April 2007

Multiple Melanomas After Treatment for Hodgkin Lymphoma in a Non-Dutch p16-Leiden Mutation Carrier With 2 MC1R High-Risk Variants

Author Affiliations

Author Affiliations: Skin Cancer Unit (Drs Figl, Ugurel, and Schadendorf) and Division of Molecular Genetic Epidemiology (Drs Thirumaran, Gast, Hemminki, and Kumar), German Cancer Research Center, Heidelberg, Germany; Department of Dermatology, University Hospital of Mannheim, Mannheim, Germany (Drs Figl, Ugurel, and Schadendorf); and Department of Biosciences, Karolinska Institute, Huddinge, Sweden (Drs Hemminki and Kumar).


Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007

Arch Dermatol. 2007;143(4):495-499. doi:10.1001/archderm.143.4.495

Background  A 19–base pair germline deletion in exon 2 of the CDKN2A (cyclin-dependent kinase inhibitor 2A) gene (Leiden mutation) has been detected in Dutch families with familial melanomas. The penetrance of CDKN2A mutations varies widely and is influenced by environmental and unrelated genetic factors such as variants in the MC1R gene.

Observations  We describe a 25-year-old German woman who developed 8 invasive melanomas and 6 in situ melanomas after radiation therapy and polychemotherapy for Hodgkin lymphoma. Genetic testing revealed a constitutional CDKN2A Leiden mutation in the proband and her sister, mother, and mother's sister. The proband also carried high-risk MC1R variant alleles R151C and R160W, which she had inherited from her father and her mother, respectively. The less affected mutation carrier sister did not have high-risk MC1R variant alleles. Analysis of DNA from paraffin-embedded tissues showed loss of heterozygosity at CDKN2A loci in all 3 melanomas studied but not in Hodgkin lymphoma. The pedigree revealed several types of cancers on both sides of the family, but no Dutch ancestors were found. No mutations in the CDK4, B-raf, and N-ras genes were detected either in the germline or in tumors from the patient.

Conclusion  This study shows the variability of the penetrance of the CDKN2A Leiden mutation within the same family, which could be due to genetic or exogenous factors.

Malignant melanoma is associated with complex genetic heterogeneity. An estimated 10% to 15% of melanomas occur in familial settings. Germline disease–segregating mutations in the CDKN2A (cyclin-dependent kinase inhibitor 2A) gene have been observed on average in up to 40% of melanoma-prone families.1 The penetrance of the CDKN2A mutations in carriers within melanoma families varies between 53% and 100% by the age of 80 years.24 However, a population-based survey has estimated an approximately 28% risk of melanoma in carriers of the CDKN2A mutation in the general population by the age of 80 years.5 The difference in penetrance rates between the latter study and previous studies is likely attributable to the different ascertainment approaches. However, variability in penetrance suggests risk modification by other genetic or environmental factors, including geographic location.4

In Dutch families with familial atypical multiple mole melanoma, a founder germline 19–base pair (bp) frame-shift deletion has been described in exon 2 of the CDKN2A gene (Online Mendelian Inheritance in Man [OMIM] 600160.0003).6 All families with this particular deletion (named the CDKN2A Leiden mutation) were reported from the same geographic area in the Netherlands. It is presumed that all mutation carriers can be traced to a Dutch founder. Outside the Netherlands, only 4 families with the germline CDKN2A Leiden mutation have been described, to our knowledge. In 2003, a heterozygous mutation was detected in a blood sample from a 54-year-old man with 3 oropharyngeal tumors. Both of his parents and his only sister died of cancer when they were very young (the mother of gynecologic cancer, the father of liver carcinoma, and the sister of leukemia).7 Other melanoma-prone families with the segregating CDKN2A Leiden mutation have been reported from North America and Australia.8,9


Constitutional DNA samples were screened for mutations in the CDKN2A, CDK4, B-raf, and N-ras genes and for variants in the MC1R gene by direct DNA sequencing. For the CDKN2A gene, exons 1β, 1α, 2, and 3 were amplified using polymerase chain reaction (PCR) primers and conditions as described elsewhere.10,11 Similarly, exon 11 and 15 of the B-raf gene, exon 1 and 2 of the N-ras gene, and the MC1R gene were amplified using primers and conditions used previously.12 Amplified exons were sequenced using a commercially available cycle sequencing kit (BigDye Kit; Applied Biosystems, Foster City, Calif) and an automated sequencer (ABI 3100; Applied Biosystems).


Mutations in DNA from paraffin-embedded tissues were detected by single-strand conformation polymorphism using PCR products with a phosphorus 32–labeled radioisotope incorporated during amplification reactions.1114 Radiolabeled PCR products were electrophoresed under nondenaturating conditions, and mutations were detected by aberrant band shifts and sequencing.


Loss of heterozygosity (LOH) in paraffin-embedded tissues was determined by amplifying dinucleotide repeat microsatellite markers D9S974 and D9S942 using primers at the flanking sequences (www.gdb.org).10 Fluorescent-labeled PCR products were electrophoresed in an automated sequencer, and fragments were analyzed using fragment analysis software (Applied Biosystems). The LOH was scored if 1 of the 2 fragments in paraffin-embedded DNA from tumor tissue showed more than 50% reduction in size compared with corresponding constitutional DNA.10


In November 2004, a 24-year-old woman was referred to the skin cancer unit, Department of Dermatology, University Hospital of Mannheim, Mannheim, Germany. At the age of 20 years, she had been diagnosed as having axillary and mediastinal Hodgkin lymphoma, stage IIA, and received polychemotherapy with doxorubicin, bleomycin, vinblastine, and dacarbazine followed by radiation therapy. One year later, she again received radiation therapy for an early relapse and was in complete remission until June 2006, when she presented with a second relapse of Hodgkin lymphoma in both axillae. At the time of the writing of this article, she was consulting with her medical oncologists regarding possible high-dose chemotherapy.

Twenty-two months after the second radiation treatment, an ulcerated invasive melanoma (Breslow thickness 2.65 mm and Clark level IV) was excised from the patient's left upper arm. An axillary sentinel lymph node biopsy showed micrometastasis, and complete lymph node dissection was performed, with no additional metastatic nodes. Within the following 2 months, 2 more invasive melanomas (Breslow thickness 0.41 mm and 0.56 mm) and 2 in situ melanomas were excised. After diagnosis of stage III melanoma, adjuvant low-dose subcutaneous interferon alfa therapy was initiated after an intravenous high-dose induction phase. However, half a year later, 4 more invasive melanomas (Breslow thickness between 0.35 and 0.50 mm) and 4 in situ melanomas were discovered and excised. The patient has lightly pigmented skin with a few freckles corresponding to skin phototype II, auburn hair, and a large number of nevi. Clinically, these nevi are small and regular rather than dysplastic.


The patient described an aunt (her mother's sister) with a history of 4 invasive melanomas (Breslow thickness <1.0 mm), 1 melanoma in situ, and ovarian cancer. Because of her unusual personal and family melanoma history, we suggested genetic testing for alterations at melanoma susceptibility loci. During these investigations, the patient developed 1 more invasive melanoma (Breslow thickness 0.6 mm), and her sister was diagnosed for the first time as having melanoma in situ. Neither of the parents has presented with malignancy so far, but reconstruction of a pedigree revealed multiple cancers on both sides of the family (Figure 1). Constitutional DNA from the patient, her only sibling sister, and their parents was screened for mutations in the CDKN2A, CDK4, B-raf, and N-ras genes. Subsequently, other relatives were also screened for germline mutations in the CDKN2A gene.

Figure 1.
Image not available

Pedigree representing a family with a CDKN2A (cyclin-dependent kinase inhibitor 2A) 19–base pair deletion. Numbers outside parentheses represent current age or age at death, as appropriate, in years. Numbers within parentheses give age at onset of cancer, also in years. The proband is indicated by an arrow. The father of the proband's mother died of a septicemia of unknown origin at the age of 39 years. Circles indicate females; squares, males; diagonal lines, deceased; shaded central dots, p16-Leiden mutation; black outer circles, melanoma; vertical lines, no Leiden mutation; and open circles and squares, mutation status unknown.

DNA from the 2 sisters and their mother and aunt contained a heterozygous 19-bp deletion in exon 2 of the CDKN2A gene, known as the CDKN2A Leiden mutation (Figure 2). No mutation was detected in any other investigated gene (Table). Furthermore, none of the other investigated relatives carried the mutation (Figure 1). Despite the 19-bp CDKN2A deletion being a founder mutation in the Netherlands, there have been no Dutch ancestors, going back at least 3 generations in the family under investigation.

Figure 2.
Image not available

Results of CDKN2A (cyclin-dependent kinase inhibitor 2A) analysis. Detection of a germline heterozygous 19–base pair deletion (Leiden mutation) in exon 2 of the CDKN2A gene in DNA from a patient with multiple melanomas (A), her sister (B), and her mother (C); DNA from her father showed a wild-type sequence (D). Analysis of dinucleotide repeat microsatellite markers located within the CDKN2A gene showed loss of heterozygosity in tumor tissues at D9S974 (E and G) and D9S942 (F and H) loci. Panels E and F show amplification of markers D9S974and D9S942 in germline DNA from a patient with multiple melanomas. Panels G and H represent amplification of DNA extracted from a tumor in the same patient and show more than 50% reduction in the wild-type allele.

Image not available
Constitutional Mutations in the CDKN2A, CDK4, B-raf, N-ras, and MC1R Genes in the Family of a Proband With Multiple Melanomas

The constitutional DNA from the family members was also genotyped for polymorphisms in the entire MC1R gene. The results showed that while the father was heterozygous for R151C and V60L polymorphisms; the mother carried the R160W variant allele. The previously unaffected sister inherited variant V60L from her father. However, the proband was heterozygous for high-risk R151C and R160W polymorphisms. The aunt (the CDKN2A Leiden deletion carrier) with a history of melanomas and ovarian cancer also carried the high-risk R151C variant allele (Table). A 57-year-old female cousin of the patient's mother and aunt was homozygous for the variant 160W allele and was not a carrier of the CDKN2A Leiden mutation. The patient and her sister, mother, and aunt share a similar skin phenotype that clearly differs from other members of the family.


Archival paraffin-embedded tissue was obtained from the Hodgkin lymphoma of the proband and from 3 of her invasive melanomas. All of these tissue samples carried the CDKN2A Leiden mutation. No mutation was detected in exons 11 and 15 of the B-raf gene or in exons 1 and 2 of the N-ras gene. Analysis of 2 microsatellite markers at the CDKN2A loci D9S942 and D9S974 in DNA from paraffin-embedded tissues showed LOH in all 3 melanoma samples at D9S942 and in 2 of the melanoma samples also at the D9S974 locus. No LOH was found in the Hodgkin lymphoma (Figure 2).


In this study, we detected a so-called CDKN2A Leiden mutation in the germline DNA of a 24-year-old patient who initially presented with Hodgkin lymphoma and who subsequently developed multiple melanomas after chemotherapy and radiation therapy. A similar analysis of DNA from her only sibling, a sister with no initial disease, showed her to be a mutation carrier. Both sisters had inherited the mutant CDKN2A allele from their malignancy-free mother. The mother's sister, who had a history of multiple primary melanomas and ovarian cancer, was shown to be a mutation carrier as well. Because the other investigated relatives were noncarriers, we were not able to determine whether the mother and the aunt inherited the mutation from their mother or their father. The variability in the penetrance of the CDKN2A mutation in carriers implies a confounding effect of other genetic and environmental factors on phenotypic expression. This supposition is further supported by 2 known human homozygous p16-Leiden mutation carriers whose phenotype is similar to that of heterozygotes.6

We also found that the affected patient carried variant alleles for R151C and R160W polymorphisms in the MC1R gene, which were inherited from her father and mother. Incidentally, the other sister carried only the V60L variant allele. Interestingly, the affected aunt carried variant R151C, which the proband had inherited from her father. The father and aunt were not related. Variants of MC1R are known to contribute to melanoma risk in families with disease-segregating CDKN2A mutations. More than 65 human MC1R variants account for variations in skin and hair color and in skin cancer incidence by controlling the relative amounts of eumelanin and pheomelanin. Investigations in 16 American and 15 Australian CDKN2A families and in 6 Dutch p16-Leiden mutation families demonstrated that MC1R variants, notably the 3 red hair color variants R151C, R160W, and D294H, increased the penetrance of melanoma in mutation carriers.1517 In the Dutch families, the MC1R-melanoma association was primarily related to the R151C variant. The concurrent occurrence of the high-risk MC1R allele and the CDKN2A Leiden mutation might be the reason that the proband and her aunt were more seriously affected than the patient's sister and mother. However, at the same time, immunosuppression due to Hodgkin disease could also be the reason for multiple melanomas in the proband. Another possible cause could be the doxorubicin, bleomycin, vinblastine, and dacarbazine treatment of Hodgkin disease.

The analysis of tumors from the patient with lymphoma and multiple melanomas showed loss of the wild-type allele of CDKN2A in melanomas but not in lymphomas. The wild-type allele of a tumor suppressor is lost mostly through deletion, mitotic recombination, nondisjunctional chromosomal loss, or gene conversion, termed LOH. However, silencing of the wild-type allele through promoter methylation has also been reported. In a more specific instance, in a case of a p16-Leiden mutation carrier with head and neck cancer, the wild-type allele showed promoter methylation in a tumor.7

The attribution of Hodgkin lymphoma to the CDKN2A mutational status is elusive, although loss of p16 expression in Hodgkin and Reed-Sternberg cells as well as non-Hodgkin lymphomas in CDKN2A mutation carriers and leukemia in a first-degree relative of a carrier have been reported.7,1820 To date, no report, to our knowledge, has shown Hodgkin lymphoma in CDKN2A Leiden mutation carriers. Furthermore, a strong statistical association has been reported between the occurrence of non-Hodgkin lymphoma and cutaneous melanoma.21

Although we analyzed both the B-raf and the N-ras genes in the germline as well as in tumors from a patient with lymphoma and multiple melanomas, no mutations were detected. Somatic mutations in the B-raf gene are of common occurrence in melanomas; however, germline mutations in the gene are rare.12,22 Somatic mutations in the N-ras gene complement mutations in the B-raf gene and occur to a certain extent in melanomas. In an earlier study, tumors from patients belonging to melanoma families and carrying the CDKN2A mutation showed a high frequency of N-ras mutations.23 The ras-raf mitogen–activated protein kinase pathway interferes with the cellular response to cyclic adenosine monphosphate–dependent growth signals in melanocytes. However, it is possible that increased melanocytic proliferation due to a germline CDKN2A mutation abrogates requirement for mutations in the B-raf and N-ras genes.

Leiden mutation carriers are also at an estimated risk of 17% to develop pancreatic cancer by the age of 75 years, although pancreatic cancer was observed in only 4 of the investigated 19 families with this specific mutation.18 The association of the pancreatic cancer of our proband's granduncle with the CDKN2A Leiden mutation remains unknown. However, based on our findings, we conclude that all mutation carriers should be advised to undergo regular clinical checkups for melanoma and pancreatic cancer.

Back to top
Article Information

Correspondence: Dirk Schadendorf, MD, Skin Cancer Unit, German Cancer Research Center Heidelberg and Department of Dermatology, University Hospital of Mannheim, Theodor-Kutzer-Ufer 1, D-68167 Mannheim, Germany (d.schadendorf@dkfz.de).

Financial Disclosure: None reported.

Accepted for Publication: April 26, 2006.

Author Contributions:Study concept and design: Figl, Ugurel, Kumar, and Schadendorf. Acquisition of data: Figl, Thirumaran, Gast, Kumar, and Schadendorf. Analysis and interpretation of data: Figl, Ugurel, Hemminki, Kumar, and Schadendorf. Drafting of the manuscript: Figl, Kumar, and Schadendorf. Critical revision of the manuscript for important intellectual content: Thirumaran, Ugurel, Gast, Hemminki, Kumar, and Schadendorf. Obtained funding: Schadendorf. Administrative, technical, and material support: Figl, Thirumaran, Ugurel, Gast, Kumar, and Schadendorf. Study supervision: Hemminki, Kumar, and Schadendorf.

Piepkorn  M Melanoma genetics: an update with focus on the CDKN2A(p16)/ARF tumor suppressors. J Am Acad Dermatol 2000;42705- 726
Cannon-Albright  LAMeyer  LJGoldgar  DE  et al.  Penetrance and expressivity of the chromosome 9p melanoma susceptibility locus (MLM). Cancer Res 1994;546041- 6044
Walker  GJHussussian  CJFlores  JF  et al.  Mutations of the CDKN2/p16INK4 gene in Australian melanoma kindreds. Hum Mol Genet 1995;41845- 1852
Bishop  DTDemenais  FGoldstein  AM  et al.  Geographical variation in the penetrance of CDKN2A mutations for melanoma. J Natl Cancer Inst 2002;94894- 903
Begg  CBOrlow  IHummer  AJ  et al.  Lifetime risk of melanoma in CDKN2A mutation carriers in a population-based sample. J Natl Cancer Inst 2005;971507- 1515
Gruis  NAvan der Velden  PASandkuijl  LA  et al.  Homozygotes for CDKN2 (p16) germline mutation in Dutch familial melanoma kindreds. Nat Genet 1995;10351- 353
Schneider-Stock  RGiers  AMotsch  C  et al.  Hereditary p16-Leiden mutation in a patient with multiple head and neck tumors. Am J Hum Genet 2003;72216- 218
Goldstein  AMStruewing  JPChidambaram  AFraser  MCTucker  MA Genotype-phenotype relationships in U.S. melanoma-prone families with CDKN2A and CDK4 mutations. J Natl Cancer Inst 2000;921006- 1010
Holland  EASchmid  HKefford  RFMann  GJ CDKN2A (P16(INK4a)) and CDK4 mutation analysis in 131 Australian melanoma probands: effect of family history and multiple primary melanomas. Genes Chromosomes Cancer 1999;25339- 348
Kumar  RSmeds  JLundh Rozell  BHemminki  K Loss of heterozygosity at chromosome 9p21 (INK4-p14ARF locus): homozygous deletions and mutations in the p16 and p14ARF genes in sporadic primary melanomas. Melanoma Res 1999;9138- 147
Kumar  RSmeds  JBerggren  P  et al.  A single nucleotide polymorphism in the 3′untranslated region of the CDKN2A gene is common in sporadic primary melanomas but mutations in the CDKN2B, CDKN2C, CDK4 and p53 genes are rare. Int J Cancer 2001;95388- 393
Kumar  RAngelini  SCzene  K  et al.  BRAF mutations in metastatic melanoma: a possible association with clinical outcome. Clin Cancer Res 2003;93362- 3368
Kumar  RLundh Rozell  BLouhelainen  JHemminki  K Mutations in the CDKN2A (p16INK4a) gene in microdissected sporadic primary melanomas. Int J Cancer 1998;75193- 198
Kumar  RAngelini  SHemminki  K Activating BRAF and N-Ras mutations in sporadic primary melanomas: an inverse association with allelic loss on chromosome 9. Oncogene 2003;229217- 9224
Goldstein  AMLandi  MTTsang  SFraser  MCMunroe  DJTucker  MA Association of MC1R variants and risk of melanoma in melanoma-prone families with CDKN2A mutations. Cancer Epidemiol Biomarkers Prev 2005;142208- 2212
van der Velden  PASandkuijl  LABergman  W  et al.  Melanocortin-1 receptor variant R151C modifies melanoma risk in Dutch families with melanoma. Am J Hum Genet 2001;69774- 779
Box  NFDuffy  DLChen  W  et al.  MC1R genotype modifies risk of melanoma in families segregating CDKN2A mutations. Am J Hum Genet 2001;69765- 773
Vasen  HFGruis  NAFrants  RRvan Der Velden  PAHille  ETBergman  W Risk of developing pancreatic cancer in families with familial atypical multiple mole melanoma associated with a specific 19 deletion of p16 (p16-Leiden). Int J Cancer 2000;87809- 811
Borg  AJohannsson  UJohannsson  O  et al.  Novel germline p16 mutation in familial malignant melanoma in southern Sweden. Cancer Res 1996;562497- 2500
Garcia  JFVilluendas  RAlgara  P  et al.  Loss of p16 protein expression associated with methylation of the p16INK4A gene is a frequent finding in Hodgkin's disease. Lab Invest 1999;791453- 1459
Lens  MBNewton-Bishop  JA An association between cutaneous melanoma and non-Hodgkin's lymphoma: pooled analysis of published data with a review. Ann Oncol 2005;16460- 465
Casula  MColombino  MSatta  MP  et al.  BRAF gene is somatically mutated but does not make a major contribution to malignant melanoma susceptibility: the Italian Melanoma Intergroup Study. J Clin Oncol 2004;22286- 292
Eskandarpour  MHashemi  JKanter  LRingborg  UPlatz  AHansson  J Frequency of UV-inducible NRAS mutations in melanomas of patients with germline CDKN2A mutations. J Natl Cancer Inst 2003;95790- 798