A fast-growing melanoma developed within 3 weeks and was the fourth to occur in patient M0881-01. Clinical picture of a 4-mm-diameter nodular lesion located on the elbow (A); dermoscopic image of the lesion showing hypopigmentation, asymmetry, unspecific pattern, atypical vessels, and blue-whitish veil (B). Under confocal microscopy, the lesion shows an ulcerated central area (C) with atypical nests in upper dermis with bright roundish nucleated cells in noncohesive nests (D). Histopathologic examination shows an ulcerated nodular melanoma (hematoxylin-eosin, original magnification ×2) (E) and nests of atypical cells and presence of mitosis (hematoxylin-eosin, original magnification ×10) (F).
The back of patient M3879-01 with 2 previous melanomas and more than 200 nevi. Six dermoscopic images show the predominant pattern, reticulated dark brown.
eTable 1. Primary Melanoma Characteristics From Patients With MITF p.E318K
eTable 2. Summary of the MITF p.E318K Reports in Melanoma Patients
Customize your JAMA Network experience by selecting one or more topics from the list below.
Potrony M, Puig-Butille JA, Aguilera P, et al. Prevalence of MITF p.E318K in Patients With Melanoma Independent of the Presence of CDKN2A Causative Mutations. JAMA Dermatol. 2016;152(4):405–412. doi:10.1001/jamadermatol.2015.4356
Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
The main high-penetrance melanoma susceptibility gene is CDKN2A, encoding p16INK4A and p14ARF. The gene MITF variant p.E318K also predisposes to melanoma and renal cell carcinoma. To date, the prevalence of MITF p.E318K and its clinical and phenotypical implications has not been previously assessed in a single cohort of Spanish patients with melanoma or in p16INK4A mutation carriers.
To evaluate the prevalence of MITF p.E318K in Spanish patients with melanoma and assess the association with clinical and phenotypic features.
Design, Setting, and Participants
A hospital-based, case-control study was conducted at the Melanoma Unit of Hospital Clinic of Barcelona, with MITF p.E318K genotyped in all patients using TaqMan probes. We included 531 patients: 271 patients with multiple primary melanoma (MPM) without mutations affecting p16INK4A (wild-type p16INK4A); 191 probands from melanoma-prone families with a single melanoma diagnosis and without mutations affecting p16INK4A, and 69 probands from different families carrying CDKN2A mutations affecting p16INK4A. A population-based series of 499 age- and sex-matched cancer-free individuals from the Spanish National Bank of DNA were included as controls. Patients were recruited between January 1, 1992, and June 30, 2014; data analysis was conducted from September 1 to November 30, 2014.
Main Outcomes and Measures
The genetic results of the MITF p.E318K variant were correlated with clinical and phenotypic features.
Among the 531 patients, the prevalence of the MITF p.E318K variant was calculated among the different subsets of patients included and was 1.9% (9 of 462) in all melanoma patients with wild-type p16INK4A, 2.6% (7 of 271) in those with MPM, and 2.9% (2 of 69) in the probands of families with p16INK4A mutations. With results reported as odds ratio (95% CI), the MITF p.E318K was associated with an increased melanoma risk (3.3 [1.43-7.43]; P < .01), especially in MPM (4.5 [1.83-11.01]; P < .01) and high nevi count (>200 nevi) (8.4 [2.14-33.19]; P < .01). Two fast-growing melanomas were detected among 2 MITF p.E318K carriers during dermatologic digital follow-up.
Conclusions and Relevance
In addition to melanoma risk, MITF p.E318K is associated with a high nevi count and could play a role in fast-growing melanomas. Testing for MITF p.E318K should not exclude patients with known mutations in p16INK4A. Strict dermatologic surveillance, periodic self-examination, and renal cell carcinoma surveillance should be encouraged in this context.
Approximately 8% to 12% of melanoma cases occur in a familial context.1 The main high-penetrance gene implicated in melanoma susceptibility is CDKN2A (OMIM code: 600160; GenBank accession number: NM_000077.4 [p16INK4A] and NM_058195.3 [p14ARF]). The gene encodes 2 tumor suppressor proteins: p16INK4A, which promotes cell-cycle arrest and plays a role in senescence, and p14ARF, which acts through p53-regulating apoptosis.2,3 Germline CDKN2A mutations are found in 20% to 40% of melanoma-prone families4 and in 8% to 16% of patients with multiple primary melanomas without other cases in the family.5 Patients with melanoma carrying CDKN2A mutations have a younger age of onset and a higher number of primary melanomas, and the presence of CDKN2A mutations is associated with dysplastic nevi.4-7
The MC1R gene (OMIM: 155555; GenBank: NM_002386) controls the pigmentation process and is a moderate-risk gene for melanoma susceptibility.8,9 Loss of function variants in MC1R impairs the ability to activate the pigmentation pathway resulting in the red-hair color (RHC) phenotype. The RHC phenotype is characterized by fair pigmentation (fair skin, red hair, and freckles) and by sun sensitivity (poor tanning response and solar lentigines).10 These variants increase the risk of melanoma with an odds ratio (OR) between 1.5 and 4.1.8,10,11 The association between MC1R variants and melanoma is stronger in individuals with dark skin or few nevi.12,13 Therefore, there might be a modest benefit to measure MC1R genotype for melanoma risk prediction, in addition to clinically measured pigmentation characteristics and nevi count.12 A rare functional variant in MITF, p.E318K (rs149617956), was identified in 2 independent studies.14,15 This variant may be considered as a moderate-risk allele in melanoma.14-17 The master regulator gene of melanocyte development and differentiation is MITF, and it is also associated with melanoma development and progression.18 The presence of MITF p.E318K predisposes to both familial and sporadic melanoma susceptibility, and/or renal cell carcinoma (RCC), and/or pancreatic cancer.14-16MITF p.E318K occurs at a conserved small ubiquitinlike modifier position, and this variant decreases the amount of small ubiquitinlike modifier–modified MITF forms.14 The small ubiquitinlike modifier of MITF represses its transcriptional activity; therefore, p.E318K increases MITF transcriptional activity and may result in the upregulation of distinct sets of genes. Furthermore, this variant promotes invasive and tumorigenic behaviors in melanoma and RCC cells and might favor a phenotypic switch of melanoma cells toward a tumor-initiating cell phenotype.14
To our knowledge, the role of the MITF p.E318K has not been previously explored in Spanish patients with melanoma. The aim of this study was to evaluate the role of the MITF p.E318K variant in Spanish patients with melanoma and assess the association of this variant with clinical and phenotypic features.
A total of 531 patients at high risk of melanoma, recruited from January 1, 1992, to June 30, 2014, at the Melanoma Unit of Hospital Clinic of Barcelona, were included in the study. Patients were grouped into 3 different subsets. The first set was composed of 271 individuals (51%) with multiple primary melanoma (MPM) (212 sporadic MPM and 59 familial MPM) who did not carry mutations in CDKN2A affecting the p16INK4A protein (hereinafter referred to as wild-type p16INK4A). One patient from this set carried a CDKN2A mutation affecting p14ARF, and 75 patients were previously included in the Bertolotto et al14 study. The second set consisted of 191 probands (36%) from melanoma-prone families, with at least 2 melanoma cases, with a single melanoma diagnosis (all wild-type p16INK4A). The third set contained 69 probands (13%) from families bearing CDKN2A mutations affecting p16INK4A, independent of the number of primary melanomas.
Clinical and phenotypic characteristics were collected for most of the 531 patients, including the number of primary melanomas (99%), age of onset (90%), melanoma subtype (76%), melanoma location (84%), Breslow thickness (74%), eye and hair color (75%), skin phototype (80%), and nevus count (72%). The familial history of pancreatic cancer was obtained from 80% of the patients, and personal history of other cancers was supplied by the patients carrying the variant.
The study was approved by the ethics committee of the Hospital Clinic of Barcelona, and patients provided written informed consent. Participants did not receive financial compensation.
A population-based series of 499 cancer-free individuals recruited at the Spanish National Bank of DNA were used as controls. The controls were sex and age matched with a group of consecutively recruited individuals with sporadic melanoma in the Melanoma Unit of Hospital Clinic of Barcelona from January 1998 to December 2013. The mean (SD) age in the control group was 52.2 (18.0) years. Overall, 269 of the 499 patients were women (53.9%) and 230 were men (46.1%).
The MITF variant p.E318K (rs149617956) was analyzed (Custom TaqMan SNP Genotyping Assays) according to manufacturer’s recommendations in all of the patients and the control group. The process was carried out using polymerase chain reaction (7900HT Fast Real Time PCR System; Applied Biosystems) and SDS, version 2.4, software (Applied Biosystems).
The prevalence of MITF p.E318K was assessed in all melanoma patients with wild-type p16INK4A and the CDKN2A mutation carrier set. In a French cancer-free and Italian control population, the frequency of carriers was 0.6% (14 of 2205).14,16 The risk conferred by MITF p.E318K to melanoma development in the Spanish patients with wild-type p16INK4A melanoma was evaluated by comparing our group of patients with the Spanish and previously reported French and Italian controls together since there were no statistically significant differences between them. Clinical and phenotypic characteristics were analyzed regarding the presence of the p.E318K variant in patients with wild-type p16INK4A. Odds ratios and 95% CIs were calculated. The 2-sided Fisher exact test was used to look for statistical significance in proportion comparison. Age of onset was tested using an unpaired, 2-tailed t test. Breslow thickness was evaluated using the Mann-Whitney test. The results were considered statistically significant at P < .05. Statistical analyses were conducted using SPSS, version 17.0 (SPSS Inc). Data analysis was conducted from September 1 to November 30, 2014.
CDKN2A gene, coding for the p16INK4A and p14ARF proteins, is a high-penetrance susceptibility in melanoma. Therefore, we calculated the prevalence of the MITF p.E318K variant in all 531 patients separated according to p16INK4A status (wild-type or mutated): 462 patients with wild-type p16INK4A and 69 patients with mutated p16INK4A. Among these, the prevalence of the MITF p.E318K variant was 1.9% (9 of 462) in all melanoma patients with wild-type p16INK4A, 2.6% (7 of 271) in those with MPM, and 2.9% (2 of 69) in the probands of families bearing a mutation in p16INK4A. All individuals with MITF p.E318K carried the variant in heterozygosis. The prevalence of the variant in a Spanish cancer-free population was 0.4% (2 of 499), with no statistically significant difference with French or Italian controls (P = .54).14,16 With results reported as OR (95% CI), the MITF variant p.E318K increased the risk of developing melanoma in all melanoma patients with wild-type p16INK4A (3.3 [1.43-7.43]; P < .01) and in those with MPM (4.5 [1.83-11.01]; P < .01), using the Mediterranean controls (Table 1). When calculating the OR using, as a control population, the European non-Finnish population from the ExAC/Broad Institute exome database (http://exac.broadinstitute.org/, which gives an MITF p.E318K allele frequency of 0. 21% [140 MITF p.E318K alleles for 66732 total allele number]), the risk of developing melanoma in MITF p.E318K carriers was 4.7 (2.40-9.35; P < .01) and the risk of developing MPM was 6.3 (2.93-13.63; P < .01). We assessed the association between clinical and phenotypic features and the presence of p.E318K. The presence of the variant was associated with a very high nevi count (>200) in all patients with wild-type p16INK4A (8.4 [2.14-33.19]; P < .01) and in those with MPM (12.4 [2.58-59.7]; P < .01). We did not find any association with other phenotypical characteristics, family history of pancreatic cancer (Table 1), or clinicohistopathologic characteristics of tumors (Table 2 and eTable 1 in the Supplement).
Table 3 reports the detailed clinical, phenotypic, and genetic characteristics of all patients with melanoma carrying MITF p.E318K. Patient M0881-01, with 3 previous melanomas, developed a fast-growing nodular melanoma that had not been present in a visit 2 months earlier that included total body photography. The patient detected a fast-growing hypopigmented lesion on the elbow that arose 3 weeks before an urgent evaluation at our unit (Figure 1). With dermoscopy, the lesion showed an unspecific pattern, with asymmetry in the distribution of colors and structures, the presence of blue-gray color, and milky-red areas with some vessels. In confocal microscopy, the lesion showed some bright, large, round cells in the upper epidermis around a central ulceration; in the dermoepidermal junction, papilla were not well demarked and not visible in some areas with dermal nests of noncohesive bright cells with large nuclei, highly suggestive of melanoma. The lesion was excised the same day with the final diagnosis of nodular melanoma with Breslow thickness of 1.3 mm, ulceration, 5 mitoses/mm2, and epithelioid cells. Wide excision was performed and sentinel lymph node biopsy identified 3 negative sentinel lymph nodes.
Patient M1340-01, in follow-up for recurrent lentigo maligna melanoma, had also developed a fast-growing melanoma, in this case amelanotic, that had not been present in a visit 4 months earlier. Under dermoscopy, the lesion also showed an unspecific pattern with remnants of pigmentation, the presence of doted and linear irregular vessels, and short white streaks. The lesion was excised and the diagnosis was superficial-spreading melanoma in a vertical growth phase with a Breslow thickness of 1.65 mm, lack of ulceration, 3 mitoses/mm2, and fusocellular morphology. Wide excision was performed, and sentinel lymph node biopsy identified 3 negative sentinel lymph nodes.
Patient M3879-01 belonged to a family with 2 cases of melanoma; thus, we assessed whether the other patient carried the variant. In this case, MITF p.E318K did not segregate with melanoma; however, the carrier was younger at the time of diagnosis (30s vs 70s). We also detected 2 healthy individuals carrying the variant in this family: M3879-03 and M3879-09 (phenotypic features of those 2 individuals are also recorded in Table 3). The nevi from carriers followed a reticular pattern and were dark brown (Figure 2).
In this study, we analyzed the prevalence of MITF p.E318K in Spanish patients with melanoma. To our knowledge, this is the first study in which a set of individuals bearing mutations in CDKN2A affecting p16INK4A was also tested for MITF p.E318K. We detected a prevalence of 1.9% of the variant in all patients with wild-type p16INK4A, which was higher in the MPM subgroup (2.6%), and a similar prevalence was found in the set of patients with the p16INK4A mutation (2.9%). Previous studies14-17,19 reported that MITF p.E318K increases the risk of developing melanoma (eTable 2 in the Supplement). We also detected this association in our set of patients. In addition to reporting increased melanoma risk, Yokoyama and colleagues15 stated that the presence of this variant was associated with a high nevi count in an Australian and UK population. We also found that MITF p.E318K is associated with a very high nevi count (>200 nevi) in a Mediterranean population. These findings suggest that MITF may be involved in nevogenesis. Twin studies20-22have revealed evidence that the nevi count is genetically determined, with an additive genetic variance of 36% to 84%, increasing with age. The nevi count is a polygenic trait determined by multiple alleles.23-27 The present study and the results in Australian and UK populations indicate that MITF p.E318K should be included in the set of known genes involved in this phenotypical trait.
To our knowledge, only one study19 has described the phenotypical features and dermoscopic pattern of nevi from MITF p.E318K carriers in Australian patients. The investigators observed that these carriers had pink or light brown nevi, suggesting that MITF p.E318K could modulate nevi pigmentation. In contrast, in our study, the nevi were dark brown, indicating that other genes may be involved in this feature. The MC1R gene may modulate the pigmentation of the nevi since melanomas from carriers of RHC variants are less pigmented.28 Otherwise, in the Australian study19 and ours, the dermoscopic pattern of nevi present in MITF p.E318K carriers was predominantly reticular. These findings are suggestive of photoinduced nevogenesis.29
Sturm and colleagues19 noticed a high incidence of amelanotic melanoma within MITF p.E318K carriers. One of our patients (M1340-01) also developed this type of melanoma. The patient carried one MC1R RHC variant: p.R151C. However, although our findings support the hypothesis that a genetic interaction between MC1R RHC and MITF p.E318K could increase the risk of developing amelanotic melanomas, it has been reported17 that the interaction of MITF and MC1R variants is not associated with melanoma pigmentation.
Ghiorzo and colleagues16 found an association between MITF p.E318K and the presence of nodular melanomas. We did not detect a significant association with any clinicohistopathologic features, probably because the sample size lacked the power to detect these possible associations. However, we noted 2 fast-growing melanomas in 2 MITF p.E318K carriers who were receiving dermatologic surveillance owing to a previous melanoma diagnosis. Dermatologic digital follow-up has been demonstrated30,31 to be relevant for detecting melanomas at early stages with a low rate of excisions in patients at high risk to develop melanoma. During 10 years of dermatologic surveillance of patients at high-risk of melanoma in our melanoma unit (from January 1, 1999, to December 31, 2008), 98 new melanomas were diagnosed in these patients; 54% were in situ melanoma and 46% were invasive melanoma. Among the invasive melanomas diagnosed, none was more than 1-mm Breslow thickness and no melanomas behaved as fast-growing melanomas.31,32 Until now in our melanoma unit, the only 2 fast-growing melanomas identified by dermatologic digital follow-up in individuals at high risk of melanoma were in MITF p.E318K carriers. Fast-growing melanomas are defined by having a growth rate of greater than 0.4 mm per month; in general, the melanoma growth rate is approximately 0.1 mm per month, and slow-growing melanomas usually have a growth rate of 0.01 mm per month.33 Furthermore, a high growth rate is associated with a worse prognosis in melanoma; thus, strategies for early detection of fast-growing melanomas are necessary.34 Although further studies should address the role of MITF p.E318K in fast-growing melanoma, the carriers of MITF p.E318K should be encouraged to perform monthly total-body self-examination of the skin and receive fast-track, urgent dermatologic visits if any new lesion appears.
Genetic counseling is increasingly being offered to patients with sporadic MPM or familial melanoma and/or to their healthy relatives.35 The genetic counseling in melanoma is focused on the screening of high-penetrance genes such as CDKN2A. Although MITF p.E318K is a moderate melanoma risk allele, it also increases the risk of developing RCC and pancreatic cancer. In our set of carriers, we observed that 42.9% (3 of 7) of the patients carrying MITF p.E318K developed cutaneous basal cell carcinoma, which is similar to previous data reported in wild-type CDKN2A MPM.36 We did not detect any association with pancreatic cancer, probably owing to the small number of carriers in the study. Reinforcing their predisposition to develop RCC, 14.3% (1 of 7) carriers had developed this kind of tumor. Thus, if genetic counseling included MITF p.E318K genetic testing, individuals carrying the MITF p.E318K variant could benefit from being included both in melanoma and RCC prevention/surveillance programs. Furthermore, the detection of MITF p.E318K may identify patients at risk of developing fast-growing melanomas. We have observed a similar prevalence of MITF p.E318K in cases with germline CDKN2A mutations as in patients with wild-type. However, further studies should be performed to assess the role of MITF p.E318K as a possible modulator of the effect of CDKN2A mutations. This result suggests that individuals with a mutation in CDKN2A might also be included in MITF p.E318K screening as the identification of this variant allows for better characterization of the risk in the family and to adapt the cancer surveillance programs accordingly.
Patients with MITF p.E318K should be encouraged to follow melanoma prevention programs, which include sun protection strategies, monthly self-examination of the skin, and dermatologic surveillance. Because MITF p.E318K has been associated with RCC,14 the use of renal ultrasonography as a safe and low-cost screening technique to detect the presence of kidney tumors should be considered. Future studies should explore the cost-efficacy and acceptance of this screening technique in MITF p.E318K carriers. It would be important to recommend that carriers avoid smoking and control their weight since smoking and obesity are important risk factors for both RCC and pancreatic cancer.37,38 Moreover, detection of MITF p.E318K in a patient leads to the possibility of extending genetic testing to other relatives, and positive cases should be encouraged to follow the same preventive measures.
Based on the results of this study, MITF (and MC1R) should be added to CDKN2A/CDK4 genetic testing based on published international recommendations for countries with low and high sun exposure.39 Genotyping for MITF (and MC1R) could be added to predictive testing for all relatives in CDKN2A-positive families.
Corresponding Author: Susana Puig, MD, PhD, Dermatology Department, Hospital Clínic de Barcelona, Universitat de Barcelona, C/Villarroel, 170, 08036 Barcelona, Spain (firstname.lastname@example.org).
Accepted for Publication: September 18, 2015.
Published Online: December 9, 2015. doi:10.1001/jamadermatol.2015.4356.
Author Contributions: Drs Puig and Potrony had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Puig.
Acquisition, analysis, and interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Potrony, Puig-Butille, Puig.
Statistical analysis: Potrony.
Obtained funding: Malvehy, Puig.
Study supervision: Puig.
Conflict of Interest Disclosures: None reported.
Funding/Support: The research at the Melanoma Unit in Barcelona is partially funded by Spanish Fondo de Investigaciones Sanitarias grants 09/01393 and 12/00840; Centro de Investigación Biomédica en Redde Enfermedades Raras of the Instituto de Salud Carlos III, Spain; Agència de Gestió d’Ajuts Universitaris i de Recerca (AGUAR) 2009 SGR 1337 and AGAUR 2014_SGR_603 of the Catalan Government, Spain; European Commission under the Sixth Framework Programme, contract No. LSHC-CT-2006-018702 (GenoMEL) and by the European Commission under the Seventh Framework Programme, Diagnoptics; the National Cancer Institute of the National Institutes of Health (CA83115); and a grant from Fundació La Marató de TV3, 201331-30, Catalonia, Spain. Ms Potrony is the recipient of a PhD fellowship from Instituto de Salud Carlos III, Spain.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: The work was carried out at the Esther Koplowitz Center in Barcelona. The Spanish National DNA Bank performed sample and data procurement of the control population. We thank the nurses of the melanoma unit: Maria Eugenia Moliner, BSs; Pablo Iglesias, BSs; and Daniel Gabriel, BSs (Melanoma Unit, Dermatology Department, Hospital Clínic de Barcelona); they received no compensation other than their salary. We also thank all of the residents and fellows who contributed to the tasks in the melanoma unit as well as the patients and their families. We thank the patients for granting permission to publish this information.