Activating BRAF and N-Ras mutations in sporadic primary ... - Nature

Oncogene (2003) 22, 9217–9224

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Activating BRAF and N-Ras mutations in sporadic primary melanomas: an inverse association with allelic loss on chromosome 9 Rajiv Kumar*,1,2, Sabrina Angelini2 and Kari Hemminki1,2 1 Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 580, Heidelberg 69210, Germany; 2Department of Biosciences, Center for Nutrition and Toxicology, Karolinska Institute, Novum, Huddinge 14157, Sweden

We searched and report mutations in the BRAF and N-ras genes in 22 out of 35 (63 percent) primary sporadic melanomas. In three melanomas, mutations were concomitantly present in both genes. In all, 10 out of 12 mutations in the BRAF gene involved the ‘hot spot’ codon 600 (In all communications on mutations in the BRAF gene, the nucleotide and codon numbers have been based on the NCBI gene bank nucleotide sequence NM_004333. However, according to NCBI gene bank sequence with accession number NT_007914, there is a discrepancy of one codon (three nucleotides) in exon 1 in the sequence with accession number NM_004333. The sequence analysis of exon 1 of the BRAF gene in our laboratory has shown that the sequence derived from NT_007914 is correct (Kumar et al., 2003). Due to the correctness of the latter, sequence numbering of codons and nucleotides after exon 1 are changed by þ 1 and þ 3, respectively.), one tandem CT1789-90TC base change represented a novel mutation and another mutation caused a G466R aminoacid change within the glycine-rich loop in the kinase domain. Mutations in the N-ras gene in 11 melanomas were at codon 61 whereas two melanomas carried mutations in codon 12 including a tandem mutation GG4AA. We observed an inverse association between BRAF/N-ras mutations and the frequency of loss of heterozygosity (LOH) on chromosome 9 at 10 different loci. Melanomas with BRAF/N-ras mutations showed a statistically significant decreased frequency of LOH on chromosome 9 compared with cases without mutations (mean fractional allelic loss (FAL) ¼ 0.2970.23 vs 0.7270.33; t-test, P ¼ 0.0001). Difference in the FAL value between tumours with and without BRAF/N-ras mutations on 33 loci on five other chromosomes was not statistically significant (mean FAL 0.1770.19 vs 0.2570.22; t-test, P ¼ 0.24). Melanoma cases with BRAF/N-ras mutations were also associated with lower age at diagnosis than cases without mutations (mean age 80.3877.24 vs 65.77719.79 years; t-test, P ¼ 0.02). Our data suggest that the occurrence of BRAF/N-ras mutations compensate the requirement for the allelic loss *Correspondence: R Kumar, Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 580 TP3, Heidelberg 69210, Germany; E-mail: [email protected] Received 30 April 2003; revised 23 June 2003; accepted 26 June 2003

at chromosome 9, which is one of the key events in melanoma. Oncogene (2003) 22, 9217–9224. doi:10.1038/sj.onc.1206909 Keywords: BRAF; N-ras; melanoma; mutation; FAL; SSCP

Introduction Cutaneous malignant melanoma is a potentially fatal neoplasm with complex and heterogeneous etiology (Chin et al., 1998). The sporadic form, which constitutes over 90% of all cases, is linked to sunlight exposure (Gilchrest et al., 1999; Hemminki et al., 2001). Though the mechanism for the casual relationship between sunlight exposure and melanoma is less clear, the epidemiological data are strongly supportive of such an association (Whiteman et al., 2001). The inherited form of melanoma is associated with the melanoma susceptibility locus on chromosome 9p21. Germline mutations in the CDKN2A gene, which encode two cell cycle inhibitors p16INK4a and p14ARF, are found in a proportion of melanoma-prone families (Bishop et al., 2002). However, somatic alterations in sporadic melanoma are heterogeneous. Allelic loss at chromosome 9p21 locus is the most prevalent genetic occurrence in sporadic melanoma, but mutations and other alterations in the CDKN2A gene, though common in melanoma cell lines, are less evident in primary tumours (Kumar et al., 1998b, 1999). The predominant oncogenic changes associated with malignant melanoma have been the activating mutations in the N-ras gene (Castellano and Parmiani, 1999; Saida, 2001). The reported frequency of mutation in the ras genes has varied; the most common mutation detected in melanomas involves codon 61 and these are reported to occur early (van Elsas et al., 1997; Omholt et al., 2002). An activated RAS stimulates a multitude of downstream signalling cascades (Shields et al., 2000). RAF serine/threonine kinases are the key signalling components in the RAS pathway (Kolch, 2000). In melanocytes, BRAF, one of the members of the RAF family, is activated in a cAMP-dependent signalling cascade, as a consequence of a-melanocyte-stimulating

BRAF/N-ras mutations in melanoma R Kumar et al

9218

hormone and related peptides binding to melanocortin receptor 1 (Busca et al., 2000; Halaban, 2000). The importance of BRAF in melanocyte biology was revealed by the detection of alterations in a high proportion of melanoma cell lines and primary tumours in the BRAF gene (Davies et al., 2002). An overrepresenting proportion of mutations detected in the melanoma and other cancers affected a single residue (V600E; previously V599E) in the kinase-activation domain of BRAF. These results have subsequently been confirmed by other studies on melanomas and melanocytic nevi (Brose et al., 2002; Kumar et al., 2003; Pollock et al., 2003). In the present study, we analysed exons 11 and 15 of the BRAF gene for mutations in 35 microdissected primary melanomas. Additionally, we also screened for mutations in exons 1 and 2 of the N-ras gene. The primary melanomas, included in this study, have previously been screened for alterations of the CDKN2A and related genes (Kumar et al., 1998a, 1999, 2001). We have also, in these tumour samples, previously determined the allelic losses at 43 different microsatellite marker loci on six different chromosomes including chromosome 9 (Kumar et al., 1999; Smeds et al., 2000).

We found mutations in the BRAF and N-ras genes in 22 melanomas including two novel tandem mutations at dipyrimidinic sites. We also found an inverse association between BRAF/N-ras mutations and the frequency of allelic losses at chromosome 9 in primary melanomas.

Results Mutations in the BRAF and N-ras genes in primary melanomas In all, 35 microdissected sporadic primary melanomas were analysed for mutations in exons 11 and 15 of the BRAF gene, and exons 1 and 2 of the N-ras gene, using PCR-single-strand conformation polymorphism (PCRSSCP) and the DNA-sequencing techniques (Figure 1). In exon 11 of the BRAF gene, a G1396C (previously G1393C) mutation was detected in one melanoma (case 35), which resulted in a novel glycine-to-arginine change in the amino acid at codon 466 (previously 465). However, in exon 15 of the same gene, we detected mutations in 11 melanomas (Table 1). In total, 10 of these mutations involved the hotspot codon 600. In nine

Figure 1 A representative autoradiograph showing SSCP analysis of exon 1 of the N-ras gene. Samples in lanes 2 and 3 show the aberrant bands due to G4A and GG4AA mutations in codon 12 in melanoma cases 31 and 35, respectively. Lane 1 contains control DNA showing wild-type bands. Wild-type bands seen in lanes 4–9 are from cases 34–32, 30 and 29. (a) Representative results from SSCP analysis of exon 15 of the BRAF gene. Lanes 1 and 2 that contain amplified DNA from the control sample and melanoma 35, respectively, show wild-type bands. The aberrant bands in lanes 3, 4 and 5 are due to T1799A (V600E) mutation in melanomas 34, 33 and 29. The two samples in lanes 6 and 7 are from the same melanoma 32, and represent GT1798-99AA (V600K) tandem mutation. The aberrant shift due to GT1798-99AA sequence alteration was also observed in the double strand (not shown). (b) Sequence analysis showing GG4AA mutation at codon 12 (shown by a thick line) in exon 1 of the N-ras gene in melanoma case 35, which results in Glyto-Asn amino-acid change. The sequence was obtained from DNA extracted from the aberrant band excised from the SSCP gel. (c) Part of exon 15 of the BRAF gene sequence from melanoma case 32 showing GT1798-99AA (V600K) sequence change (shown by a thick line). (d) Corresponding wild-type sequence of exon 1 of the N-ras gene, showing a part with codon 12 (thick line) and codon 13 obtained from a control DNA sample. (e) Part of the sequence of exon 15 of the BRAF gene obtained from a control DNA sample corresponding to wild-type sequence reported in gene databank (accession number NT_007914). A thick line indicates the position of mutated bases shown in the melanoma case 32 Oncogene


9219 Table 1

Mutations in the CDKN2A, N-ras and BRAF genes, and fractional allelic loss (FAL) at various chromosomes in microdissected primary melanomas

Melanoma 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24h 25 26 27 28 29 30 31 32 33 34 35m

FALa chromosome 9

FAL other chromosomesb

1.00 0.75 0.11 1.00 0.00 0.71 1.00 0.75 0.43 0.50 0.50 0.25 1.00 0.33 1.00 0.14 0.25 0.00 0.33 0.25 0.00 0.00 0.43 0.66 0.66 0.17 1.00 0.00 0.13 0.17 0.40 0.20 0.50 0.33 0.75

0.32 0.38 0.53 0.30 0.00 0.60 0.15 0.64 0.46 0.27 0.50 0.00 0.00 0.48 0.14 0.34 0.15 0.07 0.10 0.19 0.27 0.24 0.41 0.00 0.08 0.17 0.09 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

CDKN2Ac mutations

N-ras exon 1 mutations

N-ras exon 2 mutations

BRAF exon 15 mutations

WT WT WT WT WT WT WT WT WT WT WT WT Mutation WT Mutation WT Mutation WT WT WT WT WT Mutation WT WT WT WT WT WT WT WT WT Mutation WT Mutation

WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT GGT4AGTj WT WT WT GGT4AATn

WT CAA4AAAd CAA4CATe WT CAA4CGAf WT WT WT WT WT WT WT WT CAA4AAA WT CAA4CGA CAA4CGA CAA4AAA WT WT WT WT CAA4CGA WT CAA4CTAi CAA4CGA WT WT WT CAA4AAA WT WT WT WT WT

WT WT WT WT WT WT WT WT WT WT GTG4GAGg WT WT GTG4GAG WT WT WT WT GTG4GAG GTG4GAG WT GTG4GAG WT WT WT WT WT GTG4GAG GTG4GAG WT CTA4TCAk GTG4AAGl GTG4GAG GTG4GAG WT

a

FAL represents fractional allelic loss, and is defined as the ratio of number of markers with LOH to the total number of informative markers. FAL values for other chromosomes include data from 33 different microsatellite markers on chromosomes 1, 6, 10, 11 and 13. cMutations in the CDKN2A gene and FAL data included in this study have been described in earlier studies. dCAA4AAA mutation in codon 61 of the N-ras gene results in Gln-to-Lys amino-acid change. eCAA4CTA causes Gln-to-His change. fCAA4CGA changes Gln to Arg. gGTG4GAG mutation in codon 600 of the BRAF gene causes Val-to-Glu change in amino acid. hCase 24 carried a CC4T* mutation in codon 151–52 (exon 5) of the p53 gene. iCAA4CTA causes Gln-to-Leu change in amino acid. jGGT4AGT in codon 12 of the N-ras gene causes Gly-to-Ser change in amino acid. k CTA4TCA tandem mutation in codon 597 of the BRAF gene changes the residue Leu to Ser. lGTG4AAG tandem mutation in codon 600 of the BRAF gene causes Val-to-Lys change in amino acid. mMelanoma 35 also carried GGA4CGA mutation in codon 466 of the BRAF gene that causes Gly-to-Arg change in amino acid. nGGT4AAT tandem mutation in codon 12 of the N-ras genes causes Gly-to-Lys change in amino acid. Bold type indicates mutations and underline indicates base change b

cases, mutation in codon 600 (previously 599) was T1799A (previously T1796A) and one case carried GT1798-99AA (previously GT1795-96AA) tandem mutation causing a valine-to-lysine amino-acid change at the residue 600 (previously 599). In one melanoma, we detected hitherto unreported novel CT1789-90TC (previously CT1786-87TC) mutation changing leucine to serine at codon 597 (previously 596). Similarly, mutations in the N-ras gene were detected in 13 cases. In 11 melanoma cases N-ras mutations involved codon 61 in exon 2, and in two cases mutations effected codon 12 in exon 1, including a novel GG4AA tandem mutation. In three melanomas (cases 14, 31 and 35), mutations were present in both the BRAF and N-ras genes. The somatic nature of mutations in the BRAF/N-ras genes in melanomas was indicated by the non-occurrence of

those alterations in the surrounding tissues available from six melanoma cases. Association of BRAF/N-ras mutations in primary melanomas with CDKN2A mutations and allelic losses on chromosome 9 The present series of sporadic melanomas, in which mutations in the BRAF and N-ras genes are described in this report, had been previously studied for mutations in the CDKN2A (p16INK4a and p14ARF) and p53 genes. In addition, in these melanomas, we had also determined the frequency of loss of heterozygosity (LOH) at 43 different microsatellite loci on six different chromosomes (Kumar et al., 1998a, 1999, 2001; Smeds et al., 2000). The microsatellite markers studied were from Oncogene


9220

chromosomes 1 (six markers), 6 (12 markers), 9 (10 markers), 10 (eight markers), 11 (three markers) and 13 (four markers) (Smeds et al., 2000). Mutations in the CDKN2A (p16INK4a and p14ARF) gene, determined in our previous studies, were present in six melanomas and four of these tumours carried mutation either in the BRAF or N-ras gene; the association between mutational frequencies of the CDKN2A and BRAF/N-ras genes was not statistically significant (Table 1). Mutation in the p53 gene was present only in one melanoma (case 24) that was wild type for both BRAF and N-ras. When the frequency of LOH (in the form of fractional allelic loss (FAL)) on six chromosomes was compared between the cases with and without BRAF/N-ras mutations, the combined FAL value for markers on chromosome 9 showed statistically significant difference. The mean FAL value for cases with BRAF/N-ras mutations was 0.29 (95% CI, 0.19–0.39)70.23 compared to a value of 0.72 (95% CI, 0.53–0.90)70.33 for cases without mutations (t-test, P-value, 0.0001). The difference in frequencies of LOH between the two groups was significant only in chromosome 9. In all other chromosomes in which the frequency of LOH had been determined, no statistically significant difference between the cases with and without mutations in the BRAF/N-ras genes was observed. The combined FAL for 33 makers on the five other chromosomes (excluding chromosome 9) was 0.17 (95% CI, 0.09–0.24)70.19 for melanomas with mutations compared to 0.25 (95% CI, 0.13–0.37)70.22 for the cases without mutations (t-test, P-value, 0.24). Five of the 12 melanomas without mutations in the BRAF/N-ras genes showed a potential loss of the entire chromosome 9 (FAL value 1) compared to none in cases with mutations (Table 2). At the level of individual markers (Table 2), a statistically significant decreased frequency of LOH in melanomas with BRAF/N-ras mutations was seen at the markers D9S942 (located within the CDKN2A gene between exons 1a and b; Fisher exact test, P ¼ 0.02), D9S171 (centromeric to CDKN2A; w2 ¼ 3.89, P ¼ 0.05) and D9S280 (located on 9q22; Fisher exact test, P ¼ 0.04). The overall difference in the frequencies of LOH on markers at chromosome 9 in cases with and without BRAF/N-ras mutations was statistically significant (w2 for trend ¼ 46.1, d.f. ¼ 19, P ¼ 0.0005; odds ratio (OR) ¼ 5.07, 95% CI, 2.65–9.76). Interestingly, four of the five melanomas which showed microsatellite instability at the marker loci close to the CDKN2A carried mutations in the N-ras gene. In addition to the difference in the frequency of allelic losses on chromosome 9 between the groups with and without mutations in the BRAF and N-ras genes, we also found that the two groups had a statistically significant difference in patient age at diagnosis. The group with BRAF/N-ras mutation had a statistically significantly lower age at diagnosis than the group without mutations (mean age 65.77 (95% CI, 57.5–74.0)719.79 vs 80.38 (95% CI, 76.4–84.3)77.24 years; t-test, P ¼ 0.02). We did not detect any statistically significant association between mutations in the BRAF/N-ras genes and tumour thickness (data not shown). Oncogene

The data on patient age at diagnosis showed a weak correlation with tumour thickness (r ¼ 0.31, P ¼ 0.06).

Discussion In this study, we have combined the results of mutational analysis of the BRAF and N-ras genes with our results from previous studies on the frequency of LOH from six different chromosomes, and on mutations in the CDKN2A gene in a set of microdissected sporadic primary melanoma (Kumar et al., 1999; Smeds et al., 2000). The salient features of our results are: (a) detection of somatic mutations in either the BRAF or N-ras gene in 63 percent of primary melanomas; (b) detection of mutations in both the BRAF and N-ras genes at the dipyrimidinic sites and (c) inverse association of the activating mutations in the BRAF/N-ras genes with the frequency of monoallelic loss on chromosome 9, determined at 10 different microsatellite loci that included the CDKN2A locus. The predominant mutation detected in the BRAF gene, as in the previous studies, is a single base substitution in the kinase activation domain involving codon 600 (numbered as codon 599 in all previous studies) that probably mimics the activational phosphorylation of the adjacent residues (Zhang and Guan, 2000; Brose et al., 2002; Davies et al., 2002; Pollock and Meltzer, 2002). This residue is identical at the corresponding positions in RAF1 and ARAF1, and is conserved through the evolution with a single exception of Drosophila Raf homologue (Davies et al., 2002; Yuen et al., 2002). The mutant V600E has been reported to possess a 10-fold greater basal activity, and it induces focus formation in NIH3T3 cells with much higher efficiency than the wild-type BRAF (Davies et al., 2002). The valine-to-lysine change at codon 600 (previously 599) through GT1798-1799AA tandem mutation in one melanoma similar to the one in melanocytic nevi (reported as V599K; GT1795-96AA) and in two metastatic melanomas confirms the possibility of BRAF activation through substitution with both positively and negatively charged amino acids (Kumar et al., 2003; Pollock et al., 2003). One additional mutation in the activational segment of the BRAF gene was a novel tandem base CT4TC change at nucleotides 1789–90 (previously 1786–87), resulting in leucine-to-serine amino-acid substitution at a highly conserved residue 597 (previously 596). The only mutation detected in exon 11 changed a glycine into a basic arginine residue at codon 466 (previously 465) within the GXGXXG motif in a glycine-rich loop in the kinase domain of BRAF. Previously reported changes within this codon resulted in replacement of glycine with acidic or neutral residues (Davies et al., 2002; Naoki et al., 2002; Mercer and Pritchard, 2003). The hypothesis that the activation of RAS or BRAF leads to the same phenotype gained support by the mutual exclusivity of mutations in two genes in melanomas and other cancer types (Brose et al., 2002;


9221 Table 2

Loss of heterozygosity at different microsatellite markers located on chromosome 9 and association with activating mutations in the BRAF/N-ras genes CDKN2A/CDKN2B

Telomere

Centromere 9p22

Melanomaa 2 3 5 11 14e 16 17 18 19 20 22 23 25 26 28 29 30 31 32 33 34 35 1f 4 6 7 8 9 10 12 13 15 21 24 27 P-valuesg

IFNAb LOHd I I I I I NI NI NI NI I I LOH NI I I I I I I I LOH LOH LOH LOH NI I I NI NI LOH NI I NI NI 0.13

9p21 D9S736 LOH I I LOH NI I NI NI I I NI LOH NI I NI NI NI LOH I NI LOH LOH LOH LOH LOH LOH LOH NI LOH MSI NI LOH I NI LOH 0.19

D9S974 LOH LOH I LOH NI NI NI NI I LOH I LOH NI MSI I I I NI NI LOH NI LOH LOH LOH NI NI LOH I I NI LOH I NI NI 0.60

D9S942 LOH MSI MSI I I I I I I I I MSI LOH I I LOH I I I I LOH LOH LOH LOH LOH NI LOH I I LOH LOH LOH I I LOH 0.02

D9S171 LOH I I LOH NI LOH NI I I LOH I NI LOH NI I I I I I I I LOH NI LOH LOH LOH LOH LOH NI I NI LOH I LOH LOH 0.05

D9S104 LOH I I NI NI I LOH I NI I NI NI NI NI I I LOH NI NI NI NI LOH LOH NI LOH LOH LOH NI I NI LOH NI I NI NI 0.06

Telomere 9q22.1

9q22.3

D9S257 NI NI I NI LOH I I NI LOH I NI LOH LOH I NI I I LOH NI LOH I NI LOH NI I NI NI LOH LOH

D9S280 I I

D9S287 I I

D9S195 NI I

NI NI NI NI I NI NI NI NI NI NI

I NI I I I LOH I I I I LOH NI I

LOH LOH NI NI NI 0.15

NI LOH I NI NI 0.04

NI NI NI NI NI NI I NI I I I I NI NI NI NI LOH I I NI NI NI NI NI I I NI NI I NI NI LOH 0.50

I NI NI NI NI NI NI NI NI LOH LOH LOH

9q33

NI LOH NI I LOH NI I LOH I NI LOH I NI LOH I LOH NI 0.09

BRAF/N-rasc MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT WT WT WT WT WT WT WT WT WT WT WT WT WT w2 for trend 46.1; df 19; P 0.0005h

a

Melanomas are arranged into two categories, with and without mutations in the BRAF/N-ras gene, and case numbering is the same as in Table 1. Loss of heterozygosity in melanomas described at all markers has been conducted in studies carried out and described earlier. cDetails of mutations in the BRAF and N-ras genes are given in Table 1. dLOH are cases which showed of loss of heterozygosity in tumours; NI are the non-informative cases; I represent the cases that were informative and showed retention of heterozygosity in tumours and MSI represents microsatellite instability, where the tumour showed irregular alleles compared to benign tissues. eMelanomas 14, 31 and 35 carried mutations in the both BRAF and N-ras genes. fMelanomas 1, 4, 7, 13 and 27 showed loss of entire chromosome 9. gP-values are for differences in the frequency of LOH at individual microsatellite markers in melanomas with and without mutations in the BRAF/N-ras genes. hThis P-value represents the overall statistically significant difference in the level of frequency of allelic losses between cases with and without BRAF/N-ras mutation. Bold type indicates LOH b

Yuen et al., 2002; Pollock et al., 2003; Singer et al., 2003). In the present study, we also found mutations in the BRAF and N-ras genes to be mutually exclusive in all but three melanomas. Concomitant detection of both BRAF and N-ras mutations can probably be due to the presence of polyclonal cellular population in these melanomas, as seen previously in melanocytic nevi (Pollock et al., 2003). The ras proteins, with intrinsic GTPase activity, in the activated form stimulate several transducer pathways including RAF-MEK-ERK (Hingorani and Tuveson, 2003). The centrality of BRAF/N-ras in melanoma biology has been underlined

by detection of mutations in histologically diverse melanocytic nevi and the persistence of such mutation through metastasis (Demunter et al., 2001; Pollock et al., 2003). The oncogenic mutations in the kinase domain of BRAF in melanoma cell lines and tumour tissues result in activation of the Raf/MEK/ ERK pathway (Satyamoorthy et al., 2003; Smalley, 2003). This leads to the hypothesis, soundly supported by animal models, that activation of the RAS-RAFMAPK pathway is an important and initiating step in melanoma development (Walker and Hayward, 2002). Oncogene


9222

Mutations in the human N-ras gene have been reported mainly in melanomas on the sun-exposed sites (Jiveskog et al., 1998). The most common sequence alteration in the BRAF gene, in melanomas, is not a typical UV-associated mutation. However, our observation that nine mutations in the N-ras gene involved dipyrimidinic sites, including a novel GG4AA tandem mutation at codon 12 of the N-ras gene, suggest a molecular link between sunlight exposure and melanoma. In addition, we detected another novel dipyrimidinic TC4CT mutation at codon 597 of the BRAF gene. In nonmelanoma skin cancers, UV-induced mutations in the p53 tumour-suppressor gene are common occurrences, but in melanoma these mutations are rather rare (Kumar et al., 2001; Bolshakov et al., 2003). The overriding and dominant role of mutated BRAF and N-ras genes in melanoma is also highlighted by our discovery of a strong inverse association of mutations in these genes with the frequency of allelic loss on chromosome 9. Further association of melanoma cases with BRAF/N-ras mutations with lower patient age at diagnosis partly explains our earlier observation of associating the frequency of allelic loss with increased patient age at diagnosis (Smeds et al., 2000). Since the inverse relationship between allelic loss and occurrence of mutations was confined to chromosome 9, we hypothesize that the activating mutations in the BRAF and N-ras genes represent an important event in melanoma, which compensates the requirement for allelic loss on chromosome 9p21. The allelic loss on chromosome 9p21, independent of mutations in the CDKN2A gene, with haplo-insufficient effect, represents an important step in melanoma initiation (Funk et al., 1998; Pollock et al., 2001). Even though, at present, a search of the genome database does not reveal any such gene, the existence of some, hitherto, unidentified gene on chromosome 9 that is functional in a pathway similar to RAS-BRAF cannot be ruled out. In five melanomas, we observed tumourspecific microsatellite instability and four of these melanomas carried mutations in the N-ras gene. These data, though limited, are contrary to an earlier report on colorectal tumours, which had suggested an association of BRAF mutations with microsatellite instability (Rajagopalan et al., 2002). However, in this study, the data on mutations in the BRAF/N-ras and CDKN2A genes did not show an inverse relationship. In animal models, activated RAS has been shown to cooperate with both Arf/p53 and Ink4a/Rb pathways to induce melanomas (Yang et al., 2001; Sharpless and Chin, 2003). Even though individuals from melanoma-prone families and patients with multiple primary melanomas lack germline mutations in the BRAF gene, at the same time, a higher prevalence of activating N-ras mutations in melanomas from familial cases with germline alterations in the CDKN2A gene than in tumours from sporadic cases has been reported (Eskandarpour et al., 2003; Laud et al., 2003). These observations basically suggest the somatic nature of BRA/N-ras mutations that occur early, as indicated by Oncogene

detection of such mutations in melanocytic nevi (Pollock et al., 2003). The detection and confirmation of mutations in the BRAF/N-ras genes, in a majority of primary melanomas, is a step towards understanding genetics events in the transformation of melanocytes. Our observation of inverse association between these mutations and monoallelic loss at different loci including the CDKN2A cluster on chromosome 9 further underscores the dominant role of such mutations in melanoma biology, that merit further investigation.

Materials and methods Paraffin blocks containing specimens from primary cutaneous melanoma were collected from the archives of the Department of Pathology, Huddinge University Hospital, Sweden. The age of the patients at the time of diagnosis of primary melanomas ranged from 23 to 94 years (median 78). All selected melanomas were in the vertical growth phase according to Clark’s model of tumour progression, and tumour thickness ranged from 1.7 to 11 mm (median 3.9). From available paraffin blocks, 10-mm-thick sections were cut and put on glass slides. The sections were then carefully dissected under a microscope, separating tumour cells and the surrounding benign tissues. DNA was extracted by incubation in a digesting buffer containing Proteinase K, as described earlier (Kumar et al., 1998a). PCR and SSCP analysis Exons 11 and 15 of the BRAF gene and exons 1 and 2 of the N-ras gene were amplified for SSCP analysis using the primers described in Table 3. PCR was carried in 10 ml volume that contained 50 mm KCl, 0.11 mm of each dNTP, 1 mCi [a-32P] dCTP, 0.3 U Taq DNA polymerase, 1–2 mm MgCl2 and 0.15– 0.3 mm of each primer. The temperature for PCR was set as denaturation at 951C for 1 min, annealing (at temperatures specific for each exon as given in Table 3) for 1 min and polymerization at 721C for 1 min for three cycles, followed by 27–33 cycles at the same temperatures with the segment time of 30 s each. In order to increase the sensitivity of mutation detection, amplified fragments were electrophoresed on a 0.5 MDE gel in at least three different conditions, which included different temperature settings and inclusion of glycerol in the gel. Sequence analysis Mutations detected by SSCP in different exons of the BRAF and N-ras genes were identified and confirmed by direct DNA sequencing using Rhodamine dye terminator cycle sequencing kit (Big Dye; Applied Biosystems). Mutations were also confirmed by sequencing the DNA extracted from the aberrantly migrated bands that were excised from the SSCP gels. Individual exons containing mutations were amplified by PCR. The amplified products were purified using Sephadex microspin columns (Amersham-Pharmacia), and subjected to 26 cycles of sequencing reaction using forward or reverse primers separately (Table 3). The precipitated sequencing reaction products were electrophoresed on a denaturing polyacrylamide gel in an automated sequencer (ABI 377, Applied Biosystems), and analysed using Prism and Edit View 1.0.1 software. The sequencing data were analysed by Align software in DNA star package using the reference sequences of


9223 Table 3 Primers used for the amplification of the BRAF and N-ras genes Exon BRAF gene 11 15 N-ras gene 1 2

Primer

Sequence

Annealing temp.

Size (bp)

Forward Reverse Forward Reverse

50 CTC TCA GGC ATA AGG TAA TG 50 CAC TTT CCC TTG TAG ACT GTT 50 CCT AAA CTC TTC ATA ATG CTT 50 ATA GCC TCA ATT CTT ACC AT

53

204

52

209

Forward Reverse Forward Reverse

50 CGC CAA TTA ACC CTG ATT ACT 50 CAC TGG GCC TCA CCT CTA 50 CCC CTT ACC CTC CAC AC 50 AGG TTA ATA TCC GCA AAT GAC

56

174

55

196

the BRAF (accession no. NT_007914) and N-ras (accession no. NT_019273) genes obtained from the NCBI gene databank (http://www.ncbi.nlm.nih.gov). Statistical analysis of association between mutations in the BRAF/N-ras gene and frequency of allelic losses Frequency of LOH at 43 different polymorphic microsatellite markers in melanoma cases, investigated in the present study for mutations in the BRAF/N-ras genes, were determined previously (Kumar et al., 1999; Smeds et al., 2000). The microsatellite markers studied were from chromosomes 1 (six markers), 6 (12 markers), 9 (10 markers), 10 (eight markers), 11 (three markers) and 13 (four markers) (Smeds et al., 2000). The difference in the frequency of LOH, in the form of FAL (defined as the ratio of number of markers with LOH to total number of informative markers), at chromosomes 1, 6, 9, 10, 11 and 13 in melanoma cases with and without mutations in

the BRAF/N-ras genes were compared by using two-sided t-test. The difference in FAL between cases with and without BRAF/N-ras mutations was also determined separately for individual chromosomes. The w2 and Fisher tests were used to determine the differences in allelic losses on individual markers on chromosome 9 between cases with and without BRAF/Nras mutations. The difference in patient age at diagnosis between melanoma cases with and without BRAF/N-ras mutations was also compared by t-test. Since all melanoma in the present series were in the vertical growth phase, the data on patient age at diagnosis were correlated with tumour thickness. All the statistical tests were carried out using Statistica software. Acknowledgements We thank Dr. Barbro Lundh-Rozell for providing melanoma samples originally for the CDKN2A studies.

References Bishop DT, Demenais F, Goldstein AM, Bergman W, Bishop JN, Bressac-de Paillerets B, Chompret A, Ghiorzo P, Gruis N, Hansson J, Harland M, Hayward N, Holland EA, Mann GJ, Mantelli M, Nancarrow D, Platz A and Tucker MA. (2002). J. Natl. Cancer Inst., 94, 894–903. Bolshakov S, Walker CM, Strom SS, Selvan MS, Clayman GL, El-Naggar A, Lippman SM, Kripke ML and Ananthaswamy HN. (2003). Clin. Cancer Res., 9, 228–234. 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 P, Futreal PA, Wooster R, Stratton MR and Weber BL. (2002). Cancer Res., 62, 6997–7000. Busca R, Abbe P, Mantoux F, Aberdam E, Peyssonnaux C, Eychene A, Ortonne JP and Ballotti R. (2000). EMBO J., 19, 2900–2910. Castellano M and Parmiani G. (1999). Melanoma Res., 9, 421–432. Chin L, Merlino G and DePinho RA. (1998). Genes Dev., 12, 3467–3481. 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, PritchardJones 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. Demunter A, Stas M, Degreef H, De Wolf-Peeters C and van den Oord JJ. (2001). J. Invest. Dermatol., 117, 1483–1489. Eskandarpour M, Hashemi J, Kanter L, Ringborg U, Platz A and Hansson J. (2003). J. Natl. Cancer Inst., 95, 790–798. Funk JO, Schiller PI, Barrett MT, Wong DJ, Kind P and Sander CA. (1998). J. Cutan. Pathol., 25, 291–296. Gilchrest BA, Eller MS, Geller AC and Yaar M. (1999). N. Engl. J. Med., 340, 1341–1348. Halaban R. (2000). Pigment Cell Res., 13, 4–14. Hemminki K, Lonnstedt I and Vaittinen P. (2001). Melanoma Res., 11, 133–140. Hingorani SR and Tuveson DA. (2003). Curr. Opin. Genet. Dev., 13, 6–13. Jiveskog S, Ragnarsson-Olding B, Platz A and Ringborg U. (1998). J. Invest. Dermatol., 111, 757–761. Kolch W. (2000). Biochem. J., 351 (Part 2), 289–305. Kumar R, Angelini S, Czene K, Sauroja I, Hahka-Kemppinen M, Pyrho¨nen S and Hemminki K. (2003). Clin. Cancer Res. (in press). Kumar R, Lundh Rozell B, Louhelainen J and Hemminki K. (1998a). Int. J. Cancer, 75, 193–198. Kumar R, Sauroja I, Punnonen K, Jansen C and Hemminki K. (1998b). Genes Chromosom. Cancer, 23, 273–277. Oncogene


9224 Kumar R, Smeds J, Berggren P, Straume O, Rozell BL, Akslen LA and Hemminki K. (2001). Int. J. Cancer, 95, 388–393. Kumar R, Smeds J, Lundh Rozell B and Hemminki K. (1999). Melanoma Res., 9, 138–147. Laud K, Kannengiesser C, Avril M-F, Chompret A, StoppaLyonnet D, Desjardins L, Eychene A, Demenais F, French Hereditary Melanoma Study Group, Lenoir GM and Bressac-de Paillerets B. (2003). Cancer Res., 63, 3061–3065. Mercer KE and Pritchard CA. (2003). Biochim. Biophys. Acta, 1653, 25–40. Naoki K, Chen TH, Richards WG, Sugarbaker DJ and Meyerson M. (2002). Cancer Res., 62, 7001–7003. Omholt K, Karsberg S, Platz A, Kanter L, Ringborg U and Hansson J. (2002). Clin. Cancer Res., 8, 3468–3474. 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. Pollock PM and Meltzer PS. (2002). Cancer Cell, 2, 5–7. Pollock PM, Welch J and Hayward NK. (2001). Cancer Res., 61, 1154–1161. Rajagopalan H, Bardelli A, Lengauer C, Kinzler KW, Vogelstein B and Velculescu VE. (2002). Nature, 418, 934. Saida T. (2001). J. Dermatol. Sci., 26, 1–13.

Oncogene

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. Sharpless E and Chin L. (2003). Oncogene, 22, 3092–3098. Shields JM, Pruitt K, McFall A, Shaub A and Der CJ. (2000). Trends Cell. Biol., 10, 147–154. Singer G, Oldt III R, Cohen Y, Wang BG, Sidransky D, Kurman RJ and Shih Ie M. (2003). J. Natl. Cancer Inst., 95, 484–486. Smalley KS. (2003). Int. J. Cancer, 104, 527–532. Smeds J, Kumar R, Rozell BL and Hemminki K. (2000). Mutagenesis, 15, 257–260. van Elsas A, Scheibenbogen C, van der Minne C, Zerp SF, Keilholz U and Schrier PI. (1997). Melanoma Res., 7 (Suppl 2), S107–S113. Walker GJ and Hayward NK. (2002). J. Invest. Dermatol., 119, 783–792. Whiteman DC, Whiteman CA and Green AC. (2001). Cancer Causes Control, 12, 69–82. Yang FC, Merlino G and Chin L. (2001). Semin. Cancer Biol., 11, 261–268. Yuen ST, Davies H, Chan TL, Ho JW, Bignell GR, Cox C, Stephens P, Edkins S, Tsui WW, Chan AS, Futreal PA, Stratton MR, Wooster R and Leung SY. (2002). Cancer Res., 62, 6451–6455. Zhang BH and Guan KL. (2000). EMBO J., 19, 5429–5439.