, Woodruff JM, Antonescu CR (2007) Schwannoma.In: Louis DN, Oghaki H, Wiestler OD, Cavenee WK (eds) WHO Classification of Tumours of the Central Nervous System, 4 th ed. IARC, Lyon, p Scheithauer BW, Woodruff JM, Erlandson RA (1999) Tumors of the Peripheral Nervous System. Armed Forces Institute of Pathology, Washington DC 26. Schmoeckel C, Stolz W, Sakai LY, Burgeson RE, Timpl R, Krieg T (1989) Structure of basement membranes in malignant melanoma and nevocytic nevi. J Invest Dermatol 92: Vallat-Decouvelaere AV, Wassef M, Lot G et al (1999) Spinal melanotic schwannoma: a tumour with poor prognosis. Histopathology 35: Van Raamsdonk CD, Bezrookove V, Green G et al (2009) Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature 457: Veugelers M, Wilkes D, Burton K et al (2004) Comparative PRKAR1A genotype-phenotype analyses in humans with Carney complex and prkar1a haploinsufficient mice. Proc Natl Acad Sci U S A 101:

125

126 7 Genomic analysis of leptomeningeal and skin melanocytic tumors in neurocutaneous melanosis; potential clues for diagnosis and prognosis Adriana C.H. van Engen-van Grunsven 1, *, Katrin Rabold 1, *, Heidi V.N. Küsters-Vandevelde 2, Jos Rijntjes 1, Melika Djafarihamedani 1, Jayne Y. Hehir-Kwa 3, Michel A.A.P. Willemsen 4, Ineke van der Burgt 5, Pieter Wesseling 1,2,6,Willeke A.M. Blokx 1, Patricia J.T.A. Groenen 1 *Both authors contributed equally to this work; shared first authorship. 1 Department of Pathology, Radboud University Medical Center, Nijmegen; 2 Department of Pathology, Canisius Wilhelmina Hospital, Nijmegen; 3 Department of Human Genetics, Donders Centre for Neuroscience, Radboud University Medical Center, Nijmegen, 4 Department of Pediatric Neurology, Radboud University Medical Center, Nijmegen, 5 Department of Genetics, Radboud University Medical Center, Nijmegen, 6 Department of Pathology, VU University Medical Center, Amsterdam. Revised manuscript published in Acta Neuropathol Feb;133(2):

127 Chapter 7 ABSTRACT Neurocutaneous melanosis (NCM) is a rare congenital syndrome characterized by the presence of large or multiple congenital melanocytic nevi (CMN) of the skin and melanotic and nonmelanotic lesions in the central nervous system (CNS). It is hard to predict which patients with large or multiple CMN will develop symptomatic NCM. The prognosis for patients with diffuse leptomeningeal melanocytic deposits or CZS melanoma is currently poor as there is no effective treatment. We studied central nervous system melanocytic tumors (CNS-MTs) from five NCM patients and corresponding CMN from three of these patients by genome-wide analysis of copy number variations and analysis of (hotspot) mutations in a panel of genes. All CNS-MTs and CMN harbored a heterozygous NRAS Q61 mutation, confirming a central role for NRAS mutations in NCM-associated melanocytic neoplasms. Additionally, in the CNS-MTs, but not in the CMN, multiple large copy number variations were present. Focal LOH at 6p22.1 and focal genomic aberrations (copy neutral LOH and or copy number loss) at 10q22.1 were found in all CMN and all CNS-MTs of the pediatric patients and may thus indicate underlying NCM when found in CMN. Our findings provide clues for diagnosis and prognosis for patients with NCM. 126

128 Genomic analysis of melanocytic tumors in neurocutaneous melanosis INTRODUCTION Neurocutaneous melanosis (NCM) is a rare, congenital syndrome characterized by the association of large or multiple congenital melanocytic nevi (CMN) of the skin and central nervous system (CNS) abnormalities (1-4). The CNS abnormalities in NCM are broad and include both melanocytic as well as non-melanocytic lesions with widely varying clinical outcomes (5-7). The CNS melanocytic lesions in NCM range from mere melanin deposits in the brain parenchyma to proliferations of melanocytic cells in the CNS, mostly in or along the leptomeninges. These melanocytic cells can be cytologically benign or malignant, called melanocytosis or melanoma, respectively. Diffuse leptomeningeal melanocytic deposits are associated with poor survival (7). Neurological manifestations in NCM patients usually occur before the age of two years, and less frequently, later on in the second or third decade of life (4, 8-11). Relatively little is known about the exact risk that patients with large or multiple CMN have of developing neurological manifestations. The prevalence of radiological CNS abnormalities although increases with the severity of cutaneous phenotype, i.e. size of largest CMN and total number of nevi (4, 12, 13). Bittencourt et al. (14) prospectively followed 160 patients with large CMN. Four patients developed manifest NCM (2,5%). DeDavid et al. (15) found a higher incidence, with manifest NCM in 33 patients among a population of 289 patients with large cutaneous melanocytic nevi (11%). Recently, Bekiesinska et al. (16) reported an even higher incidence of NCM (29.2 %) in a series of 24 children with giant CMN. CNS magnetic resonance imaging has recently been proposed as a screening instrument in children with multiple CMN (6). The histologic features of the congenital nevi in NCM are the same as in congenital nevi found in patients who never develop neurologic symptoms. Acquisition of post-zygotic somatic NRAS mutations occurring during early embryonic migration of melanocytes is thought to underlie the pathogenesis of NCM (17). These mutations mainly affect codon 61 of NRAS. In a recent series, twelve of sixteen patients with NCM were shown to harbor a somatic NRAS Q61 mutation in their CMN (75%) (18). Kinsler et al. found mutations in codon 61 of NRAS in affected neurological and cutaneous tissues of 12 out of 15 patients (80%). Mutations were absent in unaffected tissues and blood consistent with NRAS mutation mosaicism (5). Other studies have also shown the presence of NRAS Q61 mutations in both the CMN and the central nervous system melanocytic tumors (CNS-MT) of NCM patients (19, 20). It is currently unclear whether BRAF mutations are also involved in the pathogenesis of NCM. Knowledge on copy number variations (CNVs) in CNS-MTs is largely lacking. In the single study that has been published on this subject, Kinsler et al. reported absence of large gains or losses in one case of NRAS-mutated melanocytosis in an 8-year-old boy with NCM (using array CGH) (5). In contrast, they detected multiple gains and 7 127

129 Chapter 7 losses in a primary NRAS-mutated melanoma in the cerebellum of a 10-year-old child with NCM that developed liver metastases. The aberrations in this melanoma included, among others, loss of whole chromosome 3, gain of chromosome 8 and loss of 9p24.3 p21.1, the latter including the CDKN2A locus (5). However, given the small number of cases studied, the diagnostic and prognostic significance of these CNVs is unclear. The aim of our study was to further elucidate the genetic alterations in melanocytic nevi and CNS melanocytic tumors of patients with NCM. Paired samples of skin nevi and CNS melanocytic tumors of five NCM patients were investigated by genome wide CNV analysis and next-generation sequencing to detect mutations in a panel of cancer genes. Results were correlated with clinical characteristics including survival. MATERIALS AND METHODS Patients Formalin-fixed, paraffin-embedded (FFPE) tissue was obtained from two NCM patients from our own hospital (Radboud UMC) and from three NCM patients through PALGA (the Dutch nationwide histopathology and cytopathology data network and archive) (31). Main patient characteristics are summarized in Table 1. Patients 1-4 are children with an age ranging from four to nineteen at the time of pathological diagnosis of the CNS-MT. All these children died of disease. Patient 5 is a 30-year-old woman alive two years after diagnosis of melanomatosis. None of the patients had clinical evidence of primary melanoma outside the CNS. Histology of all lesions was revised by WB and HK. Patients 1 and 5 had a histological diagnosis of diffuse melanomatosis, patients 2 and 4 were diagnosed with CNS melanoma, and patient 3 with melanocytosis. From three patients (cases 2, 3 and 5) FFPE tissue from both a CMN and the CNS melanocytic tumor was available. From patient 3 we had FFPE material from different stages of disease. Some data from patient 3 has been published before (20). DNA extraction For DNA extraction, areas with an estimated neoplastic cell percentage of at least 50% were selected and manually dissected. DNA was extracted by digestion in a proteinase K/Chelex-100 solution. After proteinase K inactivation, the mixture was spun for five minutes and the supernatant containing the DNA was used for DNA purification with the QIAamp DNA micro kit 50 (Qiagen, Stanford, CA, USA). DNA concentration was determined with a Qubit 2.0 Fluorometer (Life Technologies, Carsbad, CA, USA). Quality control of the extracted DNA was performed with size ladder PCR. 128

130 Genomic analysis of melanocytic tumors in neurocutaneous melanosis Table 1 Patient characteristics and genomic tests Mutations OncoScan Follow-up Sex Sample Sample collection Estimated % of neoplastic cells Patient Age at diagnosis + DOD 1 4 M Brain: melanomatosis Biopsy at 4y NRASc.182A>G (p.(gln61arg)); Q61R; 41% mutant allele + DOD 2 7 M Skin: congenital nevus Biopsy at 3y NRAS c.181c>a (p.(gln61lys)); Q61K; 42% mutant allele + Brain: melanoma Biopsy at 7y > 80 NRAS c.181c>a (p.(gln61lys)); Q61K; 40% mutant allele + DOD 3 13 M Skin: congenital nevus Biopsy at 2 months 80 NRAS c.181c>a (p.(gln61lys)); Q61K; 35% mutant allele + Biopsy at 13y 50 NRAS c.181c>a (p.(gln61lys)); Q61K; 34% mutant allele Brain: melanocytosis before treatment + Autopsy 80 NRAS c.181c>a (p.(gln61lys)); Q61K; 46% mutant allele Brain: melanocytosis after treatment - Autopsy 80 NRAS c.181c>a (p.(gln61lys)); Q61K; 25% mutant allele Skin: congenital nevus after treatment + DOD 4 19 M Brain: melanoma Autopsy 80 NRAS c.181c>a (p.(gln61lys)); Q61K; 23% mutant allele + alive 5 30 F Skin: congenital nevus Biopsy at 24y 90.NRAS c.182a>g (p.(gln61arg)); Q61R; 51% mutant allele + Brain: melanomatosis Biopsy at 30y 70 NRAS.c.182A>G (p.(gln61arg)); Q61R; 30% mutant allele Mutation analysis by NGS was performed in duplicate. For each sample, the identified mutation was detected in duplicate with similar mutant allele. OncoScan + or -: OncoScan has (+) or has not (-) been performed on the sample; DOD died of disease. Case 3 has had experimental treatment with a MEK-inhibitor; see data published before (20). DNA from unaffected skin of patient 3 and DNA from tonsillar tissue from a healthy person were used as control samples

131 Chapter 7 Mutation analysis Mutation analysis of a panel of genes was performed by next-generation sequencing using single molecule Molecular Inversion Probes (smmip) (A. Eijkelenboom et al., Reliable Next Generation Sequencing of FFPE tissue using single molecule tags. Journal of Molecular Diagnostics, In press.) (21). The cancer hotspot gene panel includes the following genes and hotspot regions:akt1 (NM_ ): codon 17, BRAF (NM_ ): codon , CTNNB1 (NM_ ): codon 19-48, EGFR (NM_ ): codon , , , ERBB2 (NM_ ): codon , GNA11 (NM_ ): codon 183 en 209, GNAQ (NM_ ): codon 183 en 209, GNAS (NM_ ): codon 201 en 227, H3F3A (NM_ ): codon 28 en 35, H3F3B (NM_ ): codon 37, HRAS (NM_ ): codon 12, 13, 59 en 61, IDH1 (NM_ ): codon 132, IDH2 (NM_ ): codon 140 en 172, JAK2 (NM_ ): codon 617, KIT (NM_ ): codon , , , , KRAS (NM_ ): codon 12, 13, 59, 61, 117 en 146, MPL (NM_ ): codon 515, MYD88 (NM_ ): codon 265, NRAS (NM_ ): codon 12, 13, 59, 61, 117 en 146, PDGFRA (NM_ ): codon , , , PIK3CA (NM_ ): codon , Using the TP53/CDKN2A gene panel sequence analysis more than>95% of the coding regions of TP53 (NM_ ) and CDKN2A (P14: NM_ en P16: NM_ ) was performed. Analysis of copy-number variations by OncoScan analysis Copy-number alterations were determined by OncoScan array analysis (molecular inversion probe (MIP) technology) (22) as we have described before (23). Due to limited amounts of material and DNA and/or limited quality of the DNA, it was not possible to analyze all the collected specimens by OncoScan analysis as can be seen in Table 1. DNA from unaffected skin from patient 3 and DNA from tonsillar tissue from a healthy person were used as control samples. Data were analyzed using Nexus Copy Number 7.5 genomic software (BioDiscovery, El Segundo, CA, USA). A copy number of 2 per individual is considered normal. CNVs were identified with the SNP-Fast Adaptive States Segmentation Technique 2 (SNP-FASST2) segmentation algorithm. Copy gains were defined with log2ratio values between 0.3 and 1.2, copy losses were defined with log2ratio values between -0.3 and Log2ratio values exceeding 1.2 and -1.2 were defined as high gains and big losses, respectively. To rule out that the found aberrations do occur in the general population, they were compared to known CNVs reported in the Database of Genomic Variants (DGV), Toronto ( comprising more than 2.5 million entries identified in more than 22,300 genomes (24). 130

132 Genomic analysis of melanocytic tumors in neurocutaneous melanosis RESULTS Mutation analysis Main patient characteristics are summarized in Table 1. Mutation analysis could be performed for all samples. All samples harbored a NRAS Q61 mutation; two patients had a NRAS Q61R and three had a NRAS Q61K mutation. Based upon comparison of the mutated allele frequency with the estimated neoplastic cell percentage, all NRAS mutations were heterozygous mutations. Remarkably, all samples of patient 3, including a sample of normal tissue, harbored a concomitant PDGFRA c.1895c>t (p.(thr632met)) variant with approximately 50% variant allele frequency in the tissues examined. This variant is described as an infrequently occurring single nucleotide polymorphism (SNP, rs in exac, broadinstitute.org). No other mutations in the regions of the genes investigated were found. Copy number variations Figure 1 shows an overview of all genome wide CNVs that were found. Table 2 provides information on the larger gains and losses (defined as 30Mb) per sample. In contrast to the CMN, all CNS melanocytic tumors showed various relatively large CNVs. The CMN contained only focal aberrations, mostly copy neutral loss of heterozygosity (cnlohs). Compared to the malignant CNS-MTs of patients 1, 2 and 4, the leptomeningeal melanomatosis of the adult female patient (patient 5) had a relative low number of large aberrations. Interestingly, this patient presented relatively late in life with neurological symptoms and has shown only slow progression of the disease so far. Patient 4 showed a loss of 1p including NRAS (located at 1p13.2). The other samples did not show any CNVs in this region. None of the samples showed CNVs in the region of CDKN2A (9p21). In all CNS-MT samples copy number gain and allelic imbalance of 6p was present. Gain of 8q was present in all CNS-MT samples as well, except for the melanomatosis of patient 1. These CNVs were absent in the nevi. From patient 3 DNA was available from different lesions at different stages of disease (i.e. brain biopsy at the time of diagnosis and brain tissue from autopsy performed five months later). Compared to the findings in the brain biopsy material, the CNVs in the autopsy material were more extensive with involvement of whole (arms of) chromosomes. In the autopsy material there was loss of 8p and gain of 8q, 13q, 15q and the entire chromosome 20 and LOH of 8p (see Figure 1 and Table 2). Some of these aberrations occurred specifically in this patient while others were also (partially) detected in CNS-MTs of the other patients. The latter comprise copy number gain and allelic imbalance of 13q and 15q, copy number loss and LOH of 8p as well as copy number gain and allelic imbalance of 8q, including MYC. In contrast to the long arms of chromosome 13 and 15, aberrations of chromosome 8 have frequently been 7 131

133 Chapter 7 Figure 1 Overview of all copy number variations detected by OncoScan analysis. Top view: Aggregated genomic view with copy number gains and losses from all samples for the different chromosomes (depicted on the x-axis) from all samples. The bald black horizontal lines delineate the aggregate % cut-off value which was set at 30%. Bottom view: Copy number aberrations of the individual samples for the different chromosomes. Bars in red: Copy number loss; Bars in blue: Copy number gain; Region indicated by purple line: Allelic imbalance; Region indicated by yellow line: Loss of heterozygosity. bef tr = before treatment; after tr = after treatment. 132

134 Genomic analysis of melanocytic tumors in neurocutaneous melanosis Table 2 All gains and losses 30Mb Patient/ sample Large copy number Coordinates changes (>30Mb) 1 melanomatosis Gain 1q q32.1* chr1:149,039, ,944,465 Gain 1q q44* chr1:207,013, ,250,621 Gain 6p p12.1* chr6:0-54,670,464 Loss 6q q27 chr6:93,921, ,115,067 Gain 20q q13.33 chr20:29,445,644-63,025,520 2 melanoma Gain 1q q32.1* chr1:151,549, ,458,099 Gain 1q q43* chr1:201,709, ,573,334 Gain 6p p21.31* chr6:0-35,406,385 Loss 6q12 - q27 chr6:63,869, ,115,067 Gain 8q q24.3 chr8:46,950, ,364,022 Loss 10p p11.1 chr10:0-39,087,061 Loss 10q q26.3 chr10:42,365, ,534,747 Loss 11q q25 chr11:85,459, ,006,516 Loss 16q q24.3 chr16:49,514,255-90,354,753 2 congenital nevus melanocytosis before treatment melanocytosis after treatment Loss 8p p11.21 chr8:0-40,514,039 Gain 8q q24.3 chr8:46,950, ,364,022 Gain 13q q34 chr13:49,183, ,169,878 Gain 15q q26.3 chr15:21,308, ,531,392 Gain 20q q13.33 chr20:29,445,644-63,025,520 3 congenital nevus melanoma Gain 6p p21.31* chr6:0-35,390,085 Gain 8p p11.1 chr8:0-43,749,726 Loss 10p p11.1 chr10:0-39,087,061 Gain 20q q13.33 chr20:29,445,644-63,025,520 5 melanomatosis Gain 2q q37.3 chr2:179,652, ,199,373 Gain 6p p12.3* chr6:0-50,006,716 Gain 8q q24.3 chr8:72,068, ,364,022 Control skin case Control tonsil A list of all large gains and losses (defined as 30Mb) found with the OncoScan analysis. = large losses absent. * CNVs also found by Kinsler et al. (5) 133

135 Chapter 7 described in various melanoma types. Aberration of 8q was present in all CNS-MT samples in our series except for case 1. Aberrations of 8q are frequent in uveal and skin melanomas we have analyzed before ((23) and unpublished data). Focal LOH at 6p22.1 was found in both the CNS-MTs and CMN of all pediatric patients but not in the lesions of the adult patient in our series. Copy neutral LOH at 10q22.1 was found in all cases except for the melanoma of patient 4 (which showed copy number loss) and the melanoma of patient 2 (which showed LOH and copy number loss). DISCUSSION Here we provide the results of whole-genome analyses of CNVs and of mutation analysis of melanocytic lesions from both the skin and the CNS of NCM patients. Our findings may help to identify genetic factors associated with the oncogenesis and progression of these lesions and to identify new targets for treatment. All melanocytic lesions from both the skin and the CNS harbored an activating NRAS codon 61 mutation, which is fully in line with data from the literature (5). The presence of NRAS mutations in most NCM cases could be helpful in future targeted treatment of NCM patients with neurological symptoms or melanoma development (20, 25, 26). While none of the samples in our study harbored a BRAF mutation, others have presented data indicating that BRAF mutations may occasionally be involved in the pathogenesis of NCM as well. Salgado et al. reported a BRAF V600E mutation in the CMN of two NCM patients (2/16, 12.5%) but the CNS-MTs of these patients were not investigated (18). Also, a minority of patients with large or giant CMN outside the context of NCM may harbor BRAF V600E instead of NRAS Q61 mutations (27), but data are contradictory since other studies have not found BRAF mutations in these CMN (28, 29). Obviously, demonstration of a BRAF mutation in the melanocytic tumors of patients with NCM may have therapeutic consequences and needs further investigation. Saldago et al. (30) have described amplification of mutated NRAS (p.q61r) in a patient with NCM who died at 17 months of age due to disseminated disease. They have hypothesized that amplification of mutated NRAS may be a genetic mechanism leading to melanoma development in the context of NCM. We did not see amplification of mutated NRAS in our series. One of our patients harbored an infrequently found single nucleotide polymorphism in platelet derived growth factor receptor alpha (PDGFRA), the relevance of which in the context of the developed melanocytic lesion is unknown. Of note, gene mutations in PDGFRA have been described in other melanoma subtypes (31, 32). Knowledge on CNVs in NCM has been largely lacking. OncoScan analysis revealed substantial differences in the number of aberrations between the CNS-MTs and the 134

136 Genomic analysis of melanocytic tumors in neurocutaneous melanosis CMN in our series. The nevi had very few and small aberrations, mostly cnlohs whereas the CNS-MTs had a high number of various large copy number variations (defined as 30Mb) (see Figure 1 and Table 2). This finding is in line with the previously described, highly significant differences in CNVs between cutaneous melanoma and nevi, with only Spitz nevi having some form of aberration (33). In the only previously published study on genomic and mutational aberrations in NCM associated neoplasms, Kinsler et al. reported multiple losses and gains in a primary melanoma in the cerebellum of a 10-year old child (5). The large aberrations (like in our own study defined as 30Mb) they described were gain of 1q21.1-q44, loss of chromosome 3, loss of 5p15.33-q11.2, gain of 6p25.3-p11.2, gain of whole chromosome 8 and loss of 9p24.3-p21.1, loss of 11p15.5-p13 and loss of chromosome 18 (5). We detected a gain in nearly exactly the same region of 1q in two of our CNS-MT cases (case 1 and 2; see Table 2). The gain of 6p described by Kinsler et al. was present in all malignant CNS-MTs of our series (cases 1, 2, 4 and 5- see Table 2). In our series, gain of 8q was detected in four of the CNS-MTs (cases 2, 3, 4 and 5- see Table 2). Gain of the short arm of chromosome 8 was seen in case 4. However, we did not see gain of whole chromosome 8. Furthermore, we describe allelic imbalance or cnloh of 9p24.3-p13.1 in case 3 but no loss in this region, whereas Kinsler et al. reported a loss of a larger region in 9p. We did not find loss of chromosome 3, 5, 11p or 18 in our cases. The different CNS-MT samples of patient 3, obtained in different phases of the disease (i.e. brain biopsy at time of diagnosis and brain tissue from autopsy performed five months later), enabled us to also study potential markers of progression. This patient received experimental treatment with a MEK-inhibitor shortly before he died as has been published (20). We did not find a significant change in mutant allele frequency in the context of the estimated percentage of neoplastic cells of the NRAS Q61k mutation in the samples before and after treatment. Kinsler et al. reported progression from heterozygosity to homozygosity for a NRAS Q61K mutation in samples of a cutaneous lesion before and after malignant progression of this lesion (5). Our case differed from the case of Kinsler et al. as in our patient no apparent change from histologically benign to malignant occurred. In addition, the only NRAS homozygous melanoma case described by Kinsler et al. was a cutaneous melanoma developing in a congenital nevus. They did not find loss of heterozygosity for NRAS mutations in malignant CNS tumors. We also did not find homozygosity for NRAS in our malignant cases in the CNS. It could be hypothesized that malignant progression in NCM might be different in cutaneous and leptomeningeal locations. OncoScan analysis of the CNS-MT in this case did show that the aberrations occurring at an earlier stage of disease were more focal and only small (defined as <30Mb), whereas they were more pronounced and sometimes extended over whole chromosome arms or even whole chromosomes after progression (see Figure 1 and Table 2). This supports the presumption that additional aberrations occur during progression. In 7 135

137 Chapter 7 the material of this patient, five CNVs may be associated with malignant progression: loss of 8p, and gain of 8q, 13q, 15q and 20 (see Table 2). As these regions harbor hundreds of genes it is not possible to indicate one or a few genes likely to be involved in the more aggressive behavior of the tumor cells. Copy number mean gains of 6p and 8q were detected as common aberrations in the CNS-MTs (see Table 2). Both aberrations have frequently been described in melanomas of different origin (23, 33-37). The 8q copy number gain was detected in all the CNS-MTs except for the melanomatosis of case 1. It was relatively high in both circumscribed lesions (case 2 and 4). Notably, regions with copy number gain of 8q were only focally observed, in short stretches of the genome, in the melanocytosis of case 3 before treatment. After treatment, copy number gain of the entire 8q arm was seen. LOH at 6p22.1 was found in both the CNS-MTs and CMN of all pediatric patients but not in the lesions of the adult patient in our series. Copy neutral LOH at 10q22.1 was found in the CMN of all the patients and in most of the CNS-MTs except for the melanoma of patient 4 (who showed copy number loss at 10q22.1) and the melanoma of case 2 (which showed LOH and copy number loss of 10q22.1). Using our in house database including 1400 samples (DNA from saliva of 1400 students), we have not detected common aberrations in these regions. For future research, it would be very interesting to test large CMN of children without NCM in order to know whether these aberrations are specifically associated with NCM. If so, testing CMN of children for these aberrations could potentially be a diagnostic test that helps to determine whether a patient has NCM with the risk of developing neurologic manifestations. In conclusion, all CMN and CNS-MT obtained from 5 NCM patients harbored an activating NRAS mutation, which may provide a potential target for therapy. The CMN had relatively few CNVs compared to the CNS-MTs. Five CNVs (loss of 8p, and gain of 8q, 13q, 15q and 20) may be associated with more aggressive behavior in the CNS-MTs. In addition, LOH at 6p22.1 and copy neutral LOH at 10q22.1 may serve as an indicator of underlying NCM when found in large congenital nevi. Conflict of interest The authors state no conflict of interest. Acknowledgements This work was supported by a research grant from the Koppie-Au foundation, Breda, The Netherlands. We are grateful to all the patients and their families who granted permission to perform this research. We thank Rolph Pfundt for technical assistance with the analysis of the OncoScan data and Annemiek Kastner-vanRaaij for performance of the mutation analysis. 136

138 Genomic analysis of melanocytic tumors in neurocutaneous melanosis REFERENCES 1. Rokitansky J. Ein ausgezeichneter Fall von Pigment-Mal mit ausgebreiteter Pigmentierung der inneren Hirn- und Ruchenmarkshaute. Allg Wien Med Z. 1861;6: Van Bogaert L. La melanose neurocutanee diffuse heredo-familiale. Bull AcadR Med Belg. 1948;13: Fox H. Neurocutaneous melanosis. Vinken PJ BG, editor. Amsterdam: North Holland Kadonaga JN, Frieden IJ. Neurocutaneous melanosis: definition and review of the literature. Journal of the American Academy of Dermatology. 1991;24(5 Pt 1): Kinsler VA, Thomas AC, Ishida M, Bulstrode NW, Loughlin S, Hing S, et al. Multiple congenital melanocytic nevi and neurocutaneous melanosis are caused by postzygotic mutations in codon 61 of NRAS. The Journal of investigative dermatology. 2013;133(9): Waelchli R, Aylett SE, Atherton D, Thompson DJ, Chong WK, Kinsler VA. Classification of neurological abnormalities in children with congenital melanocytic naevus syndrome identifies magnetic resonance imaging as the best predictor of clinical outcome. The British journal of dermatology. 2015;173(3): Ramaswamy V, Delaney H, Haque S, Marghoob A, Khakoo Y. Spectrum of central nervous system abnormalities in neurocutaneous melanocytosis. Developmental medicine and child neurology. 2012; 54(6): Di Rocco F, Sabatino G, Koutzoglou M, Battaglia D, Caldarelli M, Tamburrini G. Neurocutaneous melanosis. Child s nervous system: ChNS: official journal of the International Society for Pediatric Neurosurgery. 2004;20(1): Kasantikul V, Shuangshoti S, Pattanaruenglai A, Kaoroptham S. Intraspinal melanotic arachnoid cyst and lipoma in neurocutaneous melanosis. Surgical neurology. 1989;31(2): Feuillet L, Kaphan E, Audoin B, Witjas T, Pelletier J, Pellissier JF, et al. [Melanocytic meningitis and large congenital melanocytic naevus: neurocutaneous melanosis]. Revue neurologique. 2003;159(4): Araujo C, Resende C, Pardal F, Brito C. Giant congenital melanocytic nevi and neurocutaneous melanosis. Case reports in medicine. 2015;2015: Kinsler VA, Chong WK, Aylett SE, Atherton DJ. Complications of congenital melanocytic naevi in children: analysis of 16 years experience and clinical practice. The British journal of dermatology. 2008;159(4): Hale EK, Stein J, Ben-Porat L, Panageas KS, Eichenbaum MS, Marghoob AA, et al. Association of melanoma and neurocutaneous melanocytosis with large congenital melanocytic naevi--results from the NYU-LCMN registry. The British journal of dermatology. 2005;152(3): Bittencourt FV, Marghoob AA, Kopf AW, Koenig KL, Bart RS. Large congenital melanocytic nevi and the risk for development of malignant melanoma and neurocutaneous melanocytosis. Pediatrics. 2000;106(4): DeDavid M, Orlow SJ, Provost N, Marghoob AA, Rao BK, Huang CL, et al. A study of large congenital melanocytic nevi and associated malignant melanomas: review of cases in the New York University Registry and the world literature. Journal of the American Academy of Dermatology. 1997;36(3 Pt 1): Bekiesinska-Figatowska M, Szczygielski O, Boczar M, Madzik J, Klepacka T, Michalak E, et al. Neurocutaneous melanosis in children with giant congenital melanocytic nevi. Clinical imaging. 2014;38(2): Kusters-Vandevelde HV, Kusters B, van Engen-van Grunsven AC, Groenen PJ, Wesseling P, Blokx WA. Primary melanocytic tumors of the central nervous system: a review with focus on molecular aspects. Brain pathology. 2015;25(2): Salgado CM, Basu D, Nikiforova M, Bauer BS, Johnson D, Rundell V, et al. BRAF mutations are also associated with neurocutaneous melanocytosis and large/giant congenital melanocytic nevi. Pediatric and developmental pathology : the official journal of the Society for Pediatric Pathology and the Paediatric Pathology Society. 2015;18(1): Pedersen M, Kusters-Vandevelde HV, Viros A, Groenen PJ, Sanchez-Laorden B, Gilhuis JH, et al. Primary melanoma of the CNS in children is driven by congenital expression of oncogenic NRAS in melanocytes. Cancer discovery. 2013;3(4): Kusters-Vandevelde HV, Willemsen AE, Groenen PJ, Kusters B, Lammens M, Wesseling P, et al. Experimental treatment of NRAS-mutated neurocutaneous melanocytosis with MEK162, a MEK-inhibitor. Acta neuropathologica communications. 2014;2:

139 Chapter Hiatt JB, Pritchard CC, Salipante SJ, O Roak BJ, Shendure J. Single molecule molecular inversion probes for targeted, high-accuracy detection of low-frequency variation. Genome research. 2013;23(5): Wang Y, Cottman M, Schiffman JD. Molecular inversion probes: a novel microarray technology and its application in cancer research. Cancer genetics. 2012;205(7-8): van Engen-van Grunsven AC, Baar MP, Pfundt R, Rijntjes J, Kusters-Vandevelde HV, Delbecq AL, et al. Whole-genome copy-number analysis identifies new leads for chromosomal aberrations involved in the oncogenesis and metastastic behavior of uveal melanomas. Melanoma research. 2015;25(3): MacDonald JR, Ziman R, Yuen RK, Feuk L, Scherer SW. The Database of Genomic Variants: a curated collection of structural variation in the human genome. Nucleic acids research. 2014;42(Database issue): D Ruan Y, Kovalchuk A, Jayanthan A, Lun X, Nagashima Y, Kovalchuk O, et al. Druggable targets in pediatric neurocutaneous melanocytosis: Molecular and drug sensitivity studies in xenograft and ex vivo tumor cell culture to identify agents for therapy. Neuro-oncology. 2015;17(6): Pawlikowski JS, Brock C, Chen SC, Al-Olabi L, Nixon C, McGregor F, et al. Acute Inhibition of MEK Suppresses Congenital Melanocytic Nevus Syndrome in a Murine Model Driven by Activated NRAS and Wnt Signaling. The Journal of investigative dermatology. 2015;135(8): Dessars B, De Raeve LE, El Housni H, Debouck CJ, Sidon PJ, Morandini R, et al. Chromosomal translocations as a mechanism of BRAF activation in two cases of large congenital melanocytic nevi. The Journal of investigative dermatology. 2007;127(6): Wu D, Wang M, Wang X, Yin N, Song T, Li H, et al. Lack of BRAF(V600E) mutations in giant congenital melanocytic nevi in a Chinese population. The American Journal of dermatopathology. 2011;33(4): Charbel C, Fontaine RH, Malouf GG, Picard A, Kadlub N, El-Murr N, et al. NRAS mutation is the sole recurrent somatic mutation in large congenital melanocytic nevi. The Journal of investigative dermatology. 2014;134(4): Salgado CM, Basu D, Nikiforova M, Hamilton RL, Gehris R, Jakacki R, et al. Amplification of mutated NRAS leading to congenital melanoma in neurocutaneous melanocytosis. Melanoma research. 2015;25(5): Dai J, Kong Y, Si L, Chi Z, Cui C, Sheng X, et al. Large-scale analysis of PDGFRA mutations in melanomas and evaluation of their sensitivity to tyrosine kinase inhibitors imatinib and crenolanib. Clinical cancer research : an official journal of the American Association for Cancer Research. 2013;19(24): Wallander ML, Layfield LJ, Emerson LL, Mamalis N, Davis D, Tripp SR, et al. KIT mutations in ocular melanoma: frequency and anatomic distribution. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2011;24(8): Bastian BC, Olshen AB, LeBoit PE, Pinkel D. Classifying melanocytic tumors based on DNA copy number changes. The American journal of pathology. 2003;163(5): Kabbarah O, Nogueira C, Feng B, Nazarian RM, Bosenberg M, Wu M, et al. Integrative genome comparison of primary and metastatic melanomas. PloS one. 2010;5(5):e Stark M, Hayward N. Genome-wide loss of heterozygosity and copy number analysis in melanoma using high-density single-nucleotide polymorphism arrays. Cancer research. 2007;67(6): Jonsson G, Dahl C, Staaf J, Sandberg T, Bendahl PO, Ringner M, et al. Genomic profiling of malignant melanoma using tiling-resolution arraycgh. Oncogene. 2007;26(32): Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, et al. Distinct sets of genetic alterations in melanoma. The New England journal of medicine. 2005;353(20):

140 Genomic analysis of melanocytic tumors in neurocutaneous melanosis 7 139

141

142 8 NRAS mutations are more prevalent than KIT mutations in melanoma of the female urogenital tract a study of 24 cases from the Netherlands Adriana C.H. van Engen-van Grunsven 1, *, Heidi V.N. Küsters-Vandevelde 2, *, Joanne De Hullu 3, Lucette M. van Duijn 1, Jos Rijntjes 1, Judith V.M.G. Bovée 4, Patricia J.T.A. Groenen 1, Willeke A.M. Blokx 1 *Both authors contributed equally to this work; shared first authorship. 1 Department of Pathology, Radboud University Medical Center, Nijmegen, 2 Department of Pathology, Canisius Wilhelmina Hospital, Nijmegen, 3 Department of Obstetrics and Gynecology, Radboud University Medical Center, Nijmegen, 4 Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands Gynecol Oncol 2014;134:10-14.

143 Chapter 8 ABSTRACT Objective. The aim of this study was to evaluate a series of primary melanomas of the female urogenital tract for oncogenic mutations in KIT, NRAS and BRAF in order to identify patients who may be amenable to targeted therapy. Methods. We reviewed twenty-four cases of female urogenital tract melanomas and used Sanger sequencing analysis for the detection of oncogenic mutations in exons 9, 11, 13, and 17 of KIT; exons 2 and 3 of NRAS; and exon 15 of BRAF. Results. Twenty-four patients were included: fourteen vaginal melanomas, four cervical melanomas, five urethral melanomas and one vulvar melanoma. NRAS mutations (4/24, 21%) were more prevalent than KIT mutations (1/24, 4%), while BRAF mutations were absent. Three of four NRAS mutations were present in vaginal melanomas (21%), mainly affecting codon 61 (3/4). They were mutually exclusive with the KIT mutation. The KIT mutation was present in a vaginal melanoma and affected exon 17. Conclusions. Melanomas of the female urogenital tract relatively commonly harbor mutations in NRAS; this makes NRAS an interesting therapeutic target for these patients in the advanced setting. KIT mutations were rare in our study in contrast to some previous reports. We cannot exclude that anatomical site-related differences and/or population related differences in KIT mutation frequency exist within urogenital tract melanomas. 142

144 Mutation analysis in melanoma of the female urogenital tract INTRODUCTION Primary melanomas of the female urogenital tract are rare and account for approximately 20% of all mucosal melanomas (1). They comprise melanoma of the vulva, vagina, uterine cervix and urethra/ureter/urinary bladder, and originate from melanocytes present in the epithelium of mucosal membranes that cover the urogenital tract. They carry a poor prognosis with a 5-year survival of approximately 11% (1). Diagnosis often occurs at advanced stage (1, 2). In case of vaginal melanoma, about 50% of patients have positive regional lymph nodes, and nearly 20% have distant metastases at time of diagnosis (3). Locoregional control might be obtained by surgery and/or radiation therapy but in case of metastatic disease, few effective treatment options exist (2). It has been shown that melanomas arising from mucosa are genetically distinct from the common cutaneous melanoma. While oncogenic mutations in BRAF are present in approximately 50% of cutaneous melanoma, these mutations are very rare in mucosal melanoma (0-10%) (4, 5). Activating BRAF mutations lead to constitutive activation of the mitogen-activated protein kinase (MAPK) pathway (6), and are an effective target in the treatment of metastatic BRAF-mutated melanomas using selective BRAF inhibitors (such as vemurafenib) (7, 8). Mucosal melanomas commonly harbor activating KIT mutations (approximately 20%) (4, 9-11), that result in activation of several downstream pathways including the MAPK and PI3K/AKT pathways (12). Several case reports and Phase II trials have reported objective responses to the KIT inhibitors imatinib and sorafenib in advanced KIT-mutated melanoma (11, 13-15). Another major oncogene in melanoma is NRAS, which is mutated in approximately 20% of cutaneous melanomas and in varying frequencies in mucosal melanomas (16-19). Activating NRAS mutations cause aberrant signaling in the MAPK and the PI3K/AKT pathways, and are an interesting target for molecular therapy in advanced melanoma patients carrying these mutations (20). The KIT and NRAS mutation frequencies in mucosal melanomas seem to relate to specific mucosal sites. KIT mutations are relatively frequent in melanomas of the urogenital tract (up to 45%) compared to sinonasal melanomas (7-12.5%) (10, 21-25), while NRAS mutations are rare in anal melanomas (5%) compared to sinonasal (22%) and esophageal melanomas (38%) (25-27). For urogenital tract melanomas only limited numbers of cases have been studied for presence of oncogenic NRAS mutations. As knowledge of molecular subtypes of melanoma is important for selection of patients amenable to targeted therapy, the aim of our study was to search for oncogenic mutations in KIT, NRAS and BRAF in a series of primary melanomas of the female urogenital tract in The Netherlands (n=24)

145 Chapter 8 MATERIALS AND METHODS Patient selection For this retrospective study, formalin fixed and paraffin embedded (FFPE) tissues of twenty-four primary melanomas of the female urogenital tract were retrieved from five pathology archives in The Netherlands in an anonymized way (between 1995 and 2013). Histological diagnosis was revised by two experienced pathologists (WB, HKV). Tissue handling and data-assessment were performed according to the Code Proper Secondary Use of Human Tissue ( DNA isolation and mutation analysis Five 10μm sections were used for DNA extraction. The first and last slide of the paraffin blocks were stained with hematoxylin and eosin and reviewed by a pathologist for determination of representativity and estimation of percentage of neoplastic cells (at least 70 %, as specified in Table 1). DNA extraction was performed using a proteinase K treatment followed by affinity-purification using the QIAamp DNA Mini Kit (QIAGEN) according to the manufacturer s instructions. The concentration of the DNA samples was quantified by NanoDrop ND-1000 (Nanodrop). Testing for mutations in NRAS (exons 3 and 2), KIT (exons 9, 11, 13, and 17), and BRAF (exon 15) was carried out by PCR and Sanger sequencing. Genomic DNA ( ng) was amplified in a reaction volume of 25 μl containing 12.5 μl AmpliTaq Gold 360 PCR Master Mix (Applied Biosystems). Primer sequences used are listed in Table 2. All primers contained a M13 forward or reverse consensus sequence for sequencing the different exons. The PCR conditions were 95⁰C for 10 minutes, 38 cycles of 95⁰C for 30 seconds, 58⁰C for 30 seconds, and 72⁰C for 1 minute followed by an extension step of 72⁰C for 7 minutes. The PCR products were purified by using Multiscreen-HTS PCF Filter Plates (Merck, Millipore). One micro liter of the PCR product was used per sequence analysis (forward and reverse), using an ABI PRISM 3700 DNA analyzer (Life Technologies, Foster City, USA). RESULTS Clinicopathologic characteristics Twenty-four cases were collected (mean age of the patients 68 years): fourteen vaginal melanomas, four cervical melanomas, five urethral melanomas and one vulvar melanoma (see Table 1). The diagnosis was confirmed in all cases by an experienced dermatopathologist (WB). Mean tumor diameter was 30.5 mm and mean tumor thickness was 12.1 mm (biopsies excluded). 144

146 Mutation analysis in melanoma of the female urogenital tract Table 1 Clinicopathologic characteristics and results of mutation analysis KIT ex 9, 11, 13 and 17 NRAS ex 2,3 BRAF ex 15 Estimated load of neoplastic cells (%) Tumor thickness (mm) Tumor diameter (mm) Pat Age Primary location 1 58 Vagina Wt Wt (ex 2 na) Wt 2 77 Vagina 8 3, Wt Wt (ex 2 na) Wt 3 63 Vagina Wt c.181c>a (p.(gln61lys)) Wt 4 48 Vagina Wt Wt Wt 5 75 Vagina Wt Wt (ex 2 na) Wt 6 86 Vagina Wt c.181c>a (p.(gln61lys)) Wt 7 83 Vagina Wt Wt Wt 8 62 Vagina 4* 4* 80 Wt c.35g>t (p.(gly12val)) Na 9 82 Vagina 8* 4* 70 Wt Wt Wt Vagina Wt Wt Wt Vagina >90 Wt Wt Wt Vagina Wt Wt Wt Vagina c.2458g>t (p.(asp820tyr)) Wt Wt Vagina Wt Wt Wt Cervix Wt Wt Wt Cervix Wt Wt Wt Cervix Wt Wt Wt Cervix Wt Wt Wt Vulva Wt Wt Wt Urethra Wt c.181c>a (p.(gln61lys)) Wt Urethra 1,5 2,2 90 Wt Wt Wt Urethra Wt Wt Wt Urethra Wt Wt Wt Urethra 10* 8* 70 Wt Wt Wt Na: not assessable; *minimum size and thickness as measured in biopsy 8 145

147 Chapter 8 Table 2 Primers used for mutation analysis Gene Exon Direction Primer sequence (5-3 ) KIT 9a Forward GTATGCCACATCCCAAGTG Reverse CCTTACATTCAACCGTGCC 9b Forward CAGTGGATGTGCAGACAC Reverse GATATGGTAGACAGAGCC 11 Forward CCAGAGTGCTCTAATGAC Reverse CATGGAAAGCCCCTGTTTC 13 Forward CGCTTGACATCAGTTTGCC Reverse CAGACAATAAAAGGCAGC 17 Forward GGTTTTCTTTTCTCCTCC Reverse CTGTCAAGCAGAGAATGGG BRAF 15 Forward CTTCATAATGCTTGCTCTG Reverse GCCTCAATTCTTACCATCC NRAS 2 Forward ATTACTGGTTTCCAACAGG Reverse CTGGATTGTCAGTGCGCTTT 3 Forward CCCCAGGATTCTTACAGAAAA Reverse TCGCCTGTCCTCATGTATTG Since mutation analysis was the primary goal of the research described in this article and the material was collected in an anonymized way, we could not collect information on clinical stage and/or survival. Mutation analysis All cases were screened for mutations in exons 9, 11, 13, and 17 of KIT (Table 1). A KIT exon 17 mutation c.2458g>t (p.(asp820tyr)) was detected in one of the fourteen vaginal melanomas; this mutation (also described as pd820y) has previously been reported in other mucosal melanomas (21, 24, 28) (see Figure 1A for the chromatogram traces). None of the other cases showed a KIT mutation. In four cases NRAS mutations were identified of which three were present in vaginal melanomas and one in a melanoma of the urethra (see Table 1 and Figure 1B and 1C). Three mutations occurred in codon 61 and one in codon 12 of NRAS. They were mutually exclusive with the KIT mutation. BRAF exon 15 mutations were absent. Case 15 showed an unusual growth pattern with mainly spindle cells and partly a nested growth pattern and no epidermal component. This case was also tested for EWSR1-CREB and EWSR1-ATF1 translocations to exclude the possibility of the lesion being a clear cell sarcoma. No translocations were found (data not shown). 146

148 Mutation analysis in melanoma of the female urogenital tract A B C Figure 1 Chromatogram traces showing a KIT c.2458g>t (p.(asp820tyr)) mutation in case 13 (A), a NRAS c.35g>t (p.(gly12val)) mutation in case 8 (B) and a NRAS c.181c>a (p.(gln61lys)) mutation in case 3 (C). DISCUSSION In the past years, important progression has been made in the treatment of patients with advanced stage cutaneous melanoma by using targeted therapy directed against mutated signal transduction molecules. Targeted therapy might also importantly attribute to prognosis and survival in the rare group of urogenital tract melanomas. Genetic alterations in these melanomas have however not been extensively studied yet. To our knowledge, only a limited number of studies have searched for oncogenic NRAS mutations in urogenital tract melanomas (see Table 3 for an overview). In our 8 147

149 Chapter 8 series of 24 urogenital melanomas, we detected four NRAS mutations (4/24, 16%): three in vaginal melanomas (3/14, 21%) and one in an urethral melanoma (1/5, 20%). Aulmann et al. found NRAS mutations in both vulvar (5/42, 12%) and vaginal (2/15, 13%) melanomas (41). Omholt and colleagues reported an even higher NRAS mutation frequency in vaginal melanomas (43%), which was significantly higher than in mucosal melanomas of other sites in their series (24). In our study, mutations mainly involved codon 61 and were mutually exclusive with KIT mutations. The latter is in accordance with previous studies, although rarely coexistent (pre- and post treatment) KIT and NRAS mutations have been detected in samples of melanoma patients (28, 29, 41). Of note, as oncogenic NRAS mutations are an early event in melanomagenesis and frequently present in benign congenital melanocytic nevi (80%) (30), we could not identify any precursor lesions in the slides of our cases. Also, the cases with mutations did not show any histological characteristics different from those without a mutation. The relatively high frequency of NRAS mutations in our study, especially in the vaginal melanomas (21%), makes NRAS an interesting therapeutic target for these patients in the advanced setting. One strategy in targeting the mutant NRAS protein is the use of inhibitors that target downstream-activated cascades such as the MAPK pathway and the PI3K/AKT/mTOR pathway (31). Melanoma cell lines harboring NRAS gene mutations were shown to be sensitive to a MEK inhibitor. A recent clinical trial with MEK162, a potent MEK inhibitor, has revealed some response in patients with NRAS-mutant melanoma (32, 33). As NRAS activates multiple downstream pathways, combination of inhibitors such as MEK and PI3K/ mtor inhibitors might be more effective, and clinical trials with such combinatorial approaches are currently underway (31, 34). We detected only one KIT mutation, p.d820y, which was in a vaginal melanoma (1/24, 4%). KIT encodes a type III transmembrane receptor tyrosine kinase that consists of five distinct domains, including the juxtamembrane region encoded by exon 11 that is most often affected by mutations in both gastrointestinal stromal tumor (GIST) and melanoma (12). Recently, three phase II studies have shown that imatinib mesylate (Gleevec ) is effective in patients with advanced KIT-mutated melanomas, in addition to case reports of KIT-mutated urogenital melanoma patients showing response to KIT inhibition (11, 21, 28, 29, 35, 36). Response to imatinib mesylate depends on the region of the protein affected by mutation (12). For melanomas, mutations in exon 11 and exon 13 are sensitive to KIT tyrosine kinase inhibitors (11, 12, 28). The p.d820y KIT mutation in our series is associated with secondary resistance to imatinib mesylate in GIST (11, 37) and is also predictive of a lower probability of response to imatinib mesylate in KIT-mutated melanoma (11, 35); although recently a partial response to imatinib mesylate has been noted in a mucosal melanoma (location not further specified) harboring this specific KIT mutation (28). 148

150 Mutation analysis in melanoma of the female urogenital tract Table 3 KIT, NRAS and BRAF mutation frequencies reported in female urogenital tract melanomas KIT mutations (exons 9, 11, 13, 17) NRAS mutations (exons 2 and 3) BRAFV600 mutations Aulmann et al., 2014 (41) 7 / 39 (18%) vulvar 5 / 42 (12%) vulvar 0 / 39 vulvar 2/15 (13%) vaginal 0 / 15 vaginal Not tested Not tested 0/15 (0%) vaginal Schoenewolf et al., 2012 (10) 5 / 11 (45%) vulvovaginal (not further specified) Omholt et al., 2011 (24) 8 / 23 (35%) vulvar 0 / 7 vaginal 0 / 23 vulvar 3 / 7 (43%) vaginal* 2 / 23 (9%) vulvar 0 / 7 vaginal Abysheva et al, 2011(38) 1 / 15 (7%) urogenital (not further specified) Not tested Not tested Yun et al., 2011 (39) 1 / 7 (14%) genital (not further specified) 0 / 7 genital 0 / 7 genital Handolias et al. $, 2010 (21) 2 / 7 (29%) vulvovaginal (in 2 / 5 vulvar Not tested Not tested melanomas; absent in one cervical and one vaginal melanoma) Torres-Cabala et al. $, 2009 (40) 4 / 18 (22%) urogenital Not tested Not tested (in 3 / 11 vulvar and 1 / 4 vaginal melanomas) Satzger et al., 2008 (22) 3 / 8 (37.5%) genital (not further specified) Not tested 0 / 6 0 / 45 mucosal 9 / 37 (24%) mucosal (not further specified) Beadling et al. #, 2008 (23) 4 / 9 (44%) anorectal / urogenital (not further specified) 1 / 8 (12.5%) female genital tract (in a vulvar melanoma, but outside codon 600) Wong et al., 2005 (17) Not tested 1 / 8 (12.5%) female genital tract (in a vaginal melanoma) 0 / 24 1 / 24 (4%) vaginal 4 / 24 (17%) urogenital of which 3 / 14 vaginal (21%) and 1/ 5 urethral Van Engen-van Grunsven et al. (present study) *in this study vaginal melanomas had a significantly higher proportion of NRAS mutations than melanomas of the other sites. This study tested exons 11, 13, 15 and 17 of KIT and did not test exon 9. $ These studies only tested exons 11, 13 and 17 of KIT and did not test exon 9. # This study also tested exon 8 of KIT. No mutations were found in this exon in the 99 melanomas that were tested for this exon (of which 25 were mucosal)

151 Chapter 8 The low frequency of KIT mutations (4%) in our study is in contrast to several studies that have reported a higher frequency of KIT mutations in urogenital melanomas (range up to 45%) (See Table 3 for an overview)(10, 21-24, 38-41). However, most of these studies investigated a relatively low number of patients with exception of two larger studies by Omholt et al. (24) and Aulmann et al. (41). Furthermore, most of these studies did not specify for anatomic location, e.g. vulvar versus vaginal, so it cannot be assessed whether differences exist in the frequency of KIT mutations between vaginal and vulvar melanomas, while there are indications that this might be the case. In our study mainly vaginal melanomas were included (n=14) and only one of these harbored a KIT mutation (7%), suggesting that vaginal melanomas only rarely harbor KIT mutations. This is supported by the study of Omholt et al. who detected no KIT mutations in seven vaginal melanomas and a relatively high frequency of KIT mutations in vulvar melanomas (8/23, 35%), accounting for the relatively high KIT mutation frequency of 27% in thirty vulvovaginal melanomas in their study (24). Very recently, Aulmann et al. also described KIT mutations in seven out of thirty-nine (7/39, 18%) vulvar melanomas, but none in fifteen vaginal melanomas (41). In line with this, Handolias et al. reported two KIT mutations in five vulvar melanomas (40%), and Torres-Cabala et al. reported three KIT mutations in eleven vulvar melanomas (27%) (See Table 3) (21, 40). Obviously, further investigation in larger series is needed to determine whether true differences exist in KIT mutation frequency between vulvar and vaginal melanomas. This could reflect different pathways for tumor genesis, and could have potential clinical implications for testing and treatment of patients with advanced disease. Several studies have shown that BRAF mutations are uncommon in mucosal melanomas, with reported frequencies ranging from 0 to approximately 10% (11, 25, 26, 39, 41). In line with this, we did not detect BRAF mutations in our series of female urogenital tract melanomas. However, BRAF exon 15 mutations have been reported in some cases of vulvar melanomas (9%) (Table 3) (24). Therefore molecular testing for BRAF mutations in urogenital tract melanomas should be encouraged as selected patients might have a mutation and benefit from anti-braf therapy. In conclusion, NRAS mutations were more prevalent than KIT mutations in our series of melanomas arising from the female urogenital tract. This is important as these patients might benefit from MEK inhibition. As the present study mainly investigated vaginal melanomas and only patients from the Netherlands, it cannot be excluded that site- and population specific differences in KIT mutation frequency exist among female urogenital tract melanomas. The absence of BRAF mutations in our study is in line with results from other studies and also with the reported mutation frequency in other types of mucosal melanomas. Testing for BRAF mutations is however still encouraged as selected patients might have a mutation and might benefit from anti-braf therapy. To our opinion, urogenital melanomas should be tested for NRAS, 150

152 Mutation analysis in melanoma of the female urogenital tract KIT and BRAF mutations in the advanced setting as patients might benefit from targeted therapy. Acknowledgements We thank Paul Rombout and Annelies Klaasen (Radboud University Medical Center) and Inge Briaire-de Bruijn (Leiden University Medical Center) for their technical assistance with the molecular analysis. Conflict of interest statement The authors state no conflict of interest

153 Chapter 8 REFERENCES 1. Chang AE, Karnell LH, Menck HR. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer. 1998;83(8): Piura B. Management of primary melanoma of the female urogenital tract. The lancet oncology. 2008; 9(10): Gokaslan H, Sismanoglu A, Pekin T, Kaya H, Ceyhan N. Primary malignant melanoma of the vagina: a case report and review of the current treatment options. European journal of obstetrics, gynecology, and reproductive biology. 2005;121(2): Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT in distinct subtypes of melanoma. Journal of clinical oncology. 2006;24(26): Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, et al. Distinct sets of genetic alterations in melanoma. The New England journal of medicine. 2005;353(20): Smalley KS. Understanding melanoma signaling networks as the basis for molecular targeted therapy. The Journal of investigative dermatology. 2010;130(1): Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. The New England journal of medicine. 2011;364(26): Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, et al. Inhibition of mutated, activated BRAF in metastatic melanoma. The New England journal of medicine. 2010;363(9): Woodman SE, Davies MA. Targeting KIT in melanoma: a paradigm of molecular medicine and targeted therapeutics. Biochemical pharmacology. 2010;80(5): Schoenewolf NL, Bull C, Belloni B, Holzmann D, Tonolla S, Lang R, et al. Sinonasal, genital and acrolentiginous melanomas show distinct characteristics of KIT expression and mutations. European journal of cancer. 2012;48(12): Carvajal RD, Antonescu CR, Wolchok JD, Chapman PB, Roman RA, Teitcher J, et al. KIT as a therapeutic target in metastatic melanoma. JAMA : the journal of the American Medical Association. 2011;305(22): Postow MA, Hamid O, Carvajal RD. Mucosal melanoma: pathogenesis, clinical behavior, and management. Current oncology reports. 2012;14(5): Hodi FS, Friedlander P, Corless CL, Heinrich MC, Mac Rae S, Kruse A, et al. Major response to imatinib mesylate in KIT-mutated melanoma. Journal of clinical oncology. 2008;26(12): Lutzky J, Bauer J, Bastian BC. Dose-dependent, complete response to imatinib of a metastatic mucosal melanoma with a K642E KIT mutation. Pigment cell & melanoma research. 2008;21(4): Quintas-Cardama A, Lazar AJ, Woodman SE, Kim K, Ross M, Hwu P. Complete response of stage IV anal mucosal melanoma expressing KIT Val560Asp to the multikinase inhibitor sorafenib. Nature clinical practice Oncology. 2008;5(12): Omholt K, Platz A, Kanter L, Ringborg U, Hansson J. NRAS and BRAF mutations arise early during melanoma pathogenesis and are preserved throughout tumor progression. Clinical cancer research. 2003;9(17): Wong CW, Fan YS, Chan TL, Chan AS, Ho LC, Ma TK, et al. BRAF and NRAS mutations are uncommon in melanomas arising in diverse internal organs. Journal of clinical pathology. 2005;58(6): van t Veer LJ, Burgering BM, Versteeg R, Boot AJ, Ruiter DJ, Osanto S, et al. N-ras mutations in human cutaneous melanoma from sun-exposed body sites. Molecular and cellular biology. 1989;9(7): van Elsas A, Zerp S, van der Flier S, Kruse-Wolters M, Vacca A, Ruiter DJ, et al. Analysis of N-ras mutations in human cutaneous melanoma: tumor heterogeneity detected by polymerase chain reaction/ single-stranded conformation polymorphism analysis. Recent results in cancer research. 1995;139: Posch C, Ortiz-Urda S. NRAS mutant melanoma--undrugable? Oncotarget. 2013;4(4): Handolias D, Hamilton AL, Salemi R, Tan A, Moodie K, Kerr L, et al. Clinical responses observed with imatinib or sorafenib in melanoma patients expressing mutations in KIT. British journal of cancer. 2010;102(8):

154 Mutation analysis in melanoma of the female urogenital tract 22. Satzger I, Schaefer T, Kuettler U, Broecker V, Voelker B, Ostertag H, et al. Analysis of c-kit expression and KIT gene mutation in human mucosal melanomas. British journal of cancer. 2008;99(12): Beadling C, Jacobson-Dunlop E, Hodi FS, Le C, Warrick A, Patterson J, et al. KIT gene mutations and copy number in melanoma subtypes. Clinical cancer research. 2008;14(21): Omholt K, Grafstrom E, Kanter-Lewensohn L, Hansson J, Ragnarsson-Olding BK. KIT pathway alterations in mucosal melanomas of the vulva and other sites. Clinical cancer research. 2011;17(12): Turri-Zanoni M, Medicina D, Lombardi D, Ungari M, Balzarini P, Rossini C, et al. Sinonasal mucosal melanoma: Molecular profile and therapeutic implications from a series of 32 cases. Head & neck. 2013;35(8): Antonescu CR, Busam KJ, Francone TD, Wong GC, Guo T, Agaram NP, et al. L576P KIT mutation in anal melanomas correlates with KIT protein expression and is sensitive to specific kinase inhibition. International journal of cancer. 2007;121(2): Sekine S, Nakanishi Y, Ogawa R, Kouda S, Kanai Y. Esophageal melanomas harbor frequent NRAS mutations unlike melanomas of other mucosal sites. Virchows Archiv. 2009;454(5): Hodi FS, Corless CL, Giobbie-Hurder A, Fletcher JA, Zhu M, Marino-Enriquez A, et al. Imatinib for melanomas harboring mutationally activated or amplified KIT arising on mucosal, acral, and chronically sun-damaged skin. Journal of clinical oncology. 2013;31(26): Minor DR, Kashani-Sabet M, Garrido M, O Day SJ, Hamid O, Bastian BC. Sunitinib therapy for melanoma patients with KIT mutations. Clinical cancer research. 2012;18(5): Bauer J, Curtin JA, Pinkel D, Bastian BC. Congenital melanocytic nevi frequently harbor NRAS mutations but no BRAF mutations. The Journal of investigative dermatology. 2007;127(1): Kelleher FC, McArthur GA. Targeting NRAS in melanoma. Cancer journal. 2012;18(2): Solit DB, Garraway LA, Pratilas CA, Sawai A, Getz G, Basso A, et al. BRAF mutation predicts sensitivity to MEK inhibition. Nature. 2006;439(7074): Ascierto PA, Schadendorf D, Berking C, Agarwala SS, van Herpen CM, Queirolo P, et al. MEK162 for patients with advanced melanoma harbouring NRAS or Val600 BRAF mutations: a non-randomised, open-label phase 2 study. The lancet oncology. 2013;14(3): Posch C, Moslehi H, Feeney L, Green GA, Ebaee A, Feichtenschlager V, et al. Combined targeting of MEK and PI3K/mTOR effector pathways is necessary to effectively inhibit NRAS mutant melanoma in vitro and in vivo. Proceedings of the National Academy of Sciences of the United States of America. 2013;110(10): Guo J, Si L, Kong Y, Flaherty KT, Xu X, Zhu Y, et al. Phase II, open-label, single-arm trial of imatinib mesylate in patients with metastatic melanoma harboring c-kit mutation or amplification. Journal of clinical oncology. 2011;29(21): Woodman SE, Trent JC, Stemke-Hale K, Lazar AJ, Pricl S, Pavan GM, et al. Activity of dasatinib against L576P KIT mutant melanoma: molecular, cellular, and clinical correlates. Molecular cancer therapeutics. 2009;8(8): Liegl B, Kepten I, Le C, Zhu M, Demetri GD, Heinrich MC, et al. Heterogeneity of kinase inhibitor resistance mechanisms in GIST. The Journal of pathology. 2008;216(1): Abysheva SN, Iyevleva AG, Efimova NV, Mokhina YB, Sabirova FA, Ivantsov AO, et al. KIT mutations in Russian patients with mucosal melanoma. Melanoma research. 2011;21(6): Yun J, Lee J, Jang J, Lee EJ, Jang KT, Kim JH, et al. KIT amplification and gene mutations in acral/ mucosal melanoma in Korea. APMIS : acta pathologica, microbiologica, et immunologica Scandinavica. 2011;119(6): Torres-Cabala CA, Wang WL, Trent J, Yang D, Chen S, Galbincea J, et al. Correlation between KIT expression and KIT mutation in melanoma: a study of 173 cases with emphasis on the acral-lentiginous/ mucosal type. Modern pathology. 2009;22(11): Aulmann S, Sinn HP, Penzel R, Gilks CB, Schott S, Hassel JC, et al. Comparison of molecular abnormalities in vulvar and vaginal melanomas. Modern pathology mar 7. Epub ahead of print

155

156 Part III

157

158 9 Summarizing discussion and future perspectives

159

160 Summarizing discussion and future perspectives SUMMARIZING DISCUSSION AND FUTURE PERSPECTIVES Melanocytic neoplasms include several tumor types that are characterized by distinct clinical features, histopathological appearances, genetic aberrations, and biological behavior (1). Cutaneous melanocytic neoplasms are by far the most common type of melanocytic neoplasms and in the past decades much knowledge has been gained on the molecular mechanisms underlying the most frequent cutaneous melanocytic lesions. Much less is known about the molecular mechanism in less frequent melanocytic tumor subtypes in the skin or in melanocytic tumors developing in non-cutaneous sites. The main goal of the research described in this thesis was to learn more about the genetic aberrations present in different less frequent melanocytic neoplasms, including Spitz tumors, primary leptomeningeal melanocytic tumors, uveal melanomas, and melanomas of the female urogenital tract, in order to gain knowledge on pathogenesis, improve diagnostics and to search for treatment targets. In order to facilitate the reading of the rest of this chapter, an overview of the different genes and pathways involved is given in figure 1 (2). SPITZ TUMORS Spitz tumors are uncommon skin tumors that mainly arise in children and adolescents, but may also occur in older individuals and constitute diagnostic problems for dermatopathologists on a regular basis (3-5). Under-diagnosis as a nevus, overdiagnosis as a melanoma, and an uncertain diagnosis of Spitzoid tumor of unknown malignant potential (STUMP) all occur in this specific subset of melanocytic skin tumors. Currently, three subgroups of spitzoid melanocytic tumors have been described with specific genetic characteristics partly associated with a typical phenotype: the HRAS-mutated Spitz tumors (chapter 3), the melanocytic BAP1- associated intradermal tumors (MBAITs), and the spitzoid lesions with kinase fusions (see also chapter 2). HRAS-mutated Spitz tumors HRAS-mutated Spitz tumors are characterized by a typical phenotype. In chapter 3 twenty-four cases are described which we identified while studying a large group of Spitz tumors (93 Spitz nevi and 77 STUMPs). HRAS-mutated Spitz tumors are dermal-based tumors with a relatively low cellularity, desmoplasia and an infiltrative base. They can be easily misdiagnosed as a melanoma especially when dermal mitoses are present and when the lesions present in older patients. In our series of twenty-four HRAS-mutated Spitz tumors, seven (29%) were initially diagnosed as a melanoma or melanoma was an important differential diagnostic consideration

161 Chapter 9 Figure 1 Main intracellular pathways that are implicated in different subtypes of melanoma. Mutations in KIT are relatively frequent in mucosal melanoma (~20%) and acrolentiginous melanoma, result in activation of the MAPK and PI3/AKT/mTOR pathways, and can be treated with KIT inhibitors. Approximately 50% of cutaneous melanomas (CM) harbor mutations in BRAF at codon 600, which provide a therapeutic target for selective BRAF inhibitors. Mutations in NRAS (~20% of CM) signal to MEK/ERK through CRAF, but also to the PI3/AKT/mTOR pathway. MEK inhibitors have shown activity in some patients with NRAS mutant melanoma. Signaling pathways downstream of GNAQ/GNA11 (~80% of uveal melanomas (UMs)) include activation of PKC through the release of diacylglycerol (DAG) by β PLC, which can activate the MAPK pathway. Another downstream pathway is the Hippo YAP pathway resulting in activation and translocation of the transcriptional coactivator YAP into the nucleus through a mechanism of actin polymerization regulated by the GTPases Rho and Rac. Metastatic UMs frequently harbor inactivating mutations in BAP1, which is involved in multiple functions including regulation of the cell cycle and epigenetic modulation. Inhibition of downstream components such as MEK, PKC, BAP1 and YAP are now under (pre) clinical investigation for treatment of GNAQ/GNA11 mutated UM. MAPK = mitogen activated protein kinase; PKC = protein kinase C; PLC = phospholipase C; UM = uveal melanoma; ERK = extracellular signal regulated kinase; MEK = mitogen activated ERK kinase; PI3K/AKT/mTOR = phosphatidylinositol 3 kinase/protein kinase AKT/mammalian target of rapamycin. Adapted from (2). 160

162 Summarizing discussion and future perspectives A correct distinction with a melanoma is clinically relevant since prognosis is different; none of the patients with a HRAS-mutated Spitz tumor in our series developed recurrences or metastases (mean and median follow-up of 10.5 y) which is in accordance with findings in most other studies. HRAS mutations are encountered in <1% of cutaneous melanomas (6). The HRAS-mutated Spitz tumors account for ~15% of Spitz tumors. The typical morphology of the HRAS-mutated Spitz tumors is probably a consequence of the HRAS mutation. It has been demonstrated that overexpression of HRAS in normal human melanocytes causes increase in cell size and massive intracellular vacuolization (7). It is currently unclear why HRAS mutations occur almost exclusively in Spitz tumors and not in spitzoid melanoma. Melanocytic BAP1-associated intradermal tumors The melanocytic BAP1-associated intradermal tumors (MBAITs) (8), ~5% of Spitz tumors, are often polypoid, dermal based tumors consisting of large epithelioid spitzoidlooking cells and can have a small compound nevus component (especially in BAP1-germline mutated lesions). In one-third of cases tumor-infiltrating lymphocytes are prominent. MBAITs can occur in the setting of a BAP1 germline mutation (8), but also sporadically (9, 10). The majority of lesions carry a concomitant BRAF mutation (9, 11) and, much less frequently, a concomitant NRAS mutation (12, 13). No HRAS mutations have been described in MBAITs. The prognosis of MBAITs seems favorable without recurrences and infrequent progression to melanoma and very rarely metastases (14). At present, the most important reason for recognition of a MBAIT is that they can be a marker of an underlying BAP1-associated germline mutation. The BAP1 tumor predisposition syndrome predisposes to cutaneous and uveal melanoma (8, 15), mesothelioma (14, 16), renal cell cancer (17), and possibly several other types of cancer (18). The cutaneous melanocytic tumors in the setting of germline BAP1 mutations do not differ in histological appearance from those with sporadic mutations (12). In addition, recognition of a melanocytic tumor as a MBAIT is important to prevent overdiagnosis of melanoma. In the past, but even nowadays, not all (general) pathologists are familiar with these tumors and they are still diagnosed as melanoma or consultated to exclude melanoma. Of note, several other terms have been proposed for MBAITs, especially BAPoma and naevoid melanoma-like melanocytic proliferation (19) and Wiesner tumor. Spitz tumors with genomic rearrangements involving kinases The third group of Spitz tumors, up to 50%, are those with genomic rearrangements (translocations, fusions) of various receptor tyrosine kinases, including ALK (~10% of Spitz tumors), ROS (~10%), NTRK1 (~10%), RET (<5%) and MET (rare), or the serinethreonine kinase BRAF (~5%) (19-21). Fusion-positive Spitz tumors generally occur in patients that are younger than those with tumors without a translocation (20)

163 Chapter 9 The genomic rearrangements cause fusion of the kinase domain to a wide range of partner genes, which leads to high expression of fusion genes that subsequently causes ligand independent dimerisation and auto-phosphorylation of the kinase domain. These kinases subsequently activate multiple oncogenic signaling pathways like the MAPK/ERK, P13K/AKT/mTOR, and JAK-STAT pathways resulting in increased cell proliferation, and cell survival (8). In these tumors, kinase fusions are mutually exclusive; no combinations with BAP1 inactivation or HRAS mutations have been observed. Furthermore, these kinase fusions are present along the entire spectrum of the Spitz tumors (55% in Spitz nevi; 56% in atypical spitzoid tumors and 39% in spitzoid melanomas). Most have no specific histologic features and are difficult to identify. An in-depth description of the clinicopathological features of the different types of Spitz tumors with genomic rearrangements involving kinases is beyond the scope of this thesis, but are described in detail by Wiesner et al. (19). Recognition of Spitz tumors that carry a kinase fusion might be mainly important for therapy of malignant lesions in the advanced setting. Thus far no reports on treatment of these lesions as such are available. Metastatic spitzoid melanomas harboring kinase fusions might be candidates for treatment with kinase inhibitors, just like for instance lung cancers with ALK or ROS1 fusions (22, 23), and medullary thyroid carcinoma with RET fusions (24-26). Implications for the pathologist Thorough histopathological examination remains the gold standard for the diagnosis of Spitz tumors, but in certain cases additional molecular analysis can be useful. To our opinion, in daily pathological practice, Spitz tumors that are hard to classify should be tested for mutations in HRAS, BRAF and NRAS. BRAF and NRAS should be tested for as these are frequently mutated in adult cutaneous melanomas (see chapter 1). The presence of a BRAF or NRAS mutation therefore favors the diagnosis of a melanoma whereas the presence of a HRAS mutation indicates an indolent biological behavior. When based on morphological findings a MBAIT is considered, we advocate that BAP1 protein expression analysis is performed. In case of inconclusive BAP1 immunostaining results (a staining that is difficult to interpret) or unexpected positive protein expression, while histological suspicion for a MBAIT is high, we advise to perform additional BAP1 mutation analysis. Immunostaining for BAP1 is sometimes positive despite an underlying BAP1 mutation (own observation and (27)). In the past, also in the setting of spitzoid tumors, mutation analysis by conventional techniques such as Sanger sequencing was often hampered due to a limited amount of material or because of the presence of a low percentage (i.e. less than 30%) of neoplastic cells. Often spitzoid lesions have inflammation or are relatively small or superficial; all these factors can lead to a low neoplastic cell percentage in a specimen. By mutation detection via next-generation sequencing (NGS), the sensitivity of a 162

164 Summarizing discussion and future perspectives molecular test is greatly improved enabling mutation detection in tissues with a lower percentage of neoplastic cells. Spitzoid melanomas with metastases, in which targeted treatment is considered, should -if negative for BRAF, NRAS, or KIT, be tested for genomic rearrangements involving kinases as patients with those harboring kinase fusions (of ALK, ROS, or RET), might benefit from treatment with kinase inhibitors. In many laboratories the gene fusions are currently interrogated using in situ hybridization. Detection of gene fusions by NGS technology accompanied by robust data analysis will soon make its entrance in molecular laboratories. The classification of Spitz tumors will probably be more refined in the future due to progress of our understanding of tumor pathogenesis. By integration of clinical, pathological, and genetic criteria, more subsets of Spitz tumors will probably be defined which will lead to improved diagnosis, treatment and prognosis. UVEAL MELANOMAS Uveal melanomas (UMs) are derived from melanocytes in the choroid coat, ciliary body, and iris of the eye. Despite local control, up to 50% of patients develop metastases, with average survival less than one year after the diagnosis of metastatic disease. At the moment there is no effective treatment for metastatic disease. Chromosome 3 loss, which is often associated with a BAP1 mutation on the remaining chromosome, is strongly correlated with metastatic disease as is, to a lesser extent, gain of chromosome 8q (15, 28-30). Disomy of chromosome 3, gain of chromosome 6p, and mutations in EIF1AX and SF3B1 have been reported to be associated with a relatively good prognosis (28, 30-33). However, in a recently published series of 151 UMs, the tumors with disomy 3 and a SF3B1 mutation carried an increased risk of metastatic spread compared to those with wild type SF3B1. Metastases in patients with mutated SF3B1 occurred late whereas those in patients with BAP1 mutations occurred early (34). UMs frequently carry mutations in the G protein encoding genes GNAQ and GNA11 (45% and 32% respectively) (35-37) (see figure 2). The results of whole genome copy-number analysis of twenty-five UMs as well as mutation and MLPA analysis of these tumors, are described in chapter 4. Our study confirms previously reported GNAQ and GNA11 mutation frequencies in UMs, as well as the presence of monosomy 3 as the most relevant genetic prognostic factor indicating a poor prognosis. We demonstrated that analysis of monosomy 3 by the described MLPA assay was efficient and robust. Soon, this MPLA assay will be replaced by NGS technologies to detect copy number alterations. Our study indicates that chromothripsis (a single catastrophic genetic event leading to a genomic landscape with multiple chromosomal rearrangements), and chromosome 8q gain 9 163

165 Chapter 9 may play a role in initiation of UMs lacking a GNAQ or GNA11 mutation, while loss of chromosome 2p may be involved in metastatic disease in UMs without monosomy 3. These latter findings need to be corroborated in a larger series of tumors. There was no correlation between the various chromosomal abnormalities and histologic features. PRIMARY LEPTOMENINGEAL MELANOCYTIC NEOPLASMS In the adult setting, primary leptomeningeal melanocytic neoplasms (LMNs) generally lack mutations in BRAF, HRAS and NRAS that are commonly found in cutaneous melanocytic neoplasms, but, like UMs, frequently harbor oncogenic mutations in GNAQ and GNA11 (see figure 2 and chapter 5). We have tested twenty LMNs (eleven melanocytomas (MC), five intermediate grade melanocytomas (IMCs) and four leptomeningeal melanomas (LMM)) for hotspot mutations in GNAQ, GNA11, BRAF, NRAS and HRAS. GNAQ Q209 mutations were found in eleven neoplasms (in total 11/20 (55%); 7/11 MCs (64%); 2/5 IMCs (40%), 2/4 LMMs (50%)), and two neoplasms carried a GNA11 Q209 mutation (in total 2/20 (10%); 1/5 IMCs (20%) and 1/4 LMMs (25%)). GNA11 Q209 mutations were less frequent in the LMNs than GNAQ Q209 mutations, as has been reported (38, 39). This is in contrast to UMs where there is a more even distribution of GNAQ and GNA11 mutations, but in accordance with the reported distribution in blue nevi (35). In the reported series within this thesis, no GNAQ R183, GNA11 R183, BRAF, HRAS or NRAS mutations were found. In another more recent unpublished melanocytoma case we did, however, find a GNA11 R183 mutation, and a GNAQ R183 mutation has been reported in a melanocytoma in literature as well (40). Mutations at codon 183 of GNAQ and GNA11 thus do occur, but, as is also the case for UMs, at a much lower frequency than those at codon 209. GNA11 mutations might be associated with a more aggressive behavior There was no significant association between GNAQ/ GNA11 mutational status and cell type, location of the lesion, disease-related death, or clinical behavior. However, GNA11 Q209 mutations might have a more potent oncogenic effect than GNAQ Q209 mutations. All cases in our series harboring a GNA11 mutation were melanocytomas of intermediate grade or a melanoma (chapter 5). Recently, van de Nes et al. (39) reported comparable findings. In a series of seventeen primary LMNs, they found two GNA11 mutated cases (12%) compared to twelve cases with a GNAQ mutation (71%). Both GNA11 mutated cases were melanocytomas of intermediate grade that recurred. In UMs, GNA11 mutations also seem to be associated with a more aggressive behavior as UM metastases carry a higher frequency of GNA11 (57-60%) to GNAQ 164

166 Summarizing discussion and future perspectives Figure 2 Relative frequency of oncogenic mutations in primary LMNs in adults vs. frequencies in uveal and cutaneous melanomas. Primary leptomeningeal melanocytic neoplasms (LMNs) in adults frequently harbor oncogenic mutations in GNAQ and to a lesser extent in GNA11. The mutation profile is comparable with that of uveal melanoma, but distinct from cutaneous melanoma. GNAQ/GNA11/BRAF mutation analysis can be used as a diagnostic tool in the differential diagnosis of primary LMNs vs. metastasis of cutaneous melanoma to the central nervous system (CNS). Adapted from (2). (20-22%) mutations (35, 41), whereas primary UMs carry slightly more GNAQ (45-47%) than GNA11 (32-44%) mutations (35, 42). Moreover, GNA11 mutations are relatively infrequent in blue nevi compared to UMs. However, in UM patients, no significant association has been found between GNAQ or GNA11 mutational status and survival (35, 37, 43). Future studies are needed to determine whether GNA11 mutations are associated with a more aggressive behavior in primary LMNs as well as in UMs

167 Chapter 9 Genetic resemblance of primary leptomeningeal melanocytic neoplasms with uveal melanomas LMNs and UMs not only carry the same type of mutations. In chapter 5 we show with multiplex-ligation dependent probe analysis (MLPA) and OncoScan analysis, that they have more genetic aberrations in common (chapter 5). Detailed study of one intermediate grade melanocytoma and two leptomeningeal melanomas showed copy number variations resembling those found in UMs (i.e. gain of chromosome 6, gain of chromosome 8q and monosomy of chromosome 3) (28, 44). While these copy number variations have prognostic relevance in UMs, the prognostic relevance in LMNs still needs further investigation. Recently, we have described another similarity between LMNs and UMs (45). Mutations in either SF3B1 or EIF1AX, which are frequently mutated in UMs (31-33), were detected in eight out of twenty-four primary LMNs (8/24, 33%). This underscores even more that UMs and LMNs resemble each other genetically. Immunohistochemical nuclear staining for BAP1, which was used as a surrogate marker for the detection of inactivating mutations in the BAP1 gene, was present in all cases in our series, suggesting absence of underlying BAP1 mutations in these tumors. However, there may have been a selection bias in our case series as it mainly concerned (relatively) low-grade tumors with disomy of chromosome 3. Recently, in a series of seventeen primary LMNs, one intermediate grade melanocytoma was reported to carry an inactivating BAP1 c.178c>t (p.r60*) mutation as well as a GNAQ Q209 mutation and monosomy of chromosome 3 (39). Detailed follow-up data was not available, but the tumor did recur. De la Fourchadière et al. (46) reported a case of a primary leptomeningeal melanoma with a GNAQ R183 mutation and complete absence of nuclear BAP1 staining. Array comparative genomic hybridization analysis revealed an altered profile, including a loss at the 3p2.1 BAP1 locus. The patient had a family history of uveal melanomas and meningeomas, and both sisters of the patient turned out to have a germline BAP1 mutation. These findings suggest that BAP1 mutations may play a role in LMNs, and that these tumors might need to be added to the list of tumors associated with germline BAP1 mutations. The prognostic significance of SF3B1, EIF1AX and BAP1 mutations in LMNs is presently unclear and needs to be further elucidated. Mutation analysis in the differential diagnosis of a primary leptomeningeal melanocytic neoplasm and a melanoma metastasis Mutation analysis of BRAF, GNAQ and GNA11 can be of aid in the differential diagnosis of a primary CNS melanoma versus a melanoma metastasis in the CNS in an adult patient. In contrast to primary LMNs, mutations in GNAQ or GNA11 are absent or very rare in cutaneous melanoma (1.4%, tested in >150 cases), acral melanoma (0%) and mucosal melanoma (0% in most studies), whereas cutaneous melanomas frequently 166

168 Summarizing discussion and future perspectives (50%) harbor BRAF mutations which are absent or rare in LMNs, see figure 2 (2). A melanoma in the CNS is much more likely a metastasis than a primary CNS melanoma, as primary melanocytic LMNs are very rare and cutaneous melanomas are relatively frequent and frequently metastasize to the CNS. In autopsy studies up to 2/3 of patients who died of a cutaneous melanoma had a CNS metastasis, whereas in clinical studies 10-40% of patients develop CNS metastases (47-49). Presence of a GNAQ or GNA11 mutation in a CNS melanoma, favors a primary LMN, whereas demonstration of a BRAF or NRAS mutation in adults, favors a metastasis from a cutaneous melanoma. Of note, this is different in children where primary LMNs harbor NRAS mutations (see chapter 7). Differentiating a primary LMN from a metastasis is important because of its implications for further patient work-up, and the better prognosis of patients with a primary CNS melanoma compared to those with a melanoma metastasis in the CNS (47-54). Mutation analysis in the differential diagnosis of a primary leptomeningeal melanocytic neoplasm and a melanotic schwannoma Histologically, melanotic schwannoma can closely resemble LMNs and blue nevi (55). Additional stainings are not always of help as there is overlap in staining patterns and they can be hard to interpret, especially in small biopsies (56-59). The difficulty of the diagnosis of a melanotic schwannoma, is illustrated by the fact that in our series of twenty-five cases initially diagnosed as a melanotic schwannoma, sixteen were reclassified based on criteria from the literature (55, 60-63) (as conventional schwannoma, n= 9; melanocytoma, n= 4; lesion not further classified with certainty as melanotic schwannoma or melanocytoma, n= 2 or blue nevus, n= 1) (see chapter 6). All twenty-five cases were tested for GNAQ Q209 mutations. No mutations were found in the nine melanotic schwannomas. However, GNAQ Q209 mutations were found in one case reclassified as a blue nevus, in three cases reclassified as melanocytoma, and in one case that could not be reclassified with certainty as melanotic schwannoma or melanocytoma. Of note, all melanotic schwannomas were also tested for GNAQ R183, GNA11 Q209 and GNA11 R183 mutations; none were found (data not published). Combining these data with our previous findings in nineteen LMNs where seven contained a GNAQ Q209 mutation (64), we concluded that GNAQ mutations are very helpful for discriminating primary LMNs from melanotic schwannomas. Correct diagnosis of the latter is important to select those patients who need further evaluation for possible underlying Carney Complex, an autosomal dominant syndrome associated with myxoid tumors (especially myxomas of the heart which can be lethal), endocrine overactivity and pigmented skin lesions (65, 66)

169 Chapter 9 NEUROCUTANEOUS MELANOSIS Primary LMNs can occur sporadically, mainly in adults, but they can also occur in the setting of neurocutaneous melanosis (NCM), mainly in children. NCM is a rare, congenital syndrome characterized by the association of large or multiple congenital melanocytic nevi (CMN) of the skin and central nervous system (CNS) abnormalities. The CNS abnormalities in NCM are diverse and include both melanocytic as well as non-melanocytic lesions with widely varying clinical outcomes. Once diffuse leptomeningeal melanocytic growth or CNS melanoma is present, prognosis is very poor and there is no effective treatment. Patients with large and/or multiple CMN, are at risk for having neurocutaneous melanosis, but it is currently hard to predict which patients will develop symptomatic NCM. Recently, magnetic resonance imaging (MRI) has been proposed as a screening instrument for congenital neurological lesions in children with more than one CMN at birth (67), but to our opinion there are several disadvantages: many often very young children need to be scanned (under general anesthesia) in order to find a rare case of NCM. In the reported study (67), twenty-one percent of the children had an abnormal MRI. A negative MRI, however, does not exclude NCM developing at a later stage. Furthermore, there can be subtle changes of which the relevance is currently unknown but nevertheless cause a lot of anxiety for parents. Currently, there is only very limited knowledge on the genetic aberrations in the melanocytic lesions of patients with NCM. Chapter 7 describes whole genome copy-number variation analysis as well as mutation analysis in paired samples of primary LMNs and CMN of five patients with NCM. Results were correlated with clinical characteristics including survival. All primary LMNs and CMN harbored a heterozygous NRAS Q61 mutation, confirming a central role for NRAS mutations in NCM-associated melanocytic neoplasms. No additional mutations were found. Our study supports the current hypothesis that both skin and leptomeningeal melanocytic lesions in NCM are caused by an early post-embryonic NRAS mutation in precursors of melanocytes. Depending on the timing of this NRAS mutation during development, this leads to NCM (in case of early mutation) or only a congenital nevus (in case of a later mutation). Additionally, in the LMNs, but not in the CMN, we found multiple large copy number variations. This indicates that additional genetic aberrations cause proliferation of the leptomenigeal melanocytes. A potentially very interesting finding was focal loss of heterozygosity (LOH) at 6p22.1 and focal genomic aberrations (copy neutral LOH and or copy number loss) at 10q22.1 in all CMN and all primary LMNs of the four pediatric patients in our series. This recurrent and characteristic genetic alteration might be indicative of an underlying NCM when found in CMN, and might be a predictive marker for development of subsequent neurological disease. Such a marker is currently lacking; in patients with 168

170 Summarizing discussion and future perspectives large or several congenital nevi it is currently unclear which patients are going to develop neurological disease because of melanocytic proliferation. It would be very interesting to compare a series of CMNs from patients who have never developed NCM with a set of CMN and primary LMNs from patients with NCM for these aberrations in order to know if information on 6p22.1 and 10q22.1 can indeed be used as a diagnostic test. Due to the rarity of NCM, collection of sufficient numbers of patients and tumor tissues for systematic study however requires more time. MELANOMAS OF THE FEMALE UROGENITAL TRACT Urogenital melanomas are rare and belong to the subgroup of the mucosal melanomas. They often get diagnosed at advanced stage with little treatment options, and are associated with a poor outcome. A recent study from our hospital showed that 6% of all vulvar cancers are melanomas and that they carry a more-worse prognosis than cutaneous melanomas. The five year survival for vulvar melanoma patients was 35% (95% CI 26,7-44.4%) compared to 50% (95% CI ) for cutaneous melanoma patients with the same age, tumor ulceration stage, Breslow thickness, lymph node status and distant metastases diagnosed in the same time period (135 vulvar melanoma patients between 2004 and 2012) (log rank test: p = 0.002) (68). This difference may be caused by a difference in biological behavior, impact of tumor environment, or a combination thereof. In mucosal melanomas mainly KIT and NRAS mutations are found with frequencies varying between different mucosal sites, while oncogenic mutations in BRAF are very rare (69-76). In the study described in chapter 8 we investigated a series of twenty-four primary female urogenital tract melanomas. More precisely, fourteen vaginal melanomas, four cervical melanomas, five urethral melanomas and one vulvar melanoma were tested for mutations in KIT, NRAS and BRAF. NRAS mutations were detected in four of these tumors (4/24, 16%): three in vaginal melanomas (3/14, 21%) and one in an urethral melanoma (1/5, 20%). Only one KIT mutation was found: in a vaginal melanoma (1/24, 4%). No BRAF mutations were present. The low frequency of KIT mutations is in contrast to some other studies (for an overview see table 1 in chapter 8). However, in previous studies the number of patients was generally low. Also, the exact anatomic location (e.g. vulvar or vaginal) was often not specified in previous studies, but may well have had an impact, as KIT mutations seem more prevalent in vulvar melanomas compared to vaginal melanomas (our present study, (74, 77, 78)). Based on our results, we advocate that mutation analysis (KIT, NRAS and BRAF) should be performed in patients with metastatic urogenital tract melanoma since this may provide treatment options

171 Chapter 9 CONCLUDING REMARKS Melanomas originating from epithelia-associated vs non-epitheliaassociated melanocytes Melanocytic neoplasms range from benign to malignant lesions and represent a group of distinct tumor types. At present, different melanocytic neoplasms are considered to arise from different types of melanocytes. These different types may be intrinsically different and/or the different microenvironments may change their function and response to stimuli. One way to subgroup melanocytic neoplasms is to discriminate between those originating from epithelia-associated versus non-epithelia-associated melanocytes (1, 70, 79-81). Cutaneous melanomas, acral melanomas (originating from the glabrous skin of the palms, soles and the nail apparatus) and the mucosal melanomas belong to the first and by far largest group. For cutaneous melanomas, there is a clear difference between melanomas occurring in skin with chronic sun damage (CSD; the tumors are generally characterized by a higher mutation burden, relatively late age of onset, and mutations mainly in NRAS, NF1, KIT, and in BRAF but not the most frequent BRAF V600E mutation) versus melanomas occurring in skin with intermittent sun exposure (non-csd; these present earlier in life, are often associated with nevi, having a comparatively low mutation burden, and frequently harboring the most frequent V600E BRAF mutation). In acral melanomas UV-radiation is probably not a causal factor as they occur at similar incidences across the world. The frequency of BRAF mutations is lower than in non-csd melanomas and approximately 20% have mutations in KIT. Mucosal melanomas, as described in chapter 8, harbor frequent mutations in NRAS and KIT, but less frequently in BRAF. The melanomas originating from melanocytes not associated with epithelia are those arising in the dermis, uvea and CZS. As mentioned in chapters 4-7, melanomas belonging to this group mainly carry GNAQ or GNA11 mutations whereas those are virtually absent in the melanocytic neoplasms originating from epithelia-associated melanocytes. In contradiction with previously published literature and the above mentioned hypothesis, Sheng et al. (82) very recently have described the presence of GNAQ and GNA11 mutations in mucosal melanomas. In a series of 284 primary mucosal melanoma samples from different primary anatomic sites (head and neck, anorectum, genitourinary and esophagus), they found GNAQ mutations in 4.6% of cases and GNA11 mutations in 4.9% of cases. GNAQ/GNA11 mutations were not significantly associated with age, gender, or primary anatomic site. The GNAQ mutations involved only exon 5 codon 209 whereas the GNA11 mutations involved both exon 5 (79%) and exon 4 (21%). Mutation types were mostly the same as described for UMs, but at 170

172 Summarizing discussion and future perspectives other frequencies (i.e. GNAQ Q209L mutations were found in 92% of the GNAQ mutated mucosal melanomas and Q209P in 8% whereas mutation frequencies in GNAQ mutated primary UMs are 33% and 66% respectively ( org). GNA11 Q209L mutations were found in 43% of the GNA11 mutated mucosal melanomas and Q209R in 36% and R183C in 21% whereas mutation frequencies in GNA11 mutated primary UMs are 92%, 0% and 2-5% respectively (www. mycancergenome.org)). Van Raamsdonk et al. did not find any GNAQ or GNA11 mutation in 62 mucosal melanoma samples they tested (35). Maybe there is a difference between races as the study by Sheng et al. has been performed among a Chinese population whereas the study from van Raamsdonk et al. has been performed in the United States. It is remarkable that there was no difference in mutation types and frequencies between the different primary anatomical sites, as this seems to be true for NRAS and KIT mutations as has been mentioned in chapter 8. Most samples were also tested for mutations in KIT, BRAF and NRAS. Ten BRAFV600E mutations were found in the 188 samples tested for a BRAF mutation (5%), and only 17 KIT mutations in the 259 samples tested for a KIT mutation (7%), and only 14 NRAS mutations in the 192 tested for a NRAS mutation 7%). This makes it questionable whether all melanomas were truly primary mucosal. Systemic therapy in locally advanced or metastatic melanomas Traditional cytotoxic agents like Dacabarzine have had none to limited impact on overall survival in patients with locally advanced or metastatic melanoma. Dacabarzine has been the only chemotherapeutic agent that has been approved for the treatment of metastatic melanoma, but is associated with a low response rate (approximately 10%) and limited influence on median overall survival. The discovery that the pathways involved by mutated oncogenes can be more specifically targeted, has drastically improved therapeutic efficacy in a subset of these patients; mainly in those with a BRAF mutated cutaneous melanoma. The response rate to the first BRAF inhibitor vemurafenib (approved by the European Medicines Agency (EMA) since 2011) in patients with a BRAF-mutated metastatic cutaneous melanoma is 50%. The median overall survival in patients with advanced cutaneous melanoma treated with vemurafenib is fourteen to sixteen months whereas it is six to nine months in those treated with dacabarzine. Unfortunately, progression occurs in the vast majority of the patients after a median progression-free survival of five to seven months due to acquired resistance (83-85). The predominant pattern of resistance involves BRAFindependent reactivation of the mitogen-activated protein kinase (MAPK) pathway. However, MAPK pathway independent mechanisms may also be important (86, 87). The combined inhibition of both BRAF and MEK has been shown to improve clinical outcome by preventing or delaying the onset of resistance. Moreover, combined therapy decreases the number of secondary cutaneous cancers that occur due to 9 171

173 Chapter 9 paradoxical activation of the MAPK pathway in keratinocytes. In a trial with 495 patients with previously untreated unresectable locally advanced or metastatic BRAFV600 mutation-positive melanoma, the median progression-free survival was almost ten months in the group treated with the BRAF inhibitor vemurafenib in combination with the MEK inhibitor cobimetinib versus six months in the control group treated with vemurafenib and placebo. The rate of complete or partial response in the combination group was 68%, as compared with 45% in the control group, including rates of complete response of 10% in the combination group and 4% in the control group (88). In a more recent trail, 423 patients with previously untreated unresectable locally advanced or metastatic BRAF Val600Glu or BRAF Val600Lys mutation-positive melanoma were treated with BRAF inhibitor dabrafenib in combination with the MEK inhibitor trametinib or dabrafenib only. Median progression-free survival was eleven months in the combination group and almost nine months in the dabrafenib only group. Median overall survival was twenty-five months and almost nineteen months respectively (89). While the research described in this thesis has focused on finding genetic aberrations that could be used for targeted therapy, others have identified a variety of targets for immunotherapeutic approaches. In the context of immunotherapy, cutaneous melanoma is studied very extensively as it appears to be one of the most immunogenic cancers witnessing occurrence of spontaneous regression and the presence of tumor-infiltrating T-lymphocytes in primary tumors and metastases (90). In the recent years, immunotherapy has emerged and for metastatic cutaneous melanoma there has been a shift towards immunotherapy as first-line treatment. The currently most used immunotherapies in melanoma, target immune checkpoint receptors. Ipilimumab has been approved by the EMA for unresectable stage III and IV cutaneous melanoma since Ipilimumab is a monoclonal antibody that harnesses the immune system by blocking cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). Blocking of CTLA-4 by ipilimumab augments T-cell activation and proliferation. Treatment with ipilimumab can provide durable tumor control and long-term survival benefits. A disease control rate of 28.5% has been reported. However, a high rate of severe adverse effects have been observed (fatigue, skin toxicity, diarrhea, colitis etc.) (90). Anti-PD1 immune checkpoint inhibitors nivolumab and prembolizumab have been approved by the EMA since 2015 for unresectable stage III and IV cutaneous melanoma. Programmed death-1 (PD1) is another cell-surface small receptor on T-cells that, like CTLA-4, is inhibitory to the immune response of T-cells. In normal tissues, binding of PD-1 to its ligand PD-L1 prevents an excessive immune response and regulates tolerance to self-antigens. First results in patients with advanced melanoma have been promising and trials are ongoing (90, 91). In a recent trial, 418 patients who had a previously untreated unresectable locally advanced or metastatic melanoma without a BRAF mutation, were treated with PD-1 172

174 Summarizing discussion and future perspectives immune-checkpoint inhibitor antibody nivolumab or dacarbazine chemotherapy. At 1 year, the overall rate of survival was 72.9% in the nivolumab group, as compared with 42.1% in the dacarbazine group. The median progression-free survival was 5.1 months in the nivolumab group versus 2.2 months in the dacarbazine group. The objective response rate was 40.0% in the nivolumab group versus 13.9% in the dacarbazine group (92). Treatment with a combination of both immune checkpoint inhibitor antibodies seems even more effective. The objective-response rate and the progression-free survival among patients with advanced melanoma who had not previously received treatment were significantly greater with nivolumab combined with ipilimumab than with ipilimumab monotherapy (93). Although treatment with immune checkpoint inhibitors has led to major improvements in the survival of patients with advanced metastatic BRAF mutation-negative cutaneous melanoma, there are still several challenges. Treatment induces toxicity, is very expensive, and still only effective in a subset of patients while currently there are no biomarkers to select those patients who will benefit from therapy. Several other types of immunotherapy are under investigation like dendritic cell vaccination (which has recently shown improved overall survival when used in the adjuvant setting in cutaneous melanoma patients with regional lymph node metastases (94)), stimulation of the immune system with the use of vaccines to induce melanoma-specific responses, adoptive T cell transfer therapy, and others (91). Summarizing, the advantage of immunotherapy is the fact that more durable responses are achieved in comparison to targeted treatment and resistance does not occur. Disadvantage is that the effects are of later onset while targeted treatment often gives very rapid response. In rapidly progressive patients, therefore, immunotherapy is not applicable, and targeted treatment is still the treatment of first choice. For the future, combination of immunotherapy and targeted treatments (combined or sequential, depending on disease progression) are expected to combine the best of both treatments. There are not much data on the effectiveness of targeted therapy or immunotherapy in the less frequent melanoma subtypes this thesis focuses on. A few series have been published on the use of immune checkpoint inhibitors in patients with mucosal melanoma. These have shown that it is effective in a subset of patients and warrant further investigation. In a small case series of four female patients with large mucosal melanomas of the lower genital tract (three vaginal and one cervical), Ipilimumab has shown promising results in combination with radiation therapy in the initial treatment. One of the patients achieved complete pathological response (95). In a series of 71 Italian patients with metastatic mucosal melanoma from different primary sites treated with Ipilimumab, the response rate was 12% (i.e. partial or complete remission) and the immune-mediated disease control-rate was 36% (i.e. stable disease). Median progression-free survival and overall survival were 9 173

175 Chapter and 6.4 months, respectively (96). In a recent retrospective multicenter cohort of 35 patients with advanced mucosal melanoma, anti-pd-1 agents nivolumab or pembrolizumab showed an objective response rate of 23% and a median progression-free survival of 3.9 months. There was no detectable difference between subsites (twelve mucosal melanomas were from anorectal origin, fourteen primary vulvovaginal and nine originated from the head/ neck region). Response rates in this study were comparable to the published rates in patients with cutaneous melanoma (97). Several studies have implicated the efficiency of KIT inhibitors in the treatment of patients with advanced mucosal melanomas with KIT aberrations. In a phase II trial conducted by Carvajal et al. (98), 28 patients with mutations in or amplification of KIT, who had advanced unresectable melanoma arising from acral (10/28), mucosal (13/28) or chronically sun-damaged (5/ 28) sites, were treated with KIT inhibitor imatinib. Overall durable response rate was 16%, with a median time to progression of 12 weeks, and a median overall survival of 46.3 weeks. Two patients achieved durable complete response (lasting 94 (ongoing) and 95 weeks), two achieved durable partial response (lasting 53 and 89 (ongoing) weeks), two achieved transient partial response (lasting 12 and 18 weeks), and five achieved stable disease lasting for 12 weeks or longer. Guo et al. (99) have published a phase II trial of imatinib in 43 patients with metastatic melanoma harboring KIT aberrations, including 11 cases with mucosal melanoma. The median progression free survival was 3.5 months, and the 6-month progression free survival rate was 36.6%. 10 patients achieved partial response (23.3%), and 13 patients achieved stable disease. An overall response rate of 29% has been reported by Hodi et al. (100) among 25 patients with metastatic melanomas harboring mutationally activated or amplified KIT (17 mucosal melanomas). Response rates were higher in the patients with a KIT mutation compared to those with amplification only, and seemed more effective when the mutant allele was relatively abundant compared to the wild-type allele (98, 100). Mutations in KIT are widely distributed over the coding region, and the modest results from the studies might indicate that only select KIT alterations are truly oncogenic and indicative of an effective therapeutic target. The effectiveness of newer-generation KIT inhibitors is under investigation. Treatment with MEK inhibitors has not been proven effective in mucosal melanomas. In contrast to mutant BRAF and c-kit, at the moment, small molecules that selectively inhibit mutant NRAS, GNAQ or GNA11 are not available. For UMs, adjuvant therapy, including both targeted therapy as well as treatment with immune-checkpoint inhibitors has not been very effective with at most only limited clinical activity in small subset of patients ( ). Maybe immunotherapy is not effective in uveal melanomas because the eye is an immune privileged site. Histopathologically, uveal melanomas lack inflammatory cells in contrast to most of the cutaneous melanomas. 174

176 Summarizing discussion and future perspectives Some in vitro data have suggested that MEK inhibitors might be able to target NRAS-mutated LMNs. Experience in the use of targeted therapy in LMNs, however, is largely lacking (2). The knowledge on the pathogenesis of melanoma as well as the treatment options of advanced melanoma are evolving rapidly. With the perspective discovery of more targets, the development of new drugs and the discovery of biomarkers, we are entering a new era being able to develop a more personalized approach. The use of combined therapies seems most promising. Hopefully, in the near future, advanced melanoma can be cured in a larger subset of patients, and therapeutic options will also improve for the less frequent melanoma subtypes

177 Chapter 9 REFERENCES 1. Bastian BC. The molecular pathology of melanoma: an integrated taxonomy of melanocytic neoplasia. Annual review of pathology. 2014;9: Kusters-Vandevelde HV, Kusters B, van Engen-van Grunsven AC, Groenen PJ, Wesseling P, Blokx WA. Primary melanocytic tumors of the central nervous system: a review with focus on molecular aspects. Brain Pathol. 2015;25(2): Barnhill RL. The Spitzoid lesion: rethinking Spitz tumors, atypical variants, Spitzoid melanoma and risk assessment. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2006;19 Suppl 2:S van Dijk MC, Aben KK, van Hees F, Klaasen A, Blokx WA, Kiemeney LA, et al. Expert review remains important in the histopathological diagnosis of cutaneous melanocytic lesions. Histopathology. 2008;52(2): Barnhill RL, Argenyi ZB, From L, Glass LF, Maize JC, Mihm MC, Jr., et al. Atypical Spitz nevi/tumors: lack of consensus for diagnosis, discrimination from melanoma, and prediction of outcome. Human pathology. 1999;30(5): Network TCGA. Genomic Classification of Cutaneous Melnoma. Cell. 2015;161(7): Bansal R, Nikiforov MA. Pathways of oncogene-induced senescence in human melanocytic cells. Cell cycle. 2010;9(14): Wiesner T, Obenauf AC, Murali R, Fried I, Griewank KG, Ulz P, et al. Germline mutations in BAP1 predispose to melanocytic tumors. Nature genetics. 2011;43(10): Wiesner T, Murali R, Fried I, Cerroni L, Busam K, Kutzner H, et al. A distinct subset of atypical Spitz tumors is characterized by BRAF mutation and loss of BAP1 expression. The American journal of surgical pathology. 2012;36(6): Busam KJ, Wanna M, Wiesner T. Multiple epithelioid Spitz nevi or tumors with loss of BAP1 expression: a clue to a hereditary tumor syndrome. JAMA dermatology. 2013;149(3): Busam KJ, Sung J, Wiesner T, von Deimling A, Jungbluth A. Combined BRAF(V600E)-positive melanocytic lesions with large epithelioid cells lacking BAP1 expression and conventional nevomelanocytes. The American journal of surgical pathology. 2013;37(2): Yeh I, Mully TW, Wiesner T, Vemula SS, Mirza SA, Sparatta AJ, et al. Ambiguous melanocytic tumors with loss of 3p21. The American journal of surgical pathology. 2014;38(8): Blokx WA, Rabold K, Bovenschen HJ, Ligtenberg MJ, Mensenkamp AR, van Lijnschoten G, et al. NRAS-mutated melanocytic BAP1-associated intradermal tumor (MBAIT): a case report. Virchows Archiv : an international journal of pathology. 2015;466(1): Wiesner T, Fried I, Ulz P, Stacher E, Popper H, Murali R, et al. Toward an improved definition of the tumor spectrum associated with BAP1 germline mutations. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2012;30(32):e Harbour JW, Onken MD, Roberson ED, Duan S, Cao L, Worley LA, et al. Frequent mutation of BAP1 in metastasizing uveal melanomas. Science. 2010;330(6009): Testa JR, Cheung M, Pei J, Below JE, Tan Y, Sementino E, et al. Germline BAP1 mutations predispose to malignant mesothelioma. Nature genetics. 2011;43(10): Popova T, Hebert L, Jacquemin V, Gad S, Caux-Moncoutier V, Dubois-d Enghien C, et al. Germline BAP1 mutations predispose to renal cell carcinomas. American journal of human genetics. 2013;92(6): Abdel-Rahman MH, Pilarski R, Cebulla CM, Massengill JB, Christopher BN, Boru G, et al. Germline BAP1 mutation predisposes to uveal melanoma, lung adenocarcinoma, meningioma, and other cancers. Journal of medical genetics. 2011;48(12): Wiesner T, Kutzner H, Cerroni L, Mihm MC, Jr., Busam KJ, Murali R. Genomic aberrations in spitzoid melanocytic tumours and their implications for diagnosis, prognosis and therapy. Pathology. 2016;48(2): Wiesner T, He J, Yelensky R, Esteve-Puig R, Botton T, Yeh I, et al. Kinase fusions are frequent in Spitz tumours and spitzoid melanomas. Nature communications. 2014;5:

178 Summarizing discussion and future perspectives 21. Yeh I, Botton T, Talevich E, Shain AH, Sparatta AJ, de la Fouchardiere A, et al. Activating MET kinase rearrangements in melanoma and Spitz tumours. Nature communications. 2015;6: Kwak EL, Bang YJ, Camidge DR, Shaw AT, Solomon B, Maki RG, et al. Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. The New England journal of medicine. 2010;363(18): Bergethon K, Shaw AT, Ou SH, Katayama R, Lovly CM, McDonald NT, et al. ROS1 rearrangements define a unique molecular class of lung cancers. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2012;30(8): Leboulleux S, Bastholt L, Krause T, de la Fouchardiere C, Tennvall J, Awada A, et al. Vandetanib in locally advanced or metastatic differentiated thyroid cancer: a randomised, double-blind, phase 2 trial. The Lancet Oncology. 2012;13(9): Wells SA, Jr., Robinson BG, Gagel RF, Dralle H, Fagin JA, Santoro M, et al. Vandetanib in patients with locally advanced or metastatic medullary thyroid cancer: a randomized, double-blind phase III trial. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2012;30(2): Elisei R, Schlumberger MJ, Muller SP, Schoffski P, Brose MS, Shah MH, et al. Cabozantinib in progressive medullary thyroid cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(29): Klebe S, Driml J, Nasu M, Pastorino S, Zangiabadi A, Henderson D, et al. BAP1 hereditary cancer predisposition syndrome: a case report and review of literature. Biomarker research. 2015;3: Prescher G, Bornfeld N, Hirche H, Horsthemke B, Jockel KH, Becher R. Prognostic implications of monosomy 3 in uveal melanoma. Lancet. 1996;347(9010): Sisley K, Rennie IG, Parsons MA, Jacques R, Hammond DW, Bell SM, et al. Abnormalities of chromosomes 3 and 8 in posterior uveal melanoma correlate with prognosis. Genes, chromosomes & cancer. 1997;19(1): Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer research. 2004;64(20): Martin M, Masshofer L, Temming P, Rahmann S, Metz C, Bornfeld N, et al. Exome sequencing identifies recurrent somatic mutations in EIF1AX and SF3B1 in uveal melanoma with disomy 3. Nature genetics. 2013;45(8): Harbour JW, Roberson ED, Anbunathan H, Onken MD, Worley LA, Bowcock AM. Recurrent mutations at codon 625 of the splicing factor SF3B1 in uveal melanoma. Nature genetics. 2013;45(2): Furney SJ, Pedersen M, Gentien D, Dumont AG, Rapinat A, Desjardins L, et al. SF3B1 mutations are associated with alternative splicing in uveal melanoma. Cancer discovery. 2013;3(10): Yavuzyigitoglu S, Koopmans AE, Verdijk RM, Vaarwater J, Eussen B, van Bodegom A, et al. Uveal Melanomas with SF3B1 Mutations: A Distinct Subclass Associated with Late-Onset Metastases. Ophthalmology. 2016;123(5): Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, Vemula S, Wiesner T, et al. Mutations in GNA11 in uveal melanoma. The New England journal of medicine. 2010;363(23): Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, O Brien JM, et al. Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature. 2009;457(7229): Onken MD, Worley LA, Long MD, Duan S, Council ML, Bowcock AM, et al. Oncogenic mutations in GNAQ occur early in uveal melanoma. Investigative ophthalmology & visual science. 2008;49(12): Gessi M, Hammes J, Lauriola L, Dorner E, Kirfel J, Kristiansen G, et al. GNA11 and N-RAS mutations: alternatives for MAPK pathway activating GNAQ mutations in primary melanocytic tumours of the central nervous system. Neuropathology and applied neurobiology. 2013;39(4): van de Nes J, Gessi M, Sucker A, Moller I, Stiller M, Horn S, et al. Targeted next generation sequencing reveals unique mutation profile of primary melanocytic tumors of the central nervous system. Journal of neuro-oncology. 2016;127(3): Murali R, Wiesner T, Rosenblum MK, Bastian BC. GNAQ and GNA11 mutations in melanocytomas of the central nervous system. Acta neuropathologica. 2012;123(3): Griewank KG, van de Nes J, Schilling B, Moll I, Sucker A, Kakavand H, et al. Genetic and clinico-pathologic analysis of metastatic uveal melanoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2014;27(2):

179 Chapter Daniels AB, Lee JE, MacConaill LE, Palescandolo E, Van Hummelen P, Adams SM, et al. High throughput mass spectrometry-based mutation profiling of primary uveal melanoma. Investigative ophthalmology & visual science. 2012;53(11): Koopmans AE, Vaarwater J, Paridaens D, Naus NC, Kilic E, de Klein A, et al. Patient survival in uveal melanoma is not affected by oncogenic mutations in GNAQ and GNA11. British journal of cancer. 2013;109(2): Damato B, Dopierala JA, Coupland SE. Genotypic profiling of 452 choroidal melanomas with multiplex ligation-dependent probe amplification. Clinical cancer research : an official journal of the American Association for Cancer Research. 2010;16(24): Kusters-Vandevelde HV, Creytens D, van Engen-van Grunsven AC, Jeunink M, Winnepenninckx V, Groenen PJ, et al. SF3B1 and EIF1AX mutations occur in primary leptomeningeal melanocytic neoplasms; yet another similarity to uveal melanomas. Acta neuropathologica communications. 2016;4: de la Fouchardiere A, Cabaret O, Petre J, Aydin S, Leroy A, de Potter P, et al. Primary leptomeningeal melanoma is part of the BAP1-related cancer syndrome. Acta neuropathologica. 2015;129(6): Davies MA, Liu P, McIntyre S, Kim KB, Papadopoulos N, Hwu WJ, et al. Prognostic factors for survival in melanoma patients with brain metastases. Cancer. 2011;117(8): Bafaloukos D, Gogas H. The treatment of brain metastases in melanoma patients. Cancer treatment reviews. 2004;30(6): Sampson JH, Carter JH, Jr., Friedman AH, Seigler HF. Demographics, prognosis, and therapy in 702 patients with brain metastases from malignant melanoma. Journal of neurosurgery. 1998;88(1): Rodriguez y Baena R, Gaetani P, Danova M, Bosi F, Zappoli F. Primary solitary intracranial melanoma: case report and review of the literature. Surgical neurology. 1992;38(1): Skarli SO, Wolf AL, Kristt DA, Numaguchi Y. Melanoma arising in a cervical spinal nerve root: report of a case with a benign course and malignant features. Neurosurgery. 1994;34(3):533-7; discussion Greco Crasto S, Soffietti R, Bradac GB, Boldorini R. Primitive cerebral melanoma: case report and review of the literature. Surgical neurology. 2001;55(3):163-8; discussion Rivers JK, Bhayana S, Martinka M. Dural melanoma associated with ocular melanosis and multiple blue nevi. Journal of cutaneous medicine and surgery. 2001;5(5): Theunissen P, Spinc le G, Pannebakker M, Lambers J. Meningeal melanoma associated with nevus of Ota: case report and review. Clinical neuropathology. 1993;12(3): Mooi WJ, Krausz, T. Blue nevus and related lesions. In: Mooi WJ, Krausz, T., editor. Pathology of melanocytic disorders. 2 ed. London: Hodder Arnold; p Carlson JA, Dickersin GR, Sober AJ, Barnhill RL. Desmoplastic neurotropic melanoma. A clinicopathologic analysis of 28 cases. Cancer. 1995;75(2): Huang HY, Park N, Erlandson RA, Antonescu CR. Immunohistochemical and ultrastructural comparative study of external lamina structure in 31 cases of cellular, classical, and melanotic schwannomas. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2004;12(1): Prieto VG, Woodruff JM. Expression of basement membrane antigens in spindle cell melanoma. Journal of cutaneous pathology. 1998;25(6): Schmoeckel C, Stolz W, Sakai LY, Burgeson RE, Timpl R, Krieg T. Structure of basement membranes in malignant melanoma and nevocytic nevi. The Journal of investigative dermatology. 1989;92(5): Mooi WJ, Krausz, T. Other extracutaneous melanotic tumors. In: Mooi WJ, Krausz, T., editor. Pathology of melanocytic disorders. 2 ed. London: Hodder Arnold; p Brat DJ, Perry, A. Melanocytic lesions. In: Louis DN, Ohgaki, H., Wiestler, O.D., Cavanee, W.K., editor. WHO classification of tumours of the central nervous system. 4 ed. Lyon: IARC; p Brat DJ, Giannini C, Scheithauer BW, Burger PC. Primary melanocytic neoplasms of the central nervous systems. The American journal of surgical pathology. 1999;23(7): Scheithauer BW, Louis D.N., Hunter S., Woodruff J.M., Antonescu C.R. Schwannoma. In: Louis DN, Ohgaki, H., Wiestler, O.D., Cavanee, W.K., editor. WHO classification of tumours of the central nervous system. 4 ed. Lyon: IARC; p

180 Summarizing discussion and future perspectives 64. Kusters-Vandevelde HV, Klaasen A, Kusters B, Groenen PJ, van Engen-van Grunsven IA, van Dijk MR, et al. Activating mutations of the GNAQ gene: a frequent event in primary melanocytic neoplasms of the central nervous system. Acta neuropathologica. 2010;119(3): Rothenbuhler A, Stratakis CA. Clinical and molecular genetics of Carney complex. Best practice & research Clinical endocrinology & metabolism. 2010;24(3): Carney JA. Psammomatous melanotic schwannoma. A distinctive, heritable tumor with special associations, including cardiac myxoma and the Cushing syndrome. The American journal of surgical pathology. 1990;14(3): Waelchli R, Aylett SE, Atherton D, Thompson DJ, Chong WK, Kinsler VA. Classification of neurological abnormalities in children with congenital melanocytic naevus syndrome identifies magnetic resonance imaging as the best predictor of clinical outcome. The British journal of dermatology. 2015;173(3): Pleunis N, Schuurman MS, Van Rossum MM, Bulten J, Massuger LF, De Hullu JA, et al. Rare vulvar malignancies; incidence, treatment and survival in the Netherlands. Gynecologic oncology Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT in distinct subtypes of melanoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2006;24(26): Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, et al. Distinct sets of genetic alterations in melanoma. The New England journal of medicine. 2005;353(20): Schoenewolf NL, Bull C, Belloni B, Holzmann D, Tonolla S, Lang R, et al. Sinonasal, genital and acrolentiginous melanomas show distinct characteristics of KIT expression and mutations. European journal of cancer. 2012;48(12): Satzger I, Schaefer T, Kuettler U, Broecker V, Voelker B, Ostertag H, et al. Analysis of c-kit expression and KIT gene mutation in human mucosal melanomas. British journal of cancer. 2008;99(12): Beadling C, Jacobson-Dunlop E, Hodi FS, Le C, Warrick A, Patterson J, et al. KIT gene mutations and copy number in melanoma subtypes. Clinical cancer research : an official journal of the American Association for Cancer Research. 2008;14(21): Omholt K, Grafstrom E, Kanter-Lewensohn L, Hansson J, Ragnarsson-Olding BK. KIT pathway alterations in mucosal melanomas of the vulva and other sites. Clinical cancer research : an official journal of the American Association for Cancer Research. 2011;17(12): Turri-Zanoni M, Medicina D, Lombardi D, Ungari M, Balzarini P, Rossini C, et al. Sinonasal mucosal melanoma: Molecular profile and therapeutic implications from a series of 32 cases. Head & neck. 2013;35(8): Sekine S, Nakanishi Y, Ogawa R, Kouda S, Kanai Y. Esophageal melanomas harbor frequent NRAS mutations unlike melanomas of other mucosal sites. Virchows Archiv : an international journal of pathology. 2009;454(5): Aulmann S, Sinn HP, Penzel R, Gilks CB, Schott S, Hassel JC, et al. Comparison of molecular abnormalities in vulvar and vaginal melanomas. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2014;27(10): Rouzbahman M, Kamel-Reid S, Al Habeeb A, Butler M, Dodge J, Laframboise S, et al. Malignant Melanoma of Vulva and Vagina: A Histomorphological Review and Mutation Analysis--A Single-Center Study. Journal of lower genital tract disease. 2015;19(4): Whiteman DC, Pavan WJ, Bastian BC. The melanomas: a synthesis of epidemiological, clinical, histopathological, genetic, and biological aspects, supporting distinct subtypes, causal pathways, and cells of origin. Pigment cell & melanoma research. 2011;24(5): Shain AH, Bastian BC. From melanocytes to melanomas. Nature reviews Cancer. 2016;16(6): Shain AH, Garrido M, Botton T, Talevich E, Yeh I, Sanborn JZ, et al. Exome sequencing of desmoplastic melanoma identifies recurrent NFKBIE promoter mutations and diverse activating mutations in the MAPK pathway. Nature genetics. 2015;47(10): Sheng X, Kong Y, Li Y, Zhang Q, Si L, Cui C, et al. GNAQ and GNA11 mutations occur in 9.5% of mucosal melanoma and are associated with poor prognosis. European journal of cancer. 2016;65: Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, et al. Inhibition of mutated, activated BRAF in metastatic melanoma. The New England journal of medicine. 2010;363(9):

181 Chapter Sosman JA, Kim KB, Schuchter L, Gonzalez R, Pavlick AC, Weber JS, et al. Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. The New England journal of medicine. 2012;366(8): Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. The New England journal of medicine. 2011; 364(26): Van Allen EM, Wagle N, Sucker A, Treacy DJ, Johannessen CM, Goetz EM, et al. The genetic landscape of clinical resistance to RAF inhibition in metastatic melanoma. Cancer discovery. 2014;4(1): Wagle N, Van Allen EM, Treacy DJ, Frederick DT, Cooper ZA, Taylor-Weiner A, et al. MAP kinase pathway alterations in BRAF-mutant melanoma patients with acquired resistance to combined RAF/MEK inhibition. Cancer discovery. 2014;4(1): Larkin J, Ascierto PA, Dreno B, Atkinson V, Liszkay G, Maio M, et al. Combined vemurafenib and cobimetinib in BRAF-mutated melanoma. The New England journal of medicine. 2014;371(20): Long GV, Stroyakovskiy D, Gogas H, Levchenko E, de Braud F, Larkin J, et al. Dabrafenib and trametinib versus dabrafenib and placebo for Val600 BRAF-mutant melanoma: a multicentre, double-blind, phase 3 randomised controlled trial. Lancet. 2015;386(9992): van Zeijl MC, van den Eertwegh AJ, Haanen JB, Wouters MW. (Neo)adjuvant systemic therapy for melanoma. European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology Zhu Z, Liu W, Gotlieb V. The rapidly evolving therapies for advanced melanoma--towards immunotherapy, molecular targeted therapy, and beyond. Critical reviews in oncology/hematology. 2016;99: Robert C, Long GV, Brady B, Dutriaux C, Maio M, Mortier L, et al. Nivolumab in previously untreated melanoma without BRAF mutation. The New England journal of medicine. 2015;372(4): Postow MA, Chesney J, Pavlick AC, Robert C, Grossmann K, McDermott D, et al. Nivolumab and ipilimumab versus ipilimumab in untreated melanoma. The New England journal of medicine. 2015;372(21): Bol KF, Aarntzen EH, Hout FE, Schreibelt G, Creemers JH, Lesterhuis WJ, et al. Favorable overall survival in stage III melanoma patients after adjuvant dendritic cell vaccination. Oncoimmunology. 2016;5(1): e Schiavone MB, Broach V, Shoushtari AN, Carvajal RD, Alektiar K, Kollmeier MA, et al. Combined immunotherapy and radiation for treatment of mucosal melanomas of the lower genital tract. Gynecologic oncology reports. 2016;16: Del Vecchio M, Di Guardo L, Ascierto PA, Grimaldi AM, Sileni VC, Pigozzo J, et al. Efficacy and safety of ipilimumab 3mg/kg in patients with pretreated, metastatic, mucosal melanoma. European journal of cancer. 2014;50(1): Shoushtari AN, Munhoz RR, Kuk D, Ott PA, Johnson DB, Tsai KK, et al. The efficacy of anti-pd-1 agents in acral and mucosal melanoma. Cancer Carvajal RD, Antonescu CR, Wolchok JD, Chapman PB, Roman RA, Teitcher J, et al. KIT as a therapeutic target in metastatic melanoma. Jama. 2011;305(22): Guo J, Si L, Kong Y, Flaherty KT, Xu X, Zhu Y, et al. Phase II, open-label, single-arm trial of imatinib mesylate in patients with metastatic melanoma harboring c-kit mutation or amplification. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(21): Hodi FS, Corless CL, Giobbie-Hurder A, Fletcher JA, Zhu M, Marino-Enriquez A, et al. Imatinib for melanomas harboring mutationally activated or amplified KIT arising on mucosal, acral, and chronically sun-damaged skin. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(26): Kelderman S, van der Kooij MK, van den Eertwegh AJ, Soetekouw PM, Jansen RL, van den Brom RR, et al. Ipilimumab in pretreated metastastic uveal melanoma patients. Results of the Dutch Working group on Immunotherapy of Oncology (WIN-O). Acta oncologica. 2013;52(8): Algazi AP, Tsai KK, Shoushtari AN, Munhoz RR, Eroglu Z, Piulats JM, et al. Clinical outcomes in metastatic uveal melanoma treated with PD-1 and PD-L1 antibodies. Cancer

182 Summarizing discussion and future perspectives 103. Carvajal RD, Sosman JA, Quevedo JF, Milhem MM, Joshua AM, Kudchadkar RR, et al. Effect of selumetinib vs chemotherapy on progression-free survival in uveal melanoma: a randomized clinical trial. Jama. 2014;311(23): Choudhary MM, Triozzi PL, Singh AD. Uveal melanoma: evidence for adjuvant therapy. International ophthalmology clinics. 2015;55(1):

183

184 Part IV