CDKN2A Common Variants and Their Association with Melanoma Risk: a Population-Based Study

Research Article

CDKN2A Common Variants and Their Association with Melanoma Risk: A Population-Based Study

Tadeusz De¸bniak,1 Rodney J. Scott,4 Tomasz Huzarski,1 Tomasz Byrski,1 Andrzej Rozmiarek,5 Bogusław De¸bniak,6 Elz˙bieta Załuga,2 Romuald Maleszka,2 Jo´zef Kładny,3 Bohdan Go´rski,1 Cezary Cybulski,1 Jacek Gronwald,1 Grzegorz Kurzawski,1 and Jan Lubinski1

Departments of 1Genetics and Pathology, International Hereditary Cancer Center and 2Dermatology and Venerology, Pomeranian Medical University; 3III Department of Surgery, Pomeranian Academy of Medicine, Szczecin, Poland; 4Discipline of Medical Genetics, Faculty of Health, University of Newcastle and Hunter Medical Research Institute, Newcastle, Australia; 5Regional Hospital, Zielona Go´ra, Poland; 6Regional Hospital, Gorzow Wielkopolski, Poland

Abstract Germ line mutations of the CDKN2A have been detected with The population frequencies of the CDKN2A variants remain varying frequencies (6-48%) in small samples of French (5), U.S. (6), undetermined. In Poland there are three common variants Australian (7), Dutch (8), Italian (9), Swedish (10), English (11), and of CDKN2A: an alanine to threonine substitution (A148T), Polish melanoma-prone families (12, 13). Thus far, there have been Nt500c>g and Nt540c>t, which have been detected in other 77 causative CDKN2A variants listed in the international melanoma populations. To establish if they are associated with an mutation database, eMelanoBase. The frequency of these CDKN2A increased malignant melanoma (MM) risk we did an associa- variants in the reported populations remains undetermined; tion study based on genotyping 471 patients with MM and 1,210 however, because they result in an altered protein product they are random control subjects from the same Polish population. We considered to account for rare familial forms of the disease. Two common polymorphisms in the 3V untranslated region of found a significantly increased frequency of the A148T variant among patients with MM (7.0%) in comparison with the CDKN2A have also been described as being associated with a general population (2.9%). The incidence of the A148T variant modulation of risk (14) or disease progression (15). However, remained greater in both unselected and familial melanoma one study indicates that there is no overrepresentation of subgroups. A statistically significant positive association was the Nt500c>g and the Nt540c>t variants in the melanoma seen for unselected MM (odds ratio, 2.529; P = 0.0003), population (16). V especially in patients diagnosed under 50 years of age (odds In Poland, apart from the two polymorphisms in the 3 ratio, 3.4; P = 0.0002). The A148T carrier population untranslated region, there is one more common variant of (heterozygous G/A alleles) was more likely to have a relative CDKN2A—an alanine to threonine substitution at codon 148 with malignancy compared with the noncarrier population (A148T)—which has been estimated to be present in approxi- (57% versus 36%, respectively; P = 0.03). Further examination mately 3% to 3.5% of the population (12, 13). Functional studies of the CDKN2A promoter sequence done in 20 melanoma suggest this variant is a polymorphism, which seems to have no patients with the A148T change (heterozygous G/A alleles) and major affect on p16 function (17, 18). Previous studies have shown that the A148T polymorphism is in linkage disequilibrium 20 patients with MM without this alteration identified it was in linkage disequilibrium with a polymorphism in the promoter with a promoter polymorphism P-493, which has been shown to region at position P-493. We found no statistically significant affect gene expression (19, 20). Nevertheless, the A148T change overrepresentation of the Nt500c>g and the Nt540c>t poly- has been found to be overrepresented in melanoma kindreds morphisms in the Polish melanoma population. In conclusion, (3%) in comparison to the general population (1.8%; ref. 16). the A148T variant of the CDKN2A gene seems to be associated To establish if the A148T, Nt500c>g, and the Nt540c>t variants with an increased risk of development of MM. Additional are associated with increased melanoma risk we did an association studies are required to confirm whether this particular change study based on genotyping 471 patients with MM and 1,210 is associated with increased risk of other nonmelanoma random control subjects from the same Polish population. malignancies. (Cancer Res 2005; 65(3): 835-9) Patients and Methods Introduction Patients. The unselected case group consisted of 471 Cutaneous and ocular malignant melanomas (MM) represent patients with MM (264 females, 207 males; mean age at onset, one of the most aggressive neoplasms and their frequency is 54.5 years; age range, 20-85 years), comprising 301 unselected increasing rapidly (1, 2). The genetic basis of MM is complex and patients with MM (mean age, 54.7 years; range, 26-78 years) seems to involve multiple genes. CDKN2A (OMIM 60160) was the diagnosed in northwestern Poland between 2000 and 2003 first to be associated with MM risk and is regarded as the major (Szczecin, Gorzo´w Wlkp, Zielona Go´ra), 80 unselected consec- MM susceptibility gene (3). Its protein product p16 is a cyclin- utive MM cases (mean age, 53.2 years; range, 29-74 years) diagnosed between 2002 and 2003 in northeastern Poland dependent kinase inhibitor that suppresses cell proliferation (4). (Białystok), and 90 unselected consecutive MM cases (mean age, 54.1; age, 34-79) diagnosed between 2002 and 2003 in southwest Poland (Opole). All MM cases were taken from cancer registries in the five Requests for reprints: Tadeusz De¸bniak, Połabska 4, 70-115 Szczecin, Poland. Phone: 48-91-4661533; Fax: 48-91-4661533. E-mail: [email protected]. cities mentioned above. Participation rates exceeded 75% for I2005 American Association for Cancer Research. Białystok and Opole. In northwestern Poland (301 unselected www.aacrjournals.org 835 Cancer Res 2005; 65: (3). February 1, 2005

Table 1. Frequencies of CDKN2A variant alleles in controls and cases

A148T Nt500c>g

Frequency OR (95% CI) Frequency

Newborns (n = 500) 0 (0%) A/A 6 (1.2%) G/G 14 (2.8%) G/A 88 (17.6%) G/C 486 (97.2%) G/G 406 (81.2%) C/C Adults (n = 710) 0 (0%) A/A 9 (1.3%) G/G 21 (2.95%) G/A 143 (20.1%) G/C 689 (97.05%) G/G 558 (78.6%) C/C Total controls (n = 1,210) 0 (0%) A/A 15 (1.2%) G/G 35 (2.89%) G/A 229 (18.9%) G/C 1,175 (97.1%) G/G 966 (79.8%) C/C Allele A frequency 1.5% Allele G frequency 10.7% Unselected MM (n = 471) 0 (0%) A/A 8 (1.7%) G/G c 33 (7%) G/A 2.529 (1.552-4.121) 103 (21.9%) G/C c 438 (93%) G/G 0.395 (0.243-0.644) 360 (76.4%) C/C b Allele A frequency 3.5% 2.474 (1.528-4.005) Allele G frequency 12.6% Melanoma <50 (n = 172) 0 (0%) A/A 1 (0.6%) G/G b 16 (9.3%) G/A 3.443 (1.862-6.367) 32 (18.6%) G/C b 156 (90.7%) G/G 0.290 (0.157-0.537) 139 (80.8%) C/C b Allele A frequency 4.7% 3.324 (1.819-6.074) Allele G frequency 9.9% Melanoma >50 (n = 299) 0 (0%) A/A 7 (2.3%) G/G 17 (5.7%) G/A 2.024vx(1.117-3.665) 71 (23.7%) G/C 282 (94.3%) G/G 0.494x (0.273-0.895) 221 (73.9%) C/C Allele A frequency 2.8% 1.994c (1.109-3.585) Allele G frequency 14.2%

*P > 0.5, non significant. cP = 0.0003. bP = 0.0002. xP = 0.0315.

cases), participation rate exceeded 75%; in Szczecin (214 cases The Nt500c>g and the Nt540c>t variants were analyzed by RFLP regarded as unselected consecutive MM), the rate exceeded 50% PCR, using np16ex3f (TGAAGCCATTGCGAGAACTT) and np16ex3r but not 75% in Gorzo´w Wlkp (47 cases regarded as unselected (TCTACGTTAAAAGGCAGGAC) primers. PCR products were but nonconsecutive) and Zielona Go´ra (40 cases regarded as digested with the AvaI and HaeIII enzymes and separated in 2% unselected but nonconsecutive). to 3% agarose gels. Among 471 unselected cases, 56 patients (11.9%) had a first- or In cases positive in RFLP PCR, DNA samples were sequenced to second-degree relative affected with MM of which 14 patients confirm the presence of the mutation. Sequencing was done with (3.0%) had a first-degree relative affected with MM (familial BigDye Terminator Ready Reaction kit and analyzed in ABI PRISM melanoma cases). 377 DNA sequencer (Applied Biosystems). The control population consisted of 500 consecutive newborns Additional analysis of the P-493 polymorphism was based on from the clinical hospitals of Szczecin and 710 controls selected genomic sequencing of CDKN2A promoter sequence done using at random from the computerized patient lists of five family primers P96F (AAAGCAGGGGGCACTCATATTC) and P968R practices in Szczecin, Białystok, and Opole (363 females; mean (TCCGAGCACTTAGCGAATGT). age, 52.3 years; range, 21-77 years) and 347 males (mean age, Statistics. To evaluate whether identified CDKN2A alterations 54.7 years; range, 19-81 years). To ensure comparability with the were associated with MM we compared the frequency of control groups, the allele frequencies of the three alleles under occurrence of detected variant in our patients with the control study was computed separately for the adult and neonatal group from the general population, using two-sided Fisher’s exact control groups. test. The patient cohort was also further analyzed by dividing the Materials. DNA samples were obtained from peripheral blood MM group into those persons <50 years of age and comparing the of individuals or from umbilical chord blood of newborns. frequency of the variants observed in this group to those >50 years The A148T variant was analyzed by RFLP PCR, using np16ex2f of age. The age determinant of 50 years was chosen because those (AGGGGTAATTAGACACCTGG) and np16ex2r (TTTGGAAGCTCT- cases <50 years of age were more likely to be due to genetic CAGGGTAC) primers. PCR products were digested with the SacII influences compared with those patients >50 years of age whose enzyme and separated in 2% to 3% agarose gels. disease was more likely to be a result of environmental exposure.

Cancer Res 2005; 65: (3). February 1, 2005 836 www.aacrjournals.org

Table 1. Frequencies of CDKN2A variant alleles in controls and cases (Cont’d)

Nt540c>t

OR (95% CI) Frequency OR (95% CI)

4 (0.8%) T/T 62 (12.4%) C/T 434 (86.8%) C/C 3 (0.4%) T/T 83 (11.7%) C/T 625 (88%) C/C 7 (0.6%) T/T 145 (12%) C/T 1,058 (87.4%) C/C Allele T frequency 6.6% 1.377* (0.5796-3.269) 3 (0.6%) T/T 1.102* (0.2836-4.280) 1.199* (0.9230-1.558) 64 (13.5%) C/T 1.155* (0.8424-1.584) 0.818* (0.6348-1.057) 404 (85.7%) C/C 0.866* (0.6357-1.181) 1.206* (0.9571-1.521) Allele T frequency 7.4% 1.142* (0.8526-1.528) 0.4659* (0.0611-3.551) 2 (1.2%) T/T 2.022* (0.4164-9.816) 0.9792* (0.6497-1.476) 21 (12.2%) C/T 1.021* (0.6267-1.665) 1.064* (0.7098-1.595) 149 (86.6%) C/C 0.931* (0.5811-1.491) 0.9151* (0.6277-1.334) Allele T frequency 7.3% 1.114* (0.7193-1.727) 1.910* (0.7715-4.728) 1 (0.3%) T/T 0.576* (0.0706-4.708) 1.334* (0.9855-1.806) 43 (14.4%) C/T 1.234* (0.8549-1.780) 0.716* (0.5334-0.9602) 255 (85.3%) C/C 0.833* (0.5795-1.196) 1.382x (1.062-1.799) Allele T frequency 7.5% 1.157* (0.8205-1.632)

Possible deviation of the genotype frequencies from those To further determine the importance of the A148T change, its expected under Hardy-Weinberg equilibrium was assessed by the frequency was assessed in patients who were <50 years of age and m2 probability test (21). compared with those >50 years. The results reveal that the frequency is greater in the <50-year-old group compared with the Results >50-year-old group (see Table 1). The OR was consequently greater The frequency of the A148T change was assessed in the two in the younger age group (OR, 3.5; P = 0.0007) compared with the control populations (newborns versus adults) and the frequencies older group (OR, 2.1; P = 0.0351). There was no statistical difference were 2.8% and 2.95% (Table 1). These were not statistically different between the frequency and OR of the polymorphism between both (P = 1.0). Further verification of the similarities of the two control groups (OR, 1.6; P = 0.265). populations were undertaken by examining a common CHEK2 The clinical features of patients harboring the A148T change polymorphism. There was no statistical difference in the CHEK2 were compared with patients who did not. The mean age of A148T allele frequencies in the adult controls recruited from the Szczecin carriers was 53 years and the mean age of noncarriers was 55 years; metropolitan region compared with other Polish cities (data not this was not statistically significant (P = 0.632). shown). It was then considered justified to use the combined Among the cancers seen in the first-degree relatives of the control population frequency of 2.89% for statistical analysis. mutation-positive melanoma patients were stomach (2 cases), There were no large differences in the frequencies of the larynx (2), pancreas (2), lung (3), breast (1), female genital (3), Nt500c>g alleles and in the frequencies of the Nt540c>t alleles in bladder (3), and skin (2) cancers and single cases of colon, kidney, the two control populations (Table 1). Because these were not and liver cancer. statistically significant (data not shown), combined control To further define the association of the A148T change and its link population frequencies were used for statistical analyses of both to malignancy we examined the occurrence of cancer of any type in of these CDKN2A changes. first-degree relatives of carriers compared with noncarriers. Familial There was no evidence that the genotype frequencies of the melanoma cases were excluded from this evaluation. This analysis three CDKN2A variants deviated from those expected under revealed that the carrier population was more likely to have a relative Hardy-Weinberg for the control group or any of the melanoma with malignancy compared with the noncarrier population (57% groups (P > 0.4). versus 36%, respectively; P = 0.03). In the melanoma population under study the frequency of the Because there is doubt about the functional consequence of the A148T variant was significantly greater than that observed in the A148T change, but it is in linkage disequilibrium with a promoter control population (see Table 1) and it was associated with a polymorphism that affects gene expression levels, 20 A148T significant increase in the odds ratio (OR) of developing melanoma. heterozygous carriers and 20 A148T noncarriers were studied to www.aacrjournals.org 837 Cancer Res 2005; 65: (3). February 1, 2005

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 2005 American Association for Cancer Research. Cancer Research determine the proportion of samples that were linked. The results Because the A148T change has not been shown to alter protein revealed that all A148T heterozygous carriers were heterozygous A/ function yet seems to be associated with disease it is most likely that T carriers at position P-493 and all A148T noncarriers were A/A this polymorphism is in linkage disequilibrium with another homozygous carriers at position P-493. alteration that does affect protein function. Indeed, there are several There was no statistically significant overrepresentation of the reports indicating that the A148T polymorphisms is in linkage Nt500c>g and the Nt540c>t polymorphisms in the Polish melanoma disequilibrium with a change in the promoter region at position population (Table 1). P-493 (19, 20) and the Nt540T in the 3Vuntranslated region (16). It is beyond the scope of this study, but it would be interesting to evaluate Discussion different common polymorphisms, such as the Nt191 in the 5‘ untranslated region (28). Further examination of the CDKN2A In a previous study we identified the A148T variant (Nt442A) in a promoter sequence done in 20 melanoma patients with A148T group of familial melanoma patients and melanoma patients with change and 20 patients with MM without this alteration supports the familial aggregations of breast cancer (12). findings of others that linkage disequilibrium exists between the This is the first case-control study that shows an association A148T polymorphism and the P-493 polymorphism. These results between 148Talteration and increased risk of melanoma. The A148T indicate that the Polish population under study is not significantly variant is broadly cited as a common polymorphism on the basis of different to other European populations with respect to the A148T two observations: functional studies indicating that this amino acid change and the P-493 polymorphism. substitution does not affect the ability of the p16 to precipitate with The results of our study indicate that in the Polish population CDK4/CDK6 proteins (17, 18) and it does not segregate with the A148T variant is associated with an increased melanoma risk melanoma in melanoma-prone families (22, 23). Some authors have especially for patients diagnosed under 50 years of age (OR, 3.5). reported an increased incidence of the A148T variant in small The A148T can be added to the list of DNA variants, which are cohorts of melanoma-prone families (6, 11). It has been found to be believed to predispose to MM. Others include the melanocortin-1 present in 4% of the Utah population (24), 3% of parents of the Centre receptor (MC1R) gene variants, which were the first reported to be d’Etude Polymorphisme Humain (22), and 1.8% of the Queensland influencing melanoma risk independently of fair skin and red hair population (16). In all instances, these studies have focused on (29). Polymorphisms in the nucleotide excision repair gene XPD relatively small numbers of patients; even in the larger study done by have also been reported to be associated with an increased risk of Aitken et al. (16) the number of control subjects did not exceed 200. melanoma (30, 31). In the current study, the frequency of the A148T change in 1,210 The frequencies of the two 3Vuntranslated region polymorphisms control samples was found to be 3%. recorded in the Polish control population are similar to those Population stratification might explain significant overrepresen- reported in the Nordic healthy control population (14). However, we tation of the A148T allele in the Polish population. In a recent study found lower incidences of these polymorphisms among melanoma Acton et al. found that the frequencies of Celtic ancestry and patients in comparison with Nordic patients affected with this associated marker HLA-DRB1*04 were statistically greater in the disease. The results of our statistical analyses are consistent with melanoma cases (25). It seems that there are no major population data presented by Aitken et al. (16) in that we did not find any substructures within Poland. First, most of the cases used in the significant increase of the two polymorphisms in our melanoma present study were recruited from the northwest region, which is population and that for each polymorphism the rare allele frequency populated by ethnic Poles who migrated to the region from was higher in melanoma cases than controls. throughout Poland. Second, there were no significant differences in In conclusion, the A148T variant of the CDKN2A gene seems to be the CDKN2A allele frequencies in the three examined regions of associated with an increased risk of developing MM. Nevertheless Poland. Finally, the similar frequency of the three polymorphisms this does not imply that the A148T polymorphism is associated with identified in the newborn and adult populations indicates that the increased disease risk alone but rather suggests there are interactive polymorphisms are in genetic equilibrium. These findings and modifiers that when taken together result in malignancy. Additional those published previously by our group (26, 27) suggest that the studies are required to determine whether this particular change can Polish population is homogeneous. On the other hand, we observed be associated with the increased risk of malignancies at different a trend of higher G allele frequency, although statistically sites of origin as suggested by the increased frequency of cancer in nonsignificant, for Nt500 variant in late-onset melanoma cases first-degree relatives of A148T carriers. (P = 0.0664). This could point at small numbers of cases or possible population stratification. Thus it is necessary to determine the Acknowledgments distribution of genotypes at a number of other unrelated loci throughout Poland. To evaluate the hypothesis that A148T Received 8/30/2004; revised 11/4/2004; accepted 11/24/2004. The costs of publication of this article were defrayed in part by the payment of page aggregates with country of ancestry, other Slavic populations charges. This article must therefore be hereby marked advertisement in accordance should be examined. with 18 U.S.C. Section 1734 solely to indicate this fact.

References with multiple cutaneous melanomas. Cancer Res melanoma-prone families in France. The French 2000;60:6864–7. Familial Melanoma Study Group. Hum Mol Genet 1. Goldstein AM, Tucker MA. Genetic epidemiology of 4. Serrano M, Hannon GJ, Beach D. A new 1998;7:941. familial melanoma. Dermatol Clin 1995;13:605–12. regulatory motif in cell-cycle control causing 6. FitzGerald MG, Harkin DP, Silva-Arrieta S, et al. 2. Greene MH. The genetics of hereditary melanoma and specific inhibition of cyclin D/CDK4. Nature 1993; Prevalence of germ-line mutations in p16, p19ARF, nevi. 1998 Update. Cancer 1999;86:2464–77. 366:704–7. and CDK4 in familial melanoma: analysis of a clinic- 3. Hashemi J, Platz A, Ueno T, Stierner U, Ringborg U, 5. Soufir N, Avril MF, Chompret A. et al. Prevalence based population. Proc Natl Acad Sci U S A 1996;93: Hansson J. CDKN2A germ-line mutations in individuals of p16 and CDK4 germline mutations in 48 8541–5.

Cancer Res 2005; 65: (3). February 1, 2005 838 www.aacrjournals.org

7. Walker GJ, Hussussian CJ, Flores JF, et al. Mutations CDK4 and p53 genes are rare. Int J Cancer 2001;95: nevus/melanoma susceptibility allele. J Invest Derma- of the CDKN2/p16INK4 gene in Australian melanoma 388–93. tol 2002;119:961–5. kindreds. Hum Mol Genet 1995;4:1845–52. 15. Straume O, Smeds J, Kumar R, Hemminki K, Akslen 24. Kamb A, Shattuck-Eidens D, Eeles R, et al. Analysis 8. Gruis NA, Sandkuijl LA, van der Velden LA. Significant impact of promoter hypermethylation of the p16 gene (CDKN2) as a candidate for the PA, Bergman W, Frants RR. CDKN2 explains part of and the 540 C>T polymorphism of CDKN2A in chromosome 9p melanoma susceptibility locus. Nat the clinical phenotype in Dutch familial atypical cutaneous melanoma of the vertical growth phase. Genet 1994;8:23–6. multiple-mole melanoma (FAMMM) syndrome fami- Am J Pathol 2002;161:229–37. 25. Acton RT, Barton EH, Hollowell WW, Dreibelbis AL, lies. Melanoma Res 1995;5:169–77. 16. Aitken J, Welch J, Duffy D, et al. CDKN2A variants in Go RC, Barton JC. Ancestry reported by white adults 9. Ciotti P, Strigini P, Bianchi-Scarra G. Familial a population-based sample of Queensland families with cutaneous melanoma and control subjects in melanoma and pancreatic cancer. Ligurian Skin with melanoma. J Natl Cancer Inst 1999;9:446–52. central Alabama. BMC Cancer 2004;4:47. Tumor Study Group. N Engl J Med 1996;334:469–70. 17. Ranade K, Hussussian CJ, Sikorski RS, et al. 26. Gorski B, Jakubowska A, Huzarski T, et al. A high 10. Platz A, Hansson J, Mansson-Brahme E, et al. Mutations associated with familial melanoma impair proportion of founder BRCA1 mutations in Polish Screening of germline mutations in the CDKN2A p16INK4 function. Nat Genet 1995;10:114–6. breast cancer families. Int J Cancer 2004;110:683–6. and CDKN2B genes in Swedish families with 18. Lilischkis R, Sarcevic B, Kennedy C, Warlters A, 27. Cybulski C, Huzarski T, Gorski B, et al. A novel hereditary cutaneous melanoma. J Natl Cancer Inst Sutherland RL. Cancer-associated mis-sense and founder CHEK2 mutation is associated with 1997;89:697–702. deletion mutations impair p16INK4 CDK inhibitory increased prostate cancer risk. Cancer Res 11. Harland M, Meloni R, Gruis N, et al. Germline activity. Int J Cancer 1996;66:249–54. 2004;64:2677–9. mutations of the CDKN2 gene in UK melanoma 19. Mantelli M, Barile M, Ciotti P, et al. High prevalence 28. Pollock PM, Stark MS, Palmer JM, et al. Mutation families. Hum Mol Genet 1997;6:2061–7. of the G101W germline mutation in the CDKN2A analysis of the CDKN2A promoter in Australian 12. Debniak T, Gorski B, Scott RJ, et al. Germline (P16(ink4a)) gene in 62 Italian malignant melanoma melanoma families. Genes Chromosomes Cancer mutation and large deletion analysis of the CDKN2A families. Am J Med Genet 2002;107:214–21. 2001;32:89–94. and ARF genes in families with multiple melanoma or 20. Harland M, Holland EA, Ghiorzo P, et al. 29. Palmer JS, Duffy DL, Box NF, et al. Melanocortin-1 an aggregation of malignant melanoma and breast Mutation screening of the CDKN2A promoter in receptor polymorphisms and risk of melanoma: is the cancer. Int J Cancer 2004;110:558–62. melanoma families. Genes Chromosomes Cancer association explained solely by pigmentation pheno- 13. Lamperska K, Karezewska A, Kwiatkowska E, 2002;28:45–57. type? Am J Hum Genet 2000;66:176–86. Mackiewicz A. Analysis of mutations in the p16/ 21. Ott J. Utility programs for analysis of genetic 30. Baccarelli A, Calista D, Minghetti P, et al. XPD gene CDKN2A gene in sporadic and familial melanoma in linkage. Program HWE. 1988. Available from: http:// polymorphism and host characteristics in the associ- the Polish population. Acta Biochim Pol 2002;49: www.hgmp.mrc.ac.uk/Registered/Help/linkutil/. ation with cutaneous malignant melanoma risk. Br J 369–76. 22. Hussussian CJ, Struewing JP, Goldstein AM, et al. Cancer 2004;90:497–502. 14. Kumar R, Smeds J, Berggren P, et al. A single Germline p16 mutations in familial melanoma. Nat 31. Tomescu D, Kavanagh G, Ha T, Campbell H, Melton nucleotide polymorphism in the 3’ untranslated region Genet 1994;8:15–21. DW. Nucleotide excision repair gene XPD polymor- of the CDKN2A gene is common in sporadic primary 23. Bertram CG, Gaut RM, Barrett JH, et al. An phisms and genetic predisposition to melanoma. melanomas but mutations in the CDKN2B, CDKN2C, assessment of the CDKN2A variant Ala148Thr as a Carcinogenesis 2001;22:403–8.

www.aacrjournals.org 839 Cancer Res 2005; 65: (3). February 1, 2005

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 2005 American Association for Cancer Research. CDKN2A Common Variants and Their Association with Melanoma Risk: A Population-Based Study

Tadeusz Debniak, Rodney J. Scott, Tomasz Huzarski, et al.

Cancer Res 2005;65:835-839.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/65/3/835

Cited articles This article cites 28 articles, 3 of which you can access for free at: http://cancerres.aacrjournals.org/content/65/3/835.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/65/3/835.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/65/3/835. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.