The first 2 authors contributed equally to this paper.
−93GA polymorphism of hMLH1 and risk of primary lung cancer
Article first published online: 9 JUL 2004
Copyright © 2004 Wiley-Liss, Inc.
International Journal of Cancer
Volume 112, Issue 4, pages 678–682, 20 November 2004
How to Cite
Park, S. H., Lee, G. Y., Jeon, H.-S., Lee, S. J., Kim, K. M., Jang, S. S., Kim, C. H., Lee, W. K., Kam, S., Park, R. W., Kim, I.-S., Jung, T. H. and Park, J. Y. (2004), −93GA polymorphism of hMLH1 and risk of primary lung cancer. Int. J. Cancer, 112: 678–682. doi: 10.1002/ijc.20359
- Issue published online: 13 SEP 2004
- Article first published online: 9 JUL 2004
- Manuscript Accepted: 22 MAR 2004
- Manuscript Received: 31 OCT 2003
- National Cancer Control R&D Program 2003, Ministry of Health and Welfare, South Korea
- lung cancer;
- genetic susceptibility
Polymorphisms in DNA repair genes may be associated with differences in the repair capacity of DNA damage and may thereby influence an individual's susceptibility to smoking-related cancer. We investigated the association between the −93GA polymorphism in the hMLH1 gene and the risk of lung cancer in a Korean population. The hMLH1 −93GA polymorphism was typed in 372 lung cancer patients and 371 healthy controls that were frequency-matched for age and sex. There was no significant association between the hMLH1 −93GA genotype and the risk for adenocarcinoma or small cell carcinoma. However, the AA genotype was associated with a significantly increased risk for squamous cell carcinoma compared with both the GG genotype (adjusted OR = 2.02; 95% CI = 1.15–3.55; p = 0.014) and the combined GG and GA genotype (adjusted OR = 1.83; 95% CI = 1.24–2.71; p = 0.003). When the subjects were stratified by smoking exposure, the AA genotype was associated with a significantly increased risk for squamous cell carcinoma in lighter smokers (≤ 39 pack-years; adjusted OR = 1.95; 95% CI = 1.03–3.66; p = 0.039) compared with the combined GG and GA genotype, whereas there was no significant association in heavier smokers (> 39 pack-years; adjusted OR = 1.47; 95% CI = 0.82–2.61). These results suggest that the hMLH1 −93GA polymorphism could be used as a marker of genetic susceptibility to squamous cell carcinoma of the lung. © 2004 Wiley-Liss, Inc.
Although cigarette smoking is the major cause of lung cancer, only a small fraction of smokers develop the disease, suggesting that genetic factors are of importance in determining individual susceptibility to tobacco carcinogens. This genetic susceptibility may result from inherited polymorphisms in genes involved in carcinogen metabolism and repair of DNA damage.1, 2, 3
DNA repair systems are fundamental to the maintenance of genomic integrity in the face of replication errors, environmental carcinogens and the cumulative effects of age, and their inactivation can dramatically increase the susceptibility to cancer.4, 5 In humans, more than 70 genes are involved in the 5 major DNA repair pathways: nucleotide excision repair, base excision repair, mismatch repair (MMR), homologous recombinational repair and nonhomologous end joining.5, 6
Molecular epidemiologic studies have shown considerable interindividual variation in DNA repair capacity (DRC) in the general population. Individuals with suboptimal DRC are at increased risk of smoking-related cancers, such as lung cancer and squamous cell carcinoma (SCC) of the head and neck.7, 8 Polymorphisms in the DNA repair genes may contribute to variation in DRC in the general population. Therefore, it has been hypothesized that inherited polymorphisms in the DNA repair genes may modulate the susceptibility to lung cancer. To test this hypothesis, we previously studied the contribution of polymorphisms in the DNA repair genes to the risk of lung cancer in a Korean population.9, 10, 11, 12
A highly conserved set of MMR proteins is primarily responsible for the correction of replication errors (base-base or insertion-deletion mismatches) caused by DNA polymerase errors.13, 14 Genetic and epigenetic inactivation of MMR genes is implicated in the etiology of hereditary nonpolyposis colorectal cancer syndrome and a wide variety of sporadic tumors, such as colorectal, ovarian and endometrial cancers.15, 16 However, the pathogenic role of MMR genes in environment-induced cancers such as lung cancer has not been well defined.17, 18, 19, 20, 21 In addition to correcting DNA replication errors, the MMR proteins are also involved in a variety of other vital cellular processes, including cell cycle checkpoint activation, induction of apoptosis in response to DNA damage by methylating agents or oxidative stress22, 23, 24 and transcription-coupled nucleotide excision repair of bulky DNA adducts.25, 26 Therefore, a subtle defect in the DRC caused by functional polymorphisms in MMR genes that are neither necessary nor sufficient for the development of lung cancer could place individuals at increased risk of lung cancer.
hMLH1 plays a central role in MMR through its interaction with hMSH2 and hMSH6.13, 14 Several polymorphisms have been identified in the hMLH1 gene.27, 28, 29, 30 While the functional effects of these polymorphisms are not known, it is possible that some of these variants could have an effect on DRC, thereby modulating the susceptibility to lung cancer. We evaluated the association of the −93GA polymorphism with lung cancer, since this polymorphism is located in the putative consensus sequence for the binding of transcription factor AP-4 (nCnnCAGCTG from −102 to −9330), possibly influencing the activity of the hMLH1 promoter.
MATERIAL AND METHODS
This case-control study included 372 lung cancer patients and 371 healthy controls. Detailed descriptions of the methods used for subject enrollment and study population have been published previously.11 Eligible cases included all patients newly diagnosed with primary lung cancer between January 1998 and March 1999 in Kyungpook National University Hospital (Daegu, South Korea). There were no age, sex, histologic, or stage restrictions, but patients with a prior history of cancers were excluded. The cases included 184 (49.5%) SCCs, 123 (33.1%) adenocarcinomas, 57 (15.3%) small cell carcinomas and 8 (2.1%) large cell carcinomas. The demographics and clinical characteristics of the cases were consistent with those of a nationwide lung cancer survey conducted by the Korean Academy of Tuberculosis and Respiratory Disease in 1998.31 Controls were randomly selected from a pool of healthy volunteers who visited the general health checkup center of Kyungpook National University Hospital during the same period. Controls were frequency-matched (1:1) to cases based on sex and age (± 5 years). All cases and controls were Koreans and residents of Daegu city and the surrounding regions. A detailed questionnaire was completed for each case and control by a trained interviewer. The questionnaire included information on the average number of cigarettes smoked daily and the number of years the subjects had been smoking. For smoking status, a person who had smoked at least once a day for > 1 year in his or her lifetime was regarded as a smoker. A former smoker was defined as one who had stopped smoking at least 1 year before diagnosis in the case of patients and 1 year before the study began in the case of controls. Cumulative cigarette dose (pack-years) was calculated by the following formula: pack-years = (packs per day) × (years smoked).
Genomic DNA was extracted from peripheral blood lymphocytes by proteinase K digestion and phenol/chloroform extraction. The hMLH1 −93GA genotypes were determined by a PCR-restriction fragment length polymorphism assay. The PCR primers for the −93GA polymorphism (Genbank accession no. AC011816) were 5′-CCGAGCTCCTAAAAACGAAC-3′ (bases 142905–142924 of hMLH1) and 5′-CTGGCCGCTGGATAACTTC-3′ (bases 143290–143272 of hMHL1), which generates a 387 bp fragment. The PCR reactions were performed in a 20 μl reaction volume containing 200 ng of genomic DNA, 25 pmol of each primer, 0.2 mM dNTP, l × PCR buffer (50 mM KCl and 10 mM Tris-HCl, pH 8.3), 1.5 mM MgCl2 and 1 unit Taq polymerase (Takara Shuzo, Otsu, Shiga, Japan). The PCR profile consisted of an initial denaturation step of 95°C for 5 min followed by 36 cycles of 95°C for 30 sec, 58°C for 30 sec, 72°C for 30 sec and a final elongation step of 72°C for 10 min. The PCR products were digested overnight with 10 units of PvuII (New England BioLabs, Beverly, MA) at 37°C and then resolved on a 6% acrylamide gel. The wild-type (G) allele (i.e., −93G) yields 2 bands (207 and 180 bp) and the polymorphic (A) allele (i.e., −93A) is determined by the presence of the uncut 387 bp band (indicative of the absence of the PvuII cutting site). For quality control, the genotyping analysis was performed with blinding to case/control status and repeated twice for all subjects. The results of genotyping were 100% concordant. To confirm the genotyping results, selected PCR-amplified DNA samples (n = 2 each for the GG, GA and AA genotypes) were examined by DNA sequencing, and the results were also 100% concordant.
Cases and controls were compared using Student's t-test for continuous variables and the chi-square test for categorical variables. When multiple comparisons are made, significant associations may arise by chance. To avoid such errors, the corrected p-values (p corrected) were also calculated for multiple testing using Bonferroni's inequality method. Hardy-Weinberg equilibrium was tested by a goodness-of-fit chi-square test with one degree of freedom to compare the observed genotype frequencies with the expected genotype frequencies among the cases and controls. Unconditional logistic regression analysis was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs), with adjustment for possible confounders (sex, as a nominal variable; age and pack-years, as continuous variables).
The interaction between genotype and smoking was tested both by logistic regression model, including the interaction term between genotype and pack-years of smoking, and by stratification analysis. For stratification analysis, ever smokers were dichotomized into 2 subgroups by the median pack-years for all subjects excluding never smokers (i.e., ≤ 39 pack-years and > 39 pack-years). Stratification analyses were also performed when the subjects were divided according to the tertiles or quartiles of pack-years. All analyses were performed using Statistical Analysis Software for Windows, version 6.12 (SAS Institute, Gary, NC).
The demographics of the cases and controls enrolled in this study are shown in Table I. There were no significant differences in the mean age and sex distribution between cases and controls, suggesting that the matching based on these 2 variables was adequate. The cases had a higher prevalence of current smokers than the controls (p < 0.001). The number of pack-years in smokers was significantly higher in cases than in controls (40.8 ± 21.0 vs. 34.1 ± 16.5 pack-years; p < 0.001). These differences were controlled in later multivariate analyses.
|Variables||Case (n = 372)||Control (n = 371)|
|Age (years)||60.8 ± 9.01||60.6 ± 9.1|
|Male||305 (82.0)2||301 (81.1)|
|Female||67 (18.0)||70 (18.9)|
|Current||280 (75.3)||200 (53.9)|
|Former||26 (7.0)||79 (21.3)|
|Never||66 (17.7)||92 (24.8)|
|Pack-years4||40.8 ± 21.0||34.1 ± 16.5|
The distributions of hMLH1 −93GA genotypes among controls and cases are shown in Table II. The frequency for the hMLH1 −93A variant allele among controls was 0.53, which is comparable with its frequency in previous studies (0.50 for Koreans32 and 0.46 for Japanese30). The frequencies of the GG, GA and AA genotypes among cases (17.7%, 47.3% and 35.0%, respectively) were significantly different from those among controls (19.2%, 55.5% and 25.3%, respectively; p = 0.016). When lung cancer cases were stratified by histologic type, the frequencies of the GG, GA and AA genotypes in SCC cases (14.7%, 46.2% and 39.1%, respectively) differed significantly from controls (p = 0.004). However, the distribution of genotypes in the other histologic types was not significantly different from controls.
|Genotype||A allele frequency (%)|
|Control||71 (19.2)1||206 (55.5)||94 (25.3)||53.1|
|Case||66 (17.7)||176 (47.3)||130 (35.0)2||58.6|
|Squamous cell carcinoma||27 (14.7)||85 (46.2)||72 (39.1)3||62.2|
|Adenocarcinoma||26 (21.1)||58 (47.2)||39 (31.7)||55.3|
|Large cell carcinoma||2 (25.0)||3 (37.5)||3 (37.5)||56.2|
|Small cell carcinoma||11 (19.3)||30 (52.6)||16 (28.1)||54.4|
The GG and GA genotypes were less frequent in overall cases and SCC cases than in controls, whereas the AA genotype was more frequent in overall cases and SCC cases than in controls (Table II), suggesting that the AA genotype might be a risk genotype for lung cancer. The distribution of GG and GA genotypes was not significantly different between SCC cases and controls, while the distributions of GG and AA genotypes and GA and AA genotypes were significantly different between SCC cases and controls (p = 0.01, p corrected = 0.02; and p = 0.002, p corrected = 0.004, respectively). We therefore calculated the ORs for lung cancer using the homozygous GG genotype alone or combined with the heterozygous GA genotype as the reference groups. The association between the hMLH1−93GA genotype and the risk of lung cancer is shown in Table III. The adjusted ORs of the GA and AA genotypes for overall lung cancer were not significantly different from that of the GG reference group. However, the AA genotype was associated with a significantly increased risk for overall lung cancer when the combined GA and GG genotype was used as the reference group (adjusted OR = 1.57; 95% CI = 1.14–2.18; p = 0.006). When analyses were stratified by tumor histology, the AA genotype was associated with a significantly increased risk for SCC compared with either the GG genotype (adjusted OR = 2.02; 95% CI = 1.15–3.55; p = 0.014) or the combined GG and GA genotype (adjusted OR = 1.83; 95% CI = 1.24–2.71; p = 0.003). However, there was no significant association between the hMLH1 −93GA genotype and the risk of adenocarcinoma or small cell carcinoma.
|Genotype||All cases||Squamous cell carcinoma||Adenocarcinoma||Small cell carcinoma|
|GA||0.89 (0.59–1.32)||1.14 (0.67–1.95)||0.73 (0.42–1.27)||1.00 (0.47–2.12)|
|AA||1.44 (0.93–2.23)||2.02 (1.15–3.55)1||1.05 (0.57–1.93)||1.06 (0.46–2.46)|
|GG + GA (reference)||1.0||1.0||1.0||1.0|
|AA||1.57 (1.14–2.18)2||1.83 (1.24–2.71)3||1.32 (0.83–2.10)||1.06 (0.57–2.00)|
The association between hMLHI −93GA genotypes and SCC according to the extent of tobacco smoke exposure (≤ 39 pack-years and > 39 pack-years) is shown in Table IV. In the group of individuals having ≤ 39 pack-years of smoking, the AA genotype was associated with a borderline significantly increased risk for SCC compared to the GG genotype (adjusted OR = 2.62; 95% CI = 1.00–6.84; p = 0.05), and this genotype was associated with a significantly increased risk for SCC compared to the combined GG and GA genotype (adjusted OR = 1.95; 95% CI = 1.03–3.66; p = 0.039). In the group of individuals with > 39 pack-years of smoking, however, the distribution of hMLHI −93GA genotypes was not significantly different between SCC cases and controls. The results revealed the same trend when tertiles or quartiles of pack-years were used. For the gene-smoking interaction analysis between hMLH1 genotype and SCC risk, the interaction term between hMLH1 genotype and pack-years of smoking for the AA vs. GG + GA genotype comparison was borderline significant (p = 0.08).
|Genotype||≤ 39 pack-years||> 39 pack-years|
|Control (%)||Case (%)||Adjusted1 OR (CI)||Control (%)||Case (%)||Adjusted OR (CI)|
|GG||35 (20.5)||7 (11.7)||1.00||19 (17.6)||17 (15.5)||1.00|
|GA||89 (52.0)||28 (46.7)||1.49 (0.58–3.81)||58 (53.7)||53 (48.2)||0.91 (0.42–1.98)|
|AA||47 (27.5)||25 (41.7)||2.62 (1.00–6.84)2||31 (28.7)||40 (36.4)||1.37 (0.60–3.11)|
|GG + GA||124 (72.5)||35 (58.3)||1.00||77 (71.3)||70 (63.6)||1.00|
|AA||47 (27.5)||25 (41.7)||1.95 (1.03–3.66)3||31 (28.7)||40 (36.4)||1.47 (0.82–2.61)|
This is the first case-control study of an hMLH1 polymorphism in relation to lung cancer susceptibility. In this study, the hMLH1 −93AA genotype was associated with a significantly increased risk for SCC of the lung. The effect of the polymorphism on the risk of SCC was more evident in lighter smokers than in heavier smokers. These findings suggest that the hMLH1 −93GA polymorphism may contribute to the inherited genetic susceptibility to SCC of the lung.
In the current study, the hMLH1 −93GA polymorphism was significantly associated with the risk of SCC, but not with adenocarcinoma or small cell carcinoma. Although the reason for the observed histology-dependent difference in the risk conferred by the hMLH1 polymorphism is unknown, this difference may be attributable to the differences in the pathways of carcinogenesis among histologic types of lung cancer. Various lines of evidence suggest that the histologic type of lung cancer may be determined by the particular initiating agent to which an individual is exposed.33, 34, 35 Certain genotypes could therefore confer a greater susceptibility to a particular histologic type of lung cancer.36, 37 However, the lack of statistical significance for adenocarcinoma might be due to the small number of adenocarcinoma cases. A large study is therefore warranted to confirm this finding.
The mechanism responsible for the association between the hMLH1 −93GA polymorphism and SCC remains to be elucidated. However, the strong relationship between SCC and tobacco smoking suggests that hMLH1 may be involved in the repair of mutagenic DNA damage arising from carcinogens produced by smoking. Tobacco smoke is a complex mixture containing many potent carcinogens that produce a variety of bulky DNA adducts and reactive oxygen-induced DNA damage.33, 38 MMR genes have long been known to play a major role in the repair of errors made during DNA replication. However, recent studies have demonstrated that the hMLH1 protein functions in the repair of physical or chemical damage to the DNA and/or its nucleotide precursors as well as replication error correction.22, 23, 24 Thus, our finding supports the hypothesis, proposed by Benachenhou et al.,18 that the hMLH1 gene could be involved in lung tumorigenesis by functioning in the repair of DNA damage other than replication error correction.
In the current study, we did not observe an altered risk of lung cancer for −93GA heterozygotes compared with homozygous −93GG individuals. Although it is unclear why only −93AA genotype was a high-risk genotype, it could be due to that the −93A allele has a recessive effect on hMLH1 phenotype as in the case of the XPD Asp312Asn (23592GA of exon 10) polymorphism, where individuals with the heterozygous Asp312Asn genotype had DRC capacity and apoptotic response to UV damage similar to those with the wild-type genotype,39, 40 and only Asn312Asn genotype was associated with a significantly increased risk of lung cancer (i.e., the 312Asn allele is recessive41).
Genetic susceptibility to lung cancer may depend on the level of exposure to tobacco smoke.41, 42 Therefore, the association between tobacco smoke exposure and the distribution of hMLH1 genotypes was examined in smokers; as the number of never smokers among the SCC cases was small (n = 14), never smokers were excluded from this analysis. When we stratified the smokers by median pack-years of smoking, the AA genotype was significantly associated with an increased risk for SCC among lighter smokers, whereas the increase in risk for the AA genotype among heavier smokers was smaller and not statistically significant. These results conform to current theories that genetic susceptibility has less impact at high loads of carcinogen exposure, when environmental influences may overpower any genetic predisposition.1, 43 However, because the difference of OR between subgroups in the stratified analysis was modest and the interaction term between genotype and smoking was borderline significant, it is possible that such a finding is attributable to chance. Additional studies with more subjects will be needed to confirm this finding.
Wei et al.44 reported that the expression level of hMLH1 in blood lymphocytes shows considerable interindividual variation, and low expression of hMLH1 is associated with an increased risk of head and neck cancer. Therefore, the association between the hMLH1 −93GA polymorphism and the risk of SCC in the present study may be due to difference in the transcriptional activity of the hMLH1 promoter. To determine whether this polymorphism is associated with the regulation of gene expression, we compared the promoter activity of the 2 alleles (−93G and −93A) by a luciferase assay but found no significant difference in the promoter activity (data not shown). This suggests that the association between the −93GA polymorphism and SCC may be due to linkage disequilibrium with either another hMLH1 variant or an adjacent true susceptibility gene rather than to a direct functional effect of the GA substitution.
One must consider a number of limitations of this study. Since this study was a hospital-based case-control study, there might be some selection bias. Given that most lung cancer patients are treated at university hospital in South Korea, the demographics and clinical characteristics of the cancer patients in the current study were compatible to those of a nationwide lung cancer survey.31 Furthermore, as all the lung cancer patients diagnosed at a national university hospital were included in this study, it is reasonable to assume that the case group represents the lung cancer cases in our community. All cases and controls were Koreans and residents of Daegu city and the surrounding regions. Another selection bias may derive from controls that did not participate in this study. However, because the age and sex distribution of nonparticipating controls were similar to those of the participating controls in the current study, a self-selection bias is unlikely. By matching on age and sex, potential confounding factors might be minimized. An inadequacy in matching on smoking exposure would be controlled in data analysis with additional adjustment.
A further limitation of this study is that the distribution of hMLH1 −93GA genotypes among controls was not in Hardy-Weinberg equilibrium (by one degree of freedom; chi-square = 5.01; p = 0.025). Several factors may account for this finding. It is possible that this finding may be due to genotyping error. However, we repeated twice the genotyping analysis for all subjects, and the results were 100% concordant. To confirm the genotyping results, moreover, selected PCR-amplified DNA samples were examined by DNA sequencing and the results were also 100% concordant. Therefore, it is unlikely that such a violation is due to genotyping error. Hardy-Weinberg equilibrium can be disturbed when one or more of the assumptions of Hardy-Weinberg law (no migration, no selection, no mutation, random mating) are violated. Although there has been no proven evidence, it is possible that this polymorphism may influence mating, fertility, or fecundity and/or viability, any of which could in turn influence changes in the genotype frequencies in a population.
In conclusion, we found that the hMLH1 −93GA polymorphism is associated with an increased risk of SCC of the lung, implying that the hMLH1 −93GA genotyping may help identify individuals who are at increased risk for SCC of the lung. It is possible that our findings are attributable to chance. Therefore, additional studies with larger sample sizes are required to confirm our findings. Future studies of other sequence variants in the hMLH1 gene and their association with DRC phenotypes are also needed to understand the role of hMLH1 in determining the risk of lung cancer.
- 17A national cancer institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998; 58: 5248–57., , , , , , , , , , .
- 32Mutational analysis of promoters of mismatch repair genes hMHS2 and hMLH1 in hereditary nonpolyposis colorectal cancer and early onset colorectal cancer patients: identification of three novel germ-line mutations in promoter of the hMSH2 gene. Cancer Res 2002; 62: 38–42., , , .
- 34Cigarette smoking and adenocarcinoma of the lung: the relevance of nicotine-derived nitrosamines. J Smoking Relat Disord 1993; 4: 165–90., , , , , .