Screening of N-ras Codon 61 Mutations in Paired Primary and Metastatic Cutaneous Melanomas: Mutations Occur Early and Persist throughout Tumor Progression¹

Cutaneous malignant melanoma is the most fatal of skin cancers. The successive steps in tumor progression of cutaneous melanoma have been described (1). It has been demonstrated that superficial spreading melanomas, which may arise from dysplastic nevi, display an early radial growth phase where tumor cells are confined to the epidermis and lack invasive capacity. This may be followed by a vertical growth phase characterized by invasion of melanoma cells into the dermis. These cells have also acquired metastatic capacity. Metastatic melanoma is frequently resistant to radio- and chemotherapy and, consequently, has a high mortality rate. The disease seems to be genetically heterogeneous, and mutations in a variety of genes have been reported. Thus far, some of the most frequently found alterations are ras-activating mutations.

The human ras proto-oncogenes (H-ras, K-ras, and N-ras) encode cell membrane-associated proteins that are involved in transduction of extracellular growth and differentiation signals. The ras proteins have GTPase activity, and as such, they cycle between an inactive GDP-bound form and an active GTP-bound form. Once activated, ras stimulates several different effector pathways, one of which is the raf-MEK-ERK pathway (2).

Mutations in the ras genes frequently occur during tumorigenesis. Different tumor types are often associated with mutations in a distinct ras gene. Thus, high frequencies of K-ras alterations have been found in carcinomas of the pancreas, colon, and lung, whereas N-ras mutations are frequent in myeloid leukemias (3). The codons at which most activating ras mutations occur include codons 12, 13, and 61. Mutations in these codons reduce the intrinsic GTPase activity of the ras proteins and also make them insensitive to GTPase-activating proteins (4, 5, 6). As a result, mutated ras is locked in the GTP-bound state and continuously activates its downstream effector targets.

In sporadic malignant melanoma, the frequencies of ras mutations reported in the literature vary between 5 and 36% (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18). The discrepancy in mutation frequency may partly be attributable to the use of different screening techniques but may also reflect differences in the composition of melanoma tumors included in the studies. Indeed, ras mutations have been found more frequently in melanomas from chronically sun-exposed body sites (8, 11, 16, 17) and have also been associated with nodular melanomas in some studies (14, 16). In melanoma, the vast majority of activating ras mutations has been identified in the N-ras gene, whereas mutations in the H-ras and K-ras genes are rare. Only a single study has reported a frequent occurrence of K-ras mutations in melanoma (10). The most frequent N-ras mutations in melanoma are codon 61 mutations, whereas codon 12 and 13 mutations are less common. Recently, a novel mutation in N-ras codon 18 has been reported in a subset of primary melanomas (18).

Although it is well documented that N-ras codon 61 mutations are the most common ras alterations in malignant melanoma, there are no studies where large series of paired primary and metastatic lesions have been analyzed. Therefore, to evaluate the role of N-ras codon 61 mutations in melanoma initiation and progression, we have screened a large set of paired primary and metastatic melanomas. For this study, we identified primary and metastatic tumors from 53 patients with cutaneous melanoma. Primary tumors from an additional 19 patients whose metastases were not available for analyses were also included in the study. Our purpose was to measure the frequency of activating N-ras codon 61 mutations and to correlate mutations in the primary tumors to those in the corresponding metastases. N-ras mutations were also related to histopathological and clinical parameters, including patient survival.

Seventy-four primary cutaneous malignant melanomas from 72 patients who were followed up at the Department of Oncology, Karolinska Hospital, were included in this study. In all patients, the disease had progressed to the metastatic stage, and from 53 of the patients, a total of 88 metastases were available for analysis. All specimens were obtained as surgical excision biopsies, which had been formalin fixed and paraffin embedded. Forty-two of the primary tumors were SSMs,⁴ 29 were nodular melanomas, and 1 was a lentigo maligna melanoma. In 2 primary tumors, the histogenic type was not classifiable. The median tumor thickness was 2.55 mm (range, 0.5–25 mm), and the levels of tumor invasion according to Clark were II in 3 tumors, III in 29 tumors, IV in 33 tumors, V in 7 tumors, and unknown in 2 tumors. Of the metastases, 50 involved lymph nodes, and 37 involved skin and s.c. tissue, whereas in 1 metastasis, the site was unknown. Of the 53 patients where both primary tumors and metastases were analyzed, a single metastasis was analyzed in 36 cases. In the remaining 17 patients, multiple metastases were analyzed: 2 metastases in 8 cases, 3 in 3 cases, 4 in 4 cases, and 5 and 6 metastases were each analyzed in a single patient.

Tumor tissue was manually dissected from 20-μm sections of the paraffin blocks using parallel H&E-stained sections to localize tumor cell areas. Dissected tissue was deparaffinized in xylene and incubated in PK2 buffer [50 μg/ml proteinase K (Boehringer Mannheim, Germany), 10 mm Tris (pH 8.3), 1 mm EDTA, and 0.5% Tween 20] at 56°C overnight. After inactivation of proteinase K at 95°C for 10 min, DNA was extracted using the Wizard DNA clean-up system (Promega Corp., Madison, WI). Metastases showing lymphocytic infiltrates difficult to avoid by manual dissection were subjected to laser capture microdissection using a PixCell LCM system (Arcturus Engineering, Mountain View, CA). Likewise, in addition to manual dissection, a limited number of the primary tumors were laser capture microdissected to separate the radial and vertical growth phases as well as tumor-associated nevi. In these cases, tumor tissue was dissected from three to four consecutive 5-μm sections, and DNA was extracted according to the manufacturer’s instructions.

A 118-bp fragment of exon 2 of the N-ras gene was amplified by nested PCR. Table 1 shows the sequences of the primers used. In the first PCR, 25 ng of genomic DNA were amplified in a 10-μl mixture containing 1× PCR buffer, 2.5 mm MgCl₂, 0.2 mm of each deoxynucleotide triphosphate, 5 pmol of each primer, and 0.25 unit of AmpliTaq polymerase (Applied Biosystems, Foster City, CA). Samples were denatured at 94°C for 4 min; amplified for 30 cycles of 94°C for 30 s, 56°C for 30 s and 72°C for 30 s; and finally extended at 72°C for 7 min. The PCR products were diluted 1:20, and 1 μl of the dilution was used for the nested PCR. In the nested PCR, 3 μCi of [α-³²P]dCTP was included. The conditions for the second PCR were the same as for the first PCR, except that the annealing temperature was 60°C, and amplification was performed for 20 cycles.

For SSCP analysis, the PCR products were denatured in denaturing buffer (95% formamide, 10 mm NaOH, 0.05% bromphenol blue, and 0.05% xylene cyanol) at 92°C for 10 min. Electrophoresis was carried out on 7.5% nondenaturing acrylamide gels with 10% glycerol at 18°C. Shifted bands were cut out from the gels, reamplified, and purified using the QIAquick Gel Extraction kit (Qiagen, Hilden, Germany). Nucleotide sequence analyses were carried out using the BigDye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems) and an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). Sequence analyses were performed in both directions, using the nested primers as sequencing primers. Mutations were confirmed by two independent PCR/SSCP analyses.

To determine the lowest proportion of mutated cells that is required for SSCP to detect an N-ras codon 61 mutation, dilution series of wt DNA from the melanoma cell line 397 and DNA from either the melanoma cell line 224 (codon 61 Arg mutation) or the fibrosarcoma cell line HT-1080 (codon 61 Lys mutation) were analyzed. We found that 10% of mutated cells are enough for detection of mutations (data not shown).

To evaluate possible relationships between N-ras codon 61 mutations and histopathological or clinical parameters, either the χ² test or Student’s t test was used. Survival analyses were carried out using the life table method and Cox regression analysis. P ≤ 0.05 was regarded as statistically significant.

Seventy-four primary malignant melanomas and 88 metastases, originating from 54 of the primary tumors, were screened for mutations in exon 2 of the N-ras gene using SSCP and nucleotide sequence analyses. Twenty-one of the 74 primary tumors (28%) and 33 of the 88 metastases (37.5%) showed band shifts indicative of mutations on the SSCP-gels (Fig. 1). DNA sequence analysis confirmed the presence of mutations in all of these cases and revealed that they were all located in codon 61 of the N-ras gene.

Of the 21 mutated primary tumors, 10 (48%) had CAA(Gln)-AAA(Lys) mutations, 9 (43%) CAA(Gln)-CGA(Arg) mutations, 1 (4.5%) a CAA(Gln)-CAT(His) mutation, and 1 (4.5%) a tandem CAA(Gln)-TTA(Leu) mutation. From 20 of the mutated primary tumors, a total of 34 metastases were available for analysis, and all but one metastasis showed the same mutation as the primary tumor from which they originated (Table 2). The 33 mutated metastases included 12 metastases with the AAA(Lys) mutation, 19 with the CGA(Arg) mutation, 1 with the CAT(His), and 1 with the TTA(Leu) mutation. From 7 of the patients with mutated primary tumors, 2–6 separate metastatic tumors were analyzed. It is worth noting that in all of these patients with multiple metastases, every single metastasis had the same mutation as detected in the corresponding primary tumor.

To better determine at what stage in the primary lesions the N-ras codon 61 mutations occur, 10 of the 21 mutated primary tumors were further analyzed (Table 3 and Fig. 2). Using a laser capture microscope, the rgp and vgp, respectively, were dissected and subjected to mutational analyses. The analyses showed that the N-ras codon 61 mutations could be detected both in the rgp and vgp. In two cases, the primary tumors were associated with adjacent nevi (a dysplastic and compound nevus, respectively). In both cases, the N-ras mutation could already be detected in these lesions.

Fifty-three of the 74 primary tumors (72%) were wt for N-ras exon 2. From 34 of these primary tumors, a total of 54 metastases were available for analysis, and they all showed a wt genotype (Table 2). In many metastases, lymphocytic infiltrates were present. However, manual dissection, or in some cases laser capture microdissection, was used to obtain material enriched for tumor cells, and in all metastases showing a wt genotype, the presence of lymphocytes was reduced to ≤10%. Therefore, a non-tumor cell population should not have prevented the detection of possible mutations in the melanoma cells, but rather that the metastases are truly wt.

In 1 patient, the N-ras mutation found in the primary tumor was not detectable in the metastatic lesion. Instead, this metastasis showed a wt genotype (Table 2). This metastasis was heavily infiltrated by lymphocytes, which were impossible to eliminate by any dissection method. However, the proportion of tumor cells was still ≥10%, and a possible mutation should therefore have been detected. Thus, overall the N-ras exon 2 genotype of the primary tumor was retained in 87 of the 88 metastases (99%), and multiple metastases in the same patient always shared the same N-ras exon 2 genotype.

For mutated tumor samples, the SSCP analysis provides information about the relative abundance of wt and mutant alleles. The majority of patients displayed balanced band intensities for the wt and mutant alleles. However, 3 of the 21 patients with mutated tumors showed a marked decrease of the wt allele in comparison with the mutant allele. In two of these patients, the wt allele was lost in both the primary tumor as well as in the metastases (Fig. 1), whereas in 1 patient the wt loss was only obvious in the metastasis.

From Table 4, it can be seen that the Clark level of tumor invasion and age at diagnosis were the only clinical parameters that significantly differed between patients with tumors that were wt for N-ras codon 61 and those with tumors that were mutated. N-ras codon 61 mutations were associated with a higher Clark level of invasion (P = 0.012) and a lower age at diagnosis (P = 0.042). None of the other parameters analyzed, including gender, histogenic type, ulceration, and tumor thickness differed between the two patient groups.

The univariate survival analyses showed that N-ras codon 61 mutations had no impact on patient overall survival in our patient material (Table 5). From Table 5, it can further be seen that neither gender, ulceration, nor age at diagnosis had an impact on survival. In contrast, the univariate analyses showed that clinical stage at diagnosis, histogenic type, level of tumor invasion according to Clark, and tumor thickness according to Breslow all affected overall survival. In the multivariate analysis, tumor thickness and clinical stage were the only independent prognostic factors with respect to overall survival (Table 6).

In this study, a large series of paired primary and metastatic melanomas has been analyzed for N-ras codon 61 mutations to obtain a clearer picture of the role of these mutations in melanoma tumor initiation and progression. The study was limited to exon 2 of the N-ras gene, because several earlier studies pointed to N-ras codon 61 mutations as the most common ras alterations in human melanomas (7, 8, 9, 11, 12, 13, 14, 15, 16, 17).

We found that N-ras codon 61 mutations first appeared in the primary tumors rather than in the metastases, indicating that they are early events in melanoma tumor development. In total, 21 of 74 primary tumors (28%) had activating N-ras codon 61 mutations. Substitution of glutamine by lysine or arginine were the two most common alterations, detected in 10 and 9 primary tumors, respectively. Substitution of glutamine by histidine or leucine each occurred in one primary tumor. Thus, the frequency of N-ras mutations in our study does not diverge from previous reports on melanoma tumors. Similarly, the lysine and arginine mutations, which predominated in our patients, belong to the most common codon 61 alterations. However, the CAA(Gln)-TTA(Leu) tandem mutation identified in 1 primary tumor and its corresponding metastasis has, to our knowledge, not been detected previously in melanomas.

To better determine at what stage in the primary lesions the N-ras codon 61 mutations occur, the different growth phases of the mutated primary melanomas were analyzed. We could show that the mutations are present already in the rgp and are preserved in the vgp. Furthermore, in 2 cases, the primary tumor was associated with a nevus (a dysplastic and compound nevus, respectively), and in both cases the mutation was present already at this stage. The demonstration of N-ras mutations in melanoma-associated nevi indicates that such mutations are already present in precursor lesions and may be among the initial genetic alterations leading to melanoma development. Similar results, i.e., detection of ras mutations in the rgp of primary melanomas as well as in associated nevi, has been reported recently by Demunter et al. (18, 19).

UV light has been proven to induce activation of the N-ras gene (20). The mutations that were detected in the present study may all be explained by UV-induced DNA lesions; the CAA(Gln)-AAA(Lys) mutation may arise from mispairing of adenosine with 8-oxo-deoxyguanosine, whereas the other mutations all occurred opposite dipyrimidine sequences. Together, our results indicate that the activating N-ras codon 61 mutations occur early during melanoma development as a result of UV irradiation.

This study shows clearly that N-ras codon 61 mutations are preserved throughout melanoma tumor progression, once they have occurred in the primary lesions. This was demonstrated by the fact that metastases belonging to 19 of the 20 mutated primary tumors showed the same mutation as the primary tumor from which they originated. Furthermore, in all cases where multiple metastases in the same patient were analyzed, the mutation was detected in every single metastasis. From 1 patient as many as six metastases were available, and all of them retained the arginine mutation found in the primary tumor. Together, these results indicate that N-ras codon 61 mutations are stable during melanoma tumor progression, and furthermore, that the metastasizing cells are recruited from cell clones of the primary tumors carrying the N-ras mutations, and that in patients with multiple metastases, there is a clonal relationship between the different metastases.

In none of the patients with primary tumors that were wt for N-ras codon 61 did mutations arise in the metastases. This indicates that N-ras codon 61 mutations are not the critical event for metastasis initiation. That N-ras mutations do not play a major role in metastasis initiation has been suggested already in previous studies. In these studies, the mutation frequency between unrelated primary and metastatic melanomas were compared, and mutations were found to occur no more commonly in the metastases (9, 11, 16). Although ras mutations may be without importance for metastasis initiation, our finding that they are preserved in the metastases suggests that they may still have an important influence on tumor progression in the subgroup of melanomas in which they occur.

Our finding that in a subset of melanoma patients N-ras codon 61 mutations are early events that may be of importance for tumor progression and clonal expansion is in good agreement with the results of two experimental studies by Chin et al. (21), in which transgenic mice models have been used to determine the effect of activated ras on melanoma development. In mice that are deficient for Ink4a, melanocyte-specific expression of H-ras^G12V leads to development of nonmetastatic melanomas at high frequency (21). Furthermore, in a model where the H-ras^G12V expression was under control of doxycycline, the same investigators showed that tumors regressed upon doxycycline withdrawal (22). These experiments demonstrate that activated ras has a role not only in melanoma initiation but also in melanoma tumor maintenance.

To date, only a few other studies have analyzed paired primary and metastatic melanomas for ras mutations. In all cases, the number of patients has been limited. Ball et al. (11) did not find a relationship between ras mutations in primary tumors and metastases, whereas in two other studies, the ras mutations present in the primary tumors were also detected in the corresponding metastases (8, 18). Thus, our study of paired primary and metastatic melanomas from 53 different patients constitutes the most extensive study performed and shows clearly that there is a strong relationship between the N-ras genotype of primary melanomas and corresponding metastatic lesions.

In SSCP analyses of DNA extracts from the majority of tumors with N-ras mutations, the mutant and wt bands had similar intensities. This finding demonstrates that the N-ras codon 61 mutations are present in the majority of melanoma cells, not only in a minor fraction of them. Moreover, it indicates that most tumor cells are heterozygous and carry one mutant and one wt allele. There was evidence of loss of the wt allele only in rare cases. This corresponds well with the documented effect of mutated N-ras as a dominant oncogene.

We found that the N-ras codon 61 mutations had no significant influence on patient overall survival. However, our series of melanoma patients were selected to consist exclusively of patients who had developed metastases and were thus enriched for patients with a poor prognosis. Therefore, this set of patients is not optimal for an analysis of the possible prognostic impact of N-ras codon 61 mutations. This fact may also explain the lack of significant influence of previously documented prognostic factors, such as tumor ulceration and gender on patient survival in our study. Despite this, our results are similar to those reported by Van Elsas et al. (16); the classical N-ras codon 12, 13, and 61 mutations do not have an impact on survival. In contrast, the N-ras codon 18 mutation, which was recently detected in a subset of primary melanomas, seems to significantly improve patient outcome, because tumors carrying this mutation seem to lack metastasizing capacity (18).

In summary, by screening a large series of paired primary and metastatic melanomas for N-ras codon 61 mutations, we showed both that the mutations arise early in melanoma development, because they were detected in the rgp as well as in precursor lesions, and also that they are preserved during tumor progression into the vgp and metastatic lesions. The recent finding that B-raf mutations occur at high frequency in melanomas (23) further supports the importance of the ras-raf-MEK-ERK pathway in melanoma. This makes activated N-ras and the ras-signaling pathway attractive targets for future therapeutic trials in melanoma patients. Induction of immune responses against mutated ras (24) as well as the use of chemotherapeutic agents acting on ras (25, 26) may be successful therapeutic strategies.

We thank Bo Nilsson for help with statistical analyses.