Centers for Disease Control and Prevention Centers for Disease Control and Prevention CDC Home Search CDC CDC Health Topics A-Z site search
National Office of Public Health Genomics
Centers for Disease Control and Prevention
Office of Genomics and Disease Prevention
Site Search
 

HuGENet Review
“The findings and conclusions in this review are those of the author(s) and do not
necessarily represent the views of the funding agency.”
This HuGE Review was published in the Am J of Epidem 2005; 162(10):925-942

line

Genetic Polymorphisms in the Base Excision Repair Pathway and Cancer Risk: A HuGE Review
Rayjean J. Hung, Janet Hall, Paul Brennan and Paolo Boffetta

From the Genetic Epidemiology Group, International Agency for Research on Cancer, Lyon, France

Correspondence: Dr. Rayjean J. Hung, Genetic Epidemiology Group, International Agency for Research on Cancer, 150 cours Albert-Thomas, 69008 Lyon, France (e-mail: hung@iarc.fr).

line

 ABSTRACT

Genetic variations in DNA repair genes are thought to modulate DNA repair capacity and are suggested to be related to cancer risk. However, epidemiologic findings have been inconsistent. The authors conducted meta-analyses of associations between genes in the base excision repair pathway and cancer risk, focusing on three key genes: 8-oxoguanine DNA glycosylase (OGG1), apurinic/apyrimidinic endonuclease (APE1/APEX1), and x-ray repair cross-complementing group 1 (XRCC1). They found increased lung cancer risk among subjects carrying the OGG1 Cys/Cys genotype (odds ratio (OR) = 1.24, 95% confidence interval (CI): 1.01, 1.53), using 3,253 cases and 3,371 controls from seven studies; this is consistent with experimental evidence that this isoform exhibits decreased activity. They found a protective effect of the XRCC1 194Trp allele for tobacco-related cancers (OR = 0.86, 95% CI: 0.77, 0.95), using 4,895 cases and 5,977 controls from 16 studies; this is compatible with evidence of lower mutagen sensitivity for this allele. The XRCC1 399Gln/399Gln genotype was associated with increased risk of tobacco-related cancers among light smokers (OR = 1.38, 95% CI: 0.99, 1.94) but decreased risk among heavy smokers (OR = 0.71, 95% CI: 0.51, 0.99), suggesting effect modification by tobacco smoking. There was no association between cancer risk and the APE1/APEX1 Asp148Glu and XRCC1 Arg280His polymorphisms. Recommendations for future studies include pooling of individual data to facilitate evaluation of multigenic effects and detailed analysis of effect modification by environmental exposure.


Abbreviations: APE1/APEX1, apurinic/apyrimidinic endonuclease; CI, confidence interval; dbSNP, Database of Single Nucleotide Polymorphisms; OGG1, 8-oxoguanine DNA glycosylase; OR, odds ratio; XRCC1, x-ray repair cross-complementing group 1.

back to top

  INTRODUCTION

There is growing evidence that genetic predisposition to cancer acts via a combination of high-risk variants in a set of low- and medium-penetrance genes rather than a few high-penetrance genes. Recent genetic association studies on cancer risk have focused on identifying effects of single nucleotide polymorphisms in candidate genes, among which DNA repair genes are increasingly studied because of their critical role in maintaining genome integrity. The base excision repair pathway is one DNA repair pathway which removes various forms of base damage via a number of coordinated sequential reactions that detect and process the damage (1). Mammalian cells contain 11 different glycosylases, each with a specialized function, as reviewed by Barnes and Lindahl (2). In the first step, a DNA glycosylase, such as 8-oxoguanine DNA glycosylase (OGG1), initiates this process by releasing the modified base, which generates an apurinic/apyrimidinic site. Some glycosylases (the bifunctional glycosylases) have an associated apurinic/apyrimidinic lyase activity and further catalyze the cleavage of the sugar-phosphate chain and excision of the abasic residue, leaving a single nucleotide gap. This gap is filled by DNA polymerase ß, and the nick is sealed by the DNA ligase III/x-ray repair cross-complementing group 1 (XRCC1) complex. Certain glycosylases (the monofunctional DNA glycosylases) have no associated lyase activity. When such enzymes initiate repair, the phosphodiester bond at the 5' side of the intact apurinic/apyrimidinic site is incised by apurinic/apyrimidinic endonuclease (APE1/APEX1). DNA polymerase ß, DNA ligase III, and XRCC1 complete the repair process. The net result is the replacement of a single nucleotide; this is called short-patch base excision repair (1-5). A subpathway of base excision repair, the long-patch repair—which appears to play a crucial role in processing oxidized or reduced apurinic/apyrimidinic sites that are resistant to the apurinic/apyrimidinic lyase activity of DNA polymerase ß—results in the replacement of several nucleotides (1–5). These pathways are illustrated in Figure 1.

Figure 1 : Diagram of the short-patch and long-patch base excision repair pathways.

Sequence variants in DNA repair genes are thought to modulate DNA repair capacity and consequently are suggested to be associated with altered cancer risk (6). However, results from epidemiologic studies have been inconsistent, possibly because of 1) low statistical power for detecting a moderate effect, 2) false-positive results, 3) heterogeneity across study populations, 4) failure to consider effect modifiers such as environmental exposures, and 5) publication bias. Reliable knowledge of which sequence variants influence cancer risk may help in identifying persons at high risk of developing cancer and shed light on cancer etiology. Therefore, we conducted a systematic Human Genome Epidemiology review on associations between genes in the base excision repair pathway and cancer risk, focusing on genes encoding three key enzymes in this repair pathway: OGG1, APE1/APEX1, and the XRCC1 protein.

back to top

 GENE AND GENE VARIANTS

8-Oxoguanine DNA glycosylase
The OGG1 gene is located on chromosome 3p26.2, a region that frequently shows loss of heterozygosity in several human cancers (7, 8). It consists of seven exons and six introns and encodes a 345 amino acid, a bifunctional glycosylase.

OGG1 repairs one of the most mutagenic lesions among base damages, 8-oxoguanine, also called 8-hydroxyguanine when present in DNA in its alternative tautomeric form. 8-Oxoguanine is able to base-pair with adenine and cause G:C -->T:A transversions in repair-deficient bacteria and yeast (7). At least 20 validated sequence variants have been described to date in Internet databases. Among those, a C-->G sequence variant leading to an amino acid change from serine to cysteine at codon 326 (Ser326Cys; Database of Single Nucleotide Polymorphisms (dbSNP) no. rs1052133) has been studied most frequently. Several in vivo or in vitro studies have examined the association between OGG1 genotypes and enzyme activity, though the results have been inconsistent, as reviewed by Weiss et al. (9). Although no association was found between OGG1 genotypes and the enzyme activity of OGG1 in two studies (10, 11), Kohno et al. (8) found, by using a complementation assay of an Escherichia coli mutant defective in the repair of 8-oxoguanine, that 326Ser-containing OGG1 has a sevenfold higher activity for repairing 8-oxoguanine than 326Cys-containing OGG1.

Apurinic/apyrimidinic endonuclease
The gene encoding APE1 (also known as APEX1, HAP1, or REF1) is located on chromosome 14q11.2–q12 and encodes a 317 amino acid protein. It processes the abasic sites left from the incision of the damaged base by cleaving the DNA backbone at the 5' side to the abasic site, leaving a 3'-hydroxyl group and a 5'-deoxyribose phosphate group flanking the nucleotide gap (5, 12, 13). APE1 also processes other 3' DNA termini that impede further gap filling or religation, allowing repair to be completed; thus, it is an essential protein playing a pivotal role in the processing of abasic sites. Attempts to create Ape1 -null mice have demonstrated an early embryonic lethality. Several sequence variants were identified in this gene, including a G-->T change in exon 5 leading to an amino acid change from aspartic acid to glutamic acid (Asp148Glu; dbSNP no. rs3136820), a C-->G change in exon 3 leading to an amino acid change from glutamic acid to histidine (Gln51His; dbSNP no. rs1048945), and an A-->G substitution resulting in an amino acid change from isoleucine to valine in exon 64 (Ile64Val; dbSNP no. rs2307486). While the impact of the Gln51His and Ile64Val variants on enzyme function has not yet been elucidated, it has been suggested that Asp148Glu may be associated with hypersensitivity to ionizing radiation (14).

X-ray repair cross-complementing group 1
The XRCC1 protein is essential for mammalian viability. XRCC1 deficiency in mice results in embryonic lethality, and XRCC1 is required for the efficient repair of single-strand breaks and damaged bases in DNA. XRCC1 has no known enzymatic activity, and it is thought to act as a scaffold protein for both single-strand break repair and base excision repair activities (4). XRCC1 has been shown to physically interact with DNA polymerase ß, polyadenosine diphosphate-ribose polymerases 1 and 2, APE1/APEX1, OGG1, and proliferating cell nuclear antigen. Its absence leads to a substantial reduction in the levels of its partner ligase III (15–17). The gene is located on chromosome 19q13.2; it consists of 17 exons and encodes a protein of 633 amino acids (4). More than 60 validated single nucleotide polymorphisms in XRCC1 are listed in the Ensembl database, among which approximately 30 variants are located in exons or promoter regions. The most extensively studied single nucleotide polymorphisms are Arg194Trp on exon 6 (dbSNP no. rs1799782), Arg280His on exon 9 (dbSNP no. rs25489), and Arg399Gln on exon 10 (dbSNP no. rs25487). The Arg194Trp variant has been shown to be associated with lower bleomycin and benzo( a )pyrene diol epoxide sensitivity in vitro (18). The functional significance of Arg280His is not yet well-established; however, the 280His is located in the proliferating cell nuclear antigen-binding region and was suggested in a small study ( n = 80) to be associated with higher bleomycin sensitivity (16, 19). The 399Gln allele is located at the carboxylic acid terminal side of the polyadenosine diphosphate-ribose polymerase-interacting domain. It was shown to be associated with higher levels of aflatoxin B 1 -DNA adducts and higher bleomycin sensitivity in several studies (18, 20, 21), while another study did not find such an association (22).

Reliable knowledge on which base excision repair sequence variants are associated with cancer risk would help to elucidate the disease mechanism. Given that most of the previous studies had inadequate statistical power, we have conducted a systematic review on sequence variants in these three key players in this repair pathway.

Prevalence of gene variants
To estimate the prevalence of XRCC1, OGG1 , and APE1/APEX1 variants and their associated cancer risk, we conducted MEDLINE searches for case-control studies published up to February 1, 2005, on the associations between these genes and cancer risk. When more than one article was identified for the same study population, we included the most recent publication. When one publication reported results from more than one population, with an indication that different populations were recruited separately, we considered them to be separate study populations (23–25). Cancers studied included lung, upper aerodigestive tract ( International Classification of Diseases for Oncology codes 140–150 and 161), colorectal, bladder, liver, pancreas, prostate, breast, skin, and hematologic malignancies.

For each publication, we extracted information on the publication year, cancer site, control source, country, numbers of cases and controls, matching type and factors, ethnic composition of the population, and genotype frequency for the reported locus. We assessed departure from Hardy-Weinberg equilibrium for the control group in each study. Characteristics of individual studies are summarized in Tables 1–3, presented by gene and type of cancer. In total, 22 studies were included for OGG1 Ser326Cys, four studies for APE1/APEX1 Asp148Glu, 29 studies for XRCC1 Arg194Trp, 11 studies for XRCC1 Arg280His, and 50 studies for XRCC1 Arg399Gln.

Table 1: Study characteristics and genotype prevalences from published studies on the relation of the 8-oxoguanine DNA glycosylase (OGG1) Ser326Cys polymorphism and the apurinic/apyrimidinic endonuclease (APE1/APEX1) Asp148Glu polymorphism to cancer risk

 

Table 2: Study characteristics and genotype prevalences from published studies on the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp and Arg280His polymorphisms to cancer risk

 

Table 3: Study characteristics and genotype prevalences from published studies on the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism to cancer risk

The OGG1 Ser326Cys polymorphism is slightly more prevalent among Asians (39.4–60.6 percent carry the heterozygous variant; 13.4–38.2 percent carry the homozygous variant) than among persons of European descent (hereafter called Caucasians) (23.0–41.0 percent heterozygotes, 1.8–8.6 percent homozygotes). The genotype distributions in the control groups were under Hardy-Weinberg equilibrium in all studies, except for one Chinese study (26). The prevalences of APE1/APEX1 Asp148Glu heterozygosity and homozygosity ranged from 49 percent to 53 percent and from 14.3 percent to 25.5 percent, respectively. The XRCC1 Arg194Trp polymorphism also appeared to be more prevalent among Asians (18.7–50.0 percent carried the heterozygous variant and 2.6–13.2 percent the homozygous variant) than among Caucasians (5.4–17.6 percent heterozygous, ≤2.3 percent homozygous). The genotype distributions in the control groups were under Hardy-Weinberg equilibrium in all studies, except for one small study conducted in Finland (27). The genotype frequency of Arg280His was 6.3–26.0 percent for the heterozygous variant and ≤1.6 percent for the homozygous variant, with no apparent ethnic differences. The genotype distributions in the control groups were under Hardy-Weinberg equilibrium in all studies, except for one bladder cancer study conducted in the United States (28). XRCC1 Arg399Gln was the most common sequence variant among the three XRCC1 polymorphisms studied, and there was no major variation by ethnicity. Genotype distributions in the control groups were not under Hardy-Weinberg equilibrium in several studies, including three US studies (24, 29, 30), one United Kingdom study (31), and two Asian studies (32, 33).

back to top

 ASSOCIATIONS AND INTERACTIONS

To estimate the association between the base excision repair genetic variants and cancer risk, we conducted a meta-analysis of identified studies. We calculated the crude odds ratio and 95 percent confidence interval for each study whenever possible. The meta-analysis was performed on crude odds ratios, since the adjusted odds ratios were not comparable because of different covariates' being included in the multivariate regression models. Using persons with the homozygous common allele as the reference group, we calculated odds ratios for persons with the heterozygous and homozygous variants separately whenever possible (information available in at least three studies). Summary odds ratios were calculated for each cancer site separately and for tobacco-related cancers combined. Tobacco-related cancers included cancers of the lung, upper aerodigestive tract, bladder, stomach, liver, and pancreas, as well as myeloid leukemia (34). We aimed to investigate the effect modifications of tobacco exposures and age of onset. When the published data permitted, odds ratios were calculated for various categories of smoking (never, ever, former, current, light, and heavy smoking; defined differently in different studies) and different ages of onset (young and old persons; defined differently in different studies). Effect modification of tobacco smoking was evaluated for tobacco-related cancers only. Interaction odds ratios for smoking and age of onset were calculated using the case-only design.

The summary odds ratios were estimated using the random-effect model, which takes into account the heterogeneity among studies by adding a term to the model (35). We estimated the summary odds ratio when there were data from at least three studies available. The number of studies may appear to be different in each stratum, depending on the amount of information provided in the published articles. A test of heterogeneity was performed for each summary estimate. We evaluated publication bias by means of funnel plots and Egger's test (36).

We conducted influence analysis when there was evidence of heterogeneity or publication bias. When the study-specific odds ratios appeared to be heterogeneous, we evaluated the source of heterogeneity by Galbraith plot and by comparing the Q values. Studies contributing the most heterogeneity were excluded until the heterogeneity was reduced to a nonsignificant level. When there was evidence of publication bias, we excluded the studies that had the largest variance (from the right-hand side of the funnel plot) until the publication bias was reduced to a nonsignificant level. When a study was excluded from the calculation of summary odds ratios by cancer site, because of either its contribution to heterogeneity or publication bias, it was subsequently excluded from all relevant stratified analyses. All statistical analyses were conducted with Stata software, version 8 (Stata Corporation, College Station, Texas).

The individual odds ratios from each study for OGG1 Ser326Cys are shown in Web table 1 (posted on the Journal 's website ( http://www.aje.oxfordjournals.org )), and the summary odds ratios are shown in Table 4. Combining data from all eight studies on lung cancer, the summary odds ratios were 1.09 (95 percent confidence interval (CI): 0.86, 1.40) for Ser/Cys carriers and 1.37 (95 percent CI: 1.02, 1.82) for Cys/Cys carriers (data not shown). However, there was evidence of heterogeneity between individual studies ( p < 0.01 for Ser/Cys). One outlying study was identified as contributing the most heterogeneity (37). After removing this study, we still observed an increased risk of lung cancer for Cys/Cys carriers, finding an odds ratio of 1.24 (95 percent CI: 1.01, 1.53) based on 3,253 cases and 3,371 controls (Table 4). Combing five studies of upper aerodigestive tract cancers, we observed an odds ratio of 1.28 (95 percent CI: 0.96, 1.69) for Cys/Cys carriers; however, the results of Egger's test suggested the presence of publication bias ( p = 0.03), with the more imprecise studies reporting a positive effect. After exclusion of the study with the largest variance (38) and an outlying study (39) (heterogeneity p < 0.01), the odds ratio for Cys/Cys was reduced to 1.15 (95 percent CI: 0.90, 1.46) (Table 4). Given the ethnic differences in the allele frequency of this sequence variant, we evaluated the effect of the 326Cys allele in Asian and Caucasian populations separately (data not shown). However, we did not observe a difference: The summary odds ratio was 1.16 (95 percent CI: 0.98, 1.40) for Asian populations and 1.15 (95 percent CI: 0.90, 1.46) for Caucasian populations. On the basis of limited data, we did not observe effect modification by smoking.

Table 4: Summary odds ratios for the relation of the 8-oxoguanine DNA glycosylase (OGG1) Ser326Cys polymorphism to cancer risk

Web table 2 shows the individual and summary odds ratios for APE1/APEX1 Asp148Glu. On the basis of 1,359 cases and 1,686 controls, we did not observe an association between the 148Glu allele and aerodigestive tract cancer. When we analyzed lung cancer separately, the summary odds ratio was 0.94 (95 percent CI: 0.77, 1.14; p for heterogeneity = 0.56; p for Egger's test = 0.08).

Web table 3 shows the odds ratios from individual studies for the XRCC1 Arg194Trp, Arg280His, and Arg399Gln polymorphisms. Table 5 shows the summary odds ratios for the XRCC1 Arg194Trp polymorphism by cancer site, tobacco smoking, and age of onset. We did not observe an association between the 194Trp allele and the risk of lung or upper aerodigestive tract cancers. However, when we combined data from 16 studies on tobacco-related cancers (4,895 cases and 5,977 controls), the 194Trp allele was associated with a decreased risk of tobacco-related cancers (odds ratio (OR) = 0.86, 95 percent CI: 0.77, 0.95) (Figure 2). There were 10 case-control studies on breast cancer and the XRCC1 Arg194Trp polymorphism. The study-specific odds ratios for the Arg/Trp and Trp/Trp genotypes appeared to be heterogeneous ( p = 0.04), mainly because of one outlying study (40). After removal of this study, the summary odds ratio for breast cancer was 0.95 (95 percent CI: 0.78, 1.16; p for heterogeneity = 0.20) (Table 5). When data were stratified by smoking status for tobacco-related cancers, there seemed to be a small protective effect among ever smokers (OR = 0.77, 95 percent CI: 0.58, 1.01). Given the ethnic difference in the allele frequencies, we also analyzed Asian and Caucasian populations separately (data not shown). However, the effect was not modified by ethnicity: The summary odds ratio for having at least one 194Trp allele was 0.90 (95 percent CI: 0.78, 1.03) from nine studies conducted in Asian populations and 0.89 (95 percent CI: 0.80, 1.00) from 15 studies conducted in Caucasian populations.

Table 5: Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp polymorphism to cancer risk

 

Figure 2: Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp polymorphism (Arg/Trp and Trp/Trp vs. Arg/Arg) and risk of tobacco-related cancer among all subjects.

Summary odds ratios for XRCC1 Arg280His are shown in Table 6. We did not observe an association between the 280His allele and either lung cancer, upper aerodigestive tract cancer, or tobacco-related cancers combined.

Table 6: Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg280His polymorphism (Arg/His or His/His) to cancer risk

Summary odds ratios for XRCC1 Arg399Gln, stratified by cancer site, tobacco smoking, and age of onset, are shown in Table 7. On the basis of 6,120 lung cancer cases and 6,895 controls from 13 studies, we did not observe an association between the 399Gln allele and lung cancer risk. We identified eight studies on upper aerodigestive tract cancers. The study-specific odds ratios for the Gln/Gln genotype from these eight studies appeared to be heterogeneous ( p = 0.01), mainly because of one outlying study (41). After excluding this outlying study, we observed a nonsignificant decreased risk of upper aerodigestive tract cancer among carriers of the 399Gln allele (OR = 0.88, 95 percent CI: 0.78, 1.01) (Table 7). We identified a total of five studies on bladder cancer and the Arg399Gln polymorphism. There was evidence of publication bias in the summary odds ratio for bladder cancer for carriers of the Arg/Gln genotype (Egger's test: p = 0.04). After exclusion of the study with the largest variance (42), the p value from Egger's test was increased to 0.10, and the summary odds ratio was 1.12 (95 percent CI: 0.95, 1.33) for carriers of the Arg/Gln or Gln/Gln genotype. Combining data from 11 studies on breast cancer, we did not observe an association between the presence of the 399Gln allele and risk of breast cancer; neither did we find an association between the 399Gln allele and risk of skin cancer from three studies on skin cancer.

Table 7: Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism to cancer risk

When data were stratified by tobacco smoking for tobacco-related cancers, the presence of the 399Gln allele seemed to be associated with an increased risk of tobacco-related cancers among light smokers (OR = 1.20, 95 percent CI: 1.03, 1.40) (Figure 3), whereas it was associated with a decreased risk among heavy smokers (OR = 0.81, 95 percent CI: 0.64, 1.04) (Figure 4 ), with the effect being more prominent among carriers of the Gln/Gln genotype (OR = 0.71, 95 percent CI: 0.51, 0.99) (Table 7). The interaction odds ratio from case-case analysis was 0.73 (95 percent CI: 0.51, 1.04) for the Gln/Gln genotype, which suggests an interaction less than multiplicativity between the Gln/Gln genotype and heavy tobacco smoking. We did not observe effect modification by age of onset.

Figure 3: Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism (Arg/Gln and Gln/Gln vs. Arg/Arg) and risk of tobacco-related cancer among light smokers.

 

Figure 4: Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism (Arg/Gln and Gln/Gln vs. Arg/Arg) and risk of tobacco-related cancer among heavy smokers.

back to top

 DISCUSSION

We conducted a systematic literature review to evaluate the associations between sequence variants in three base excision repair genes and cancer risks. We also evaluated possible effect modifications by tobacco smoking and age of onset. In summary, we found an increased risk of lung cancer among subjects carrying the OGG1 Cys/Cys genotype, which is consistent with experimental evidence that this isoform exhibits decreased base excision repair activity (8, 43). Epidemiologic studies on the OGG1 Ser326Cys polymorphism were reviewed previously; however, no meta-analysis was conducted (9). We also found a protective effect of the XRCC1 194Trp allele for tobacco-related cancers, which is compatible with the evidence of lower mutagen sensitivity for this allele (18). However, the effect of XRCC1 194Trp alleles was mainly observed in heterozygotes but not in homozygotes; therefore, the results should be treated with caution. We did not find an association between cancer risk and APE1/APEX1 Asp148Glu or XRCC1 Arg280His.

We observed modification of the effect of the XRCC1 399Gln/399Gln genotype by tobacco smoking. Studies showed that the 399Gln allele may be associated with higher mutagen sensitivity and higher levels of DNA adducts (18, 21). An increased risk of tobacco-related cancers among light smokers who carried the Gln/Gln genotype is consistent with the functional data. Conversely, the mechanistic basis of a decreased risk among heavy smokers who carried the Gln/Gln genotype remains to be elucidated. Compatible with our findings, Matullo et al. (21) reported that Gln/Gln carriers had higher DNA adduct levels than Arg/Arg carriers among never smokers but lower DNA adduct levels among current smokers, though the differences were not statistically significant. It is possible that the resulting increased levels of DNA damage from heavy tobacco smoking might give rise to enhanced apoptosis at the time of cell division and would be manifested as a reduced risk of exposure-induced cancer. Such a model has been proposed to explain reduced risks of sunburn-related nonmelanoma skin cancer in homozygotic carriers of the XRCC1 codon 399 glutamine variant (44). Alternatively, tobacco smoking might induce DNA repair capacity in response to DNA damage. In support of this hypothesis, Wang et al. (18) noted a trend towards lower levels of chromosome breaks in healthy heavy smokers (>42 pack-years) as compared with never smokers following treatment with bleomycin or benzo( a )pyrene diol epoxide, though the association with genotype has not been examined. Lower levels of 8-oxoguanine in lymphocytes of smokers as compared with nonsmokers, which could be explained by the presence of efficient repair processes for the oxidative damage induced by smoking, have also been reported in some studies (45, 46).

There are some limitations inherent in meta-analysis. Each study had different eligibility criteria for subjects and different sources of controls; this should be taken into account when interpreting the summary estimates. For example, the effect of OGG1 Cys/Cys on lung cancer risk was predominantly detected in population-based studies: When studies were stratified by control source, the lung cancer odds ratio for the OGG1 Cys/Cys genotype was 1.78 (95 percent CI: 1.22, 2.59) using data from population-based studies and 1.09 (95 percent CI: 0.86, 1.31) using data from hospital-based studies. This indicates that the allele distribution of the OGG1 polymorphism in the hospital control groups might not have been representative of the general population. On the other hand, the effects of the XRCC1 Arg194Trp polymorphism on tobacco-related cancers from population-based studies (OR = 0.88, 95 percent CI: 0.76, 1.03) and hospital-based studies (OR = 0.82, 95 percent CI: 0.71, 0.94) were comparable. The differences among studies were estimated by tests of heterogeneity and influence analysis, and only the summary estimates without significant heterogeneity at the 0.05 level were considered valid, although the standard test of heterogeneity generally has low statistical power.

The effect of a single common sequence variant might not be detectable in population association studies, and the combination of multiple sequence variants in the same gene and in genes functioning in the same biochemical pathway might be more important in carcinogenesis. A segregation analysis of breast cancer in United Kingdom families suggested that the residual family clustering in noncarriers of BRCA 1 / 2 mutations may be explained by a large number of low-penetrance sequence variants (47, 48). Likewise, another recent segregation analysis suggested a multigenic model for lung cancer susceptibility (49). However, it is difficult to assess the effect of the combination of multiple sequence variants via meta-analysis. A large-scale investigation of the multigenic model of cancer through pooling of individual data would be feasible.

Similarly, although we attempted to evaluate effect modification by age of onset and tobacco smoking, only a few investigators reported such results. In addition, interpretation of the results is limited, since the definition of each stratum varied among studies. Moreover, the cancer sites shown to be related to tobacco smoking were combined (34) to obtain sufficient statistical power for further stratified analysis on smoking. However, the magnitudes of the associations with smoking vary across these cancer sites, and the modification of effects on sequence variants might also differ. A more appropriate investigation requires the pooling of individual data; such a coordinated effort can be achieved via collaborative arrangements such as consortia.

Consortia that may be able to undertake pooled analysis have been created for lung cancer (e.g., the International Lung Cancer Consortium (ILCCO)) and head and neck cancer (e.g., the International Head and Neck Cancer Epidemiology Consortium (INHANCE)). Their aims are to increase statistical power to detect low-risk genetic variants and gene-environment interactions and to focus on special subgroups such as early-onset cases, nonsmokers, or patients with tumors of rare histology. Consortia also provide an opportunity for researchers in the field to rapidly test the repeatability of results, as well as discuss markers that may be most relevant for a specific cancer site. With regard to genetic association studies, consortia and international collaborations may be a way to maximize study efficiency and overcome the limitations of individual studies.

back to top

 ACKNOWLEDGEMENTS

This study was funded by a National Cancer Institute R01 grant (contract CA 092039-01A2), a National Cancer Institute R03 grant (contract CA 119704-01), and an Association for International Cancer Research grant (contract 03-281). Dr. Rayjean Hung worked on this study under the tenure of Special Training Awards from the International Agency for Research on Cancer.

Conflict of interest: none declared.

back to top

 TABLES AND FIGURES

Tables

  • Table 1 - Study characteristics and genotype prevalences from published studies on the relation of the 8-oxoguanine DNA glycosylase (OGG1) Ser326Cys polymorphism and the apurinic/apyrimidinic endonuclease (APE1/APEX1) Asp148Glu polymorphism to cancer risk
  • Table 2 - Study characteristics and genotype prevalences from published studies on the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp and Arg280His polymorphisms to cancer risk
  • Table 3 - Study characteristics and genotype prevalences from published studies on the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism to cancer risk
  • Table 4 - Summary odds ratios for the relation of the 8-oxoguanine DNA glycosylase (OGG1) Ser326Cys polymorphism to cancer risk
  • Table 5 - Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp polymorphism to cancer risk
  • Table 6 - Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg280His polymorphism (Arg/His or His/His) to cancer risk
  • Table 7 - Summary odds ratios for the relation of the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism to cancer risk

Figures

  • Figure 1 - Diagram of the short-patch and long-patch base excision repair pathways
  • Figure 2 - Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg194Trp polymorphism (Arg/Trp and Trp/Trp vs. Arg/Arg) and risk of tobacco-related cancer among all subjects
  • Figure 3 - Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism (Arg/Gln and Gln/Gln vs. Arg/Arg) and risk of tobacco-related cancer among light smokers
  • Figure 4 - Odds ratios for the relation between the x-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism (Arg/Gln and Gln/Gln vs. Arg/Arg) and risk of tobacco-related cancer among heavy smokers

back to top

 REFERENCES

List of References

back to top

 INTERNET SITES

back to top

This reference links to a non-governmental website
 Provides link to non-governmental sites and does not necessarily represent the views of the Centers  for Disease Control and Prevention.
Page last reviewed: November 15, 2005 (archived document)
Page last updated: November 2, 2007
Content Source: National Office of Public Health Genomics