Malignant Melanoma

Clinical Features:
The incidence of melanoma has been rapidly rising over the last 50 years in the US with a lifetime risk increasing from 1 in 1500 in 1935 to 1 in 75 at present (Rigel et al., 1996). In contrast to many other solid tumors, melanoma may affect a younger population and is the leading cause of death due to malignancy among females between the ages of 20 and 30 (Johnson et al. 1998). While the mortality has risen, it has been at a lower rate than the incidence. This may be due to the earlier detection and treatment of melanomas today.

Malignant melanoma is classified into several subtypes based on clinical characteristics and histopathological features (Langley et al., 1999). Superficial spreading melanoma (70% of cases) typically present as a flat or slightly elevated brown/black lesions that vary in color and have irregular outlines dermatology.uiowa.edu. The back is the most common site of primary melanomas in males, while the legs are most frequently involved in females. Nodular melanomas (15% of cases) appear as small but rapidly enlarging nodules that tend to ulcerate dermatology.uiowa.edu. Acral lentiginous melanoma (8% of cases) occurs on the hands or feet and bear some resemblance to superficial spreading melanoma dermatology.uiowa.edu Lentigo maligna melanoma (5% of cases) tends to arise from pigmented lesions on the sun-damaged skin of elderly patients.

Melanoma prognosis is affected by several variables, including patient age, gender, primary tumor site, tumor thickness, tumor ulceration, mitotic rate of tumor cells, the presence of lymphovascular or perineural invasion, the presence of tumor infiltrating lymphocytes, and the presence of metastasis to lymph nodes, the skin, or visceral organs (Langley et al., 1999). Of these, the tumor thickness (Breslow thickness), the presence of ulceration, and the presence of metastases have been the most useful prognostic indicators (Balch et al., 2001b). The prognosis for thin non-ulcerated melanomas (less than 1 mm in thickness) is uniformly good, with a ~ 90% 15 year survival. In contrast, the prognosis for patients with thick melanomas with ulceration (greater than 4 mm) is poor (30% 10 year survival). The ability to predict outcome of individual patients with intermediate thickness melanomas (1-4 mm) is less accurate with the 15 year survival ranging from 50 - 70% (Balch et al., 2001b). Improved prognostic indicators for intermediate thickness melanomas would be of great clinical value. The number of lymph nodes positive for metastatic melanoma and presence of visceral metastases are important prognostic indicators and are incorporated in the AJCC melanoma staging classification (Balch et al., 2001a).

Metastatic melanoma usually involves draining lymph nodes and occasionally adjacent skin first , but eventually metastasizes to distant visceral sites. The lung is most commonly involved followed by brain, liver, bone marrow, and intestine. A wide variety of other sites are involved less frequently. In about 5% of cases, metatstatic melanoma is detected in the absence of an identifiable primary tumor (Chang et al., 1982). In these cases it is generally believed that the primary tumor has regressed.

Treatment of melanomas initially consists of wide excision of the primary tumor with 1 to 2 cm margins, usually after a diagnostic biopsy has already been performed. For most thin melanomas, wide excision is curative, however for intermediate and thick melanomas, sentinel lymph node biopsy is performed to better predict outcome. This procedure involves injecting a small amount of radioactive tracer at the site of the primary tumor and monitoring the radioactive signal until it accumulates in the first draining lymph node and excising it. Completion lymphadenectomy is frequently performed in cases with a positive sentinel lymph node.

Treatment of metastatic melanoma is difficult. High-dose interferon alpha has been shown to have statistically significant efficacy in stage III melanoma (Kirkwood et al., 1996), however the increase in overall survival does not appear to be significant (Lens, 2002). Melanoma is notoriously resistant to conventional chemotherapy regimens. Radiation therapy is occasionally used for palliation. Interestingly, a small percentage of high stage melanoma patients have spontaneous remissions by unexplained mechanisms (Wang et al., 1999). This observation and the occasional finding of immune-mediated regression of primary melanomas has driven substantial effort to develop a cancer vaccine for melanoma (Jager et al., 2001).

Pathology

Evalution of the histopathology of melanoma biopsies and excisions is of vital importance to determine prognostic indices such as tumor thickness and ulceration, adequacy of margins, whether sentinel lymph node biopsy is indicated, and whether metastatic tumor is present. Most melanoma appears to arise de novo, without an identifiable precursor. Roughly 35 % of melanomas arise in the context of a dysplastic nevus, which shares some clinical and histologic features of melanoma (Rhodes, 1999). Occasionally melanoma arises within a giant congenital nevus or adjacent to a common dermal or compound nevus.

Figure 1. Invasive melanoma arising in a dysplastic nevus. Invasive melanoma (right portion of image, full thickness) is present adjacent to residual dysplastic nevus (left, nested proliferation of melanocytes at dermal/epidermal junction).
Figure 2. Pagetoid spread. Severely atypical melanocytes (with relatively clear cytoplasm) are seen migrating upwards with the epidermis. This finding is called Pagetoid spread because of its histologic resemblance to Paget's disease of the nipple.
The great majority of melanomas have an in situ (epidermal) component, which is typically composed of severely atypical melanocytes that migrate upward into the epidermis with a proliferation of single and nested atypical melanocytes at the dermal/epidermal junction. This component of melanomas is sometimes referred to as the radial growth phase, as the lesions tend to grow laterally (radially) rather than invade the underlying dermis. The vertical growth phase is the portion of the melanoma that invades the dermis and occasionally the subcutis.
Advanced melanoma is frequently amelanotic and the cells can mimic a wide range of poorly differentiated carcinomas, sarcomas, and even some lymphomas. The histopathologic features of melanoma subtypes can vary dramatically.
Figure 3. The invasive component of superficial spreading and acral lentiginous melanomas consist of nests and individual epithelioid to somewhat spindled cells with vesicular nuclei and prominent nucleoli.
Figure 4. Nodular melanoma has an exophytic growth pattern and tends to be composed of epithelioid cells with a high mitotic rate.
Figure 5. Lentigo maligna melanoma arises in the context of a linear proliferation of atypical melanocytes that extends down hair follicles in sun-damaged skin.
Figure 6. Desmoplastic/neurotropic melanoma also can be found underlying a lentiginous confluent proliferation of atypical melanocytes, but is very spindled and usually present in collagenous stroma. This melanoma subtype has a propensity for perineural invasion and recurs locally more frequently than other subtypes.

Biology
The study of familial melanoma has provided several insights into the genetics of both inherited predisposition to melanoma and sporadic melanoma (Pollack and Trent, 2000). The CDKN2A locus on chromosome 9p21 is mutated in a substantial portion (roughly 20-50 %) of famial melanoma kindreds (OMIM#600160) . This complex locus encodes two gene products: p16INK4A and p14ARF. These two gene products are unrelated at the protein level, but both function as tumor suppressors (Chin et al., 1998b). p16INK4A functions in the RB pathway, binding to and reducing the kinase activity of CDK4 or hCDK6/cyclin D complexes. p14ARF functions predominantly in the p53 pathway, binding to MDM2 and preventing it from inactivating p53. Most germline mutations in CDKN2A involve exon 2, which is shared between both gene products in alternative reading frames, or involve exon 1a, which is p16INK4A -specific (Pollack et al., 1996, Fargnoli et al., 1998). Rare germline mutations in exon 1b, which would be predicted to only inactivate p14ARF , have been demonstrated in two cases of patients with familial or multiple primary melanomas (Randerson-Moor et al., 2001, Rizos et al., 2001). Linkage to chromosome 12q13 has been documented in some kindreds and activating mutations have been demonstrated in CDK4 which prevent p16 binding and inactivation of the CDK4/cyclin D complexes (Zuo et al., 1996) (OMIM #123829) .

The importance of ultraviolet radiation as an etiologic factor in skin cancer in emphasized by the high incidence of melanoma, squamous cell carcinoma, and basal cell carcinoma in xeroderma pigmentosum patients (Cleaver and Crowley, 2002). In additiona to xeroderma pigmentosum patients, other inherited disorders have an increased incidence of melanoma. Retinoblastoma patients have a higher risk for developing melanoma than the general population (OMIM #180200) . The Familial The Familial Atypical Mole-Malignant Melanoma (FAMMM)/B-K Mole syndrome is characterized by familial melanoma associated with an increased number of atypical melanocytic nevi. Initially linkage to chromosome 1 was established (Bale et al., 1989), however subsequent studies have not validated this result (van Haeringen et al., 1989, Cannon-Albright et al., 1990).

Activating mutations in ras genes occur in ~25% of cutaneous melanomas (Herlyn and Satyamoorthy, 1996). Mutations in N-ras are most common and usually involve codon 61. H-ras is the second most frequently mutated family member in melanoma and is also mutated and amplified in a subset of Spitz nevi, which are benign melanocytic lesions that histologically resemble melanoma (Bastian et al., 2000a). A recent study has identified a high rate (~75%) of activating mutations in the BRAF serine-threonine kinase in primary melanomas and melanoma cell lines (Davies et al., 2002). These findings suggest that the RAS-RAF-MAPK pathway plays an important role in melanoma biology.

Tyrosine kinases appear to play a role in both the survival and migration of melanocytes as well as in melanoma formation and progression. Inactivating mutations in the c-kit tyrosine kinase or kit ligand result in inherited forms of focal non-pigmentation: piebaldism in humans (OMIM #172800) or mice (White and Steel loci). Mouse models of melanoma have been produced that involve activation of either the Ret or Met tyrosine kinases (Iwamoto et al., 1991, Otsuka et al., 1998).

The chromosomal alterations that occur during the progression of malignant melanoma have been extensively characterized. Karyotyping of melanoma cell lines and comparative genomic hybridization of tumors have demonstrated recurrent loss of chromosome 1p, 6q, 9p, 10q, 11q, and 12q and gains of chromosome 1q, 6p, 7, 5p, 8q, 11q, 17q, and 22q (Bastian et al., 1998) (Pollack and Trent, 2000). Some melanoma subtypes appear to have a higher incidence of alterationas in specific loci. Cyclin D1 (chromosome 11q13) has been shown to be preferentially amplified in acral lentiginous melanoma (Bastian, 2000b) and the neurofibromatosis 1 gene (NF1) undergoes loss of heterozygosity in a relatively high proportion of desmoplastic melanomas (Gutzmer et al., 2000).

Mouse models of melanoma and comparsion with human melanoma

The anatomic location of melanocytes in mice overlaps with that of humans but is somewhat different. Human melanocytes are predominantly located at the junction of epidermis and dermis and also are present within the hair follicle. Murine melanocytes are predominantly associated with hair follicles or are present within the interfollicular dermis, and only rarely are present at the dermal/epidermal junction. Early human melanoma is characterized by upward spread of atypical melanocytes within the epidermis (Pagetoid spread).

Figure 7. Pagetoid spread in human melanoma. Intraepidermal spread (Pagetoid spread is a prominent feature in human melanoma.	Figure 8. Pagetoid spread in a mouse model of melanoma. Prominent intraepidermal spread of melanoma cells in seen in this mouse model of melanoma produced by perinatal UV irradiation of metallothionein-hepatocyte growth factor transgenic mice (Noonan et al., 2001).

Pagetoid spread is uncommon in mouse melanoma models and may be a reflection of the follicular/dermal location of murine melanocytes. Transgenic expression of mitogenic or survival signals by keratinocytes has resulted in increased numbers of melanocytes at the dermal/junction (Kunisada et al., 1998) and combined with perinatal ultraviolet irradiation, has resulted in a mouse model with striking Pagetoid spread (Noonan et al., 2001)

Several mouse models of melanoma studied to date have involved expression of a transgene under the control of the melanocyte-specific tyrosinase promoter, including the SV40 early region (Bradl et al. 1991) and activated alleles of H-ras (Broome Powell et al., 1999, Chin et al, 1997). Alternatively, melanomas have developed in the context of more ubiquitous transgene expression, such as the expression of ret (Iwamoto et al., 1991) or hepatocyte growth factor (Otsuka et al., 1998) under the control of the metallothionine promoter. In one model, histologically benign-appearing melanocytic lesions spontaneously arise with very short latency and high frequency in a transgenic strain of mice (Zhu et al., 1998). These lesions frequently progress to invasive lesions and/or metastasize. This highly penetrant phenotype appears to be integration site-specific and is incompletely understood at present. Murine melanomas have also arisen at increased rates in knockout mice in the context of chemical carcinogenesis protocols, typically involving 7,12-dimethylbenz(a)anthracene (DMBA) as an initiator and 12-O-tetradecanoylphorbol-13-acetate (TPA) as a promoter.

Mice that have the genetic changes observed in human familial melanoma have been constructed and observed for melanoma formation. Knockout mice lacking both p16INK4A and p19ARF (Ink4aD2/3)develop melanomas at a high rate in the setting of tyrosinase-driven activated H-ras expression, but rarely develop melanomas in the absence of a second genetically engineered change. Mice containing a knock-in activating mutation of the CDK4 gene develop melanomas at a high frequency after a two step chemical carcinogenesis protocol (Sotillo et al., 2001).