Abstract
Exposure to UV radiation is a major risk factor for the development of malignant melanoma. DNA damage caused by UV radiation is thought to play a major role in carcinogenesis induction. Multiprotein pathways involved in repairing UV-DNA damage are the base excision, the nucleotide excision, and the homologous double-stranded DNA repair pathways. This study used a sequence-specific primer PCR (PCR-SSP) genotyping method to investigate the association between polymorphisms in DNA repair genes from these pathways with the development of malignant melanoma. The patient cohort was comprised of 125 individuals with malignant melanoma with lesions or staging suggesting a high risk of relapse or metastatic disease. The control population consisted of 211 individuals. We found the presence of a T allele in exon 7 (position 18067) of the XRCC3 gene was significantly associated with melanoma development (P = 0.004; odds ratio, 2.36; relative risk, 1.74). This gene codes for a protein involved in the homologous pathway of double-stranded DNA repair, thought to repair chromosomal fragmentation, translocations, and deletions. These results may provide further insights into the pathogenesis and the mechanism of UV-radiation induced carcinogenesis as well as having a role in prevention.
Introduction
Malignant melanoma is a neoplastic lesion arising from epidermal melanocytes. It is a highly invasive and aggressive cancer with a high mortality rate. Its incidence has increased rapidly over the past 30 years. The highest risk is associated with individuals who are fair-skinned, and who are exposed to intense, intermittent periods of sunlight (1) . Extensive epidemiological and experimental data suggests that UV radiation is an important environmental carcinogen involved in the initiation and progression of skin cancer (2) . UVBradiations at short wavelengths, 290–320 nm, induces damage in the form of cyclobutane pyrimidine dimers and pyrimidine 6–4(6–4) photoproducts. At ultraviolet A wavelengths (320–400 nm), it causes single-stranded breaks, DNA-protein crosslinking, and generates free-radicals that cause oxidative damage (3) . Cells respond to this damage through the activation of various DNA repair pathways.
DNA repair systems are responsible for maintaining the integrity of the genome and have a critical role in protecting against mutations that can lead to cancer (4 , 5) . Absent or incorrect repair can initiate carcinogenesis through the activation of oncogenes, the inactivation of tumor-suppressor genes, or the loss of heterozygosity. Repair of damaged DNA involves many proteins performing functions directly upon damaged DNA as well as the interaction and interplay with proteins involved in the regulation of DNA replication and progression through the cell cycle (6) .
Studies have shown that genes directly involved in DNA repair and the maintenance of genome integrity, or genes indirectly involved in the repair of DNA damage through the regulation of the cell cycle, are critical for protecting against the mutations that lead to cancer (4 , 5) .
Inter-individual variation in DNA repair capacity has been shown through the use of lymphocyte assays. These assays usually measure chromosome damage rather than specific biochemical pathways and are difficult to reproduce. But findings have shown that individuals with a repair capacity of 65–80% of the population mean are more often in the cancer cohorts (7, 8, 9, 10, 11, 12, 13, 14) . Reduced DNA repair capacity constitutes a statistically significant risk factor for development of breast and lung cancer with ORs 3 ranging from 1.6 to 10.0 (8, 9, 10, 11, 12 ,, 14) .
Evidence suggests that the difference in DNA repair capacity among individuals is genetically determined. The phenotype of reduced repair capacity for one pathway is independent of the phenotype for another pathway (15) ; this is consistent with DNA repair being genetically regulated. Measurement of repair capacity in twins (16) and the elevated frequency of individuals with reduced repair capacity among relatives of cancer patients is further evidence that repair capacity is a genetic trait (8 , 10 , 11 , 17) .
This variation in DNA repair capacity has characteristics expected of cancer susceptibility genes; the proteins encoded by these alleles exhibit reduced function rather than absence of function, which causes disease. They exist at polymorphic frequency in the general population, and they exhibit incomplete penetrance (5 , 18) . Thus malignant melanoma in part may be caused by intermittent intense UV exposure of skin genetically ill-adapted to deal with it.
One mechanism that may lead to this inter-individual variation in DNA repair capacity is genomic variation within the DNA repair genes. Recently a number of polymorphisms of genes that encode for DNA repair proteins have been described (18) . These genes, XRCC1 of the base excision pathway; ERCC1, XPD, and XPF of the nucleotide excision pathway; and XRCC3 of the homologous double-stranded repair pathway, encode for enzymes involved in separate DNA repair pathways. Previous studies have not substantiated the specific biochemical pathways involved in DNA repair of UV-induced damage, but these repair pathways are thought to play a role. Many of the variants in these genes result in amino acid substitutions and exist at polymorphic allele frequencies (i.e., allele frequencies >0.05). Given the known relationship of DNA repair to cancer, the polymorphic variants have the potential to be population cancer risk factors because of the large number of individuals affected. In view of the complexity and number of enzymes involved in DNA repair, it is likely however, that the genetic component of this complex phenotype is the aggregate of many “minor” gene effects.
An individual’s total genetic risk for developing UV-induced malignant melanoma may thus result from their combination of gene polymorphisms. For this reason we have developed a unified PCR-SSP method to simultaneously amplify a large number of DNA repair gene polymorphisms under identical conditions and have used this to test the hypothesis that interactive genetic factors modulate the risk of developing malignant melanoma.
Materials and Methods
Selection of Candidate Genes.
Ten polymorphisms in the DNA repair genes XRCC1, XPD, ERCC1, XPF, and XRCC3 were studied ( 18 ; Table 1 ⇓ ). Polymorphisms within the 5′UTR and those resulting in amino acid changes were studied. Two other polymorphisms were included, as they may be in linkage disequilibrium with uncharacterized functional polymorphisms.
Genes selected for analysis, with positions of single-nucleotide substitutions according to GenBank accession no. (in parentheses), with resulting amino acid change
Patient Cohorts.
The patient cohort was compromised of 125 individuals with histologically confirmed malignant melanoma who were referred to the regional cancer center because they had lesions or staging suggesting a high risk of relapse or metastatic disease (Table 2) ⇓ .
Clinical data for malignant melanoma cohort
The control population consisted of 211 cadaveric renal transplant donors. The representative nature of this cohort of the United Kingdom population has been previously demonstrated in HLA typing reports (19) . All patients were Caucasian.
Genotyping Assay.
The PCR technique (PCR-SSP) was used to genotype the polymorphisms of interest under universal conditions. This methodology has been applied previously to other polymorphic genes including those of the HLA-complex (20) . Genomic DNA extraction (21) , PCR amplification, and gel electrophoresis (20) were conducted as described previously. DNA amplification was carried out using primers specific for the selected DNA repair polymorphisms (Table 3) ⇓ .
PCR-SSP primer sequences, specificities, and concentrations
Statistical Analysis.
Phenotype frequencies were obtained by counting the number of individuals in a population positive for an allele. Allele frequencies were obtained by directly counting the number of chromosomes bearing an allele. Alleles at each polymorphism were said to be in Hardy-Weinberg equilibrium if the observed homozygote and heterozygote frequencies did not differ significantly (P > 0.05) from those expected when analyzed by χ2 test.
Associations were assessed using contingency table analysis and theχ 2 test (with Yates’ correction) using KnowledgeSEEKER (Angoss Software Corporation, Toronto, Canada). The OR and the relative risk were calculated, and a Bonferroni correction of 10 (the number of polymorphisms investigated) was used to correct for multiple comparisons.
Results
Phenotype, allele, and genotype frequencies were calculated for the 10 DNA repair gene polymorphisms in the melanoma and control cohorts (Table 4) ⇓ .
Phenotype, allele, and genotype frequencies of DNA repair polymorphisms in controls and malignant melanoma cohorts
Analysis of the DNA repair phenotype revealed an association with the exon 7 T/C polymorphism at position 18067 of the XRCC3 gene and the development of melanoma. Those individuals with the T allele had a significantly increased risk of developing melanoma compared to those without this allele [OR 2.36 (1.44–3.86); P = 0.0004 (adjusted by Bonferroni’s correction factor to P = 0.004)].
In addition, an association was found with the development of melanoma and the TT genotype in position 2063 of the 5′UTR of the XPF gene [OR 1.65 (1.03–2.66); P = 0.038 (adjusted to P = 0.38)] and the presence of the exon 11 TT genotype in position 30028 of the XPF gene [OR 0.62 (1.01–2.60); P = 0.045 (adjusted to P = 0.45)]. The associations were reduced in significance once the heterozygote genotype frequencies were included and, as shown, did not remain significant after Bonferroni correction. There did, however, seem to be a strong additive effect of the XRCC3 exon 7 T allele and the XPF exon 11 allele or XPF 5′UTR T alleles on the development of melanoma (Fig. 1) ⇓ . No associations were found with any of the other DNA repair gene polymorphisms studied.
The additive effect of the XRCC3 and XPF alleles for the total study group. Vertical bars, number of individuals carrying 0–4 risk alleles; black bars, cases with malignant melanoma; white bars, controls. P gives the statistical significance level of the risk of developing malignant melanoma between those individuals carrying zero to two alleles and those carrying three to four “risk” alleles.
Discussion
This study assessed whether polymorphisms in five DNA repair genes involved in the base excision, nucleotide excision, and homologous double-stranded DNA repair pathways are related to the development of malignant melanoma. We found an association of development of malignant melanoma with the XRCC3 gene.
The protein XRCC3 acts in the homologous pathway of double-stranded DNA repair. This pathway is of great importance in preventing chromosomal fragmentation, translocations, and deletions, which can lead to carcinogenesis (22) . XRCC3 is structurally related to Rad51, a critical component of recombination repair. The recombination repair pathway restores DNA through recombination between the damaged strand and a homologous sequence present on the second copy of the gene contained in diploid cells (22) . The recruitment of an undamaged copy of DNA requires strand exchange activity mediated by Rad51, which polymerizes onto DNA to form a nucleo-filament that searches for homologous DNA. XRCC3 is required for the assembly and stabilization of Rad51 (23) .
Cells which are XRCC3-deficient fail to form Rad51 foci after radiation damage and exhibit genetic instability and increased sensitivity to DNA-damaging agents such as UV light. There is an established precedent for associations between mutations in Rad51-associated proteins and carcinogenesis. The product of BRCA2 and BRCA1 has been shown to bind to Rad51 (24 , 25) , and mutations in these genes are associated with the development of breast and ovarian cancer (26 , 27) .
These findings suggest mutations in this pathway provide a unifying mechanism for the genetic basis of not only malignant melanoma, but of other types of cancer as well.
To date there are no studies relating to how the XRCC3 polymorphism in position 18067 of exon 7 affects the overall function of the protein. The polymorphism results in an amino acid substitution, which converts threonine to methionine. This residue may be important in maintaining normal protein structure and integrity. Conversion from a hydroxyl amino acid to one with a sulfhydryl group could represent a substantial change in protein characteristics.
XPF, XPD, and ERCC1 are members of a complex of more than 20 proteins within the nucleotide excision repair pathway. This pathway removes photoproducts from UV-light damaged DNA such as bulky adducts and thymidine dimers by excising a 24–32-nucleotide single-strand oligomer containing the damaged lesion (28 , 29) . XPD opens up the DNA structure and XPF forms a complex with ERCC1 that incises DNA at the 5′ side of a bulky adduct lesion (30) . Noncoding changes could influence mRNA stability or structure and effect levels of protein expression. Mutations in these genes have been shown to be associated with xeroderma pigmentosum (31) , a disease characterized by sensitivity to UV-light that results in an increased incidence of skin cancer of a thousand times greater than that of the general population (32) .
It is possible to hypothesize the role of XRCC3 in the development of malignant melanoma by considering the epidemiology of melanoma. Development of malignant melanoma is associated with intense intermittent exposure to sunlight, especially exposures that induce sunburn. Data predicts that a high-dose first exposure to the sun after a prolonged period of sun avoidance will cause substantial damage to melanocytes, and that cancer progression is strongly influenced by the amount of dose/exposure (33) .
This indicates that DNA damage involved in the initiation of carcinogenesis is extensive and may therefore involve pathways such as the double-stranded DNA repair pathways. In addition, apoptopic loss following sun-exposure has never been reported in melanocytes as opposed to keratinocytes, which undergo extensive apoptosis. This indicates that melanocytes are able to tolerate a greater amount of UV-induced DNA damage (33) . Our data indicates that double-strand DNA repair is likely to play an important role in preventing melanoma. A reduction in any of the involved repair pathways may increase cancer frequency.
This study used a candidate-gene approach to analyze germline genetic factors involved in disease susceptibility. A major problem of this approach is that an association will be found by chance. We addressed this error by only considering an association significant if it maintained a P < 0.05 after correction for multiple comparisons. The association remained strong after we made the Bonferroni correction.
Although the candidate genes analyzed in this study clearly have significance in the pathogenesis of malignant melanoma, further work will be required to deduce the functional effects of these polymorphisms. Analysis of these pathways in lymphoid cell lines from the different genotypes is planned.
This study has identified a genetic factor that may determine an individual’s susceptibility to malignant melanoma. These results may provide further insights into the pathogenesis of malignant melanoma and the mechanism of UV-radiation-induced carcinogenesis. The ability to reduce the impact of a high-risk genotype/environment interaction will depend on identification of “at-risk” individuals so that a subsequent reduction in environmental factors can be advised, i.e., reduction in exposure to sunlight and the application of sunscreen. These individuals may be primary candidates for experimental therapies such as the topical application of DNA repair enzymes (34) and, possibly, gene therapy.
Acknowledgments
We thank the members of the Tissue Typing team, M. Barnardo, J. Proctor, J. Cook, and C. Logan at the Transplant Centre, for their help and support.
Footnotes
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
↵1 Financial support was provided by the Dunhill Medical Trust, Royal College of Surgeons, the Medical Research Council, and the Imperial Cancer Research Fund.
-
↵2 To whom requests for reprints should be addressed, at Medical Oncology Unit, ICRF, Churchill Hospital, Oxford OX3 7LJ, United Kingdom.
-
↵3 The abbreviations used are: OR, odds ratio; PCR-SSP, PCR-sequence-specific primers; UTR, untranslated region.
- Received March 17, 2000.
- Accepted September 1, 2000.
- ©2000 American Association for Cancer Research.
References
Volume 60, Issue 20
