Primary Melanoma Tumors from CDKN2A Mutation Carriers Do Not Belong to a Distinct Molecular Subclass

Abstract

American Joint Committee on Cancer cutaneous malignant melanoma immunohistochemical principal component analysis Cutaneous malignant melanoma (CMM) is one of the most common cancers in the Western world, with a rapidly increasing incidence. Clinically, CMM is a highly heterogeneous disease that is currently staged according to the American Joint Committee on Cancer (AJCC) system, which is based on the thickness of the tumor, ulceration status, mitotic rate, and node involvement (Balch et al., 2009Balch C.M. Gershenwald J.E. Soong S.J. et al.Final version of 2009 AJCC melanoma staging and classification.J Clin Oncol. 2009; 27: 6199-6206Crossref PubMed Scopus (3662) Google Scholar). Molecular testing has explained some of this heterogeneity. On the basis of the frequent BRAF mutations in CMM (Davies et al., 2002Davies H. Bignell G.R. Cox C. et al.Mutations of the BRAF gene in human cancer.Nature. 2002; 417: 949-954Crossref PubMed Scopus (8302) Google Scholar), and their virtually mutually exclusive occurrence with NRAS mutations, tumors are today molecularly subgrouped by BRAF and NRAS mutation status. This mutation status is believed to correlate with histological and clinical tumor phenotypes (Curtin et al., 2005Curtin J.A. Fridlyand J. Kageshita T. et al.Distinct sets of genetic alterations in melanoma.N Engl J Med. 2005; 353: 2135-2147Crossref PubMed Scopus (2133) Google Scholar). Molecular classification of CMM based on gene expression profiles was recently described, identifying four distinct phenotypes (proliferative, high-immune, normal-like, and pigmentation) independent of BRAF or NRAS mutations and associated with patient outcome independent of other prognostic factors (Jonsson et al., 2010Jonsson G. Busch C. Knappskog S. et al.Gene expression profiling-based identification of molecular subtypes in stage IV melanomas with different clinical outcome.Clin Cancer Res. 2010; 16: 3356-3367Crossref PubMed Scopus (189) Google Scholar; Harbst et al., 2012Harbst K. Staaf J. Lauss M. et al.Molecular profiling reveals low- and high-grade forms of primary melanoma.Clin Cancer Res. 2012; 18: 4026-4036Crossref PubMed Scopus (79) Google Scholar). Familial CMM accounts for ∼10% of all CMM cases (Hayward, 2003Hayward N.K. Genetics of melanoma predisposition.Oncogene. 2003; 22: 3053-3062Crossref PubMed Scopus (254) Google Scholar). The most common high-penetrance melanoma susceptibility gene is CDKN2A, located at chromosome band 9p21. Distinct CDKN2A founder mutations exist, including the Swedish p.R112_L113insR and the Dutch c.225-243del19 mutations (Hayward, 2003Hayward N.K. Genetics of melanoma predisposition.Oncogene. 2003; 22: 3053-3062Crossref PubMed Scopus (254) Google Scholar). Additional germline mutations in CDK4, MITF, TERT, and POT1 have also been described in a few CMM families (Yokoyama et al., 2011Yokoyama S. Woods S.L. Boyle G.M. et al.A novel recurrent mutation in MITF predisposes to familial and sporadic melanoma.Nature. 2011; 480: 99-103Crossref PubMed Scopus (331) Google Scholar; Horn et al., 2013Horn S. Figl A. Rachakonda P.S. et al.TERT promoter mutations in familial and sporadic melanoma.Science. 2013; 339: 959-961Crossref PubMed Scopus (1271) Google Scholar; Puntervoll et al., 2014Puntervoll H.E. Yang X.R. Vetti H.H. et al.Melanoma prone families with CDK4 germline mutation: phenotypic profile and associations with MC1R variants.J Med Genet. 2014; 50: 264-270Crossref Scopus (0) Google Scholar; Robles-Espinoza et al., 2014Robles-Espinoza C.D. Harland M. Ramsay A.J. et al.POT1 loss-of-function variants predispose to familial melanoma.Nat Genet. 2014; 46: 478-481Crossref PubMed Scopus (258) Google Scholar). In this study, we aimed to determine whether primary melanomas from CDKN2A mutation carriers exhibit a distinct gene expression signature. To this end we collected formalin-fixed paraffin-embedded primary CMMs from five centers in Europe and Australia, including 237 sporadic cases and 43 tumors from CDKN2A carriers, with the addition of 6 normal skin and 11 nevus specimens (n=297 in total) (Table 1). Among CDKN2A carriers, 31 out of 43 CDKN2A carriers harbored the Swedish founder mutation p.R112_L113insR and five harbored the Dutch founder mutation c.225-243del19. The sporadic CMMs consisted of samples from a consecutive Swedish cohort (thickness >0.8 mm to ensure sufficient DNA and RNA, n=225) collected at the Department of Pathology, Lund University, as previously described (Harbst et al., 2012Harbst K. Staaf J. Lauss M. et al.Molecular profiling reveals low- and high-grade forms of primary melanoma.Clin Cancer Res. 2012; 18: 4026-4036Crossref PubMed Scopus (79) Google Scholar), and 12 sporadic cases from Brisbane, Leiden, and Stockholm. From the 297 samples RNA and DNA were extracted, and global gene expression data were obtained using the WG-DASL technique as described (Harbst et al., 2012Harbst K. Staaf J. Lauss M. et al.Molecular profiling reveals low- and high-grade forms of primary melanoma.Clin Cancer Res. 2012; 18: 4026-4036Crossref PubMed Scopus (79) Google Scholar). The Sequenom Oncocarta panel v1 was used to determine oncogene mutation status in a subset of 166 tumors as described (Harbst et al., 2012Harbst K. Staaf J. Lauss M. et al.Molecular profiling reveals low- and high-grade forms of primary melanoma.Clin Cancer Res. 2012; 18: 4026-4036Crossref PubMed Scopus (79) Google Scholar). p16 nuclear and PTEN immunohistochemical (IHC) analysis was performed on 13 CDKN2A and 152 sporadic cases, respectively (Supplementary Information online). The study was approved by the Ethics Review Boards of each participating center according to the Declaration of Helsinki guidelines.Table 1Clinicopathological and molecular features of sporadic melanomas and melanomas from CDKN2A mutation carriersFeatureCDKN2A mutation carriers (n=43)1Number of cases with percentage in brackets, except range in brackets for thickness and age.Sporadic cases (n=237)1Number of cases with percentage in brackets, except range in brackets for thickness and age.P-value2Univariate logistic regression.P-value3Multivariate logistic regression. All P-values adjusted for thickness and age at diagnosis.GenderMale22 (55)113 (51)0.63—Female18 (45)109 (49)Median thickness (mm)1.97 (0.4–14.2)3.05 (0.6–25)0.060.27Site4Other sites were excluded in analysis because of only one case in the CDKN2A cohort.Head/neck9 (24)30 (14)0.020.11Trunk16 (39)127 (60)Extremities14 (34)37 (17)Other1 (2)18 (8)ClarkII10 (26)6 (3)0.00020.0003III13 (33)84 (42)IV13 (33)97 (48)V3 (8)15 (7)Histology5For histology only SSM and NM are included in P-value calculation.SSM25 (74)107 (49)0.08—NM9 (26)79 (36)Other—33 (15)Age at diagnosis47 (26–85)64 (14–101)<0.001<0.001Mutation statusBRAF mutation18 (54)38 (29)0.030.06NRAS mutation3 (9)24 (18)Wild type13 (37)70 (53)Gene expression phenotypeHigh-immune4 (9)64 (27)0.020.13Normal-like18 (41)59 (25)Pigmentation15 (36)87 (37)Proliferative3 (7)22 (9)Unclassified3 (7)5 (2)Nuclear p16 IHCAbsent13 (100)125 (82)0.99—Present0 (0)27 (18)PTEN IHCAbsent6 (43)36 (24)0.08—Present7 (57)116 (76)Abbreviations: IHC, immunohistochemistry; NM, nodular melanoma; SSM, superficial spreading melanoma.P-values<0.05 are indicated as bold.1 Number of cases with percentage in brackets, except range in brackets for thickness and age.2 Univariate logistic regression.3 Multivariate logistic regression. All P-values adjusted for thickness and age at diagnosis.4 Other sites were excluded in analysis because of only one case in the CDKN2A cohort.5 For histology only SSM and NM are included in P-value calculation. Open table in a new tab Download .pdf (.4 MB) Help with pdf files Supplementary Information Abbreviations: IHC, immunohistochemistry; NM, nodular melanoma; SSM, superficial spreading melanoma. P-values<0.05 are indicated as bold. As expected, age at diagnosis (Goldstein et al., 2007Goldstein A.M. Chan M. Harland M. et al.Features associated with germline CDKN2A mutations: a GenoMEL study of melanoma-prone families from three continents.J Med Genet. 2007; 44: 99-106Crossref PubMed Scopus (308) Google Scholar) and Clark’s level of invasion and Breslow thickness (borderline significant) differed between the CDKN2A and sporadic cohorts (Table 1). Moreover, 54% and 9% of CDKN2A carriers had mutations in BRAF and NRAS, respectively, compared with 29% and 18% in sporadic cases (P=0.03). However, this difference in mutation frequency between the groups was lost when adjusting for thickness and age at onset, representing features previously associated with BRAF mutation (Scolyer et al., 2011Scolyer R.A. Long G.V. Thompson J.F. Evolving concepts in melanoma classification and their relevance to multidisciplinary melanoma patient care.Mol Oncol. 2011; 5: 124-136Crossref PubMed Scopus (113) Google Scholar). Notably, these results are consistent with a study by Zebary et al., 2014Zebary A. Omholt K. van Doorn R. et al.Somatic BRAF and NRAS mutations in familial melanomas with known germline CDKN2A status: a GenoMEL study.J Invest Dermatol. 2014; 134: 287-290Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar reporting no difference in BRAF and NRAS mutation frequencies between CDKN2A carriers and matched sporadic cases. To investigate whether CDKN2A mutation carriers displayed gene expression patterns different from sporadic CMM, we applied principal component analysis (PCA) to the expression data from the 297-sample cohort, finding that CDKN2A carriers were firmly intermingled with sporadic cases (Figure 1a). In addition, PCA showed that the main variation in gene expression in CMM was related to specific clinicopathological factors (tumor thickness and histology) and gene expression phenotypes, whereas CDKN2A mutation status contributed little to the overall variation (Supplementary Figure 1 online). To further investigate the molecular landscape of melanomas from CDKN2A carriers, we classified the 280 familial and sporadic tumors according to our reported gene expression phenotypes (Jonsson et al., 2010Jonsson G. Busch C. Knappskog S. et al.Gene expression profiling-based identification of molecular subtypes in stage IV melanomas with different clinical outcome.Clin Cancer Res. 2010; 16: 3356-3367Crossref PubMed Scopus (189) Google Scholar; Harbst et al., 2012Harbst K. Staaf J. Lauss M. et al.Molecular profiling reveals low- and high-grade forms of primary melanoma.Clin Cancer Res. 2012; 18: 4026-4036Crossref PubMed Scopus (79) Google Scholar). Here, we found overall similar frequencies of gene expression phenotypes in the CDKN2A and sporadic cohorts, but with a slightly higher frequency of normal-like melanomas in the CDKN2A cohort (P=0.02, Table 1, Figure 1b). However, after adjustment for thickness and age at onset, this statistical difference was lost. Finally, we searched for differentially expressed genes between CDKN2A carriers and sporadic CMMs. On the basis of significance analysis of microarrays (SAM) only seven genes showed a significant difference (q-value=0) (Figure 1c). Notably, the maximum fold change of these genes was <1.4, suggesting very subtle changes not inconsistent with random observations. Moreover, to determine the validity of the differences in these seven genes, an independent CMM cohort including CDKN2A mutation carriers needs to be analyzed. Although gene expression differences stratified on somatic CDKN2A mutation status have been observed in other tumor types (Jardin et al., 2010Jardin F. Jais J.P. Molina T.J. et al.Diffuse large B-cell lymphomas with CDKN2A deletion have a distinct gene expression signature and a poor prognosis under R-CHOP treatment: a GELA study.Blood. 2010; 116: 1092-1104Crossref PubMed Scopus (102) Google Scholar), our gene expression analyses suggest that CMMs from CDKN2A germline mutation carriers do not constitute a distinct gene expression subtype and that only subtle, if any, gene expression differences characterize melanomas from CDKN2A carriers. Results from the gene expression analyses are further corroborated by p16 nuclear IHC analysis of 152 sporadic cases from our cohort, wherein 125 demonstrated a complete loss of p16 nuclear staining (Table 1 and Figure 1d). This suggests that loss of nuclear p16 protein is an early event in melanomagenesis and may partly explain why melanomas from CDKN2A carriers are molecularly almost indistinguishable from sporadic cases. Moreover, in melanoma tumors the MAPK and the PI3K pathways are commonly altered (Hodis et al., 2012Hodis E. Watson I.R. Kryukov G.V. et al.A landscape of driver mutations in melanoma.Cell. 2012; 150: 251-263Abstract Full Text Full Text PDF PubMed Scopus (1850) Google Scholar). PTEN, a regulator of the PI3K pathway, is a major tumor suppressor in melanoma, and PTEN gene alterations are commonly found in CMM. However, we found no difference in PTEN protein expression by IHC between CDKN2A carriers and sporadic cases (P=0.08, Table 1 and Figure 1d), indicating that deficiencies of the PI3K pathway occur in both sporadic and familial melanoma. In conclusion, this is the only study, to our knowledge, investigating genome-wide molecular changes in primary melanomas from CDKN2A mutation carriers. Although larger numbers of familial tumors could potentially identify additional subtle molecular alterations between familial and sporadic cases, our results strongly suggest that melanomas from CDKN2A mutation carriers are not associated with a distinct gene expression signature and that overall they have similar molecular characteristics to sporadic melanomas. We thank Kristina Lövgren and Björn Nodin for assistance with TMA construction and immunohistochemical staining. This project received support from the European Commission under the 6th Framework Programme, Contract: LSHC-CT-2006-018702 (GenoMEL), the Swedish Cancer Society, the Swedish Research Council, the Nordic Cancer Union, the Berta Kamprad Foundation, the Gunnar Nilsson Cancer foundation, the Gustav Vth Jubilee foundation, BioCARE, the National Health and Medical Research Council of Australia, and governmental funding for medical health research (ALF). Supplementary material is linked to the online version of the paper at http://www.nature.com/jid