Main

Malignant melanoma is the most aggressive form of skin cancer and has a multifactorial etiology. It is assumed that the disease progresses toward aggressive metastatic disease via the accumulation of nonrandom multiple genetic aberrations.1

Earlier, using comparative genomic hybridization (CGH) analysis, we and others found frequent amplification of the 11q13–q21 chromosome band in primary melanomas.2, 3, 4 The Cyclin D1 (CCND1) gene has been widely considered to be a target gene in the region because its overexpression is commonly observed in several human cancers, including breast, head and neck, and bladder cancers.5, 6, 7 Beside the CCND1, several other oncogenes and/or cancer-related genes reside within this amplicon core, such as TAOS1, FGF3, FGF4, FGF19 and EMS1.8, 9, 10, 11, 12, 13

CCND1 is the regulator subunit of cyclin-dependent kinase 4 (CDK4) and cyclin-dependent kinase 6 (CDK6). These enzymes are responsible for phosphorylating the retinoblastoma (Rb) protein and thus promoting cell entry into mitosis. Transcription of CCND1 is stimulated by the MAP kinase pathway, and this is the connection point between the RAS-MAPK and p16-CCND1/CDK4-Rb pathways.14, 15 Elevated CCND1 gene copies were found in a large series of acral melanoma subtypes, occasional amplification was described in lentigo maligna melanomas and superficial spreading melanomas (SSM), and only sporadic amplification in the nodular melanoma subtype (NM).16, 17

The contribution of other genes in the 11q13 amplicon to several cancers has been investigated.8, 9, 10, 11, 12, 13 The TAOS1 gene, which is located approximately 12 kb distal to CCND1, was found to be a novel marker for advanced esophageal squamous cell carcinoma and was overexpressed only if its gene amplification was present in the oral squamous cell carcinoma cell lines.9, 13 Amplification of the EMS1 gene through the overexpression of cortactin protein has been shown to contribute to the invasive potential of tumor cells.11, 18 This protein is described as an actin-associated scaffolding protein that regulates the formation of actin-based structures closely associated with cell motility.19 Overexpression of EMS1 causes enhanced cell proliferation, predicts early recurrence and reduced survival in squamous cell carcinoma of the head and neck and in esophageal adenocarcinoma.8, 20, 21 CCND1 and EMS1 are also thought to be likely to play pathogenic roles in the 11q13 amplicon in bladder tumors.22 Coamplification of CCND1, EMS1, FGF3 and FGF4 genes was significantly associated with increased tumor stage and grade in a large series of urinary bladder cancer.12 However, the possible role of these candidate genes within the 11q13 amplicon has not yet been investigated in primary melanoma.

Activating mutations in codon 600 of BRAF and in codon 61 of NRAS are, so far, the most common single mutations detected in human cutaneous melanoma.23, 24 Drugs targeting this pathway are, therefore, the most attractive clinical agents for the disease.25 A subset of BRAF mutant melanomas are sensitive to these agents, examples of which are SB590885 (GlaxoSmithKline, Collegeville, PA, USA) and PLX-4032/PLX-4720 (Plexxikon, Berkley, CA, USA).26, 27, 28 Identifying concomitant genetic alterations may help predict the response or resistance to BRAF and NRAS inhibitors.

One of the objectives of our study was to investigate gene copy number alterations in the 11q13 amplicon in primary and metastatic melanomas by performing Q-PCR analysis. First we focused on the detection of CCND1 copy number alterations. All samples with amplified CCND1 were further examined for copy number alteration of TAOS1, FGF3, FGF4, FGF19 and EMS1. Although it is well documented that there is a high prevalence of BRAF (V600) and NRAS (Q61) mutations in melanoma, there are a few studies with discordant findings. These studies examined CCND1 copy number alterations in conjunction with BRAF and NRAS mutational status of the disease. Another aim of our study was to elucidate the possible relationship between these oncogenes and their combined incidence and clinicopathological parameters. Furthermore, we compared CCND1 copy number alterations as classified by real-time quantitative PCR (Q-PCR) with those classified by interphase fluorescence in situ hybridization (FISH) at a single cell level.

Materials and methods

Tumor Samples and DNA Isolation

Tumors were collected in the Department of Dermatology, Medical and Health Science Center, University of Debrecen, Hungary. The study was approved by the Regional and Institutional Ethics Committee, Medical and Health Science Center, University of Debrecen, and conducted according to regulations. Written informed consent was always obtained from the patients. Tumor diagnosis was carried out on formalin-fixed paraffin-embedded tissue sections using hematoxylin and eosin staining. Melanoma tumor staging was determined according to the new TNM staging system.29 The distinction between tumor groups arising from chronically sun-exposed site or intermittently sun-exposed site was based on the presence or absence of solar elastosis of the dermis surrounding the melanomas.

Clinicopathological data from the 68 primary tumors are summarized in Table 1. DNA from peripheral blood cells of healthy individuals and DNA from a melanoma cell line (M24) were also included into the analysis. Tumor DNA was extracted using the G-spin Genomic DNA Extraction Kit following the instructions of the manufacturer (Macherey-Nagel, Düren, Germany). NucleoSpin Extract II kit (Macherey-Nagel) was used to purify DNA if needed. Only high quality DNA template was accepted, with an absorbance ratio range of 1.8–1.9 (260–280 nm) or 1.7–2.5 (260–230 nm), as measured by the NanoDrop 1000 instrument (NanoDrop Technologies, Wilmington, De, USA). Control DNA was extracted from peripheral blood cells using the Nucleo Spin Blood mini kit according to the manufacturer's instructions (Macherey-Nagel). All DNA samples were stored at −20°C.

Table 1 Associations of CCND1 amplification, CCND1/TAOS1 coamplification, BRAF and NRAS mutation with clinicopathological parameters of patients
Full size table

FISH Analysis of CCND1

FISH was performed on tumor imprint preparations using the LSI Cyclin D1 (11q13) SpectrumOrange/CEP 11 SpectrumGreen Probe (Vysis, Downers Grove, USA) as described previously.30 This is a mixture of two probes: the CCND1 probe is approximately 300 kb long, containing the CCND1 gene, and the second probe is specific for the D11Z1 alpha satellite centromeric repeat of chromosome 11. CCND1 copy number alterations were determined with both FISH and Q-PCR methods in 35 melanoma samples.

Primer Design

The primers were designed for the genes GNS, UBE2E1, FGF3, FGF4, FGF19, EMS1 and TAOS1 using Primer Express 2.0.0 software (Applied Biosystems, Foster City, USA) and Primer3 (Whitehead Institute, Cambridge, USA; http://biotools.idtdna.com/primerquest/; Table 2). To avoid the secondary structures, we used the web-based MFOLD version 3.2 software (http://www.bioinfo.rpi.edu/applications/mfold/).31 The sequences of the primers for the CCND1 gene were downloaded from the RTPrimerDB database (http://medgen.ugent.be/rtprimerdb/).32 Primers were purchased from Biocenter Kft. (Szeged, Hungary).

Table 2 Sequences of oligonucleotides used for this study
Full size table

Real-Time Quantitative PCR Analysis

Quantification of gene copy number was performed on 68 primary melanoma tissues and 6 melanoma metastases using Q-PCR. We aimed to quantify the relative amounts of the target genes, all located in the 11q13 amplicon core, using two reference genes for normalization, GNS (12q14.3) and UBE2E1 (3p24.2). These reference genes have not yet been reported as having genetic abnormalities in melanomas analyzed by array CGH. To screen the copy number changes, two reference genes are needed to produce a robust, reliable and accurate quantification.33, 34, 35, 36 Reactions were carried out using an ABI Prism 7000 sequence detector (Applied Biosystems).

The amplification mixtures (25 μl) contained 1 μl template DNA (∼10 ng/μl), 12.5 μl Power SYBR-Green PCR Master Mix (Applied Biosystems) and 100 nM of each primer for TAOS1, FGF3, FGF19, UBE2E1 and FGF4, 200 nM of the primer for CCND1 and 300 nM of the primers for EMS1 and GNS. The reactions were performed under the following conditions: 10 min of polymerase activation at 95°C then 40 cycles at 95°C for 15 s and 60°C for 1 min. Assay for each gene included: (1) no template control (in duplicate), (2) 10 ng of calibrator DNA (Applied Biosystems; in triplicate) and (3) approximately 10 ng of tumor DNA (in triplicate). A melting curve analysis was run after the amplification was completed and consisted of a 20 min slow ramp from 60 to 92°C using an ABI PRISM 7000 Sequence Detection System (Applied Biosystems). The derivative melting curves showed single melting peaks, which confirmed the high Q-PCR specificity (absence of primer dimers and other nonspecific products).

Quantification and Data Analysis

Quantification was performed using the Pfaffl method.37 This requires the efficiency to be known, which is determined from standard curves. To evaluate more accurately the PCR reaction efficiency, five-point standard curves from three different fourfold dilutions (ranging from 80.0 to 0.3 ng; in triplicate) were set up using melanoma cell line DNA (M24) for the different target genes and control DNA for the reference genes. To calculate the PCR amplification efficiency (E=10−1/slope−1) we calculated the slope value from the three separated standard curves for the target genes and reference genes as well. Only standard curves R2 value >0.99 were accepted (Table 3). Instead of interpolating unknown samples from a standard curve, we calculated the relative copy number (Ratio) solely based on the observed CT values (equation (1))

Etarget gene is the efficiency of the PCR reaction for the target gene, Ereference gene is the efficiency of the PCR reaction for the reference gene, ΔCT target gene is the difference in threshold cycle value between the test sample and calibrator sample for the target gene, ΔCT reference gene is the difference in threshold cycle value between test sample and calibrator sample for reference gene.

Table 3 The calculated PCR amplification efficiencies
Full size table

The equation was applied to calculate the standard error (s.e.) of the relative copy number of the normalized target gene, as described Hoebeeck et al.35 To determine whether the copy numbers of the investigated genes were significantly different from those of the controls, we determined a tolerance interval (TI) for the relative gene copy number, using the mean standard deviation (s.d.) of the ΔCT values for the target and reference genes in 12 healthy individuals according to the equation TI=2±2 × s.d.ΔCT, as described elsewhere.38 The TI range was 1.28–2.72 for CCND1, 1.42–2.58 for TAOS1, 1.18–2.82 for FGF3, 1.10–2.90 for FGF4, 1.13–2.87 for FGF19 and 1.04–2.96 for EMS1. If the calculated copy number±s.e. exceeded the calculated upper limit, the gene(s) in the tumors were considered to be amplified.

Detection of BRAF and NRAS Mutations

Analysis of mutations in the BRAF codon 600 and in the NRAS codon 61 was performed on LightCycler real time PCR System (Roche Diagnostics, GmbH, Mannheim, Germany) by melting curve analysis using fluorescent probes. All primers and probes were purchased from TIB Molbiol (Berlin, Germany; Table 2). Amplification was performed in glass capillaries using 50 ng of sample DNA in a 10 μl volume containing 1 μl 10 × LightCycler FastStart DNA Master HybProbe (Roche Diagnostics), 0.8 μl 25 mM MgCl2, 0.75 μl DMSO, 1 μl (5 μM) forward and reverse primer and 1 μl (2 μM) of the anchor and sensor hybridization probe.

The reaction was performed under the following condition: initial denaturation at 95°C for 10 min and 45 cycles of amplification consisting of denaturation at 95°C for 0 s, annealing at 52°C for 10 s and amplification at 72°C for 20 s. Melting curve analysis was done as follows: PCR products were denatured for 1 min at 95°C then cooled down to 40°C for 1 min and warmed up to 72°C (ramping at 0.1°C/s) with continuous fluorescence detection before by a final cooling step at 40°C for 30 s. The accuracy of the method was confirmed by direct sequencing (BIOMI Kft., Gödöllő, Hungary) of PCR products that showed deviation from the wild-type (WT) genomic DNA melting peak (Figure 1).

Figure 1
figure 1

LightCycler melting curves and sequence traces of different BRAF types. (a) Typical results of melting curve analysis, (b) BRAF WT sequence (codon 600 is underlined) (c) V600R mutation (GT1798-99AG double base pair substitution), (d) V600E mutation (T1799A base pair substitution), (e) V600K mutation (GT1798-99AA double base pair substitution).

Full size image

Statistical Analysis

Statistical tests were performed using STATA version 9.0 (StataCorp LP, TX, USA). Categorical variables were compared by Fisher's exact test and continuous variables by the Mann–Whitney U-test. Correlations between the results of Q-PCR and FISH analysis were defined by calculating the Spearman's rank correlation (rs). Results were considered significant if P<0.05.

Results

Tumor Samples

A total of 74 fresh or frozen tissue samples were obtained from 70 patients (34 men and 36 women), of which 68 were primary melanomas and 6 were melanoma metastases, including 4 primary tumor-metastasis pairs. The number of patients with at least 3 years of follow-up was 57. If the follow-up period was less than 3 years, clinical data for metastasis formation were not included in Table 1, which summarizes the clinicopathological data.

Confirmation of Reference Genes Suitable for Normalization

To check the validity of the genes used as reference genes (GNS/UBE2E1), we determined the copy number ratio of the genes in 14 normal control DNA samples and 74 tumor samples. The ratio for normal and tumor DNA were 0.92 (s.d.=0.072) and 0.96 (s.d.=0.205), respectively. The copy number ratios of the two reference genes in the normal and tumor DNA were similar (P=0.7236, Mann–Whitney test).

Q-PCR Intra- and Interassay Variability

To determine the intra- and interassay variability of the Q-PCR assay, the mean coefficient of variance (CV) and mean standard deviation (s.d.) were determined for each gene, calculated from samples run in triplicate using the transformation of CT to the linear 2 - C T value. Overall, the intraassay variability was 8% (s.d.=0.11) and the interassay variability was 25% (s.d.=0.38; Table 4). Samples were retested if the CV was higher than 20%.

Table 4 Coefficient of variance (CV) and standard deviation (s.d.) for triplicate readings for intra- and interassay variability using 2 - C T values
Full size table

CCND1 Amplification in Melanoma Samples Detected by Q-PCR

A total of 68 primary and 6 metastatic lesions were tested to identify CCND1 amplification by Q-PCR relative to the two reference genes (GNS and UBE2E1). Using this technique, we found CCDN1 amplification in 22 primary melanomas (32%) and 1 metastasis (Table 1). We were able to compare the CCND1 gene amplification pattern in the primary and its metastatic tumor in four cases: the CCND1 gene was amplified in all the primary lesions (case numbers 5, 22, 8 and 13; Table 5) but we could detect CCND1 amplification only one metastasis (data not shown ).

Table 5 Coamplification of TAOS1, FGF3, FGF19, FGF4 and EMS1 with CCND1 detected by Q-PCR and BRAF V600, NRAS Q61 mutations in primary melanomas
Full size table

Comparing Gene Copy Number for CCND1: Q-PCR versus FISH

When we compared the CCND1 gene copy number alterations obtained by Q-PCR with those detected by FISH, we found a good agreement between the two methods (Figure 2). The average CCND1 copy number varied from 1.9 to 15.0 by FISH and from 1.8 to 13.4 by Q-PCR. We were able to detect amplification by Q-PCR in 8 out of 10 cases in which the frequency of amplified cells was more than 30% (copy number ≥5/cell) and in all cases in which the gene/centromere ratio was above 2 (Table 6). We did not observe amplification by Q-PCR if the disomic cell population was above >50% (data not shown). The Spearman's rank correlation coefficient showed strong correlations between FISH and Q-PCR results (r=0.6, P<0.001). In sample 7, a high-level amplification was seen by FISH, but the average copy number of the CCND1 gene was only 3.2 by Q-PCR. The reason for this discrepancy may be that the touch preparation and the DNA were from slightly different parts of the tumor or a high copy number heterogeneity existed within the sample.

Figure 2
figure 2

Comparison of copy number alterations of the CCND1 gene detected by FISH and Q-PCR analysis. CCND1 gene specific probe was labeled with Spectrum Orange (appearing as red signals) and chromosome 11 centromeric probe was labeled with Spectrum Green (appearing as green signals), cell nuclei were labeled with blue fluorescent dye (DAPI). (a) Image represents a high-level amplification for the CCND1 gene whereas (b) image shows tetrasomy for both centromeric and gene regions.

Full size image
Table 6 Comparison of copy number alterations of the CCND1 gene obtained by FISH and Q-PCR analysis
Full size table

Evaluation of TAOS1, FGF3, FGF4, FGF19 and EMS1 Gene Copy Numbers by Q-PCR

Samples with CCND1 amplification (22 primary melanomas) were further characterized for the TAOS1, FGF3, FGF4, FGF19 and EMS1 gene copy number alterations by Q-PCR. Among these 22 samples, 18% (4/22) showed amplification for all of these additional five genes (Table 5). In 46% of tumors (10/22) CCND1 was coamplified with TAOS1; 36% (8/22) of CCND1-amplified tumors also displayed amplification of FGF3. Coamplification of CCND1 and EMS1 was observed in 7 lesions (32%). The nine primary melanoma samples which exhibited higher CCND1 gene copy numbers (range 4.2–13.4) tended to display TAOS1 (7/9), FGF3 (6/9), FGF19 (6/9), FGF4 (6/9) and EMS1 (5/9) amplifications as well (see details in Table 5).

BRAF and NRAS Mutation Frequencies

BRAF mutations at codon 600 were found in 25 out of 68 primary melanomas (37%) and in 2 out of 6 metastases. Eighteen (27%) primary and two metastatic lesions had BRAF exon 15, codon V600E mutations. This mutation is a single-base T → A transition (T1799A), resulting in a valine to glutamine change and causing substitution with a negatively charged residue. Seven primary tumors (10%) with V600 mutations were double base pair tandem substitutions; five (7%) had a V600K (GT1798-99AA) mutation, causing a valine to lysine change in the amino acid; two (3%) had a V600R tandem mutation (GT1798-99AG), causing a valine to arginine change, both of which mutations resulted in a substitution of valine for a positively charged amino acid (Figure 1).

NRAS mutations at codon 61 were found in 13 out of 68 primary melanomas (19%) and in 2 out of 6 metastases. The two most frequently observed mutations were a CAA to AAA transversion, which occurred in seven primary lesions (10%) and two metastases (resulting in a Q61K change) and a CAA to CGA transition, which occurred in five primary tumors (7%; resulting in a Q61R change). A CAA to TTA tandem mutation was observed in one lesion (2%; resulting in a Q61L change). Furthermore, 56% of primary melanomas and 3 metastases had either BRAF or NRAS mutations, but both mutations were never simultaneously present. CCND1 was amplified in 34% of BRAF or NRAS mutated primary melanoma samples (Table 5).

In four patients, both primary and metastatic tumors were analyzed. Mutations in the primary tumors (one carrying a V600E mutation, two harboring a Q61K mutation) also occurred in the corresponding metastatic lesions (data not shown).

Correlation of Gene Alterations with Clinicopathological Parameters

We did not find correlations between CCND1 gene amplification status and any of the patients’ clinicopathological parameters.

Tumors with CCND1 and TAOS1 coamplification were classified with higher Clark's level (IV–V). Ulcerated tumors had a statistically significant association with CCND1 and TAOS1 coamplification (P=0.017; Table 1). In addition, coamplification of CCND1 with TAOS1 (P<0.001), FGF3 (P=0.001), FGF19 (P<0.001), FGF4 (P=0.002) and EMS1 (P=0.004) genes, respectively, were more frequently found in thick (≥9 mm Breslow thickness) melanomas. Younger age at diagnosis was significantly associated with coamplification of these five oncogenes (CCND1, TAOS1, FGF3, FGF19 and FGF4; P=0.026, Mann–Whitney test; data not shown).

NRAS codon 61 mutations were significantly more frequent in tumors originating from chronically sun-exposed sites (P=0.005; Table 1). Other clinical parameters did not show significant association with the prevalence of BRAF or NRAS mutations. However, an increased CCND1 gene copy number in conjunction with either BRAF or NRAS activation mutations was significantly more common in primary tumors with ulcerated surfaces (P=0.028; data not shown).

Discussion

We have elaborated a relative Q-PCR assay to assess the copy number profile of candidate genes located within the 11q13 chromosomal segment in primary malignant melanomas. This Q-PCR assay resulted in accurate measurement of DNA copy number alterations and can be used as an independent, sensitive method in parallel with other techniques such as Southern blot and FISH. The advantages of the technique are low cost and a requirement for only a very small amount of DNA, meaning that early-stage lesions of small size can be also analyzed. In addition, this PCR assay can be designed and validated for any loci with a known sequence in the human genome within a short period of time.

We had to consider a number of important parameters during the design of the Q-PCR assays for the quantification of the target genes within the 11q13 amplicon core. For normalization and to increase the reliability of the assay, we chose two reference genes (GNS on 12q14.3 and UBE2E1 on 3p24.2), which are not thought to have copy number alterations in melanoma. The validity and reliability of both genes as reference genes were revealed through the detection of the copy number ratio in normal DNA and melanoma tumor samples as described earlier by De Preter et al.33 The gene copy number ratio of the tumor samples is frequently calculated using the method described by Pfaffl et al.37 When using this method, the relative gene copy number ratio is calculated only from the Q-PCR efficiencies and the crossing-point deviation of an unknown sample versus a control sample; it needs no calibration curve, as control levels are included in the model. However, to evaluate the reproducibility of the Q-PCR assays we calculated the coefficient of variation (CV) using the transformation of CT to the linear 2 - C T value (calculating CV based on raw CT values produces incorrect results because of its logarithmic nature39). Using our approach, the overall intra- and interassay variability of the experiments was 8 and 25%, respectively. These values are in good agreement with those reported by others.40, 41, 42 The calculated low CVs were critical for the reliability of the designed Q-PCR assay and were the result of the accurate PCR reaction setup and the high quality of the DNA templates used during the entire experiment.

We were able to compare FISH and Q-PCR data for the CCND1 gene in case of 35 primary melanomas and found good concordance. Q-PCR was sensitive enough to detect gene amplification if the copy number was ≥5 in more than 30% of tumor cells. The rare discrepancies between Q-PCR and FISH were probably resulted from higher normal cell contamination or could be explained by genetically heterogeneous tumor cell clones in the analyzed samples.33

In addition, we evaluated the CCND1 gene copy number in all melanoma samples. Thirty-two percent of the primary lesions exhibited amplification. In contrast to a previous study,16 but in confirmation of a recently published data,17 we found that CCND1 amplification was similarly frequent in SSM and NM. We did not observe any associations between CCND1 gene alterations and clinicopathological parameters. Comparing the CCND1 gene copy numbers in four primary and metastatic tumor pairs, we found gene amplification in all primary tumors but only one metastasis.

Recently it was found that CCND1 is coamplified with other genes located within the 11q13 region in several malignancies, and it is likely that these coamplifications play a pathogenic role in those cancers.12, 13, 22 However, no data are available for melanoma. To define the coamplification pattern of genes in the 11q13 amplicon, we designed Q-PCR assays for the TAOS1, FGF3, FGF19, FGF4 and EMS1 genes. The coamplification of CCND1 with TAOS1 was the most frequent event and had a significant association with the presence of ulceration (P=0.017), a clinical feature that can predict poor prognosis. TAOS1 has been described as a possibly important gene that might drive the 11q13 amplification in oral squamous cell carcinoma and associated with poor prognosis.10, 13 RNA interference method predicts that TAOS1 would participate in cell-cycle control and regulate cell proliferation, similarly to CCND1.10 We have shown that coamplifications of each of these genes with CCND1 are characteristic for thick melanomas and are present in patients of younger age. Based on our data we assume that coamplification of these candidate genes can contribute to a more aggressive phenotype than CCND1 amplification alone. It is also possible that the coamplifications are merely a result of genetic instability that increases during tumor progression but do not contribute directly to the phenotypic alterations.43

The BRAF V600E mutation frequency in our samples was 26%, which is in concordance with other studies44, 45 but the occurrence of this mutation was lower than it has been reported elsewhere.4, 46 We detected two other BRAF mutations that were distinct from the V600E mutation. In each case, subsequent sequencing revealed the presence of V600K or V600R tandem mutations with rates of 7 and 3%, respectively. The two most frequently observed NRAS mutations were Q61K and Q61R, with rates of 10 and 7%, respectively, and these were associated exclusively with tumors derived from chronically sun-exposed sites, which agrees with past studies.46, 47 Metastatic lesions harbored the same mutations as the primary tumor from which they originated, which supports the idea that these mutations are preserved throughout melanoma progression.48

A recent study analyzing 126 melanomas found that lesions with increased CCND1 expression level had either mutation in BRAF or NRAS or increased copy number of the CCND1 gene, without exception. The study reported a strong inverse correlation between BRAF mutations and increased CCND1 copies. This suggests that an increased level of CCND1 protein (the downstream component of the RAS/BRAF/MAP kinase pathway), as a result of either mutations in upstream genes or increased gene dosage, represents a crucial event driving melanoma progression.4 Evaluating the relationship between increased CCND1 gene dosage and the frequency of BRAF and NRAS mutations, we found that 34% of the primary melanomas harboring one of these activating mutations also had CCND1 alterations. Among these tumors, melanomas with ulcerated surfaces were significantly more frequently present (P=0.028), indicating the clinical relevance of this finding. This association is supported by a recent study in which CCND1 was found to be amplified and overexpressed in 17% of BRAF V600E-mutated human metastatic melanomas and contributed to increased BRAF inhibitor resistance in melanomas.49

In conclusion, Q-PCR is a fast, reliable and accurate method for detecting amplification present in the 11q13 amplicon core in malignant melanomas. We found that neither the increased CCND1 gene dosage nor the BRAF or the NRAS mutations alone contributed to more aggressive phenotype. However, we assume that coamplification of these candidate genes in the 11q13 region or carrying a CCND1 alteration along with either the activating BRAF or NRAS mutation may be more important for prognosis in subgroups of aggressive melanomas than the presence of these alterations alone.