CDKN2A/CDK4 Status in Greek Patients with Familial Melanoma and Association with Clinico-epidemiological Parameters

Approximately 5–10% of melanoma cases occur in a familial context. CDKN2A/CDK4 were the first high- penetrance melanoma genes identified. The aims of this study were to evaluate CDKN2A/CDK4 variants in Greek familial melanoma patients and to correlate the mutational status with specific clinico-epidemiological characteristics. A cross-sectional study was conducted by genotyping CDKN2A/CDK4 variants and selected MC1R polymorphisms in 52 melanoma-prone families. Descriptive statistics were calculated and comparisons were made using the X2 test, Fisher’s exact test and Student’s t-test for statistical analysis, as appropriate. CDKN2A variants were detected in 46.2% of melanoma-prone families, while a CDK4 variant was found in only one family. This study confirmed that, in the Greek population, the age at melanoma diagnosis was lower in patients carrying a variant in CDKN2A compared with wild-type patients. No statistically significant associations were found between CDKN2A mutational status and MC1R polymorphisms.

melanocytic naevi, and the presence of clinically atypical naevi, are associated with an increased risk of developing melanoma (4).A personal history of melanoma increases the risk of developing a second melanoma by 5-8% (5,6), while a family history has been associated with a 1.74 relative risk for melanoma (7), supporting the role of genetic risk factors.A recent study of familial cancer risk in twins from Nordic countries estimated that cutaneous malignant melanoma (CMM) had the highest heritability (58%; 95% confidence interval (CI): 43-73%) among all common cancers (8).
Approximately 5-10% of melanoma cases occur in a familial context (2).Although the underlying genetic basis in the majority of melanoma-prone families is unknown, certain highly penetrant genes, such as the cyclin-dependent kinase inhibitor 2A (CDKN2A) gene (9,10) and cyclin-dependent kinase 4 (CDK4), have been implicated (11)(12)(13).CDKN2A is responsible for melanoma susceptibility in approximately 10% of 2-case melanoma families and 30-40% of families with 3 or more cases of melanoma up to third-degree relatives (14), whereas its incidence among sporadic melanomas is very low (less than 3%) (15).It encodes 2 distinct proteins: p16INK4A (p16) and p14ARF (p14), both of which function in cell cycle regulation (15).CDKN2A mutation carriers are also at increased risk of the development of pancreatic cancer compared with the general population (14,16,17).
Variants in the CDK4 gene have been observed in only a small number of melanoma families (11)(12)(13).All variants
The aims of this study were to report the incidence of CDKN2A and CDK4 variants in a series of Greek familial melanoma families and to examine their association with epidemiological and clinical factors, as well as MC1R polymorphisms.

Description of familial melanoma cases
The study group consisted of patients with a diagnosis of histologically confirmed cutaneous invasive melanoma in subjects with a confirmed history of a first, second or thirddegree relative also affected by histologically confirmed invasive melanoma.Patients were recruited consecutively at A. Sygros Hospital, a large referral centre for melanoma and skin cancer in Athens, Greece, and collaborating centres from 2000 to 2016.
Patients recruited for the study were first evaluated for familial occurrence of melanoma using a questionnaire to interview probands about their relatives.Melanoma families were identified if at least 2 melanoma-affected members existed within the family.All patients were informed about the aims and limits of the study and provided written informed consent prior to participation.The study protocol was approved by the Scientific and Ethics Committee of A. Sygros Hospital.The variables included in the analyses were: age of onset, number of primary melanomas, number of melanoma cases within the family and the presence of internal cancers in first and seconddegree relatives of the patients with melanoma.The different types of cancers reported within melanoma-prone families are depicted in Fig. S1 1 .The pedigree of each patient, as well as a family and personal history of cancer, was obtained from a dermatologist upon melanoma diagnosis or during the patient follow-up.Non-melanoma skin cancers were excluded from the analysis.Phenotypic characteristics were also recorded.Eye and skin colour were evaluated by direct inspection and were subsequently classified into different categories based on a standardized colour system (blue, green, light-brown, dark-brown for eye colour; white, light-brown and dark for skin colour of the inner part of the upper arms).Hair colour at the age of 18 years was assessed on a 5-category scale based on a colour sample chart and separated into 5 categories (blonde, red, light-brown, dark-brown, black).Skin phototype was determined by an interview based on Fitzpatrick's classification into 4 categories, i.e. (i) always burns, never tans, (ii) always burns, tans lightly, (iii) seldom burns, tans well, and (iv) never burns, tans deeply.In addition, all subjects underwent a complete skin examination by experienced dermatologists for the presence of common and atypical melanocytic naevi.

Statistical analysis
The collected data were: age of onset, phenotypic characteristics, number of primary melanomas, number of melanoma cases within the family, and the presence of internal cancers in first and second degree relatives of the melanoma patients.Data were analysed with Stata 13 software (StataCorp.2013.Stata Statistical Software: Release 13.College Station, TX: StataCorp LP).Differences in age at onset between CDKN2A + vs CDKN2A -status were assessed using a Student's t-test, whilst differences in categorical data (sex, multiple primary melanomas and phenotypic characteristics) were assessed using either a 2sided χ 2 test or a Fisher's exact test (the latter when any of expected cell counts was lower than 5).CDK4 variants were discovered in only one family; therefore no further statistical analysis was performed.A total of 11 independent statistical tests were performed and a Bonferroni-derived p-value threshold of 0.05/11 = 0.005 was applied to call a result statistically significant.

RESULTS
The total number of familial melanoma patients was 66, belonging to 52 families.Since related patients share similar genetic background, only one affected patient (the index case) from each family was included in the analysis.Of these 52 patients, 17 were men and 35 women, with a male-to-female ratio of 1:2.Mean age was 44.6 years for men and 49.4 years for women.Table I presents the distribution of sex, age, history of multiple primary melanomas and phenotypic characteristics among CDKN2A + and CDKN2A -patients.Age at diagnosis was statistically significantly lower in patients harbouring a CDKN2A variant compared with wild-type patients (41.1 vs 53.6 years, p = 0.001) (Fig. 1).The presence of multiple primary melanomas was not statistically significantly different among 1 https://doi.org/10.2340/00015555-2969www.medicaljournals.se/actaCDKN2A + compared with CDKN2A -patients.There was a trend towards higher number of naevi and more multiple primary melanomas in the CDKN2A carriers but due to the small sample size, this did not reach statistical significance.In addition, there were no statistically significant differences in the phenotypic characteristics (eye, hair and skin colour, tanning, phototype and naevi) between the 2 groups.CDKN2A variants were identified in 36.1% of families with 2 affected members and in 80% of families with at least 3 melanoma-affected members (Table SI 1 ).

MC1R status among melanoma-prone families
Forty-two out of 52 familial melanoma patients had been genotyped for 6 polymorphisms of MC1R (rs11547464, rs1805007, rs1805009, rs1805006, rs1805005, and rs2228479) in a previous study by our team (23).Table III presents the MC1R status in correlation with CDKN2A variants among familial melanoma families.A MC1R mutation was found in 15 of 18 (83.3%)CDKN2A patients and there was no association of MC1R status with CDKN2A status   after Bonferroni correction (p = 0.083).Distribution of major MC1R variants R (rs11547464, rs1805007, rs1805009, rs1805006) and minor MC1R variants r (rs1805005, rs2228479) polymorphisms did not differ statistically significantly among CDKN2A mutant and wild-type families (Table SII 1 ), although CDKN2A mutation carriers had a trend of an increased frequency of R polymorphisms compared with CDKN2A wild types (p = 0.164).

DISCUSSION
This study presents the epidemiological, clinical and genetic characterization of the largest reported series of Greek melanoma-prone families so far.There is a high diversity in the prevalence of CDKN2A variants among families across different geographical areas (14).In the study by GenoMel the highest frequency of variants was observed in Europe (57%) and the lowest in Australia (20%) (14).In areas with high melanoma incidence a combined effect of moderate or low-penetrance susceptibility genes and higher levels of sun exposure may influence the occurrence of familial cases (14).The high percentage (46.2%) of CDKN2A variants detected in our families is consistent with previous findings in European populations, albeit higher than expected for a low incidence population, underscoring the importance of genetic risk factors in the low incidence Greek population.Studies have shown that awareness of CDKN2A variant status among melanoma patients and relatives increases their compliance in sun protection and rigorous mole screening (24).With respect to early diagnosis of internal cancers in patients with CDKN2A variants, there are some encouraging investigational screening protocols for the early detection of pancreatic cancer in high-risk groups (25); however, there is no conclusive evidence to suggest that current methods of pancreatic screening offer any survival benefit (26).Thus, the overall impact of a positive genetic testing for CDKN2A/CDK4 in the early detection and improved outcome of internal cancers and, in particular, of pancreatic cancer remains undetermined.Our current study did not reveal any new variants or mutations of CDKN2A.A148T and R24P variant were the most frequent variants in the Greek familial melanoma patients.A148T has been cited as a common polymorphism in the general population predominantly of European ancestry in many studies (27)(28)(29)(30), while, in some geographical areas, it has also been strongly correlated with familial melanoma (29,31).In a previous study (32) we identified this polymorphism in 22/304 of sporadic cases (7.2%) and in 3/9 of familial cases (33.3%).R24P has also been reported in other European populations, mainly British (14,33), but shows a higher frequency in our Greek familial cases.In our familial cases we identified 2 distinct p.G101 mutations, p.G101R and p.G101E, whereas all other studies report a W protein change at the same position (1,15,34).Overall, these changes pinpoint that this position could be a hotspot for mutagenesis.
The age of onset was lower in CDKN2A + patients compared with CDKN2A wildtype, as expected.Specifically, the mean age at melanoma diagnosis was shifted a decade earlier in carriers of CDKN2A variants, highlighting the need for intense vigilance in these individuals from a young age.Presence of MC1R variations was more frequent among CDKN2A + patients, but not at a statistically significant level.It has been shown recently that MC1R variants increase the penetrance of CDKN2A variants, especially with respect to multiple MC1R variants and to the presence of R alleles (35,36).It is unclear whether other common variants that have been associated with melanoma, particularly other pigmentation genes, such as SLC45A2, TYR or TYRP-1, may also exhibit a similar modifying effect CDKN2A penetrance similar to MC1R.
Our study has certain limitations.Although it presents the most extensive and best characterized series of Greek melanoma probands published to date, the number of patients and corresponding families is relatively small compared with other published studies on familial melanoma.In addition, we did not genotype the rest of the known high-risk genes involved in familial melanoma in order to investigate the full spectrum of melanoma-related variants in our series.Nevertheless, a subset (49 cases) of these probands has undergone germline whole exome sequencing and no variants in the other high-risk loci were discovered (37).Finally, we did not include in our study other high-risk individuals, such as patients with early onset of disease, as the goal of this study was to present a comprehensive analysis of familial melanoma cases.A more detailed genetic characterization of our expanding melanoma-prone patient series will be our goal for the near future.
In conclusion, our study delineates the important role of CDKN2A variants in Greek familial melanoma families, as almost half of our studied population carried a diseasecausing variant in this gene.Intense vigilance and early intervention are important, since age of diagnosis is low, especially among CDKN2A mutants, while medical advice on healthy lifestyle and prompt screening for other cancers, such as pancreatic cancer, is essential.

ACKNOWLEDEGMENTS
The authors acknowledge the generous assistance and support of the patients in this study.

Fig. 1 .
Fig. 1.Difference in distribution at the age of melanoma onset according to CDKN2A status.
Funding sources: This research was financed by Greek national funds through the Operational Program "Education and Lifelong Learning" of the National Strategic Reference Framework

Table I . Demographic characteristics, history of multiple primary melanomas and phenotypic characteristics according to CDKN2A status
a Always burn, minimal tan.b Burn then tan well.c No burn, tan well.*Fisher's exact test.SD: standard deviation.