Dna Repair Gene Polymorphism, Genotype - Phenotype Correlation With Susceptibility To Lung Cancer The Relationship Between Molecular Epidemiological Studies

Lung cancer remains the leading cause of cancer-related deaths inthe world and its incidence rate has been increasing significantly in thelast two decades in China. It has been estimated that tobacco smoke isresponsible for more than 80% of all lung cancers; yet only less than 10%smokers develop lung cancer, suggesting that there is an inter-individualvariation in genetic susceptibility to lung cancer in the general population.Tobacco smoke contains more than 60 carcinogens, among whichpolycyclic aromatic hydrocarbons (PAH) and the tobacco-specificnitrosamine (TSN) are likely to play major roles in lung carcinogenesis.These carcinogens can form adducts or induce oxidative damages on ourDNA, resulting mutations and genome instability. There are several partly overlapped DNA repair pathways in humanfor repairing smoking-related damages, including direct repair (DR),base-excision repair (BER) and nucleotide-excision repair (NER).Reduced DNA repair capacity (DRC) may confer to the geneticsusceptibility to lung cancer. DNA repair gene polymorphisms,particularly single nucleotide polymorphisms (SNP), are the mostcommon form of genetic variations among human beings, which may bethe underlying molecular mechanisms of the inter-individual variation ofDRC. Therefore, we hypothesized that (1) functional polymorphisms incore DNA repair genes are associated with lung cancer susceptibilityindividually or jointly and gene-gene/gene-environment interactions playimportant roles; (2) DRC phenotype is genetically based, and correlatedwith the functional polymorphisms in some core DNA repair genes.To test our hypothesis (1), we selected polymorphisms in 10representive genes involved in DR, BER and NER repair pathways, using both 'sequence-based' and 'mapping-based' selection approaches, andtested their associations with lung cancer risk in several case-controlstudies. We also evaluated joint and interactional effects between thesepolymorphisms and smoking on lung cancer susceptibility. In the sametime, we illustrate the molecular mechanism of the positive associationbetween one promoter SNP XRCC1 -77T＞C and lung cancer risk usingan in vitro transcription assay. To test our hypothesis (2), we detectedDRC in 430 cancer-free controls using BPDE-induced host cell reaction(HCR) assay, and studied the phenotype-genotype correlations withseveral potentially functional polymorphisms in NER genes.In conclution, our study contributes to the knowledge on lung cancersusceptibility and DNA repair gene polymorphisms, and multiplegene-gene/gene-environment interactions. Provide a "proof-of-principle"approach for molecular epidemiological studies in cancer susceptibility.The identified risk genotypes/haplotypes could be used, if validated, asmolecular markers to improve our ability to define high-risk populationswho may be susceptible to smoking-related lung cancer in the Chinesepopulation.Partâ… DNA Repair Gene Polymorphisms and Lung Cancer RiskCharpterâ… : Functional SNPs in XRCC1 and Lung Cancer RiskXRCC1 is one of the major DNA repair proteins involved in BERpathway, which is responsible for repairs of cigarrett smoking resultedoxidative damages and N⁷-methy Guanine. Associations between XRCC1nsSNPs (e.g. Arg194Trp and Arg399Gln) and susceptibility to cancershave been extensively investigated. Recently, a novel T-to-C transitionlocated at nucleotide-77 (rs3213245) in the promoter region of XRCC1 has been identified by re-sequencing the XRCC1 gene, and this promoterT＞C substitution was associated with risk of oesophageal squamous cellcarcinoma in a Chinese population.In the current study of 710 incident histopathologically diagnosedlung cancer patients and 710 cancer-free controls frequency-matched tothe cases on age, sex and residential area, we simultaneously genotypedthe three polymorphisms (i.e. XRCC1-77T＞C, Arg 194Trp andArg399Gln) and evaluated their associations with lung cancer risk.Multiple logistic regression analyses were used to estimate the maineffects of XRCC1 variants on lung cancer risk and the multiplicativegene-environment interaction, and bootstrapping tests were used toevaluate the additive interaction between gene and smoking. To examinethe impacts of-77T＞C variant on the basal transcription activity ofXRCC1 promoter region, the T and C allelic reporter constructs wereprepared by amplifying the 1246-bp XRCC1 promoter region (from-1152 to +93 relative to the translation start site) and cloned into thepGL3-Basic vector. Human mammary gland adenocarcinoma cell lineMCF7, colon cancer cell line HCT116 (p53 positive and negative), andlung carcinoma cell line NCI-H1299 were transiently transfected with thetwo constructs, respectively, and the pGL3-basic vector without an insertwas used as a baseline to calculate the relative luciferase activity.The XRCC1-77T＞C genotype frequencies were significantlydifferented between cases and controls (P=0.0007). Logistic regressionanalysis revealed that compared with the -77TT wild-type, subjectscarrying the -77CT heterozygote had a significantly 1.51-fold increasedrisk of lung cancer (95ï¼…CI=1.17-1.94) and those carrying the -77CChomozygote had a 2.98-fold elevated risk with borderline significance(95ï¼…CI=0.93-9.59). In the joint effect analyses of cumulative smoking and XRCC1-77T＞C variant, compared with the non-smokers having the-77TT wild-type genotype, we observed a 9.82- and 3.28-fold increasedrisk of lung cancer in heavy (＞30 pack-years) and light (≤30 pack-years)smokers, respectively, carrying the -77CT/CC genotypes, but only a 4.07-and 2.66-fold increased risk for heavy and light smokers not carrying thevariant, suggesting a possible gene-smoking interaction. However, theinteractive effects of XRCC1-77CT/CC variant genotypes with smokingwere statistically evident only on an additive scale (P=0.027 for XRCCT-77C variant and smoking status), not on a multiplicative scale. In the invitro transcription assay, we found that relative luciferase activities drivenby the mutant C aUelic XRCC1 promoter were 57-64ï¼…of those driven bythe wild-type T-allelic XRCC1 promoter in the four types of cell linesexamined, which is consistent with our association study.In conclusion, our study provides evidence that the newly identifiedpolymorphism of XRCC1 -77T＞C contributes to the etiology of lungcancer and the C allelic XRCC1 promoter was associated with a reducedtranscriptional activity.Charpterâ…¡: Tagging SNPs in NER Core Genes and LungCancer RiskNER is the major repair pathway for removing DNA damage causedby tobacco smoke (i.e., BPDE-DNA adducts) and deals with a wide classof helix-distorting lesions that interfere with base pairing and generallyobstruct transcription and normal replication. Several critical genes(so-called core factors) participate in NER process, and deficient in theseproteins may result in the recessive syndrome xeroderma pigmentosum(XP) characterized by extraordinary sensitivity to ultraviolet light and anup to 2000 fold increased risk of sunlight-induced skin cancer. Accumulating evidences suggested that polymorphisms in these NERgenes might contribute to genetic susceptibility to lung cancer. However,most of the previous studies were designed to analyze single locus/geneand lacked statistical power because of the limited study sample size.To comprehensively investigate the roles of the polymorphisms inthe NER pathway in the development of smoking-related lung cancer, weconducted a case-control study of 1010 incident lung cancer cases and1011 age and sex frequency-matched cancer-free controls. In this study,we selected and genotyped 41 tagging SNPs in 8 core NER genes basedon the linkage disequilibrium (LD) analysis and tested the hypothesis thatgenotypes/haplotypes/diplotypes of the NER genes may contribute tosusceptibility to lung cancer.As a result, a total of eight SNPs in five genes (i.e., ERCC1, ERCC4,ERCC5, XPC and DDB2) were significantly associated with lung cancerrisk in the single locus analysis. Distributions of haplotypes TCCCATTin ERCC1, ACGAA in ERCC2, TA in ERCC3 and GG in DDB2 weresignificant or borderline significant (P＜0.1) different between cases andcontrols. Haplotypes TCCCATT in ERCC1, TA in ERCC3, ACCCA inXPC and GG in DDB2 were also associated with a substantially alteredrisk for lung caner compared with the other haplotypes and the effects ofthese haplotypes were in a co-dominant model. However, althoughhaplotype TGG in ERCC4 and CCCCAA in ERCC5 did not show a cleardifferent distribution between cases and controls, their diplotypes doshow significant main effects in dominant or recessive disease model.Together with haplotype AA in XPA, we have one haplotype in each geneto build the eight dipolotypes to represent the genetic variation in the 8genes. In the pathway-based combined analysis, lung cancer risk wasincreased in a dose-response manner with the increasing of risk diplotypes (P＜0.0001). Subjucts with adverse diplotypes in more than 3genes were associated with a 1.63-fold (95ï¼…CI=1.35-1.98) increasedrisk for lung cancer, compared with those with adverse diplotypes in 4 orless genes. This increased risk was slightly more evident in positivefamily cancer history carriers and non-smokers. Furthermore, there are asignificant multiplicative interaction (P=0.021) between dichotomizedNER combined diplotypes and cumulative smoking and an additiveinteraction (P=0.046) between dichotomized NER combined diplotypesand family history of cancer.Our study is among the few that used the pathway-based candidategene approach to examine the association between SNPs of multiplegenes in a specific DNA repair pathway and risk of the complex diseasessuch as lung cancer. These findings provide further evidence to supportthe important role of genetic susceptibility conferred by variants of theNER pathway in the etiology of lung cancer. Further studies withfunctional evaluation of the SNPs are warranted to replicate and extendthe significance of these findings.Charpterâ…¢: Tagging SNPs in MGMT and Lung Cancer RiskApart from PAH, the tobacco-specific nitrosamine4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is likely to playmajor roles in smoking-related lung carcinogenesis. NNK can formmethyl-adducts, including those at the N⁷- and O⁶-positions of guanine.The repair of alkylation damage to the O⁶-position of guanine isperformed by O⁶-alkylguanine-DNA alkyltransferase (MGMT), whoseexpression is considerable variated among individuals. Recently,Margison et al. reported that genetic variation in MGMT resulted inallelic expression imbalance in peripheral blood mononuclear cells, which may explain why there are different expression levels and/or functionalactivities of MGMT across different populations. To elucidate the role ofsequence variation in MGMT in the etiology of lung cancer, we conducteda comprehensive association study with a focus on LD structure ofcommon variations across MGMT sequence and its modification effect onsmoking-related lung cancer risk.Our study including 500 newly histopathologically diagnosed lungcancer patients and 517 cancer-free controls. We choose SNPs from theHapMap, dbSNPs databases and also published papers on MGMTpolymorphisms based on the HapMap block region coverage, minar allelefrequency (MAF), and functional relevance. Overall, 39 SNPs acrossMGMT were genotyped by using the Illumina SNP genotyping BeadLabplatform and PCR-RFLP method. We then reconstructed the LD blockswithin MGMT based on the genotype data of our 517 southeast Chinesecontrols by the method of Gabriel et al. and selected a minimum set ofhaplotype tagging SNPs (htSNPs) within each block to ensure a R_h² of atleast 0.80 to capture all possible haplotypes that had a frequency of atleast 5ï¼…by using the algorithm described by Stram et al..Overall, no significant main effects was observed for each singlegenetic variant in MGMT on lung cancer risk, and only the diplotypecarrying 1 copy but not 2 copies of the block 3 '12' haplotype wasassociated with a significantly decreased lung cancer risk as comparedwith the diplotype carrying 0 copies of the '12' haplotype. In addition, thedistrubutions of '22' haplotype in block 1 and 2, '22' haplotype in block 4,'12' haplotype in block 5 and '12' haplotype in block 6 were significantor borderline significant (P＜0.1) different between cases and controlsamong different cumulative smoking dose, which may suggest potencialinteraction with smoking. Then, we chose one haplotype per block to build dipolotypes. By using a dipolotype-based multifactordimensionality reduction (MDR) analysis, we found that there weresignificant more-than-multiplicative interaction between diplotypes inblock 5 and cumulative smoking and additive interaction betweengenotypes of pre-block SNP rs 1625649 and smoking status in relation onlung caner risk. Diplotypes in block 3 and block 5, genotypes ofrs 1625649 and trichotomized cumulative smoking are the four factorsincluded in the MDR-defined best model on lung cancer risk. When thesevariables were combined and dichotomized, we found that subjectscarrying the combined risk stratum had a significantly 4.10-fold (95ï¼…CI=3.12-5.37, P=2.09×10^-24) increased risk for lung cancer.Our findings suggested that genetic variants in MGMT maymodulate the risk of smoking-related lung cancer. This haplotype basedinteraction analysis might provide a "proof-of-principle" approach forstudying candidate genes in cancer susceptibility.Partâ…¡Correlation of DRC Phenotype and Genotypes ofSelected Functional NER SNPsBenzo[a]pyrene (B[a]P) is a classic DNA-damaging carcinogenfound in tobacco smoke. Its bioactivated form, BPDE, can causeirreversible damage to DNA by forming DNA adducts through covalentbinding. In living cells, BPDE-DNA adducts can block transcription of anessential gene if they are not repaired efficiently by the NER pathway.Based on this relative-specific repair model, Athas et al. developed aHCR assay that measures report gene expression levels to meet the extentto which the cellular NER repair system removed the lesions in DNA. Ingeneral, phenotypic studies were good at predicting cancer risk but hadrelatively large variations in risk estimates because of large assay variation in addition to relatively small numbers of subjects included inthe studies. But it can serve as an intermediate end-point, reflecting asummation of the effects of all possible functional SNPs in related repairgenes and provides more homogenous environment that will lead to more'repeatable' results than disease as an end-point did. Therefore, weattempted to identify underlying genotypes in relation to suboptimal DRCby conducted a genotype-phenotype correlation analysis in 430cancer-free no-hispanic Caucasians with 7 selected potencially functionalpolymorphisms in the NER pathway.DRC was measured by using the BPDE-induced HCR assay. SNPswere genotyped by using the SNPlex middle throughput assay and thePCR-RFLP method. We identified all non-synonymous SNPs andregulate SNPs with MAF＞0.05 in the core NER genes, inclduing7 eligible SNPs in 5 NER genes: G-4A (5'UTR, rs1800975) in XPA,Ala499Val (rs2228000) and Lys939Gln (rs2228001) in XPC; Asp312Asn(rs1799793) and Lys751Gln (rs13181) in ERCC2, Hisl 104Asp (rs17655)in ERCC5 and C8092A (3'UTR, rs3212986) in ERCC1.As a result, age, gender, pack-years of smoking, ethanoal-years ofdringking and family history of cancer were not correlated with DRCphenotype, suggesting a genetic basis of DRC. In the individual SNPanalysis, a significant difference between subgroups of genotypes wasevident for ERCC2 Lys751Gln, ERCC5 His 1104Asp and XPC Lys939Glnloci in a dominant effect genetic model. In addition, significantdose-dependent effects of the variant alleles on mean DRC phenotypewere found for ERCC2 Lys751Gln, XPG His1104Asp and XPCLys939Gln in a co-dominant genetic model. In the combined analysiswith the 7 loci in dominant models, the dose-dependent effects were moreevident (continuous DRC: P for trend＜0.0001; dichotomized DRC: P for trend=0.0001), and subjects with 2 or more combined riskgenotypes had a significantly increased (111ï¼…) probability of beingassociated with a lower DRC phenotype (adjusted OR=2.11, 95ï¼…CI=1.36-3.26).Our results showed the rethionale to identify underlying genotypesin relation to suboptimal DRC, although which may require a much largersample than that in the current study and a large number ofrepresentiveSNPs. BPDE-induced HCR assay can partly mimic the tobaccorelated-lung cancer model, which may contribute to the identification ofsusceptible populations.