Introduction

BRAF gene encodes a serine/threonine kinase regulated by binding to RAS protein. BRAF acts in the MAPK pathway by transducing regulatory signals from RAS to MEK1/2. Davies et al. (2002) found BRAF mutations in 59% of melanoma cell cultures, 80% short-term cultures and 66% of uncultured melanomas. Other studies confirmed a high frequency of mutations, ranging from 40 to 62%, in melanoma cell lines, primary and metastatic tissues with the notable exception of early melanomas (Brose et al., 2002; Dong et al., 2003; Gorden et al., 2003; Pollock et al., 2003). Since a single substitution (V599E) account for 80% of these mutations, it represents a genetic trait for about 50% of melanomas. Mutated RAS oncogenes have been shown in about 20% of melanomas. Mutations of NRAS are more frequently detected than those of HRAS and KRAS, although multiple RAS mutations have also been reported; high frequency has been observed in relation to sunlight exposure, in the nodular subtype and in familial cases (Hayward, 1999; Eskandarpour et al., 2003).

Besides the involvement of the MAPK pathway, melanoma pathogenesis appears to involve tumour suppressor genes, like the CDKN2A locus products p16 and ARF, PTEN and TP53. Studies in gene knockout mouse models indicated that deficiency of each of these tumour suppressors determines a favourable genetic background for the development of melanoma (Walker and Hayward, 2002).

In contrast to most human neoplasias and especially when compared to other skin tumours, melanoma has been reported to be rarely mutated at the TP53 gene, with frequencies ranging from <10 to 25%. However, the high frequency of protein overexpression indicates that pathologic stabilization of TP53 protein due to mechanism other than point mutation may occur in melanoma (Hussein et al., 2003).

Despite their established role in melanoma predisposition (Hayward, 2003), mutations of CDKN2A and CDK4 genes have been detected only in 10% sporadic melanoma lesions, though, LOH at the CDKN2A region has been detected in 50% melanomas and p16 has been shown inactivated in about 100% melanoma cell lines, thus suggesting a possible underestimation of the real frequency of CDKN2A mutations in sporadic melanoma (Castellano and Parmiani, 1999).

Also PTEN gene maps at a chromosomal region (10q23) that is frequently altered in melanoma. Its role in melanoma is still controversial since mutation rates range between 29 and 43% in cell lines though not detectable in specimens (Wu et al., 2003).

A role for other genes in melanoma development remains to be confirmed or discovered. It can be hypothesized that different melanoma subtypes exist that are characterized by different tumour suppressor and oncogene mutational profiles and by a different clinical behaviour. In line with this view, and preliminary to an analysis of global gene expression profiles, we have characterized the genetic alterations and protein expression of the major melanoma-associated genes in a series of early passage melanoma cell lines obtained from surgical specimens.

Results and discussion

BRAF and RAS oncogenes

We evaluated 41 short-term melanoma cell lines (Table 1) for BRAF mutations in exons 11 and 15 by sequence analysis. Exon 15 BRAF mutations were identified in 29 of 41 melanomas, 26 of 29 showing V599E substitution (Table 2). In addition, a double mutation in exon 15 leading to a substitution of leucine by serine at position 596 (L596S, Figure 1a) was detected in three cell lines obtained from different lesions of the same melanoma patient (20842P, 20842M1, 20842M2); this mutation was confirmed in the microdissected original tumour specimens while absent in the nontumoral DNA obtained from the autologous lymphoblastoid cell line (LCL). The L596S substitution, recently reported in primary melanoma also by Kumar et al. (2003b), may affect the protein conformation by altering the hydrophobic residue (leucine) in a hydrophilic-neutral residue (serine) next to a phosphorylation site of kinase domain. Three other double mutations have been reported in primary and metastatic cell lines derived from the same melanoma patient (Davies et al., 2002), and in two metastatic melanomas (Gorden et al., 2003; Kumar et al., 2003a).

Figure 1.

Novel mutations identified by sequence analysis. (a) BRAF exon 15 CT1786-7TC tandem mutation, determining L596S amino-acid change, detected in 20842P, 20842M1 and 20842M2 autologous cell lines, also reported by Kumar et al. (2003a). (b) PTEN gene C737T base substitution determining P246S mutation in melanoma 2211M. (c) PTEN gene T313C+314delT determining C105fsX112 mutation in melanoma 30966M. (d) p16 gene T197C base substitution in exon 2, determining L65P amino-acid change, detected in melanoma 4023M. (e) p16 gene IVS2-2A>G base substitution in intron 2, altering its splice acceptor site, detected in melanoma 26396M. (f) CDK4 gene A796T mutation in exon 2, determining K22R substitution in melanoma 4686M

Full figure and legend (505K)

Table 1 - Origin of melanoma cell lines and disease characteristics.

Full table

Table 2 - Summary of results of mutation analysis and gene and protein expression studies.

Full table

RAS genes analysis revealed four melanomas (5810P, 4405P, 8959M, 4473M) harbouring a Q61R (A182G) NRAS exon 2 mutation, and two melanomas (9923P, 9923M), one primary and the matched metastasis, showing a G12S (G34A) NRAS exon 1 mutation and a Q61L (A182T) KRAS exon 2 mutation. The analysis of three cellular clones obtained from the metastatic cell line (9923M) reproduced the concomitant NRAS G12S and KRAS Q61L mutations, thus suggesting that the two mutations were concomitant at the cell level. No HRAS gene mutations were detected. Thus, RAS mutations were observed in 6/41 (14%) melanoma cell lines, NRAS Q61R being the most frequent mutation detected, as formerly reported. None of the 29 melanomas with BRAF mutations contained a RAS mutation, thus BRAF and RAS mutations were mutually exclusive. These data confirm previous proposal that BRAF and RAS mutations have biologically equivalent effects in tumorigenesis, although few exceptions have been reported (Davies et al., 2002; Gorden et al., 2003).

Different levels of BRAF protein were detected by Western blotting (Figure 2) independently of BRAF mutations. In addition, both the 72 and 95 kDa isoforms previously described (Eychène et al., 1995) were detected in all the samples; by RT–PCR, no alternatively spliced exons 8b and 10a were found (not shown).

Figure 2.

BRAF, pERK 1-2, ERK 1-2, PTEN and p53 protein expression as detected by Western blotting analysis. Diphosphorylated ERK 1-2 (pERK 1-2) is detected at similar levels in melanoma 20842P, 20842M1, 20842M2 (all carrying a L596S-mutated BRAF gene), 14464M melanoma (with a V599E-mutated BRAF gene), 26258M and 3962M (with wild-type BRAF/RAS genes) melanomas, while scant levels of pERK 1-2 protein expression but equivalent amounts of ERK 1-2 are detected for melanomas 1007P, 879M and 13923M also carrying wild-type BRAF/RAS genes. Absence of PTEN protein expression is shown for melanomas 26258M and 3962M carrying wild-type gene and displaying gene expression by RT–PCR. p53 protein overexpression is shown in the case of melanoma 879M, carrying a wild-type p53 gene

Full figure and legend (90K)

The possible activating effect of the L596S BRAF mutation was assessed by Western blot of the diphosphorylated ERK 1-2 levels, which can reveal a constitutive MEK activation (Satyamoorthy et al., 2003). pERK1-2 protein levels were similar in melanomas bearing the L596S or the V599E mutation, while low levels or absence of pERK1-2 protein were shown for melanomas with wild-type RAS/BRAF genes (1007P, 879M, 13923M) (Figure 2). Although melanoma cells have a number of autocrine growth factor loops which may also activate ERK and additional studies are required to determine the effect of the L596S BRAF mutation, these results suggest that it can be equivalent to V599E mutation in constitutively activating ERK phosphorylation.

PTEN gene

The expression of PTEN gene and its pseudogene was analysed by RT–PCR and restriction enzyme assay. PTEN transcript was detected in 40 out of 41 melanomas, both PTEN and PTEN pseudogene expression were detected in two samples (26258M, 5810P), and lack of PTEN transcript with positive expression of the pseudogene was shown in one case (3962M). cDNA sequencing revealed PTEN point mutations in two samples (4686M, 2211M), showing a P38S (C112T) and a P246S (C737T) amino acid substitution, respectively (Table 2). For the P38S variant, already reported (Guldberg et al., 1997), only the mutated allele was detected, suggesting that the wild-type allele has been deleted. The P246S missense mutation is a novel variant (Figure 1b); it affects a region next to phosphate acceptor sites (residues 233–240) possibly altering the PTEN protein function (Deichmann et al., 2002). A third PTEN point mutation leading to a substitution of cysteine by arginine at position 105 was detected in one sample (30966 M) in association with a deletion (T313C+314delT) inducing a frame shift and a novel stop codon in position 112 (C105fsX112). This mutation, a novel variant (Figure 1c), as well as the P38S and P246S variants, were confirmed in genomic DNA. In addition, the melanoma lacking the transcript revealed a wild-type PTEN gene when sequence analysis of the genomic DNA was performed, thus indicating that the gene was not deleted but transcriptionally inactivated, as formerly reported (Zhou et al., 2000).

Expression of the PTEN protein, analysed by Western blotting, was detectable in 37 melanomas, including two showing gene mutations (4686M, 2211M), whereas it was undetectable in four samples (26258M, 5810P, 3962M, 30966M) (Figure 2), two of which showed gene amplification by RT–PCR and one showing the C105fsX112 truncating mutation. This suggests that mutations as well as post-transcriptional events can regulate PTEN protein expression. In addition, in two melanomas showing PTEN protein loss but absence of RAS/BRAF mutations (26258M, 2962M), pERK1-2 protein levels were similar to that observed in melanomas bearing the V599E mutation. These results suggest that PTEN loss results in activation of the MAPK pathway, as suggested also by reciprocity of PTEN and RAS mutations reported by Tsao et al. (2000) in long-term melanoma cell lines.

On the whole, three melanomas were found to harbour point mutations and three a lack of protein expression, for a total of 6/41 (14%) samples showing loss of normal PTEN expression. BRAF was mutated in 3/6 melanomas with PTEN alterations, as also detected in melanoma cell lines by Tsao et al. (2004). The low frequency of PTEN inactivation in our samples is in contrast with most studies reporting mutations in about 30% melanoma cell lines (Wu et al., 2003), lack of expression of the gene and low protein expression, respectively, in 15 and 50% melanoma specimens (Zhou et al., 2000). Our data suggest that low passage melanoma cell lines may better represent the original tumour lesions, and/or that population genetic differences may exist which determine a different involvement of PTEN gene in melanoma pathogenesis.

TP53 gene

TP53 gene mutation analysis was performed by sequencing exons 5–8, the most frequently mutated exons. The results showed that 10/41 melanomas carried TP53 gene mutations (Table 2), including 2/7 primary melanomas and 7/33 metastases. In addition, the polymorphic variant R213R, not altering the amino-acid sequence and occurring at a frequency of 11% in the human population, was detected in one sample (1007P). Two cell lines, obtained from the primary tumour (1402P) and from the local recurrence (1402R) excised 2 months later, were shown to carry the same point mutation Y236H (T706C), confirmed on the microdissected tissue from the two paired tumour specimens. This mutation has been detected in other tumour types, while in melanoma has been reported with a different nucleotide substitution (Ragnarsson-Olding et al., 2002). Two other paired cell lines, derived from the primary lesion (20842P) and from an autologous cutaneous metastasis excised 6 months later (20842M2), showed the same mutation G187S (G559A), already described in melanoma (Zerp et al., 1999). Interestingly, a third cell line obtained from the regional synchronous nodal metastasis of the same patient (20842M1) displayed a wild-type TP53 gene. The results were confirmed on microdissected tissue from the corresponding tumour specimens. The evidence of a selection of a mutated subpopulation in blood spreading metastasis, with respect to nodal metastasis harbouring wild-type TP53, strongly indicates an association between TP53 gene mutations and increasing disease aggressiveness. Consistently, an association between mutated TP53 and worse disease outcome was evident in stage III patients, where three out of nine melanomas clustering in the short survival group (<2 years) carried TP53 mutations, while no mutations were present in five patients showing long survival (>6 years). Two of three mutated cell lines present in the short survival group (4686M, 2211M) carried a S127F mutation (CC380-1TT), already reported with low frequency in other tumours types (Harris, 1996) but not in melanoma. Interestingly, these two melanomas, as well as 30966M, presented also PTEN gene mutations. Different studies demonstrated that mutations in the PTEN gene inversely correlate with an altered TP53 status and that these mutations are often mutually exclusive in carcinomas (Trotman and Pandolfi, 2003). To the best of our knowledge, the simultaneous presence of PTEN and TP53 alterations in melanoma has not been previously reported. Although the functional significance of the S127F mutation is not known, the presence of alterations of both TP53 and PTEN genes within the group with worse prognosis strongly suggests cooperation in determining an aggressive phenotype.

In addition, the following TP53 point mutations involving the DNA-binding domain were detected by sequence analysis: R175H (G524A), associated in melanoma 30966M with PTEN point mutation, detected in other tumour types but not in melanoma; P278S (C872T), already described in melanoma but with a different substitution (16396M); Y234C (A701G), detected in melanoma 17697M, described in carcinoma; E258K (G772A), detected in melanoma 14362M and already described in this tumour type (Papp et al., 1996). These mutations have been detected in the germ line of patients with the Li–Fraumeni syndrome (http//www.iarc.fr/p53), which in the case of melanomas 30966M, 16396M and 17697M was excluded as the LCL showed a wild-type TP53 gene.

TP53 protein expression assessed by Western blotting showed that ten out of 41 melanoma cell lines had high levels of p53 protein (Figure 2). Four cell lines harbouring missense mutations in TP53 gene did not show protein overexpression, and four cell lines harbouring wild-type TP53 showed p53 overexpression. By immunohistochemistry (IHC), the protein expression pattern was similar in cell lines and the original tumour specimen from which the cell lines were derived (Figure 3a, b).

Figure 3.

p53 and ARF protein expression in cell lines and in the original melanoma specimens as assessed by IHC. (a) Cytospin preparation showing p53 protein overexpression in 17697M cells harbouring the Y234C mutation. (b) IHC staining in the original nodal metastasis from which 17697M cell line was derived, showing focal p53 overexpression. (c) Cytospin preparation of 4023M cells showing positive cytoplasmic and nuclear staining for ARF. (d) IHC staining for ARF in the original nodal metastasis from which 4023M cell line was derived, showing a similar pattern of expression

Full figure and legend (255K)

In conclusion, these results showed superimposible patterns of TP53 alterations in primary tumours and metastases along with an association between TP53 gene mutations and a poor disease outcome. Interestingly, for the first time in melanoma, TP53 mutations were shown to be associated with PTEN mutations and always associated with BRAF alterations.

p16, ARF and CDK4 genes

To detect exon specific homozygous deletions of the CDKN2A genes, exons 1, 1, 2 and 3 were coamplified with control genes in multiplex PCR (Figure 4) and the amplified exons were sequenced. As shown in Table 3, the majority of melanomas harboured CDKN2A gene mutations, most displaying homozygous deletions with different deleted regions. Partial deletions were also found: one melanoma (4473M) displayed codon 15–25 deletion in exon 1, and two melanomas obtained from the same patient (1402P, 1402R) showed two partial deletions in exons 1 and 2, which were confirmed by Southern blotting (not shown). These samples showed also identical mutations in the TP53 gene (Table 2).

Figure 4.

Multiplex amplifications of CDKN2A exons. Exons 1, 1 and 3 were coamplified with DPC4 control gene, and exon 2 CDKN2A with -catenin control gene. Melanomas 9923P showing amplification of all exons, 10538P showing absence of exon 1, 15392M showing absence of all exons, 20706M showing loss of exons 1 and 2, and 13923M lacking exons 1, 2 and 3 are displayed

Full figure and legend (56K)

Table 3 - CDKN2A gene mutations.

Full table

Sequence analysis revealed four melanomas (4405P, 14362M, 4023M, 6854M) carrying exon 2 point mutations. Owing to its dual utilization, mutations in exon 2 may alter both p16 and ARF amino-acid sequences. Mutations found were W110stop/G125R, R80stop/P94L, L65P/A79A (Figure 1d) and P114H/A128A for p16/ARF, respectively. In addition, a novel intronic mutation IVS2-2A>G was detected (26396M), not formerly reported, which alters the splice acceptor site of intron 2 possibly altering the p16 transcript and protein function (Figure 1e). Moreover, two melanomas (4686M, 3988M) showed the A148T exon 2 polymorphism, a relatively common variant with no effect on p16 function or on melanoma susceptibility (Bertram et al., 2002). Mutations were somatically acquired and homozygous, thus suggesting the loss of the other allele has occurred. The W110stop and R80stop p16 truncating mutations have been frequently detected in melanomas and shown to impair its function (Ruas and Peters, 1998), while the concomitant ARF mutations have not been evaluated functionally. However, it has been shown that different exon 2 mutations can affect cellular localization of ARF and its cell cycle control activity (Rizos et al., 2001; Hashemi et al., 2002). In fact, in the carboxy terminal region encoded by exon 2, a nucleolar localization domain not essential for p53 stabilization or nuclear localization, but potentiating ARF activity, has been demonstrated (Zhang and Xiong, 1999). The L65P/A79A mutation has not been formerly reported; it alters an amino-acid residue in the fourth conserved ankirin repeat and may impair it functionally, while the 217 T>C base change gives rise to a silent mutation in ARF codon 48. The P114H/A128A mutation was formerly reported in carcinoma (Ku et al., 1999), while different substitutions were identified in the germ line of melanoma patients.

RT–PCR showed that some melanomas (9923P, 9923M, 1568M, 16396M) lacked p16 or ARF transcripts (5810P, 10538P), or both (14362M, 3988M, 26396M) (Table 2). Methylation of the CpG island extending from the promoter into exon 1 has been shown to determine p16 gene silencing in melanoma (Castellano and Parmiani, 1999). On the contrary, although ARF promoter methylation has been shown to occur in other tumour types such as in colon carcinoma (Esteller et al., 2000), no evidence has been reported in melanoma. By assessing promoter hypermethylation by methylation specific-PCR, we detected methylation of p16 promoter but not of ARF promoter in the melanomas lacking transcripts (not shown), thus suggesting that other mechanisms may regulate ARF expression in melanoma.

By Western blotting two additional melanomas (5810P, 9460M) were found to be p16 negative, as confirmed by FACS analysis (Figure 5), suggesting that the regulatory mechanisms include post-transcriptional events. Likewise, ARF expression was lacking in one melanoma (9460M). By IHC, the staining pattern of ARF was found to be predominantly nuclear, sometimes with a strong nucleolar accentuation; similar results were obtained in the original tumour specimens from which the melanoma cell lines were derived (Figure 3c, d).

Figure 5.

p16 protein expression as detected by FACS analysis. Intracellular staining was obtained after cell permeabilization and incubation with PE-conjugated anti-p16 mAb (grey area) or with an unrelated PE-conjugated mAb (empty area). A p16 positive and a negative melanoma are shown (melanoma 4686M and 9460M, respectively)

Full figure and legend (34K)

CDK4 point mutations determining R24C or R24H substitutions detected in the melanoma-prone families have also been found in sporadic melanoma, together with K22Q and N41S mutations. Alterations in codon 22 have been shown to result in the loss of p16-binding capacity (Ceha et al., 1998). For these reasons, CDK4 exon 2 was sequenced in melanomas carrying wild-type p16 genes. In one (4686 M) out of 14 melanomas tested (7%), a K22R substitution was detected, not previously reported and absent in the autologous LCL (Figure 1f and Table 2).

In conclusion, we found that normal p16 gene/protein expression was lacking in 36/41 (87%) melanomas, while loss of ARF was less frequent (29/41, i.e. 70% samples), independently from BRAF mutations. The frequency of CDKN2A gene alterations shown here (28/41, 68%) is closer to that of LOH at 9p21 detected in melanoma specimens than to that observed in long-term melanoma cell lines. To specifically address this issue, we examined the available matched original tumour specimens for loss of heterozygosity (LOH) with three different microsatellites mapping in CDKN2A locus (D9S1748, D9S942 and D9S974). In the six samples giving informative results (18656M, 14464M, 4023M, 13923M, 4473M, 26258M, Table 2), at least one of the microsatellite marker showed allelic loss, thus indicating the in vivo occurrence of CDKN2A gene deletion.

General conclusions

By considering the dysfunction of the different tumour suppressors and oncogenes studied as single individual traits, the 41 melanomas result defined by 12 different mutational profiles (Figure 6). The p16/ARF+BRAF^V599E pattern is the most represented, as it was detected in 15 melanomas including primary tumours, nodal and cutaneous metastases. Notably, the melanoma cell lines obtained from nodal metastases of long survivors (n=5, see Table 1) are the only homogeneous group with the single p16/ARF+BRAF^V599E pattern represented. Conversely, complex profiles including BRAF mutations and/or alterations of multiple tumour suppressors are associated in nodal metastases to short-term survival. Furthermore, by considering all the 15 patients with melanoma showing the p16/ARF+BRAF^V599E profile, 12/15 cluster in the group with survival >2 years from the day of lymph node surgery; in contrast, of the 22 patients with melanoma showing other profiles 7/22 showed survival >2 years (P=0.004 by ²). Of 17 patients with melanoma showing the p16/ARF+BRAF/RAS profile, 13 survived >2 years, compared to 6/20 in the group of melanomas with other profiles. It is noteworthy that three out of six of these cases showed absence of PTEN expression associated with high pERK levels indicating constitutive MAPK activation (see Figure 2). In addition, it can be speculated that for the four patients bearing p16/ARF+BRAF/RAS with survival <2 years, alterations at other genes not studied here has occurred. An association between BRAF mutations and longer disease-free survival and also to a shorter duration of response to treatment has been recently reported by Kumar et al. (2003a). Our results underline the importance of considering multiple gene alterations although a higher number of cases are necessary to verify their potential impact on disease outcome.

Figure 6.

Distribution of mutational profiles in primary and metastatic melanomas. All the different combinations of affected genes are shown in the legend

Full figure and legend (185K)

A role for other genes not assessed here can be proposed for few samples that do not bear alterations in the genes studied (like for 1007P) or acting in cooperation with BRAF activation or with p16/ARF loss (like for 20842M1, or 13923M and 879M). No relation to the sporadic/familial origin of the disease or to the chronically UV-exposed body sites appears, although too few samples were analysed. Nonetheless, C>T mutations at dipyrimidine sites, considered fingerprints for UV light-induced mutations, were observed in about half of the total detected gene mutations.

Our results strongly suggest that short-term cultures represent a reliable model to trace back the molecular events of melanoma pathogenesis. In fact, although a complete comparison has not been performed, cell lines and the original tumour specimens displayed the same gene mutations or protein expression patterns whenever they were analysed in parallel. In fact, the TP53 and/or BRAF gene mutations detected in the cell lines 20842P, 20842M1, 20842M2, 1402P and 1402R were confirmed to be present in the original microdissected tumours, and LOH at CDKN2A was found in six melanomas displaying p16/ARF gene mutations. p53 and ARF staining results revealed comparable protein localization and overexpression in cell lines and tumour specimens (Figure 3). Furthermore, the detection of identical mutations in two pairs of autologous cell lines included in the panel (9923P and 9923M, 1402P and 1402R) indirectly indicated the in vivo occurrence of the genetic alterations.

Although important differences in the frequency of gene alterations between melanoma cell lines and tumour biopsies have been reported for p16 and PTEN genes, we observed alterations in these genes at frequency closer to those reported in melanoma specimens than in long-term melanoma cell lines (Pollock and Trent, 2000). It is possible that the in vitro selection of melanoma cells with peculiar genetic alterations that better adapt to culture conditions has occurred. Since melanomas are usually small tumours, often highly pigmented and heavily infiltrated by stroma and lymphocytes, the use of short-term melanoma cell lines allows the identifications of gene mutations which can be masked in the absence of tissue microdissection, and the availability of cellular samples for extensive investigations.

Materials and methods

Cell lines

Melanoma cell lines were derived from eight primary cutaneous melanomas, 20 lymph node, 11 cutaneous metastases and two visceral surgically excised from 31 sporadic melanoma patients, four patients having a first-degree relative with melanoma and two multiple primary melanoma cases. Information relative to the lesions and patients from which the cell lines were originated are reported in Table 1. Cells were obtained from surgical melanoma specimens by both mechanical and enzymatic treatment and cultured in RPMI medium supplemented with 10% FCS, HEPES buffer, glutamine and antibiotics (Biowhittaker). Short-term cultures, not exceeding 10–15 passages, were used. Their malignant melanoma origin was verified by staining with HMB45 antibody or by tyrosinase detection by RT–PCR. Most of the patients were genetically HLA typed, thus enabling to confirm the origin of cell lines derived from these patients. Autologous LCL derived by EBV infection of peripheral blood lymphocytes were used as source of nontumoral DNA.

PCR analysis

Genomic DNA was isolated using the 'Salting out' procedure (Miller et al., 1988). Amplifications were carried out as previously described for BRAF gene exons 11 and 15 (Davies et al., 2002), NRAS, KRAS and HRAS exons 1–2 (Albino et al., 1989; Lupetti et al., 1994), TP53 gene exons 5–8 (Donghi et al., 1993) and for CDK4 exon 2 (Wolfel et al., 1995). For PTEN, the following primers were used to amplify the genomic region flanking the identified mutated sequences: PTENEX7S, 5'-ATCCTCAGTTTGTGGTCTGCC-3'; PTENEX7AS, 5'-GCATCTTG TTCTGTTTGTGGAAG-3'; PTENINTR1S, 5'-TCCTTAACTAAAGTACTCAG-3'; PTENINTR2AS, 5'-GCATTCTTACCTTACTACAT-3'. The amplifications were performed as follows: denaturation at 94°C for 8 min, followed by 35 cycles of PCR (94°C for 1 min, 55°C for 1 min, 72°C for 1 min), and a final extension at 72°C for 1 min. To verify PTEN gene presence in genomic DNA, the following additional primers were used: PTENINTR2S, 5'-ATTAGGAAAAAGAAAATCTGTC-3'; PTENINTR3AS, 5'-AACAAGCAGATAACTTTCAC-3'; PTENINTR3S, 5'-TGTG TCACATTATAAAGATTCAGGC-3'; PTENINTR4AS, 5'-ACTCGATAATCTGGAT GACTCATTA-3'; PTENINTR4S, 5'-CAGTGTTTCTTTTAAATACCTGTTAAGTTT-3'; PTENINTR5AS, 5'-CTGTTTTCCAATAAATTCTCAGATCCAG-3'; PTENINTR7S, 5'-ATCATTAATTAAATATGTCA-3'; PTENINTR8AS, 5'-CATAAGTTAAAAACTTG TCA-3'. For CDKN2A genes, amplifications were carried out as described by Castellano et al. (1997) for exon 1, exon 2, exon 3 and multiplex amplifications and as Fargnoli et al. (1998) for exon 1.

RT–PCR

Total RNA was extracted from melanoma cells by RNAzol B (Tel-Test Inc.). The first strand was synthesized using oligo(dT) primer and SuperScript II RNase H^- Reverse Transcriptase (Life Technologies) at 42°C for 1 h. For PTEN, multiplex PCR amplifications with the housekeeping gene -actin and restriction analysis to discriminate between RT-PCR products of the functional PTEN gene and PTEN pseudogene were performed as previously described (Frisk et al., 2002). PCR amplification of PTEN exons 6–9 cDNA fragment was performed using the following primers: PTEN-3 5'-ACCAGTGGCACTGTTGTTTCAC-3' and PTEN-5 5'-GTATGCTGATCTTCATCAAAAGG-3'. For p16 and ARF, previously described conditions were used (Castellano et al., 1997; Fargnoli et al., 1998).

DNA sequencing

The PCR products were purified using the Microcon YM-50 (Millipore), and purified PCR products were directly sequenced by ABI PRISM 377 automatic sequencer (Applied Biosystems). Sequences were analysed with Sequencing Analysis software and compared with those in the GenBank database using Sequence Navigator software or VECTOR NTI SUITE V.7 software (Informax). The following GenBank accession numbers sequences were used for comparison: NM_004333 (BRAF), AF493919 (NRAS), M54968 (KRAS), U93051 (PTEN/MMAC1), U94788 (TP53), BC021998 (p16), NM_058195 (ARF) and U37022 (CDK4).

LOH analysis

The polymorphic microsatellite markers D9S1748, D9S942 and D9S974 were amplified according to the manufacturer's instruction. The PCR products were run on 5% polyacrylamide gel containing 6 M urea and 1 Tris-borate-EDTA buffer in an ABI PRISM 377 sequencer, and the collected data were analysed by Gene Scan 3.1 software program (Applied Biosystems). LOH was defined when at least 20% signal reduction intensity of one allele was observed in neoplastic tissue compared with the matched allele in the normal tissue.

Western blotting, IHC and FACS

For Western blotting, 50 g protein samples were separated on 10% Bis-Tris precast gels (Invitrogen) and transferred to Hybond ECL Nitrocellulose membrane (Amersham). Membranes were probed with primary antibodies and the signal was detected using peroxidase-conjugated anti-mouse (Becton Dickinson) or anti-rabbit secondary antibodies (Amersham) and developed using ECL (Amersham). IHC was performed on cytospins or paraffin-embedded melanoma tissues by the avidin–biotin complex method (DAKO). Cytospins were fixed in formalin, while tissue sections were deparaffinized, rehydrated and treated for antigen retrieval with 5 nM sodium citrate buffer (pH 6) for 15 min at 95°C. Biotinylated rabbit anti-goat or anti-mouse secondary antibodies and 9-amino 3-ethyl carbazole, for the chromogenic reaction, were used. Eventually, samples were counterstained with Carazzi's haematoxylin. The following antibodies were used: anti-p53 mAb DO.7, anti-PTEN mAb 6H2.1 (Cascade Bioscience), goat polyclonal anti-ARF and anti Raf-B (Santa Cruz), anti-diphosphorylated ERK 1–2 mAb and anti-ERK 1-2 rabbit serum (Sigma). FACS analysis of p16 expression was performed after cell permeabilization with Fix Perm and staining with PE-conjugated anti-p16 mAb (Becton Dickinson). Samples were analysed with a FACScan (Becton Dickinson) by considering 10 000 events.

References

1. Albino AP, Nanus DM, Mentle IR, Cordon-Cardo C, McNutt NS, Bressler J and Andreeff M. (1989). Oncogene, 4, 1363–1374. | PubMed | ISI | ChemPort |
2. Bertram CG, Gaut RM, Barrett JH, Pinney E, Whitaker L, Turner F, Bataille V, Dos Santos Silva IJ, Swerdlow A, Bishop DT and Newton Bishop JA. (2002). J. Invest. Dermatol., 119, 961–965. | Article | PubMed | ISI | ChemPort |
3. Brose MS, Volpe P, Feldman M, Kumar M, Rishi I, Gerrero R, Einhorn E, Herlyn M, Minna J, Nicholson A, Roth JA, Albelda SM, Davies H, Cox C, Brignell G, Stephens PP, Futreal A, Wooster R, Stratton MR and Weber BL. (2002). Cancer Res., 62, 6997–7000. | PubMed | ISI | ChemPort |
4. Cascinelli N, Belli F, MacKie RM, Santinami M, Bufalino R and Morabito A. (2001). Lancet, 358, 866–869. | Article | PubMed |
5. Castellano M and Parmiani G. (1999). Melanoma Res., 9, 421–432. | Article | PubMed | ISI | ChemPort |
6. Castellano M, Pollock PM, Walters MK, Sparrow LE, Down LM, Gabrielli BG, Parsons PG and Hayward NK. (1997). Cancer Res., 57, 4868–4875. | PubMed | ISI | ChemPort |
7. Ceha HM, Nasser I, Medema RH and Slebos RJ. (1998). Biochem. Biophys. Res. Commun., 249, 550–555. | Article | PubMed |
8. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Dicks E, Ewing R, Floyd Y, Gray K, Hall S, Hawes R, Hughes J, Kosmidou V, Menzies A, Mould C, Parker A, Stevens C, Watt S, Hooper S, Wilson R, Jayatilake H, Gusterson BA, Cooper C, Shipley J, Hargrave D, Pritchard-Jones K, Maitland N, Chenevix-Trench G, Riggins GJ, Bigner DD, Palmieri G, Cossu A, Flanagan A, Nicholson A, Ho JW, Leung SY, Yuen ST, Weber BL, Seigler HF, Darrow TL, Paterson H, Marais R, Marshall CJ, Wooster R, Stratton MR and Futreal PA. (2002). Nature, 417, 949–954. | Article | PubMed | ISI | ChemPort |
9. Deichmann M, Thome M, Benner A, Egner U, Hartschuh W and Näher H. (2002). Br. J. Cancer, 87, 1431–1436. | Article |
10. Dong J, Phelps RG, Qiao R, Yao S, Benard O, Ronai Z and Aaronson SA. (2003). Cancer Res., 63, 3883–3885. | PubMed | ISI | ChemPort |
11. Donghi R, Longoni A, Pilotti S, Michieli P, Della Porta G and Pierotti MA. (1993). J. Clin. Invest., 91, 1753–1760. | PubMed | ISI | ChemPort |
12. Eskandarpour M, Hashemi J, Kanter L, Ringborg U, Platz A and Hansson J. (2003). J. Natl Cancer Inst., 95, 790–798. | Article | PubMed | ChemPort |
13. Esteller M, Tortola S, Toyota M, Capella G, Peinado MA, Baylin SB and Herman JG. (2000). Cancer Res., 60, 129–133. | PubMed | ISI | ChemPort |
14. Eychène A, Dusanter-Fourt I, Barnier JV, Papin C, Charon M, Gisselbrecht S and Calothy G. (1995). Oncogene, 10, 1159–1165. | PubMed | ISI | ChemPort |
15. Fargnoli MC, Chimenti S, Keller G, Soyer HP, Dal Pozzo V, Hofler H and Peris K. (1998). J. Invest. Dermatol., 111, 1202–1206. | Article | PubMed | ISI | ChemPort |
16. Frisk T, Foukakis T, Dwight T, Lundberg J, Höög A, Wallin G, Eng C, Zedenius J and Larsson C. (2002). Genes, Chromosomes & Cancer, 35, 74–80.
17. Gorden A, Osman I, Gai W, He D, Huang W, Davidson A, Houghton AN, Busam K and Polsky D. (2003). Cancer Res., 63, 3955–3957. | PubMed | ISI | ChemPort |
18. Guldberg P, thor Straten P, Birck A, Ahrenkiel V, Kirkin AF and Zeuthen J. (1997). Cancer Res., 57, 3660–3663. | PubMed | ISI | ChemPort |
19. Harris CC. (1996). Br. J. Cancer, 73, 261–269. | PubMed | ISI | ChemPort |
20. Hashemi J, Lindstrom MS, Asker C, Platz A, Hansson J and Wiman KG. (2002). Cancer Lett., 180, 211–221. | Article | PubMed | ISI | ChemPort |
21. Hayward NK. (2003). Oncogene, 22, 3053–3062. | Article | PubMed | ISI | ChemPort |
22. Hayward NK. (1999). Molecular Pathology of Early Cancer, Srivastava S, Henson DE, Gazdar A. (eds.). IOS Press: Amsterdam, pp. 207–231.
23. Hussein MR, Haemel AK and Wood GS. (2003). Eur J Cancer Prev., 12, 93–100. | Article | PubMed | ISI | ChemPort |
24. Ku JL, Kim WH, Lee JH, Park HS, Kim KH, Sung MW and Park JG. (1999). Laryngoscope, 109, 976–982. | Article | PubMed | ISI | ChemPort |
25. Kumar R, Angelini S, Czene K, Sauroja I, Hahka-Kemppinen M, Pyrhönen S and Hemminki K. (2003a). Clin. Cancer Res., 9, 3362–3368. | PubMed | ISI | ChemPort |
26. Kumar R, Angelini S and Hemminki K. (2003b). Oncogene, 22, 9217–9224. | Article | PubMed | ISI | ChemPort |
27. Lupetti R, Sensi M, Mortarini R, Anichini A, Clemente C and Parmiani G. (1994). Melanoma Res., 4, 11–19. | PubMed | ISI | ChemPort |
28. Miller SA, Dykes DD and Polesky HF. (1988). Nucleic Acids Res., 16, 1215–1221. | Article | PubMed | ISI | ChemPort |
29. Papp T, Jafari M and Schiffmann DJ. (1996). Cancer Res. Clin. Oncol., 122, 541–548.
30. Pollock PM and Trent JM. (2000). Clin. Lab. Med., 20, 667–690. | PubMed | ISI | ChemPort |
31. Pollock PM, Harper UL, Hansen KS, Yudt LM, Stark M, Robbins CM, Moses TY, Hostetter G, Wagner U, Kakareka J, Salem G, Pohida T, Heenan P, Duray P, Kallioniemi O, Hayward NK, Trent JM and Meltzer PS. (2003). Nat Genet., 33, 19–20. | Article | PubMed | ISI | ChemPort |
32. Ragnarsson-Olding BK, Karsberg S, Platz A and Ringborg UK. (2002). Melanoma Res., 12, 453–463. | Article | PubMed | ISI | ChemPort |
33. Rizos H, Darmanian AP, Holland EA, Mann GJ and Kefford RF. (2001). J. Biol. Chem., 276, 41424–41434. | Article | PubMed | ISI | ChemPort |
34. Ruas M and Peters G. (1998). Biochim. Biophys. Acta, 1378, F115–F177. | Article | PubMed | ISI | ChemPort |
35. Satyamoorthy K, Li G, Gerrero MR, Brose MS, Volpe P, Weber BL, Van Belle P, Elder DE and Herlyn M. (2003). Cancer Res., 63, 756–759. | PubMed | ISI | ChemPort |
36. Trotman L and Pandolfi P. (2003). Cancer Cell, 3, 97–99. | Article | PubMed | ISI | ChemPort |
37. Tsao H, Zhang X, Fowlkes K and Haluska FG. (2000). Cancer Res., 60, 1800–1804. | PubMed | ISI | ChemPort |
38. Tsao H, Goel V, Wu H, Yang G and Haluska FG. (2004). J. Invest. Dermatol., 122, 337–341. | Article | PubMed | ISI | ChemPort |
39. Walker GJ and Hayward NK. (2002). J Invest Dermatol., 119, 783–792. | Article | PubMed | ISI | ChemPort |
40. Wolfel T, Hauer M, Schneider J, Serrano M, Wolfel C, Klehmann-Hieb E, De Plaen E, Hankeln T, Meyer zum Buschenfelde KH and Beach D. (1995). Science, 269, 1281–1284. | Article | PubMed | ISI | ChemPort |
41. Wu H, Goel V and Haluska FG. (2003). Oncogene, 22, 3113–3122. | Article | PubMed | ISI | ChemPort |
42. Zerp SF, van Elsas A, Peltenburg LT and Schrier PI. (1999). Br. J. Cancer, 79, 921–926. | Article | PubMed | ISI | ChemPort |
43. Zhang Y and Xiong Y. (1999). Mol. Cell, 3, 579–591. | Article | PubMed | ISI | ChemPort |
44. Zhou XP, Gimm O, Hampel H, Niemann T, Walker MJ and Eng C. (2000). Am. J. Pathol., 157, 1123–1128. | PubMed | ISI | ChemPort |

Acknowledgements

We gratefully acknowledge the contributions of Marina Castellano, Michela Sosio and Marco Asioli in p16/ARF analysis; Licia Rivoltini and Chiara Castelli for melanoma and LCL lines; Claudia Lombardo for genetic HLA typing; Donata Penso for sequence analysis; Ilaria Bersani and Barbara Vergani for histochemical stainings, Simona Frigerio for MSM analysis, Graziella Pasquini for TP53 sequencing and Grazia Barp for editorial assistance. This study was supported by Italian Association for Cancer Research (AIRC, Milan), CNR-MIUR 'Progetto Strategico Oncologia' (02.00385.ST97 to M.A. Pierotti) and the Cariplo Foundation (Milan).