🗂 總目錄 ｜ 📖 英文原文（本篇） ｜ 📝 完整翻譯 ｜ ⭐ 精華筆記

Molecular classification of melanoma — Part 2

(SLG&RQWRXU

6RODU(ODVWRVLV

&HOO6KDSH

&HOO6L]H

$JH

loss.20 The latter mutations reflect the importance of replicative senescence and maintaining G1/S checkpoint control in preventing nevi from progressing to melanoma. UV signature mutation profiles implicate UV radiation as the driving factor in this transformation of nevi to melanoma.20

Melanoma with low cumulative sun damage (low-CSD) Low-CSD melanomas occur most frequently in patients under 50 on the trunk or extremities and show a low to moderate degree of solar elastosis histopathologically. BRAFV600E mutations are very common (approximately 70%) in low-CSD melanomas, representing the main driver mutation, followed by NRAS mutations (approximately 20%).19,23 BRAF mutant melanomas frequently exhibit specific histomorphological features such as increased upward scatter of intraepidermal melanocytes (pagetoid scatter), a predominance of nests over single cells, thickening of the involved epidermis, a sharper demarcation toward the adjacent uninvolved epidermis, and constituent tumor cells that are larger, rounder, and more heavily pigmented (Fig. 26.134).24 Simple combinations of these independently associated features predicted BRAF mutation status with 90% accuracy in one study.24 These histomorphological criteria also performed better in predicting mutation status than the traditional WHO melanoma classification, indicating the potential for developing improved classification schemes that define more biologically homogeneous subsets. This is increasingly relevant with the increased use of targeted therapies to specific genetic alterations including BRAF inhibitors25 and KIT inhibitors.26–28

Progression from in situ to invasive melanoma involves additional genetic alterations required for unrestrained growth outside the confines of the epidermis. Ability to grow and survive in the dermis alone is not indicative of malignant transformation, as the cells of melanocytic nevi already have this property. Loss of cell cycle checkpoint regulation is critical in unleashing or increasing the proliferative capacity associated with invasive melanoma as evidenced by frequent homozygous loss of the CDKN2A gene encoding both p16 and p14ARF in the transition from in situ to invasive melanoma. Additional mutations in genes encoding critical tumor suppressor proteins such as TP53 and PTEN occur later in the progression to more advanced stage disease.20 Additionally, invasive melanomas must evade antitumor immune surveillance through mechanisms such as induction of immune tolerance.

The shared high incidence of BRAFV600E mutations in melanocytic nevi and low-CSD melanomas supports a common etiology with regard to UV radiation and timing of exposure. Approximately one-fourth to one-third of melanomas arise within histopathologically identifiable precursor nevi, with low-CSD melanomas representing the main melanoma type in which precursor nevi can be identified.29 Frequently only melanoma in situ is present with the associated nevus, the latter of which remains identifiable in the underlying dermis. The melanoma in situ harbors the same initiating mutation as the precursor nevus (typically BRAF or NRAS), with additional genetic aberrations such as TERT promoter mutation or heterozygous CDKN2A

The vast majority of melanomas exhibit some degree of genomic instability in the form of chromosomal gains and losses. This instability seems to coincide with the transition from benign melanocytic nevi to melanoma in situ, as nevi tend to not have chromosomal copy number changes, whereas they can be detected in melanoma in situ.20 The instability may be related to telomere shortening and end-to-end fusion of sister chromatids with eroded telomeres that result in dicentric chromosomes that can rupture during mitosis. The selective advantage of TERT promoter mutations early during melanoma progression suggests that replicative senescence becomes an early barrier to transformation. TERT promoter mutations only lead to marginal telomerase expression, which manages to preserve critically short telomeres but cannot fully suppress telomere fusions resulting in instability that generates DNA copy number changes.30 This explains why the telomeres of melanomas with TERT promoter mutations remain short.30,31

While there are some similarities in the patterns of the chromosomal gains and losses across melanoma subtypes, significant differences exist with

Gains 0.6 0.2

-0.2 -0.6

Non-CSD CSD

Losses

0.2 0.0

Amplifications

Homozygous deletions

-0.2

1 3 5 7 9 11 13 15 17 19 21

1357 Molecular classification of melanoma

.,7 .,7DQG 156 .,7DQG %5) 15$6$)



3HUFHQW$EHUUDQW

0.6 0.2

Gains

-0.2 -0.6

Losses

0.2 0.0

Amplifications

Homozygous deletions

-0.2

1 3 5 7

9 11 13 15 17 19 21



1RQ&6’ &6’ $FUDO 0XFRVDO 

By contrast, activating mutations in NRAS activate both the MAP kinase and the PI3 kinase pathway, therefore obviating the need for a loss of PTEN in tumors with NRAS mutations.32 Point mutations in the catalytic subunit of PI3 kinase have also been found in a small minority of melanomas.33

Receptor Tyrosine Kinase (e.g., KIT)

PTEN NRAS BRAF PI3K

MEK AKT

Gains of chromosome 7 are more common in low-CSD melanomas.19 BRAF resides at 7q34 and the copy number increase reflects duplication of the mutant BRAF allele, which provides a proliferative advantage.34,35 Copy number increases of CCND1 encoding cyclin D1 at chromosome 11q13 represent an additional distinguishing feature as they occur infrequently in low-CSD melanomas, while present in up to 50% of high-CSD melanoma.19,36 These genetic changes demonstrate how both the major proliferative and survival pathways are activated during melanoma progression. This can be accomplished through different combinations of mutations and chromosomal aberrations such as activating mutations in upstream tyrosine kinase receptors such as KIT or NRAS which activate both the MAP kinase pathway and PI3K pathway, or through combinations of activating mutations in BRAF combined with loss of PTEN (Fig. 26.136).

ERK

Survival

p16

CCND1 CDK4 CDK6

Proliferation

Melanoma with high cumulative sun damage (high-CSD) High-CSD melanomas typically occur in Caucasians over age 50 mostly on the head and neck and display marked solar elastosis histopathologically. In contrast to low-CSD melanomas, high-CSD melanomas typically have no associated precursor nevus, arising ‘de novo’ in the epidermis. These melanomas are often characterized by an extended in situ stage, in which the neoplasm can grow to the size of several centimeters over many years prior to becoming invasive. Genomic data on the in situ stage of high-CSD melanoma are limited. Approximately 20% of such in situ melanomas have BRAF mutations, although BRAFnon-V600E mutations (e.g., BRAFV600K) are more frequent than BRAFV600E mutations in this group.37 In addition to the striking difference in BRAF mutations between high-CSD and low-CSD melanomas (15% vs 70%),34,38 additional genetic differences have also been found. For instance, some studies report up to 30% of high-CSD melanomas harbor mutations or DNA copy number increases of KIT, whereas KIT mutations are rare in low-CSD melanomas (Fig. 26.137).39 Differences in the mutation frequency of NRAS have been reported by some, but do not appear to be a consistent finding.19,40

respect to tumors based on the degree of UV exposure.19 Gain of chromosome 6p and loss of 6q are common in both low-CSD and high-CSD melanomas, but chromosome 10 loss occurs in 40% of low-CSD melanomas and fewer than 10% of CSD melanomas (Fig. 26.135). The region commonly lost on chromosome 10 contains the tumor suppressor PTEN, which encodes a negative regulator of the PI3 kinase pathway and is considered to represent the major selective force for chromosome 10 loss. BRAF mutations only activate the MAP kinase pathway, which explains the advantage of concurrent PTEN loss to activate the PI3 kinase pathway (Fig. 26.136).32

Desmoplastic melanoma represents a specific variant of melanoma that occurs in chronically sun-damaged skin and has a high mutation burden that typically exceeds that of all other melanomas (median 62 mutations/ Mb compared to approximately 15/Mb in low-CSD melanoma).41 The mutational profile of desmoplastic melanoma has a strong UV radiation profile and is notable for the complete absence of BRAF and NRAS hotspot mutations and a high prevalence of inactivating NF1 mutations, which can also be detected in the overlying in situ component.41,42 In addition, there are recurrent mutations in the promoter of NFKBIE, which encodes an inhibitor of nuclear factor NF-κB signaling, and inactivating mutations in TP53, CDKN2A, ARID1A, ARID2, and RB1. TERT promoter mutations are also

1358 Melanoma

Gene Function Associated with melanoma susceptibility

ACD/POT1/TERF2IP Nuclear proteins that bind telomeres and prevent inappropriate recombination events Yes

ARID1/ARID2 Chromatin remodeling

BAP1 Deubiquitinase that binds to BRCA1 and is involved with chromatin dynamics Yes

BRAF Protein kinase in the MAP kinase pathway

CDKN2A Encodes p16, an inhibitor of cyclin-dependent kinases, and p14, an inhibitor of MDM2 Yes

CDK4 Cyclin-dependent kinase that phosphorylates Rb Yes

CTNNB1 Encodes beta-catenin, an adherens junction protein that signals in the WNT pathway

ERBB4 EGFR family of tyrosine kinase receptors involved in differentiation and proliferation

GNAQ/GNA11 G-protein coupled receptor that activates phospholipase C-beta

KIT Tyrosine kinase receptor activating multiple pathways in cell survival and proliferation

MC1R Encodes the melanocortin 1 receptor that binds melanocyte stimulatory hormone Yes

MGMT DNA repair protein Yes

MITF Transcription factor in melanocyte development Yes

NF1 Encodes neurofibromin that negatively regulates RAS signaling

NRAS Tyrosine kinase receptor that activates the MAP kinase and PI3K pathways

POLE DNA polymerase subunit involved in DNA repair Yes

PTEN Lipid phosphatase antagonist of PI3K pathway Yes

RAC1 GTPase involved in cell growth

SLC45A2 Transporter protein in melanin synthesis Yes

TERT Encodes telomerase which maintains telomere length; can be amplified or have promoter mutation

TP53 Encodes tumor suppressor p53

Yes

XP genes Xeroderma pigmentosum genes encode nucleotide excision repair proteins Yes

common in desmoplastic melanomas.41 Desmoplastic melanomas have fewer chromosomal copy number alterations than most other melanomas, though focal amplifications of oncogenes such as EGFR, CDK4, CCND1, MDM2, TERT, and MAP3K1 and focal deletions in tumor suppressor genes such as CDKN2A occur.41

corneum providing additional UV protection. The nail matrix is protected by the proximal nail fold and the nail plate. With the exception of the lips and the bulbar conjunctiva, mucosal sites in which melanoma arise are not exposed to sunlight. Genomic analyses with CGH and DNA sequencing in melanomas from these UV-protected sites show a high degree of genomic instability reflected by frequent chromosomal aberrations, with focused gene amplifications in particular and a low mutation burden (Fig. 26.138).19,50–53 The mucosal example in Fig. 26.138 in particular shows copy number changes affecting virtually every chromosome. Although mucosal melanomas are typically detected with considerable latency and therefore tend to be thicker than the primaries of the other melanoma categories, this high degree of genomic instability can already be demonstrated in thinner lesions. In other cancers, gene amplifications usually arise late during progression and often are a sign of adverse prognosis.54 By contrast, in acral melanoma these amplifications arise very early during progression and can already be detected at the in situ stage.50

Fig. 26.134 Morphological and clinical features of 302 primary melanomas by mutation status: the heat map shows the features that are significantly associated with mutation status for three groups of melanomas: BRAF mutant, NRAS mutant, or neither mutation. The scores for the features in the following rows range from shades of blue (low score) to gray (intermediate scores) to yellow (high scores): pigmentation of neoplastic melanocytes, upward scatter of intraepidermal melanocytes, nesting of melanocytes, degree of solar elastosis, cell size, and patient age. For circumscription, yellow indicates an abrupt transition from involved to adjacent uninvolved skin, gray scores represent a continuous transition, and blue scores a discontinuous transition. Epidermal contour ranges from yellow (acanthotic) to blue (atrophic) of the epidermis involved by the melanoma. Cell shapes range from blue (round) to yellow (spindled). Samples are listed in columns and are ordered within each category using agglomerative hierarchical clustering. The color codes for WHO types are superficial spreading melanoma (SSM) green, lentigo maligna melanoma (LMM) yellow, nodular melanoma (NM) red, and acral lentiginous melanoma (ALM) orange. Unclassifiable samples are in black.

Fig. 26.135 Frequencies of DNA copy number changes in non-CSD and CSD melanomas: the upper histogram for each melanoma type shows the frequencies of gains (green) and losses (red), and the lower histogram shows amplifications (green) and homozygous deletions (red). The x axis represents the genomic position from chromosome 1 to 22. Vertical solid blue lines mark the boundaries between chromosomes, and the vertical dashed gray lines mark the position of the centromeres.

Fig. 26.136 Simplified schematic of pathways important in melanoma: downstream from certain receptor tyrosine kinases lies Ras which bifurcates to the BRAF/ERK and PI3K/AKT pathways leasing to proliferation and survival. It appears that activation of both of these pathways (and others) is required for melanoma development and these are grouped in a rational fashion. For instance, activating mutations in KIT, NRAS, and BRAF are mutually exclusive since they activate the same pathways and loss of PTEN is seen with BRAF, but not NRAS mutation, because NRAS activates the PI3K pathway while BRAF does not. These combinations will likely have therapeutic relevance as more specific inhibitors are developed.

Fig. 26.137 Frequency distribution of genetic alterations in BRAF, NRAS, and KIT among four groups of melanoma: non- CSD, melanomas on skin without chronic sun-induced damage; CSD, melanomas on skin with chronic suninduced as evidenced by the presence of marked solar elastosis; acral, melanomas on the soles, palms, or subungual sites; mucosal, melanomas on mucosal membranes. One CSD melanoma had a KIT and an NRAS mutation, and one acral melanoma had a KIT and a BRAF mutation.

Fig. 26.138 Differences in the degree of genomic instability among melanoma types: representative examples of array CGH profiles of non-CSD, CSD, acral, and mucosal melanoma. The y axis represents the copy number for each array element (average value of a triplicate measurement) expressed as the log2 of the ratio of the tumor to reference fluorescence intensities. Values between +0.25 and –0.25 would be considered normal; values above that range are gains and below that range are losses. Values above 0.9 would be considered amplifications. In practice, a more complex assessment that takes into account the signal-to-noise ratio for each individual case is used.33 The x axis represents the genomic position from chromosome 1p to 22.