Biomarkers for Melanoma Diagnosis and the Technologies Used to Identify Them

Melanoma is a malignant tumor originating from melanocytes (pigment-producing cells). Although the tumor is mainly detected in skin (cutaneous melanoma), it can also be detected in the eye (uveal melanoma), gastrointestinal (GI) tract and oral mucosa and genital tract (mucosal melanoma) (Landreville et al., 2008; Akaraviputh et al., 2010; Bakalian et al., 2008; Rigel et al., 2010; Seetharamu et al., 2010). Melanoma can be classified as belonging to one of four subtypes: superficial spreading, nodular, lentigo maligna, and acral lentiginous melanoma. These subtypes are characterized based on prognosis, incidence of metastasis and the frequency of gene mutations (e.g., BRAF and NRAS) (Saldanha et al., 2006; Jaeger et al., 2007; Markovic et al., 2007; Jönsson et al., 2010). Superficial spreading melanoma is the most common form of melanoma found in Caucasian populations, while the acral lentiginous melanoma is frequently detected in Asian and African populations (Cress and Holly, 1997; Weyers et al., 1999). Several important risk factors that have been linked to the development of melanoma have been identified. Of these risk factors, most can be considered to be either environmental factors, such as exposure to ultraviolet (UV) radiation, especially in childhood, or other host factors such as family history and melanocytic nevi (Markovic et al., 2007; Schulman and Fisher, 2009), but other cancer risk factors such as smoking (Osterlind et al., 1988), diet (Osterlind et al., 1988; Veierod et al., 1997) or hormone therapy (Naldi et al., 2005) have not been found to be associated with an increased risk of melanoma. The risk of developing melanoma is higher in Caucasian than in Asian or African populations. This is closely related to skin pigmentation as melanin has been shown to have a protective function for UV-induced melanoma and Caucasian populations show low levels of melanogenesis (Lens and Dawes, 2004; Hu et al., 2008; Jemal et al., 2010). In general, the melanocortin-1 receptor (MC1R), which is a G-protein-coupled receptor (GPCR), stimulates melanogenesis through the activation of adenylate cyclase and protein kinase A (PKA) (Jordan and Jackson, 1998; Rouzaud et al., 2003). Its genetic variants are associated with melanoma


391
Phosphorylation of RAF kinase by RAS activates downstream targets the MAPK extracellular signal-regulated kinase-1 and -2 (MEK1 and MEK2), which causes the phosphorylation of extracellular signal-regulated kinase-1 and -2 (ERK1 and ERK2). Activated ERK1 and 2 has been found to modulate the gene expression necessary for survival and proliferation of melanoma cells, and has been linked to the increase resistance of melanoma cells to apoptosis by inhibiting the activation of caspase 8 (Becker et al., 2006;Sekulic et al., 2008). The BRAF mutation is necessary for ERK-mediated survival and proliferation and participates in the reduction of proapoptotic proteins (e.g., Bcl-2 family), while RAF genes consist of ARAF, BRAF, and CRAF (also known as Raf-1) (Becker et al., 2006;Cartlidge et al., 2008;Sekulic et al., 2008). The BRAF mutation, predominantly V600E (substitution of glutamate to valine; previously known to V599E), frequently occurs in melanoma and is strongly related to exposure to UV radiation (Tsao et al., 2004;Wan et al., 2004;Gray-Schopfer et al., 2005;Becker et al., 2006;Sekulic et al., 2008).  1) These receptors also stimulate melanoma angiogenesis.
2) The β-catenin functions as a cell adhesion-associated molecule. 3) Other studies suggested a decrease in the expression of antiapoptotic Bcl-2 in metastatic melanoma (Fecker et al., 2006;Zhuang et al., 2007). 4) The angiogenesis-associated molecules take part in melanoma metastasis. 5) The chemokine receptors also play an important role in melanoma growth. 6) The E-cadherin also functions as a tumor suppressor-associated molecule. 7) The antigens are found mainly in metastatic melanoma. Table 1. (continues) Tissue biomarkers (immunohistochemical biomarkers) for the diagnosis of melanoma and changes in their expression levels Moreover, phosphorylation of AKT kinase by phosphatidyl inositol 3-kinase (PI3K) stimulates the mammalian target of rapamycin (mTOR), which leads to the survival and proliferation of melanoma cells. The AKT/mTOR pathway suppresses apoptosis by decreasing the levels of pro-apoptotic proteins (e.g., BAD and caspase-9). Among the three AKT members (AKT1, AKT2, and AKT3), it is AKT3 that is often overexpressed in melanoma (Stahl et al., 2004) and is regulated by the phosphatase and tensin homolog (PTEN), which degrades the products of PI3K (Wu et al., 2003;Becker et al., 2006;Sekulic et al., 2008). In addition, many antigens, which have immunostimulatory or activator roles in tumorigenesis, have been identified in metastatic melanoma. Important antigens include the melanocyte lineage/differentiation antigens [e.g., tyrosinase-related protein-1 (TRP1)/gp75, TRP2, Melan-A/MART-1] (Thomson et al., 1985;Winder et al., 1994;Murer et al., 2004;Rad et al., 2004) and cancer/testis antigens (e.g., BAGE family, GAGE family, MAGE family, and NY-ESO-1) (Chen et al., 1998;Simpson et al., 2005;Barrow et al., 2006). Melanocyte lineage/differentiation antigens are associated with the production of melanin pigments and have been identified in both normal melanocytes and melanoma (Thomson et al., 1985;Houghton et al., 1988). Cancer/testis antigens are abundant in normal tissues during development, but in mature cells, their expression is restricted to the male germ cells in the testis and to various tumors (Simpson et al., 2005). As indicated, several gene mutations such as NRAS (Q61K/R), BRAF (V600E), PTEN, and CDKN2A mutation play an important role in the occurrence of melanoma (Tsao et al., 2004;Wan et al., 2004;Gray-Schopfer et al., 2005;Becker et al., 2006;Sekulic et al., 2008). These mutations are excellent targets for the diagnosis of melanoma. Moreover, in the context of melanoma prognosis, melanoma cells show an increase in the expression of survival or proliferation-associated molecules, angiogenesis-associated molecules, and in the expression of antigens, but a decrease in the expression of tumor suppressor-associated proteins (e.g., PTEN) or proapoptotic proteins (e.g., Bax and Bak) is observed (Table 1). Thus, these molecules, which show altered expression levels in melanoma relative to normal cells, are useful tissue biomarkers (immunohistochemical biomarkers) for melanoma diagnosis. However, in spite of these discoveries, these markers are not specific to melanoma and there are few melanoma-specific tissue biomarkers, excluding melanocyte lineage/differentiation antigens and melanoma-associated antigens (e.g., TRP1/gp75 and 2, Melan-A/MART-1, and TA-90) that are overexpressed in metastatic melanoma.

Serologic biomarkers for melanoma diagnosis
Many researchers have identified putative serologic biomarkers for melanoma diagnosis (Table 2), which play a key role in the growth and survival of melanoma cells. Typically, these markers activate survival and/or proliferation-associated and angiogenesis-associated signal transduction pathways after binding with their receptors. The most important receptors of this kind are GFRs (e.g., VEGFR and FGFR) and GPCRs (e.g., MC1R, Wnt/frizzled receptor, and chemokine receptor) (Halaban, 1996;Lee et al., 2008). The primary antigens that have been observed in the serum of melanoma patients are melanocyte lineage/differentiation antigens and melanoma-associated antigens. The existence of these antigens in serum is closely associated with melanoma progression and low survival rates. Thus, these types of markers are useful as prognostic biomarkers and have potential to act as therapeutic targets.
In the absence of vascularization, the growth of melanoma is limited to 0.2 -0.3 cm due to the limited diffusion of oxygen and nutrients into the tumor. For additional growth, angiogenesis is essential for providing adequate blood supply to the growing lesion. Angiogenesis is regulated by proangiogenic factors, such as VEGF, FGF, tumor necrosis factor (TNF), and interleukin-8 (IL-8) and by antiangiogenic factors, such as interferons (IFNs) and angiostatin. An increase in vascular density provides a greater supply of oxygen and nutrients to cells, leading to melanoma growth (Folkman, 2007;Mahabeleshwar and Byzova, 2007). High levels of proangiogenic factors in the serum of melanoma patients can be used as an indicator of melanoma at diagnosis. Multiple cytokines (e.g., 4,6,8,10 and 14), which are correlated with melanoma growth, angiogenesis and metastasis, have been observed in the serum of melanoma patient at both the protein and/or mRNA level. Serum levels of these cytokines are increased in metastatic melanoma patients, suggesting that they can be used as an indicator of melanoma progression (Porter et al., 2001;Varney et al., 2006;Yurkovetsky et al., 2007). Moreover, the serum concentration of the soluble IL-2 receptor is elevated in patients with metastatic melanoma and elevated serum IL-2 receptor levels are associated with lowered survival rates (Boyano et al., 1997;Ottaiano et al., 2006). Interestingly, IFNs are soluble cytokines, but possess antiangiogenic and antitumor activities. An increase in the melanoma progressionassociated cytokines leads to a reduction in IFNs levels, but there is a decrease in the serum levels of melanoma progression-associated cytokines in melanoma patients following immunomodulatory therapy with IFNs (mainly IFN-α2b) (Singh and Varner, 1998;Jonasch and Haluska, 2001;Yurkovetsky et al., 2007;Dummer and Mangana, 2009;Hofmann et al., 2011). Thus, the analysis of a number of different serum cytokines may be a useful means of monitoring the efficacy of immunomodulatory therapy.

Urinary biomarkers for melanoma diagnosis
Urinary biomarkers for melanoma diagnosis have received much greater interest because of the relative ease of sample collection and handling compared with the analysis of blood or tissue samples, but this form of sample may lack the sensitivity required for a diagnostic biomarker. Of the urinary biomarkers of melanoma already identified (Table 3), 5SCD and 6-hydroxy-5-methoxyindole-2-carboxylic acid (6H5MI2C), are intermediate metabolites in melanin pigment formation, and have been the most extensively studied. An increase in urine levels of these markers has been associated with tumor progression and low survival rates (Kärnell et al., 1997;Bánfalvi et al., 2000;Wakamatsu et al., 2002). In healthy patients, the urinary levels of these markers are influenced by age (Meyerhöffer et al., 1998), sex (Morishima and Hanawa, 1981;Kågedal et al., 1992), skin color  and season (Ito et al., 1987), but not by pregnancy (Carstam et al., 1985). Although both 5SCD and 6H5MI2C have been detected in the urine of melanoma patients, because of the higher levels of 5SCD, this marker is considered a more reliable urinary biomarker for melanoma than the 6H5MI2C (Kärnell et al., 1997(Kärnell et al., , 2000Wakamatsu et al., 2006). Moreover, the 90-kDa glycoprotein (TA-90) (Rote et al., 1980;Euhus et al., 1989), S100A7 (Brouard et al., 2002) and β-human chorionic gonadotropin (Carter et al., 1995) have also been identified in the urine of patients with melanoma (Table 3). Of these urinary biomarkers, 5SCD, 6H5MI2C, and S100A7 can be considered the most melanoma-specific of the urinary biomarkers (Kärnell et al., 1997;Bánfalvi et al., 2000;Brouard et al., 2002;Wakamatsu et al., 2002).

Biomarkers Functions
Proangiogenic factors VEGF 1) VEGF receptor (VEGFR) ligand FGF 1) FGF receptor (FGFR) ligand EGF 1) EGF receptor (EGFR) ligand Placental growth factor (PIGF) 1) Neuropilin-1 and -2 receptor ligand TNF 1) GPCR ligand IL-8 (CXCL8) 1) GPCR (specially, chemokine receptor CXCR1 and 2) ligand Laminin- 5 2) Laminin receptor (e.g., integrin α6β4 and α7β1) ligand Osteopontin 2) Integrin αvβ3 ligand uPA 2) uPAR ligand Antigens 3) Melanocyte lineage/Differentiation antigens Tyrosinase Regulator enzyme in melanin synthesis; Increased in metastatic prognosis gp100/pmel-17 Melanin synthesis-associated melanosomal matrix glycoprotein; Increased in metastatic prognosis S100 proteins (e.g., S100B) Cell division and differentiation-associated acidic calcium-binding protein; Increased in metastatic prognosis MIA A small soluble protein; Increased in metastatic prognosis L-dopa/L-tyrosine ratio An index of tyrosinase functional activity; Increased in metastatic prognosis Melanoma-associated antigens TA-90 Potential immunostimulator or antineoplastic activator; Increased in metastatic prognosis Survivin (one of apoptosis inhibitors gene family) Apoptosis inhibition; Increased in metastatic prognosis Cytoplasmic/high-molecular-weight melanoma-associated antigen (CYT-MAA/HMW-MAA) Unknown exactly, but may relate to melanoma progression 1) These proangiogenic factors also function as an important stimulator for melanoma growth. 2) These proangiogenic factors also take part in melanoma metastasis. A β-galactoside-binding protein; Increased in metastatic prognosis Synovial sarcoma X breakpoint-2 (SSX-2) A family of highly homologous synovial sarcoma X (SSX) breakpoint proteins and repressive gene regulator Gangliosides (GM2, GD2, GM3, and GD3) Group of glycosphingolipids; Relate to interactions between melanoma cells Cytokines and cytokine receptors 4) IL-1 Survival or proliferation-associated factor IL-4 Survival or proliferation-associated factor IL-6 Survival or proliferation-associated factor IL-10 Survival or proliferation-associated factor IL-12 Survival or proliferation-associated factor Soluble IL-2 receptor Survival or proliferation-associated factor Other serologic biomarkers YKL-40 Unknown exactly, but may function as a survival or proliferation-associated factor C reactive protein (CRP) Unknown exactly, but may relate to tumorassociated inflammatory response Lactate dehydrogenase (LDH) An indicator for liver metastasis; a prognostic indicator in metastatic melanoma Glypican-3 (GPC3) Unknown exactly, but may function as a survival or proliferation-associated factor PKCα A survival or proliferation-associated protein 5SCD A precursor of melanin; Increased in metastatic prognosis 6H5MI2C A precursor of melanin; Increased in metastatic prognosis Serum amyloid A (SAA) A superfamily of acute-phase proteins and proinflammatory adipokine Cystatin C A potent inhibitor of cysteine proteases; Increased primary and metastatic melanoma 3) Antigens that are found in melanoma tissues become good immunohistochemical biomarkers. 4) The cytokines also take part in melanoma metastasis.

Biomarkers for early melanoma diagnosis
The early diagnosis of melanoma is closely related to an increase in survival rate. Although many prognostic biomarkers (mainly metastatic prognosis biomarkers) of melanoma have been reported, there are very few capable of allowing an early diagnosis. Glypican-3 (GPC3) is a membrane-bound heparin sulfate proteoglycan which is overexpressed in several tumors. It has been suggested that GPC3 may be a useful early stage biomarker for patients with the early stages of the disease (0 -II) (Nakatsura et al., 2004;Ikuta et al., 2005). Moreover, cyclooxygenase-2 (COX-2) (Chwirot and Kuźbicki, 2007), serum amyloid A (SAA) (Mian et al., 2005;Findeisen et al., 2009) and DNA methylation profiling (Conway et al., 2011) can be used to distinguish between early melanomas and benign nevi.

Screening techniques of melanoma biomarkers
Before the development of high throughput proteomics techniques, biomarker candidates were identified based on known melanoma molecular pathways and validated by traditional techniques such as western blotting, ELISA and immunohistochemical analysis. However, the recent development of proteomics has enabled novel biomarkers to be screened from across a much larger section of the proteome. The most widely used technique for this form of screening is one whereby the samples are first separated by 2-DE and then each protein is identified by mass spectrometry (MS). Recently, simple gel-free techniques such as shotgun proteomics (Liu et al., 2002) and surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) (Petricoin et al., 2002) have been developed. These techniques employ simple separation procedures such as capillary chromatography and surface chromatography prior MS analysis.

Screening for tumor tissues and cell lines
Biomarkers discovered from tissue samples and cultured cell lines also have utility for the development of diagnostic and prognostic assays. Tissue microarray is a high throughput technique in which many tissue samples can be screened simultaneously. This technique is suitable for the validation of candidate biomarkers, which are first obtained by other proteomics techniques. Several biomarkers of melanoma such as HSP 90 (McCarthy et al., 2008), ING3 (Wang et al., 2007), and the epidermal growth factor receptor family member HER3 (Reschke et al., 2008) were validated using this technique. For comprehensive screening of biomarkers from lysates collected from tumor tissue and cultured cells, 2-DE combining MS is used as a standard method. The representative attempts of biomarker screening of melanoma lysates are summarized in Table 4. Since tumor tissue is heterogeneous mixture of several cell types, cultured cell lines may be preferable for the screening of biomarkers. To search for biomarkers useful for the diagnosis of metastases, primary and metastatic cell lines were compared and several candidate biomarkers were successfully discovered (Table 4) (Bernard et al., 2003;Zuidervaart et al., 2006;Al-Ghoul et al., 2008). Proteomics analysis of melanoma-associated fibroblast stromal cells revealed the aberrant expression of several proteins not detected in normal fibroblasts (Paulitschke et al., 2009). These proteins may promote tumor progression. However, the passages number of the melanoma cell lines tested was found to produce changes in the proteome, which may underscore the invasive character observed in melanoma cells line that have been passaged many times (Pardo et al., 2006).

Screening biomarkers obtained from serum and secreted from cultured cell lines
Because serum samples are far less invasive to obtain than biopsied material, the discovery of serological biomarkers has received a great deal of attention. Well-defined biomarkers enable early detection, allow the appropriate classification of tumor types (which provides the clinician insight into the best choice of therapy), and enable the patient to be more thoroughly monitored for progression and regression. However, there are several difficulties with screening serological biomarkers: (1) the presence of abundant blood proteins, which may inhibit the detection of biomarkers, and (2) the low serum concentration of the biomarker after it is secreted from tumor tissue and diluted in the bloodstream (Simpson et al., 2008). Hence, processes by which the abundant blood proteins are removed from the sample or the target proteins are concentrated are essential to address these problems. As a result of these difficulties inherent in serum proteomics, the secretome has received much attention recently. Table 5 summarizes a number of different reports in which comprehensive screening of biomarkers from serum and the secretome occurred. By comparing the secretome between melanoma and normal melanocyte, several potential biomarkers were successfully discovered (Pardo et al., 2007;Paulitschke et al., 2009). The traditional serological biomarkers of melanoma such as LDH, S100B and CRP lack sensitivity as early stage melanoma biomarkers. To search for early stage biomarkers, Findeisen et al. extensively analyzed the serum proteome of about 600 melanoma patients at each stage of the disease (stages I -IV) by SELDI-TOF-MS technique (Mian et al., 2005;Findeisen et al., 2009). This analysis led to the discovery of a new biomarker, SAA, which was found to be highly sensitive for detecting early stage melanoma.

Summary and overall conclusions
Melanoma biomarkers have the capacity not only to diagnose melanoma, but also to allow patients to be monitored for recurrence after surgical resection and to allow the effect of anticancer drug treatments to be evaluated. Advanced technologies (e.g., high throughput technologies in genomics or proteomics) have contributed much to the hunt for melanoma biomarkers in tissue samples or body fluids and have typically been MS-based or arraybased technologies. Several immunohistochemical, serologic and urinary biomarkers have been reported to be very useful diagnostic and prognostic biomarkers. However, there is a paucity of data on melanoma-specific biomarkers, with the exception of the melanocyte lineage/differentiation antigens and melanoma-associated antigens, and some urinary biomarkers such as 5SCD, 6H5MI2C and S100A7. Moreover, several melanoma biomarkers with prognostic capabilities, mainly for the detection of metastatic disease, have been applied to clinical use. However, there is s still a great need to identify melanoma early-stage melanoma biomarkers, as early detection of the