Xeroderma pigmentosum genes and melanoma risk
Correspondence to: Katarzyna Paszkowska-Szczur, Department of Genetics and Pathology, International Hereditary Cancer Center, Pomeranian Medical University, Polabska 4, 70-115 Szczecin, Poland, Tel.: +48-91-4661532, Fax: +48-91-4661533, E-mail: firstname.lastname@example.org
Xeroderma pigmentosum is a rare autosomal recessive disease that is associated with a severe deficiency in nucleotide excision repair. The presence of a distinct the nucleotide excision repair (NER) mutation signature in melanoma suggests that perturbations in this critical repair process are likely to be involved with disease risk. We hypothesized that persons with polymorphic NER gene(s) are likely to have reduced NER activity and are consequently at an increased risk of melanoma development. We assessed the association between 94 SNPs within seven XP genes (XPA–XPG) and the melanoma risk in the Polish population. We genotyped 714 unselected melanoma patients and 1,841 healthy adults to determine if there were any polymorphisms differentially represented in the disease group. We found that a significantly decreased risk of melanoma was associated with the Xeroderma pigmentosum complementation (XPC) rs2228000_CT genotype (odds ratio [OR] = 0.15; p < 0.001) and the rs2228000_TT genotype (OR = 0.11; p < 0.001) compared to the reference genotype. Haplotype analysis within XPC revealed the rs2228001_A + G1475A_G + G2061A_A + rs2228000_T + rs3731062_C haplotype (OR = 0.26; p < 0.05) was associated with a significantly decreased disease risk. The haplotype analysis within the Xeroderma pigmentosum group D (XPD) showed a modest association between two haplotypes and a decrease in melanoma risk. There were no major differences between the prevalence of the XP polymorphisms among young or older patients with melanoma. Linkage disequilibrium of XPC: rs2228001, G1475A, G2061A, rs2228000 and rs3731062 was found. The data from our study support the notion that only XPC and XPD genes are associated with melanoma susceptibility.
auxin response factors
cyclin-dependent kinase 4
cyclin-dependent kinase inhibitor 2A
cytochrome P450 2D6
glutathione S-transferase Mu 1
melanocortin 1 receptor
the nucleotide excision repair
vitamin D receptor
- XP genes
Xeroderma pigmentosum genes
Malignant melanoma is one of the most aggressive neoplasms afflicting mankind. It has been reported that melanoma incidence in Caucasians has increased rapidly (10-fold) over the last 50 years. Environmental factors combined with a genetic susceptibility lead to melanoma development. A major risk factor for melanoma is ultraviolet (UV) radiation as it causes severe DNA damage.[2, 3] To ameliorate the effects of UV-light-induced DNA damage, the nucleotide excision repair (NER) system has evolved and is the major defense against the mutagenic effects of UV light. NER system detects and repairs DNA damage such as UV-light-induced DNA photoproducts, the removal of large DNA adducts and crosslinks. Malfunction of NER results in a reduced DNA repair capacity and a propensity to accumulate mutations. The NER system involves products of about 30 genes that include seven genes (XPA–XPG) associated with the NER disorder Xeroderma pigmentosum, a disease characterized by the development of excessive skin malignancies including melanoma. There are many reports of the relationship between polymorphisms in XP genes and melanoma risk, but most of them evaluated only single-nucleotide polymorphisms (SNPs) among XP genes or they were performed on small numbers of patients. Overall, recent findings indicate that melanomas harbor a wide variety of mutation signatures but by far the most important are those occurring as a result of UV-light exposure and a failure in NER. Possible associations between XPA and melanoma risk have been reported but the data remain inconclusive.[8, 9] Xeroderma pigmentosum complementation (XPC) and its association with melanoma risk has been more widely reported[8-13]; however, there is little consistency in the effect sizes. A correlation between XPC and melanoma was confirmed in two of them[8, 13] but not by others.[9-12] A much larger series of studies have been reported for the Xeroderma pigmentosum group D (XPD) and melanoma risk. A positive association between SNPs in XPD and melanoma risk has been observed in five studies that mainly concern the correlation between melanoma and Lys751Gln and Asp312Asn polymorphisms in XPD. Similar to other XP gene association studies, the role of XPD polymorphisms and disease risk remains controversial. Some reports indicate an association with these common variants and melanoma development,[11, 14-17] whereas others have not observed any correlation.[8-10, 18] Studies on the association of XPF polymorphisms and melanoma risk have revealed conflicting results, with only one positive association reported. None of the SNPs examined in XPF was statistically associated with melanoma risk.[9, 10] XPG was evaluated by six groups[8-12, 19] and found by the majority not to be associated with disease risk and only one report suggested an association. XPE and XPB have not been previously examined for their association with melanoma.
The genetic background of melanoma is complex and associated with many susceptibility genes that include high-penetrance genes CDKN2A, ARF, CDK4 and low-penetrance ones MC1R, CYP2D6, VDR and GSTM1. There are several lines of epidemiological and genetic evidence that implicate DNA NER genes in melanoma development. The association between melanoma risk and XP genes is, however, not well established and requires validation in larger more comprehensive studies. We genotyped herein 94 SNPs (nonsynonymous or located on exon/intron boundaries or in UTR sequences) in seven XP genes (XPA–XPG) in series of 714 unselected melanoma patients and 1,841 healthy controls from the Polish population. This is the first report to undertake such an extensive study of NER gene polymorphisms and their association with melanoma risk.
Material and Methods
The melanoma patient group consisted of 714 unselected participants, 451 women (mean age, 63 years) and 263 men (mean age, 65.5 years) from Poland. Patients were diagnosed between 2002 and 2006 and identified from cancer registries in five Polish cities (Szczecin, Gorzow Wielkopolski, Zielona Gora, Bialystok and Opole). The registries capture more than 95% of all diagnosed melanomas. Participation rates were more than 75% for all centers.
Controls consisted of 1,841 healthy adults: 860 women (mean age, 64 years) and 981 men (mean age, 67 years) with a negative cancer family history. The healthy adults were assessed as having a negative cancer family history (first- and second-degree relatives included) after answering a questionnaire about their family's medical history which was part of a population-based study of the 1.5 million residents of West Pomerania to identify familial aggregations of malignancies performed recently by our center. During the interview, the goals of the study were explained, informed consent was obtained, genetic counseling was given and a blood sample was taken for DNA analysis. Individuals affected with any malignancy or with cancers diagnosed among first- or second-degree relatives were excluded from our study control group.
The study conformed with the Declaration of Helsinki and all participants signed an informed consent document prior to donating a blood sample. The study was approved by the institutional review board of the Pomeranian Medical University.
DNA samples were obtained from peripheral blood of melanoma patients and healthy adults according to the method of Miller et al.
MassARRAY MALDI-TOF MS (Sequenom reaction)
All nonsynonymous or exon/intron boundary or located in UTR sequences SNPs (total number, 94) located in the seven DNA repair genes (XPA: 8 SNPs, XPB: 10 SNPs, XPC: 16 SNPs, XPD: 11 SNPs, XPE: 5 SNPs, XPF: 16 SNPs XPG: 28 SNPs) were selected from http://snpper.chip.org/ and genotyped among 190 cases using the MassARRAY MALDI-TOF MS platform for genotyping (Sequenom, San Diego, CA). MALDI-TOF uses a primer extension reaction to detect and determine the SNP allele. Reactions were performed according to the manufacturer's instructions. This analysis allowed the identification of all XP polymorphic SNPs present in Polish population. Monomorphic SNPs were excluded from the further analysis.
Out of 94 SNPs, 27 were polymorphic and these were selected for further analyses. In all, 24 SNPs were genotyped among 714 melanoma cases (including 180 cases used for the first step) and 1,841 healthy controls using the MassARRAY MALDI-TOF MS platform for genotyping.
Owing to technical reasons, 3 (XPA rs1800975, XPD rs1799793 and XPG rs4150360) out of 27 SNPs mentioned above were not genotyped by MassARRAY MALDI-TOF MS, but by real-time PCR using TaqMan genotyping assays (LightCycler 480, Roche, Penzberg, Germany). TaqMan assays (Applied Biosystems, Foster City, CA) contained two specific primers and two TaqMan fluorescent probes for the detection of SNPs.
Random DNA samples were sequenced to verify the results of the MassARRAY genotyping and real-time PCR data (data not shown). Sequencing was performed with universal primers using the ABI PRISM BigDye Terminator Cycle (Applied Biosystems, Foster City, CA).
Minimum threshold for minor allele genotype frequency was set at 5%. Out of 27 polymorphic SNPs, 16 were used in final analyses.
In this case–control study, several features were included for a multivariate statistical analysis. Year of birth and sex were considered as independent variables (not covariates) along with genotype data. Genotype data were categorical and not risk per allele, but the risk for particular genotypes was calculated. There were three genotypes, except for, XPC G1475A, where no homozygotes for the minor allele could be determined in the sample set. The most common genotype among controls was always taken as the reference genotype.
The disease risk associated to each factor was calculated with a logistic regression model on an R software environment (version 2.15.0). Interaction analysis was intentionally avoided owing to the limited statistical power for the present sample size and the large amount of variables considered. Estimation of haplotype frequencies and their potential association with disease risk were performed using the haplo.stats CRAN package (version 1.5.5) by Sinnwell and Schaid for an R environment.
Linkage disequilibrium between SNPs for a given haplotype was calculated using the software JLIN by Carter et al.
A list of all SNPs with their prevalence is provided in Supporting Information Table 1.
Association between SNPs and melanoma risk: Multivariate analysis
From all 16 SNPs included in the final analysis, XPC rs2228000 as found to be significantly associated with melanoma risk (Table 1). Logistic regression indicated a significantly decreased melanoma risk for the XPC rs2228000 CT genotype (odds ratio [OR] = 0.15; 95% confidence interval [CI] = 0.07–0.29; p < 0.001) and for genotype TT (OR = 0.11; 95% CI = 0.03–0.37; p < 0.001) compared to the reference genotype (CC). Analogously, XPC G1475A_AG genotype and XPD rs238406_AC genotype were associated with significantly increased risk of melanoma (OR = 3.54; 95% CI = 1.18–12.6; p < 0.05 and OR = 2.38; 95% CI = 1.14–4.96; p < 0.05, respectively) in comparison with the reference genotypes.
Table 1. Association of common variants of XP genes with melanoma risk
|XPA||rs1800975||GG||306 (45.1%)||240 (40.8%)||–||–||–|
|AG||294 (43.4%)||255 (43.4%)||0.60||1.17||0.65–2.13|
|AA||78 (11.5%)||93 (15.8%)||0.35||1.52||0.65–3.77|
|XPC||G2061A||GG||370 (56%)||827 (54.8%)||–||–||–|
| || ||AG||244 (36.9%)||592 (39.2%)||0.23||0.67||0.34–1.92|
| || ||AA||47 (7.1%)||91 (6%)||0.17||0.40||0.11–1.52|
| ||G1475A||GG||565 (90.9%)||1197 (91.1%)||–||–||–|
| || ||AG||57 (9.1%)||116 (8.9%)||0.03||3.54||1.18–12.6|
| ||rs2228000||CC||245 (47.2%)||548 (42.6%)||–||–||–|
| || ||CT||240 (46.2%)||563 (43.7%)||<0.001||0.15||0.07–0.29|
| || ||TT||34 (6.6%)||177 (13.7%)||<0.001||0.11||0.03–0.37|
| ||rs2228001||AA||227 (35.7%)||480 (36%)||–||–||–|
| || ||AC||314 (49.5%)||647 (48.4%)||0.97||0.99||0.50–1.95|
| || ||CC||94 (14.8%)||209 (15.6%)||0.46||0.65||0.21–2.03|
| ||rs3731062||CC||624 (97.7%)||1,484 (94%)||–||–||–|
| || ||CT||34 (5.2%)||93 (5.9%)||0.70||1.23||0.43–3.76|
| || ||TT||1 (0.1%)||1 (0.1%)||NA||NA||NA|
|XPD||rs13181||AA||245 (35.6%)||592 (36.2%)||–||–||–|
| || ||AC||325 (47.1%)||767 (46.9%)||0.27||0.63||0.28–1.41|
| || ||CC||119 (17.3%)||276 (16.9%)||0.93||1.06||0.32–3.60|
| ||rs1799793||GG||215 (45.2%)||614 (50.2%)||–||–||–|
| || ||AG||180 (37.8%)||460 (37.7%)||0.62||0.80||0.33–1.91|
| || ||AA||81 (17%)||148 (12.1%)||0.56||1.49||0.39–5.68|
| ||rs238406||CC||164 (28.6%)||356 (28.7%)||–||–||–|
| || ||AC||307 (53.6%)||672 (54.3%)||0.02||2.38||1.14–4.96|
| || ||AA||102 (17.8%)||211 (17%)||0.99||0.99||0.34–2.86|
|XPE||rs1050244||CC||574 (88.3%)||1,190 (87%)||–||–||–|
| || ||CT||74 (11.4%)||167 (12.2%)||0.24||0.61||0.27–1.42|
| || ||TT||2 (0.3%)||10 (0.7%)||NA||NA||NA|
|XPF||rs762521||GG||354 (56.5%)||741 (58.7%)||–||–||–|
| || ||AG||228 (36.5%)||440 (34.8%)||0.90||0.97||0.54–1.73|
| || ||AA||44 (7%)||82 (6.5%)||0.84||1.13||0.37–4.02|
|XPG||rs1047768||CC||214 (33.8%)||465 (35%)||–||–||–|
| || ||CT||291 (46%)||623 (46.8%)||0.66||0.87||0.46–1.62|
| || ||TT||128 (20.2%)||242 (18.2%)||0.86||1.08||0.47–2.52|
| ||rs1047769||AA||613 (94.3%)||1,259 (92.8%)||–||–||–|
| || ||AG||37 (5.7%)||95 (7%)||0.36||0.57||0.17–2.08|
| || ||GG||0 (0%)||2 (0.2%)||NA||NA||NA|
| ||rs17655||GG||412 (64.4%)||869 (64%)||–||–||–|
| || ||CG||200 (31.2%)||404 (29.7%)||0.20||1.64||0.78–3.61|
| || ||CC||28 (4.4%)||85 (6.3%)||NA||NA||NA|
| ||rs2227869||GG||567 (89.2%)||1,168 (87.6%)||–||–||–|
| || ||CG||67 (10.5%)||162 (12.2%)||0.36||0.60||0.20–1.84|
| || ||CC||2 (0.3%)||2 (0.2%)||NA||NA||NA|
| ||rs4150360||TT||139 (26.8%)||167 (25.8%)||–||–||–|
| || ||CT||255 (49.1%)||339 (52.2%)||0.51||0.67||0.21–2.23|
| || ||CC||125 (24.1%)||143 (22%)||0.82||1.22||0.23–7.31|
The genotype frequencies of the polymorphisms found to be significant in XPC and XPD were further analyzed to confirm the individual contribution of the polymorphism to disease risk. The results of the logistic regression were compared to the results of a Fisher's Exact Test for XPC rs2228000 assuming CC to be the reference genotype. In this case, there was a significant decreased risk for genotype TT (OR = 0.43; 95% CI = 0.28–0.64; p < 0.001) even after correction for multiple testing (after Bonferroni correction: p < 0.05), but not significant for the XPC genotype CT (OR = 0.94; 95% CI = 0.76–1.18; p = 0.62). Disease associations for XPC G1475A and XPD rs238406 were not confirmed using this approach, the OR of being a risk factor for multiple myeloma (MM), along with its 95% CI and the calculated p-value (Supporting Information Table 2).
Age and melanoma risk
Age stratification analysis failed to reveal any association of the 16 SNPs with and melanoma risk (data not shown). Year of birth did not influence disease risk although both lifetime exposure to sun and life-style factors associated with sun exposure are different for each age cohort.
Gender and melanoma risk
Stratified analysis by gender showed no association of any SNPs variants with significantly increased melanoma risk (data not shown).
Compound genotypes analysis of Lys751Gln and Asp312Asn variants in the XPD gene
No association between compound heterozygous carriers of rs18131 (Lys751Gln) and rs1799793 (Asp312Asn) variants and melanoma risk was observed (Supporting Information Table 3).
Xeroderma pigmentosum group C
XPC was genotyped at five different sites (XPC, rs2228001, G1475A, G2061A, rs2228000 and rs3731062), and hence the apparent disagreement in the results could be owing to a haplotype effect.
Haplotype analysis of melanoma development revealed that rs2228001_A + G1475A_G + G2061A_A + rs2228000_ T + rs3731062_C haplotype is associated with a significantly decreased disease risk (OR = 0.26, p = 0.32 × 10−2) compared to the reference haplotype: rs2228001_C + G1475A_ G + G2061A_G + rs2228000_C + rs3731062_C. Moreover, the C allele at rs2228000 within the same haplotype has a neutral effect (OR = 1.08, p > 0.05), but rs2228000_C in another block: rs2228001_A + G1475A_G + G2061A_ G + rs2228000_C + rs3731062_C (OR = 1.25, p = 0.02) increased melanoma risk.
For haplotype rs2228001_A + G1475A_G + G2061A_ A + rs2228000_C + rs3731062_C, results were inconclusive. In this context, rs2228000_T and G1475A_G together decreased melanoma risk (OR = 0.26, p = 0.32 × 10−2), whereas rs2228000_T and G1475A_A had a neutral effect (OR = 0.99, p > 0.05).
For the remaining haplotypes, no significant differences in disease risk between cases and controls could be detected (Table 2).
Table 2. Haplotype frequency of examined XPC variants between cases and controls
|A||G||A||T||C||0.32 × 10−2||0.26||0.07–0.91|
It is important to note that the list of haplotypes is not equivalent to the list of all theoretically possible haplotypes. Most of the latter were estimated by the software as too improbable (as there were too few cases with these haplotypes) to be considered in the analysis.
The Xeroderma pigmentosum group D
Statistic analysis showed that two haplotypes in XPD gene decrease melanoma risk (Table 3). The rs18131_C + rs1799793_G + rs238406_C haplotype decrease risk (OR = 0.618, p = 0.04) compared to referenced rs18131_A + rs1799793_G + rs238406_A and similarly rs18131_C + rs1799793_G + rs238406_A haplotype (OR = 0.547, p = 3.8 × 10−2) compared to rs18131_A + rs17 99793_G + rs238406_A.
Table 3. Haplotype frequency of examined XPD variants between cases and controls
Xeroderma pigmentosum group G
There was no significant difference for any of the XPG haplotypes among melanoma cases and healthy adults (Table 4).
Table 4. Haplotype frequency of examined XPG variants between cases and controls
The linkage disequilibrium plot (Supporting Information Fig. 1) was calculated on the basis of the genotype frequencies in our control population. The entire fragment between the five analyzed SNPs in XPC is in linkage disequilibrium and most of them with D′ values close to 1.0, except G2061A ↔ rs2228000 (D′ value range, 0.4–0.6) and rs2228001 ↔ rs2228000 (D′ value range, 0.6–0.8).
Initially, 16 SNPs within the XPC gene were examined from which five were found to be sufficiently polymorphic. Logistic regression showed that the rs2228000_CT genotype (OR = 0.15; p < 0.05) and the rs2228000_TT genotype (OR = 0.11; p < 0.05) decreased melanoma risk. Multivariate analysis of the polymorphism G1475A_AG genotype in XPC revealed that it could be associated with increased disease risk; however, Fisher's exact test showed no association of this polymorphism with melanoma development, suggesting that its effect is apparent only when other factors are taken into account. Haplotype frequency indicated the A allele at G1475A could be discriminating the protective effect of T allele at rs2228000.
Together, haplotype analysis and logistic regression indicate that the T allele at rs2228000 (XPC Ala499Val) confers protection against MM. This evidence appears to be consistent with a dominant model of protection as the ORs are almost identical for the presence of one or two copies of the T allele (0.15 and 0.11, respectively). XPC is required for DNA damage detection and initiation of the repair process. The change C → T at rs2228000 causes an Ala → Val substitution in at codon 499, which modifies the XPC protein structure and its function.
There have been only three previous reports examining a relationship between the XPC Val499Ala variant on melanoma risk[9, 11, 12] and all three failed to demonstrate any association. Our finding of a protective association against the development of melanoma does not explain the previous findings in other diverse populations. These studies may indicate allele frequency differences owing to population stratification.
We found no association between the remaining variants of XPC and melanoma risk. So far, only two reports have identified a positive association of XPC and melanoma development.[8, 13] Both studies showed an increased risk of disease associated with three polymorphisms in XPC (PAT+; IV-6A and 939Gln). These SNPs represented as a haplotype appear to increase melanoma risk as opposed to each individual polymorphism. This evidence was based on 294 melanoma patients/375 healthy controls and 202 cases/210 controls, respectively. We genotyped only XPC Lys939Gln and did not find any association with melanoma, confirming the results from the previous reports.[9, 11, 12] A German study failed to reveal an association of another SNPs: rs3731151 (XPC Arg687Arg) and rs2227999 (XPC Arg492His) with melanoma risk. Moreover, no correlation was found between rs2227999 and disease risk among 700 melanoma patients and 516 controls from France. The results of our study confirm these findings.
Out of 11 genotyped SNPs within XPD, only three were found to be polymorphic in our population. Similar to our previous report, we found no association between rs18131 or rs1799793 with melanoma development. Haplotype analysis of XPD, however, revealed a modest association between two haplotypes (rs18131_C+ rs1799793_G + rs238406_C [OR = 0.618, p < 0.05] and rs18131_C + rs1799793_ G + rs238406_A [OR = 0.547, p < 0.05]) and a decreased melanoma risk compared to the reference haplotype. No influence of the rs238406 allele on disease development and or any clear association of rs18131_C and rs1799793_G with melanoma risk was observed in our study. The partial discordance in the results of the logistic regression and the analysis of the haplotypes suggests a potential influence of other factors involved in MM development. The rs1799793 is associated with a G → A substitution in XPD at codon 312 causing an Asp → Asn substitution in a conserved section of the XPD protein. The rs13181 polymorphism results in a Lys751Gln substitution at codon 751 in the C-terminal region of the protein.[27, 28] These substitutions potentially lead to a change of XPD protein structure, function and TFIIH complex activity, which may be associated with melanoma development. Contradictory results of rs13181 and rs1799793 have been reported. A meta-analysis that included four studies (1,397 cases and 2,873 controls) showed no association between rs1799793 and melanoma risk. A second larger meta-analysis, conducted on data derived from eight reports (2,308 cases and 3,698 healthy controls), suggested that rs13181 is associated with an increased melanoma risk. It has been suggested that the association of these SNPs with melanoma is more likely to be linked with disease development.[11, 16, 17] Li et al. suggested that XPD 312Asn/Asn or 312Asp/Asn and XPD 751Gln/Gln or 751Lys/Gln genotypes are involved in melanoma development. However, Kertat et al. showed a positive correlation of the XPD 751Gln/Gln genotype in males who consisted of 244 cases and 251 healthy controls. Smaller studies have also revealed associations; for example, Baccarelli et al. observed an increased risk of melanoma in carriers of XPD 751Gln/Gln or 312Asn/Asn for more than 50 years of age among 179 patients with melanoma and 179 healthy controls. In contrast, there are several reports indicating no correlation with MM.[8-10, 18] However, an inverse correlation of XPD codons 751 and 312 with MM has also been reported. Furthermore, there are some reports indicating a 751Lys/Lys over-representation in various haplotypes of XPD among melanoma patients.[19, 26]
Multivariate analysis of XPD genotypes in our study revealed an association with rs238406 increases melanoma risk. SNP analysis using Fisher's exact test suggested otherwise, whereas haplotype analysis appeared to be associated with decreased disease risk. Evidence reported for rs232406 suggests little if any[9, 14, 19] association between XPD rs238406 and increased melanoma risk, which has been confirmed by a meta-analysis.
In our study, we confirmed the previous reports about the role of XPA in that polymorphisms in this gene are relatively rare and were not associated with melanoma risk.
We genotyped ten SNPs within XPB and five in XPE that have not been studied previously in relation to melanoma risk. Only rs1050244 in XPE could be studied as none of the other polymorphisms was represented in the Polish population at sufficient frequency to allow any meaningful analysis. Together, both XPB and XPE do not appear to contribute significantly to disease risk in the Polish population.
We examined 16 SNPs in XPF and found only one (rs762521) that was polymorphic in the Polish population, but it was not related to melanoma.
XPG has not previously been examined with respect to its role in melanoma susceptibility. Initially, 28 SNPs were selected but only five SNPs were used. None of them showed a significantly different allele distribution between cases and healthy controls. The remaining 23 were not polymorphic in the Polish population. These results confirm the previous meta-analyses and reports[9-12, 19, 29] and suggest that the only report indicating an association may have been owing to chance.
In conclusion, summarizing the results of the haplotype analysis, the logistic regression (and the supplementary Fisher's exact tests), we conclude that CT and TT genotypes at rs2228000 (XPC Ala499Val) could be a protective factor against MM risk. Moreover, XPC G1475A and XPD rs18131_ C + rs1799793_G + rs238406_C and rs18131_C + rs1799793_ G + rs238406_A haplotypes may modify melanoma risk. The other XP genes do not appear to be associated to any significant degree with melanoma development. The lack of association of polymorphisms in NER genes and melanoma risk should not be taken to suggest that there is no relationship between melanoma and perturbations in NER. Certainly, the control of NER genes is complex and it remains to be precisely determined and what aspects of this process are implicated in melanomagenesis. Finally, as UV-light exposure varies quite considerably from one latitude to another, we cannot rule out the possibility that with more extremes of exposure, NER variance may be associated with MM in one region but not observed in another. An assessment of actual UV-light exposure is required to fully establish the relationship between polymorphisms in XP and melanoma risk. Further studies taking into account gene environmental interactions are required to confirm our results.