Introduction

Defective DNA repair may lead to increased risk of skin cancer since patients with the rare recessive disorder xeroderma pigmentosum (XP) have a >1000-fold increase in sunlight-induced skin cancers including cutaneous melanoma.1 The defects in XP patients fall into seven nucleotide excision repair (NER) complementation groups: XPA to XPG and a separate group XP variant (XPV) with a defect in DNA polymerase η. XP complementation group C (XPC) is the most prevalent form among North Americans and Europeans. The XPG (ERCC5) gene product has been reported to be involved not only in NER but also in the repair of oxidative DNA damage.1 We hypothesize that variations in DNA repair due to inherited polymorphisms of DNA repair genes might be associated with cancer susceptibility in the general population. Wei et al2 recently demonstrated that reduced DNA repair capacity is an independent risk factor for cutaneous melanoma. Variants of the XPD gene were demonstrated to influence the risk for cutaneous melanoma.3 Vodicka et al4 found variations in different lymphocytic DNA repair functions in association with the XPG C3507G polymorphism.

Materials and methods

Study subjects

The patients were recruited from the melanoma clinics at the Departments of Dermatology in Munich and Goettingen. Healthy unrelated control individuals were recruited from the blood donor registries and from local health care personnel. The following parameters were registered for the patients and the controls: sex, age, number of nevi (on both forearms, diameter >2 mm), hair color, eye color, skin type, family history of melanoma, primary tumor thickness, and multiple primary melanomas. The studies were approved by the local ethics committees and informed consent was obtained from all patients and controls.

Genotyping

The XPC codon 492 and 499 genotypes were determined by cycle sequencing (ABI Prism 310, Darmstadt, Germany) using S80neu 5′-AGC CTC TGA TCC CTC TGA TGA T-3′ and S81neu 5′-TCT TTT TTC TGC CTT CTC ACC ATC G-3′. The PCR programme was initiated 3 min at 95°C followed by 35 cycles at 95°C for 30 s, 63°C for 30 s, 72°C for 45 s and a final elongation step at 72°C for 5 min. The XPC codon 687 genotypes were determined employing a PCR-restriction-fragment-length-polymorphism (RFLP) assay using S82 s 5′-GAC TGA GTT ACC TTT GTG TCC-3′ and S83as 5′-TAC TTC TCC AAG CCT CAC C-3′. The PCR programme was initiated 3 min at 95°C followed by 35 cycles at 95°C for 45 s, 61°C for 45 s, 72°C for 60 s and a final elongation step at 72°C for 5 min. BmeI cuts in the presence of ‘G’ and divides the 110 bp PCR product into two 65 and 45 bp fragments. The XPG codon 1104 genotypes were determined likewise using 3330–3349 FWD 5′-TTT GAG CTA CTT GAT AAG GC-3′ and 3624–3607 REV 5′-CAC CTC CAT TCT TCA CGG-3′. The PCR programme was initiated 3 min at 95°C followed by 35 cycles at 95°C for 30 s, 55°C for 45 s, 72°C for 60 s and a final elongation step at 72°C for 5 min. MboII cuts in the presence of ‘G’ and divides a 136 bp fragment into two 107 and 29 bp fragments. For control purposes, results were evaluated by SB and PL independently and PCR products were initially sequenced. We ensured that in every RFLP experiment at least one sample showed a restriction enzyme digestion, serving as a positive control.

Statistical analysis

Observed and expected genotype frequencies were compared by a Monte-Carlo goodness-of-fit test. Odds ratios and exact 95% confidence intervals were calculated. Genotype frequencies were investigated using the Cochrane–Armitage trend test.

Results

This analysis included 294 unrelated Caucasian patients from Germany with histopathologically confirmed cutaneous melanoma and 375 healthy cancer-free control individuals from the same area. Applying logistic regression analysis frequencies of the phenotypic melanoma risk factors number of nevi and skin type as well as age were significantly different in the melanoma group compared to the control group confirming that these variables are independent risk factors for the development of cutaneous melanoma and are involved in the etiology of melanoma in this study population (all P<0.0001, respectively).

The four polymorphisms were not in linkage disequilibrium. No significant deviation of the observed genotype distributions from the expected genotype distributions as predicted by the Hardy–Weinberg theory were detected. We found no association of the homozygous genotypes with increased risks of melanoma (Table 1). The same results were obtained while adjusting for the possible risk factors age, gender, skin type and eye and hair color (data not shown). Exploratory analyses of subgroups of melanoma patients compared to all controls indicated no association of these genotypes with increased risks for the development of multiple primary melanomas (n=28), a negative family history for melanoma (n=277), melanomas in individuals with less than 50 nevi (n=273), melanomas in individuals older than 55 years (n=142), and melanomas thicker than 1 mm (n=126).

Table 1 Analysis of XPC and XPG genotype frequencies, ORs, and 95% CIs for melanoma and allele frequencies
Full size table

Discussion

To our knowledge, this is the first case–control study of XPC and XPG polymorphisms in relation to melanoma. The allele frequencies for the XPC G1580A and G2166A polymorphisms were previously unknown. Two of the four polymorphisms may not alter DNA repair: the G2166A base change in XPC exon 10 does not lead to an amino-acid change and Gozukara et al5 reported that both the XPC 1601T and 1601C alleles had similar functional activities using a host cell reactivation-based complementation assay. However, they may be linked to other functionally relevant genetic variations. For example, both alleles of an XPC exon 15 polymorphism (A2920C; Lys939Gln) showed similar DNA repair activities in the complementation assay but further studies revealed that it is associated with the risk for developing bladder cancer.6 Another Poly (AT) insertion polymorphism within XPC intron 9 is normally spliced out during XPC mRNA processing but was found to be associated with the development of squamous cell cancer of the head and neck.7 Subsequently, it was found that both polymorphisms are in linkage disequilibrium to yet a third XPC splice site polymorphism that exhibits functional relevance.8 The XPC exon 8 (G1580A, Arg492His) polymorphism has not been evaluated for its effect on DNA repair. The allele frequencies for the XPC T1601C and XPG C3507G polymorphisms found in the present European population match our previously reported frequencies found in 84 and 91 anonymous donors from the US.5,9 In two distinct Scandinavian populations from Sweden and Finland, similar allele frequencies for the XPG C3507G polymorphism and an association with the occurrence of bladder and breast cancer were reported.6,10 Interestingly, in a Korean study of 310 lung cancer patients and 311 healthy controls the XPG 3507G allele was found at a frequency between 0.55 and 0.5 rather than 0.8 and 0.79 and was associated with a decreased risk of lung cancer.11 This may indicate that the XPG codon 1104 polymorphism varies between ethnic groups. Further studies are necessary to clarify the role of DNA repair gene polymorphisms in melanoma development.