Skip Navigation


Carcinogenesis Advance Access originally published online on January 10, 2008
Carcinogenesis 2008 29(2):342-350; doi:10.1093/carcin/bgm285
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
29/2/342    most recent
bgm285v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, D.
Right arrow Articles by Lu, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chen, D.
Right arrow Articles by Lu, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Genetic variants in peroxisome proliferator-activated receptor-{gamma} gene are associated with risk of lung cancer in a Chinese population

Dan Chen1,{dagger}, Guangfu Jin2,{dagger}, Ying Wang3, Haifeng Wang3, Hongliang Liu1, Yanhong Liu1, Weiwei Fan1, Hongxia Ma2, Ruifeng Miao2, Zhibin Hu2, Weiwei Sun3, Ji Qian1, Li Jin1,3, Qingyi Wei4, Hongbing Shen2, Wei Huang3,* and Daru Lu1,*

1 The State Key Laboratory of Genetic Engineering and The MOE Key Laboratory of Contemporary Anthropology, School of Life Sciences, Fudan University, Shanghai 200433, China
2 Department of Epidemiology and Biostatistics, Cancer Research Center of Nanjing Medical University, Nanjing 210029, China
3 Department of Genetics, Shanghai-MOST Key Laboratory of Health and Disease Genomics, Chinese National Human Genome Center, Shanghai 201203, China
4 Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, TX 77030, USA

* To whom correspondence should be addressed. Tel/Fax: +86 21 65642799; Email: drlu{at}fudan.edu.cn Correspondence may also be addressed to Wei Huang. Tel: +86 21 50801795; Fax: +86 21 50801922; Email: huangwei{at}chgc.sh.cn


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 Funding
 References
 
Accumulating evidence indicates that activation of the peroxisome proliferator-activated receptor-{gamma} (PPAR-{gamma}) dampens the inflammation cascade and inhibits tumor growth of the lung, suggesting that it has tumor suppressor functions. We performed a case–control study of 500 incident lung cancer cases and 517 age- and sex frequency-matched cancer-free controls in a Chinese population to investigate the role of 11 selected single nucleotide polymorphisms (SNPs) of PPAR-{gamma} in the etiology of lung cancer. We found that decreased lung cancer risk was statistically significantly associated with seven SNPs (P = 0.0004 for rs13073869 and 0.0130 for rs1899951 in a dominant model; P = 0.0310 for rs4135247 in a log-additive model; and P = 0.0468 for rs2972162, 0.0175 for rs709151, 0.0172 for rs11715541 and 0.0386 for rs1175543 in an overdominant model). Consistent with these results of single-locus analysis, both the haplotype and the diplotype analyses revealed a protective effect of the haplotype ‘AGA’ and ‘AAA’ of rs13073869, rs1899951 and rs4135247. Furthermore, we observed a statistically significant interaction between the rs1899951 and cigarette smoking. Our results indicate that PPAR-{gamma} polymorphisms and their interaction with smoking may contribute to the etiology of lung cancer. These findings need to be validated in larger, preferably population-based, studies including different ethnic groups.

Abbreviations: PPAR-{gamma}, peroxisome proliferator-activated receptor-{gamma}; SNP, single nucleotide polymorphism; OR, odds ratio; CI, confidence interval; MAF, minor allele frequency; LD, linkage disequilibrium; HWE, Hardy-Weinberg equilibrium


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Supplementary material
 Funding
 References
 
Lung cancer is the leading cause of cancer deaths in the world with a poor prognosis and an overall 5-year survival rate of <15% (1). The epidemic of lung cancer is directly attributable to cigarette smoking that accounts for 87% of lung cancer cases (2). However, only a small fraction of smokers (usually <20%) develop lung cancer in their lifetime (3), suggesting that genetic susceptibility plays a role in the development of lung cancer.

Exposure to cigarette smoke activates an inflammatory cascade in the airway epithelium. For example, tobacco smoke generates reactive oxidant species, leading to damage to the lung epithelium and its altered permeability, induction of goblet cell hyperplasia and mucus production as well as recruitment of macrophages and neutrophils to the airway (4–7). Chronic inflammation causes prolonged irritation and activated local host response, which ultimately promote cell proliferation (8). As a result, sustained cell proliferation facilitates tumor formation and progression in an environment abundant in inflammatory cells, growth factors, activated stroma and enhanced angiogenesis (9,10). Indeed, it has been estimated that cancer is preceded by chronic inflammation in up to a third of all cases (11). Case–control studies have shown an increased risk of lung cancer in patients with inflammatory airway phenotypes, such as asthma, bronchitis and emphysema (12–14). Recent data suggest that cigarette smoke stimulates airway epithelial cells and immune cells to release proinflammatory cytokines, such as tumor necrosis factor-{alpha}, interleukin-4, interleukin-6, interleukin-8 and cyclooxygenase-2. The peroxisome proliferator-activated receptor-{gamma} (PPAR-{gamma}), a ligand-activated transcription factor belonging to the nuclear receptor superfamily, has been shown to counteract inflammation by inhibiting the expression of proinflammatory factors, suppressing the generation of reactive oxidant species and nitrogen species, inhibiting cigarette smoke-induced mucin production and dampening migratory responses (15).

In addition to its anti-inflammatory effect, PPAR-{gamma} also plays a critical role in regulating diverse processes in lung stromal/parenchymal cells, including cell-cycle control, differentiation, apoptosis and carcinogenesis (16–18). PPAR-{gamma} is known to express in both human small-cell lung cancer and non-small-cell lung cancer cells (19–21). Several studies have demonstrated that the activation of PPAR-{gamma} inhibited proliferation and growth of lung cancer cells (22–24). PPAR-{gamma}-selective agonists, such as 15-deoxy-delta12-14-prostaglandin J2, and the synthetic thiazolidinedione compounds, certain non-steroid anti-inflammatory drugs, lead to growth arrest, induction of apoptosis and promotion of differentiation in cell lines (18,20,24). Besides, the daily oral administration of either troglitazone or pioglitazone to the severe combined immunodeficient mice significantly reduced the number of lung metastases and restricted tumor progression in vivo (25). Taken together, these observations suggest that PPAR-{gamma} may act as a tumor suppressor in the pathogenesis and progression of human lung cancer.

The present work was motivated by the biological plausibility that genetic variation in the PPAR-{gamma} gene could alter its expression level or biochemical functions and consequently may have an impact on individual risk of lung cancer. To test this hypothesis, we conducted a case–control study of 500 incident lung cancer cases and 517 age- and sex frequency-matched cancer-free controls in a Chinese population and genotyped 11 single nucleotide polymorphisms (SNPs) of PPAR-{gamma} using an Illumina high-throughput genotyping platform. We also investigated potential interactions between polymorphisms of the PPAR-{gamma} gene and cigarette smoking in lung cancer risk.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Supplementary material
 Funding
 References
 
Study population
The study design and subject recruitment have been described previously (26). Briefly, 500 lung cancer patients and 517 cancer-free controls were genetically unrelated ethnic Han Chinese and were from Nanjing City and surrounding regions in southeastern China. Patients with histopathologically confirmed incident lung cancer were consecutively recruited between July 2002 and December 2004 at the Cancer Hospital of Jiangsu province (Nanjing) and the First Affiliated Hospital of Nanjing Medical University, Nanjing, China, with a response rate of 90.5%. Cancer-free controls were randomly selected from 10 500 individuals who participated in a community-based screening program for non-infectious diseases conducted in Jiangsu province during the same period when the cases were recruited, with a response rate of 83.8%. All the control subjects had no history of cancer and were frequency matched to the cases by age (±5 years), sex and residential area (urban or rural areas). Each participant was scheduled for an interview after a written informed consent was obtained, and a structured questionnaire was administered by interviewers to collect information on demographic data and environmental exposure history including tobacco smoking. Those who had smoked less than one cigarette per day and <1 year in their lifetime were defined as non-smokers, otherwise they were considered smokers. Those smokers who had quit for >1 year were considered former smokers. Pack-years smoked [(cigarettes per day/20)x years smoked] were calculated to indicate the cumulative smoking dose. Family history of cancer was defined as any self-reported cancer in first-degree relatives (parents, siblings or children). After the interview, an ~5 ml venous blood sample was collected from each participant. The study was approved by the institutional review boards of Fudan University.

Selection of SNPs of PPAR-{gamma}
The human PPAR-{gamma} gene is ~146 kb in size, located on chromosome 3p25. Differential promoter use and alternative splicing of the gene generate three isoforms: PPAR-{gamma}1 and PPAR-{gamma}3 protein encoding the same protein product and the PPAR-{gamma}2 protein containing an additional N-terminal 28 amino acid exon named exon B (27,28). PPAR-{gamma}1 is expressed in a broad spectrum of tissues, whereas PPAR-{gamma}2 is restricted to adipose tissue and PPAR-{gamma}3 is abundant in the macrophages, large intestine and white adipose tissue (16). In the gene model (mRNA alignment) of PPAR-{gamma}1 (NM_005037 [GenBank] .5), it has seven exons and six introns with >650 SNPs located from 2 kb upstream to 2 kb downstream as listed in the dbSNP database with a density of ~1 SNP per 200–300 bp. Because this study was initiated in January 2005 when the SNP density of phase I HapMap SNP database was not adequate, we chose SNPs from both the HapMap and the dbSNP databases. An algorithm to score SNPs across the gene was developed and a set of SNPs were selected based on their final scores. The procedure was described in detail in the supplementary material available at Carcinogenesis online. In brief, for a certain SNP, the following criteria were considered: (i) interdistance between two adjacent SNPs; (ii) heterozygosity; (iii) functional relevance and (iv) compatibility with the genotyping platform. As a result, 11 SNPs were chosen for genotyping in this study.

Genotyping assays
The 11 SNPs were genotyped by using the Illumina SNP genotyping BeadLab platform, which is a highly efficient genotyping assay with a combination of the Illumina Golden GateTM assay, SentrixTM array matrices and SherlockTM scanner technology (Illumina Corp., Foster City, CA), performed at the Chinese National Human Genome Center at Shanghai. The information on assay conditions and the primers and probes is available upon request. More detailed description of quality control method and each of the steps performed by the Illumina facility is available in our previously published paper (26) and the HapMap Web site (http://www.hapmap.org/downloads/protocols_overview.html).

Statistical analyses
Differences in selected demographic variables, smoking status, pack-years of smoking, family history of cancer and frequencies of the PPAR-{gamma} alleles and genotypes between the cases and controls were evaluated by using the {chi}2-test. Goodness of fit to the Hardy–Weinberg equilibrium expectation in control was also assessed by the {chi}2-test for each SNP. Akaike's information criteria were used to select the most parsimonious genetic model for each SNP (29). Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by unconditional logistic regression analysis with adjustment for age, sex, pack-years of smoking and family history of cancer. Stratification analyses were also performed by variables of interest, such as age, sex, smoking status, family history of cancer and histologic types.

The pairwise linkage disequilibrium (LD) among the SNPs was examined using Lewontin's standardized coefficient D' and LD coefficient r2 (30), and haplotype blocks were defined by the method of Gabriel et al. (31) in the publicly available Haploview software (http://www.broad.mit.edu/personal/jcbarret/haplo/) with default settings (the CI for a strong LD was minimal for upper 0.98 and low 0.7 and maximal for a strong recombination of 0.9, and a fraction of strong LD in informative comparisons was at least 0.95). Each common haplotype [minor allele frequency (MAF) ≥0.03] was compared between all cases and the controls and in each stratum of cumulative smoking dose to see whether smoking influenced the risk associated with PPAR-{gamma} variants by using Haplo.stats (http://www.mayo.edu/stagen). In addition, PHASE 2.1 Bayesian algorithm (32) was used to validate the haplotype frequencies estimated by Haplo.stats and infer diplotype frequencies based on the observed genotypes. Diplotype (haplotype dosage, an estimate of the number of copies of the haplotype) was the most probable haplotype pair for each individual. Unconditional logistic regression analyses were conducted to estimate ORs and 95% CIs for participants carrying one to two copies versus zero copy of each common haplotype for the dichotomized diplotypes. The issue of multiple tests was controlled by using 10 000 time permutation tests.

To explore the potential interaction between the SNPs and cigarette smoking, we used multiple approaches to evaluate consistency of the results, including analyses in specific categories of cumulative smoking exposure (i.e. pack-years), genotype–smoking joint-effects and interaction models that considered both discrete [non-smokers, light smokers (≤30 pack-years) and heavy smokers (>30pack-years)] and continuous (square root of pack-years) variables for cumulative smoking exposure. Statistical analyses were all performed using the SPSS15.0 software, unless indicated otherwise.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Supplementary material
 Funding
 References
 
Characteristics of the study population
The characteristics of the 500 lung cancer patients and 517 cancer-free controls have been described elsewhere (26). Overall, the lung cancer cases and controls appeared to be adequately matched on age and sex (P = 0.661 and 1.000, respectively). More smokers, however, were present among the cases than among the controls (77.0% versus 51.8%; P < 0.0001). Specifically, cigarette smoking was associated with increased risk of lung cancer among light smokers (OR = 1.39 and 95% CI = 1.04–1.87) and heavy smokers (OR = 2.75 and 95% CI = 2.01–3.78) (data not shown). Furthermore, lung cancer cases were significantly more likely than the controls to report a family history of cancer in their first-degree relatives (24.4% versus 16.8%; P = 0.003), which accounted for a significantly 60% increased lung cancer risk (OR = 1.60 and 95% CI = 1.17–2.17) (data not shown). Among the 500 lung cancer cases, 466 (93.2%) were classified as non-small-cell lung cancer (229 adenocarcinoma, 141 squamous cell carcinoma and 96 large-cell, mixed-cell or undifferentiated carcinoma) and only 34 (6.8%) as small-cell lung cancer.

Association between individual SNP and risk of lung cancer
As shown in Table I, genotype frequency distributions of 11 SNPs in the controls were all consistent with those expected from the Hardy–Weinberg equilibrium model (all P > 0.05). One SNP (rs1175541) in this Chinese population represented an MAF of 6.2% lower than that reported in the dbSNP database, whereas the other SNPs represented an MAF at least 4.1% higher than those reported in the HapMap SNP database, which may reflect either population difference or frequency bias due to small sample sizes from which the databases derived. Allele frequencies of three SNPs were shown significant differences between the cases and the controls (P = 0.0004 for rs13073869, P = 0.0106 for rs1899951 and P = 0.0270 for rs4135247).


View this table:
[in this window]
[in a new window]

  		Table I. Information ~11 genotyped SNPs of PPAR-{gamma} gene

 
Significant associations were observed for two SNPs (P = 0.0004 for rs13073869 and P = 0.0130 for rs1899951) in a dominant model, one SNP (P = 0.0310 for rs4135247) in a log-additive model and four SNPs (P = 0.0468 for rs2972162, P = 0.0175 for rs709151, P = 0.0172 for rs11715541 and P = 0.0386 for rs1175543) in an overdominant model, based on the best fit of the Akaike's information criteria. The difference for two SNPs (rs13073869 and rs1899951) remained significant after applying 10 000 time permutation tests (P value from empirical distribution of minimal P values = 0.0130).

Multivariate logistic regression analyses revealed that after adjusting for confounding factors, compared with wild-type carriers in a dominant model, a significantly decreased lung cancer risk was associated with the variant genotypes of rs13073869 G/A (adjusted OR = 0.65 and 95% CI = 0.51–0.85 for GA/AA genotypes) and rs1899951 G/A (adjusted OR = 0.58 and 95% CI = 0.37–0.90 for GA/AA genotypes). Moreover, compared with the combined homozygous genotypes in an overdominant model, a significantly reduced lung cancer risk was associated with the heterozygous genotypes of rs2972162 CT (adjusted OR = 0.78 and 95% CI = 0.60–1.00), rs709151 GA (adjusted OR = 0.73 and 95% CI = 0.57–0.94), rs1175541CA (adjusted OR = 0.73 and 95% CI = 0.57–0.94) and rs1175543AG (adjusted OR = 0.76 and 95% CI = 0.59–0.98) (Table II).


View this table:
[in this window]
[in a new window]

  		Table II. Genotype frequencies of selected SNPs of PPAR-{gamma} among cases and controls and their association with lung cancer risk

 
We further evaluated the associations of the rs13073869 GA/AA and rs1899951 GA/AA variant genotypes with lung cancer risk stratified by selected variables and histological types. As shown in Table III, compared with the common wild-type homozygous genotype, the protective effect of rs13073869 GA/AA was more evident in former smokers (adjusted OR = 0.43 and 95% CI = 0.21–0.88), whereas the protective effects in other subgroups were mostly diminished due to reduced sample sizes. Interestingly, the decreased risk associated with the rs1899951 GA/AA variant genotypes was more pronounced in young individuals (adjusted OR = 0.38 and 95% CI = 0.20–0.74), female subjects (adjusted OR = 0.29 and 95% CI = 0.09–0.92), non-smokers (adjusted OR = 0.28 and 95% CI = 0.12–0.66) and those with a family history of cancer (adjusted OR = 0.31 and 95% CI = 0.11–0.89).


View this table:
[in this window]
[in a new window]

  		Table III. Stratified analyses between the PPAR-{gamma} rs13073869 and rs1899951 genotypes and lung cancer risk

 
Association between haplotypes/diplotypes and risk of lung cancer
The reconstructed LD plot of 11 SNPs in 517 controls is shown in Figure 1. Two blocks were defined by 10 SNPs and the remaining one SNP located in the downstream of block 2. Global score test showed statistically significant differences in haplotype frequency distribution between the cases and controls for block 1 (global statistics = 19.5900, df = 4, P value = 0.0039, P sim = 0.0027) but not for block 2.


Figure 1
View larger version (58K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Graphical representation of the SNP locations and LD structure of PPAR-{gamma} using 11 genotyped SNPs in 517 southeast Han Chinese controls. The exact SNP positions are listed in Table I. Two haplotype blocks (colored) were defined by the Haploview program using the approach given in Gabriel et al. (31) with default settings (CI minima for strong LD: upper 0.98, low 0.7; upper CI maximum for strong recombination, 0.9; fraction of strong LD in informative comparisons must be at least 0.95). The rs number (top; from left to right) corresponds to the SNP name and the numbers in squares are D' values (|D'| x 100). The measure of LD (D') among all possible pairs of SNPs is shown graphically according to the shade of red where white represents very low D' and dark red represents very high D'.

 
The logistic regression analyses revealed that the risk of lung cancer was significantly decreased among individuals carrying the haplotype ‘AGA’ (adjusted OR = 0.80 and 95% CI = 0.65–0.98) and ‘AAA’ (adjusted OR = 0.57 and 95% CI = 0.37–0.87), compared with those carrying the most common haplotype ‘GGG’ in block 1 (Table IV). Notably, the ‘AGA’ haplotype harbored the rs13073869 A allele and ‘AAA’ harbored the rs1899951 A allele as well as the forgoing one, and these two alleles were both associated with a significant decreased risk of lung cancer in the single-locus analysis. Moreover, the stratified analyses showed that the risk of lung cancer was further reduced among non-smokers carrying the haplotype ‘AAA’ (adjusted OR = 0.27 and 95% CI = 0.12–0.63), but increased among heavy smokers carrying the haplotype ‘GGA’ (adjusted OR = 1.83 and 95% CI = 1.01–3.32). The adjusted ORs of the ‘GGA’ versus ‘GGG’ and ‘AAA’ versus ‘GGG’ increased significantly as pack-years increased, suggesting a possible interaction between ‘GGA’ and smoking, and between ‘AAA’ and smoking.


View this table:
[in this window]
[in a new window]

  		Table IV. Association between PPAR-{gamma} common haplotypes in each block and lung cancer risk in overall population and subpopulation stratified by pack-years of smoking

 
We also used the haplotypes within each block to build dichotomized diplotypes (0 and one to two copies of the haplotype). Consistent with the haplotype analyses, subjects carrying one to two copies of the haplotype ‘AGA’ had a 28% reduced lung cancer risk (adjusted OR = 0.72 and 95% CI = 0.56–0.93), and those carrying one to two copies of ‘AAA’ had a 42% reduced lung cancer risk (adjusted OR = 0.58 and 95% CI = 0.37–0.90) compared with their respective non-carriers (supplementary Table II available at Carcinogenesis online).

Gene–smoking interaction analysis
As shown in Table V, we first classified cumulative smoking dose into discrete variable (non-smokers, light smokers and heavy smokers) to avoid the issue of potential misclassification of subjects with respect to smoking exposure. The adjusted ORs of the rs1899951 GA/AA versus GG genotypes increased significantly as pack-years increased in both specific categories of cumulative smoking exposure analysis and genotype–smoking joint-effects analysis, although the analyses for light smokers did not reach statistical significance. Interestingly, the variant genotypes GA/AA conferred a decreased risk of lung cancer among non-smokers but an increased risk among heavy smokers, with the adjusted OR increasing from 0.28 (95% CI = 0.12–0.66) to 3.45 (95% CI = 1.43–8.58) compared with the wild-type genotype GG. When we defined non-smokers with GA/AA as the reference group, the heavy smokers with the same genotypes had the greatest risk for lung cancer (OR = 17.09 and 95% CI = 5.03–58.09), representing a significant multiplicative interaction between the rs1899951 polymorphism (GG versus GA/AA) and trichotomized cumulative smoking dose with P = 0.009. Consistent and robust result was also found when considering smoking as continuous cumulative smoking dose (square root of pack-years) with P = 0.016. However, smoking did not modify the effects of other polymorphisms of PPAR-{gamma} when considering either discrete or continuous cumulative smoking dose (P > 0.05).


View this table:
[in this window]
[in a new window]

  		Table V. Interaction analyses of rs1899951 genotypes and cumulative smoking dose

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Supplementary material
 Funding
 References
 
In this lung cancer case–control study in a southeastern Chinese population, we investigated the role of multiple common variants of PPAR-{gamma} and their interaction with cigarette exposure on the risk of lung cancer. We found that seven SNPs (rs13073869, rs1899951, rs4135247, rs2972162, rs709151, rs1175541 and rs1175543) of the 11 selected SNPs showed a significant association with lung cancer risk. Besides, both the haplotype and diplotype analyses consistently revealed a protective effect of the haplotype ‘AGA’ and ‘AAA’ derived from rs13073869, rs1899951 and rs4135247. Moreover, we observed consistent evidence for a statistically significant interaction between the rs1899951 and cigarette smoking measured as either discrete or continuous variable. These findings support our hypothesis that PPAR-{gamma} polymorphisms and their interaction with smoking may contribute to the etiology of lung cancer. This is, to the best of our knowledge, the first study to assess the association between a broad spectrum of genetic variants individually and collectively as haplotypes of the PPAR-{gamma} gene and lung cancer risk.

The most outstanding finding in this study was the consistency in genotypic, haplotypic and diplotypic associations of rs13073869 and rs1899951 with lung cancer risk. In the single-locus association analysis, variant genotypes of these two SNPs exhibited a statistically significant reduced risk of lung cancer individually even after 10 000 time permutation tests. The protective haplotype ‘AGA’ harbored the rs13073869 A allele, whereas ‘AAA’ harbored the rs1899951 A allele as well as the rs13073869 A allele. Furthermore, ‘AAA’ accounted for a 73% reduction in lung cancer risk among non-smokers, which was consistent with the effect of variant genotypes of rs1899951 among the same subgroup, suggesting that the protective effect of ‘AGA’ was probably driven by the rs13073869 A allele and that of ‘AAA’ was probably driven by the rs1899951 A allele. Both of these two SNPs were located in the haplotype block 1, which showed a significant reverse association with lung cancer risk. Of note, the block 1 corresponds to the region of intron 1. Accumulative evidence has indicated that the sequence in intron 1 of human genes plays an important role in transcriptional regulation, and some genetic variants in this region could affect the level or timing of gene expression by altering intronic enhancer binding sites (33–36).

It is biologically plausible that the region of intron 1 of PPAR-{gamma} may be involved in the etiology of lung cancer. PPAR-{gamma} expression has been suggested as a potential marker for progression of lung cancer, and the expression levels of PPAR-{gamma} protein appear to correlate with maturational stage, differentiated phenotype and tumor histological type and grade in lung adenocarcinoma (19,37). For example, decreased PPAR-{gamma} expression has been correlated with poor prognosis in patients with lung cancer, suggesting that the gene expression may be further diminished as lung cancer progresses (37). Although functional relevance of rs13073869 and rs1899951 is unknown, it is possible that they may increase the affinity of transcription activators or decrease that of transcription suppressors to the intronic enhancer, thus upregulating the expression levels of PPAR-{gamma}. However, we have no experimental evidence to think that minor alleles of these two SNPs, rather than any of the other alleles shared uniquely by the two haplotypes, are the causal alleles. The question still remains as to whether they are indeed etiologic or simply markers in LD with other untyped functional loci. For example, we found that rs13073869 polymorphism was in perfect LD (D' = 1.00, r2 = 1.00) with a C/G substitution (rs10865710) in the PPAR-{gamma}3 regulatory region at position –681 in PPAR-{gamma}3-specific exon A2 (the A allele being associated with the –681 G allele) according to the HapMap data on 45 Han Chinese in Beijing, northern China (genotype data of PPAR-{gamma}, released in phase II of the International HapMap Project). In vitro, the –681 G allele completely abolished the binding of STAT5B to the cognate promoter element as well as the transactivation of PPAR-{gamma}3 by the growth hormone/STAT5B pathway (38). Furthermore, rs1899951 is in perfect LD (D' = 1.00, r2 = 1.00) with the non-synonymous polymorphism Pro12Ala (rs1801282) in PPAR-{gamma}2-specific exon B (the A allele being associated with the Ala12 allele). This amino acid is located in a PPAR-{gamma} domain that enhances ligand-independent activation, and the Pro-to-Ala exchange may cause a conformational change in the protein, thus affecting its activity (39). However, because PPAR-{gamma}2 exists exclusively in adipose tissue (16), the Pro12Ala SNP may not play an important role in the etiology of lung cancer. The exact location and biological functions of the real causal SNPs of the gene are of great interest and warrant further investigation.

Our study also provided some new evidence that the effect of the rs1899951 appeared to be strongly modified by the cumulative cigarette smoking. Interestingly, the variant genotypes were protective factors among non-smokers but risk factors among heavy smokers compared with their respective wild-type GG genotype. For example, heavy smoking (>30 pack-years) alone only conferred a 2.75-fold increased risk of lung cancer, but in the presence of the variant GA/AA genotypes, the effect of heavy smoking was more than six times stronger under a simple multiplicative model, indicating a risk-enhancing relationship between smoking and the variant genotypes of rs1899951. The underlying mechanism involved in the interaction of PPAR-{gamma} and smoking is not clear. It is worth noting that the expression of PPAR-{gamma} can be upregulated by interleukin-4 in lung epithelial cells and macrophages, a proinflammatory cytokine whose release can be stimulated by tobacco smoke (40,41). In this way, it is probably that the expression level of PPAR-{gamma} can be induced significantly as well by smoking through this possible mechanism. Thus, in the normal situation, it is possible that the variant allele A of rs1899951 leads to a higher level of basic expression of PPAR-{gamma} than that of the wild-type G. As PPAR-{gamma} acts as a tumor suppressor for lung cancer, the variant allele A provides a greater protective effect than that of the wild-type allele among the non-smokers. Although the cells are stimulated by environmental factors such as cigarette smoking, inducible expression system is triggered in which the level of inducible expression is much higher than that of basic expression. Therefore, the subjects carrying the rs1899951 variant allele A may have a lower risk in developing lung cancer under healthy conditions but a higher risk when the A allele is in strong LD with a variant allele of another gene (such as interleukin-4) that may induce expression of PPAR-{gamma} in response to heavy smoking. Nevertheless, such speculation needs further support from additional research and functional studies.

In the single-locus analysis, we found that the protective effect appeared to be significant in the heterozygotes but not in homozygotes. Although current knowledge does not provide a convincing explanation for such a finding, there are several possibilities that may lead to the seemingly peculiar finding. Technically, although such a bias toward the heterozygotes could result from a systematic error, particularly in the genotyping using restriction enzymes, this is unlikely to occur in the Illumina assay for genotyping using variant-specific probes. In a strictly genetic sense without any selection bias, such data apparently well fit an overdominant model, in which heterozygous genotype confers a higher fitness compared with the corresponding homozygous genotypes. Molecularly, this is consistent with such a hypothesis—that is, if PPAR-{gamma} performs its function within an optimal expression range as determined by the heterozygosity, the function may be weakened when the expression level is above or below a normal threshold because of homozygotes. However, this hypothesis needs to be further tested in in-depth molecular mechanistic studies in the future. It is also possible that the variant allele may have a very strong dominant effect so that there is little difference between the effects of the variant homozygotes and heterozygotes. Statistically, because of the relatively small numbers of the variant homozygotes observed in both cases and controls, the effect of the variant homozygotes might more likely be subject to any selection bias or other unfavorable genotypes than that of the heterozygotes that often present in a much larger number of observations, or simply there is no enough statistical power to detect any real effect among the variant homozygotes, which can only be corrected by much larger studies in the future (42).

There are three main strengths of this study. First, several previous studies have assessed single variants in PPAR-{gamma} for associations with diseases, but mostly focusing on only one or two variants in the coding region or promoter region and none of the polymorphisms genotyped in our study has been investigated in reported association studies. As lung cancer is a multifactorial disease probably involving multiple SNPs in multiple genes, we evaluated a broader spectrum of PPAR-{gamma} variants individually as alleles and collectively as haplotypes, which may be more powerful than that of analyzing a single allele or locus. Second, all diagnoses of lung cancer were confirmed histologically, and complete smoking data were collected systematically. The adjusted ORs in both stratified and joint-effect analyses for different pack-year categories of smoking were similar in magnitude and direction to the point estimates obtained from fitted ORs of the interaction models. The consistency of these results suggests that our findings are unlikely due to chances. Finally, an investigation of a candidate gene needs many SNPs for individual association analysis (43,44). However, unless the selected SNPs are all in complete LD with each other, such a multiple testing will increase the false-positive (type I error) rate under nominal significance thresholds (e.g. {alpha} = 0.05). On the other hand, when background LD exists between SNPs but they are assumed to be completely independent, then the popular Bonferroni correction would cause overcorrection for the inflated false-positive rate, resulting in a reduction in study power (45). For calculating the significance of SNPs in LD with each other, a permutation test has been suggested to adjust for multiple testing while preserving the correlation structure among linked markers (46–49). By using this method, the false-positive rate for a large number of tests was satisfactorily controlled in our study.

Despite the strengths and biologic plausibility of the associations observed in our study, inherited biases in the present study may have led to spurious findings. First of all, the lung cancer cases were enrolled from the hospitals and the controls were selected from the surrounding communities; inherent selection bias cannot be completely excluded. However, by matching the controls to the cases on age, sex and residential area (urban or rural), the potential confounding factors may be minimized. Second, the sample size of our study may not be large enough either to detect a small effect from very low-penetrance SNPs or to identify significant associations of the effect in different strata in subgroup analysis adequately. Third, except for cigarette smoking, the information on other factors such as occupational exposure and certain dietary components, which might interact with PPAR-{gamma} genotypes or act as potential confounding factors, was not available in our study. Possible interactions between PPAR-{gamma} genotypes and these risk factors should be thoroughly investigated in future studies. Finally, since the phase II HapMap was not accomplished when we selected SNPs for genotyping, our genotyping efforts were limited because we chose SNPs from both HapMap public SNP database of phase I and dbSNPs database available to us and the 11 SNPs in PPAR-{gamma} we genotyped were not enough to capture or represent all genetic variants of this gene based on our current knowledge. In the newly released data in phase II of the International HapMap Project (HapMap Data Rel 22/phase II Apr 07), 259 SNPs (rs1175541 not included) of PPAR-{gamma} are genotyped in 45 Han Chinese in Beijing, northern China, accounting for a density of one SNP per 560 bp. Among these SNPs, 73 SNPs have an MAF ≥5%, with at least a 90% genotyping rate and a P value of the Hardy–Weinberg equilibrium test ≥0.05. However, the 11 SNPs we selected are able to capture 100% of these informative alleles with a mean r2 of 0.809 and 74% of captured alleles have pairwise LD with r2 >0.8.

In conclusion, our study provided evidence for the first time that the PPAR-{gamma} polymorphisms and their interactions with cigarette smoking may contribute to the etiology of lung cancer in a Chinese population. Furthermore, we also showed that a genetic susceptibility, coupled with a modifiable lifestyle factor (such as cigarette smoking), appeared to have conferred a significantly higher risk of lung cancer than either factor alone. However, these findings need to be substantiated by larger studies with diverse ethnic populations.


   Supplementary material
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
Funding
 References
 
Supplementary material can be found at http://carcin.oxfordjournals.org/.


   Funding
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 Funding
References
 
China National Key Basic Research Program (2002CB512902 to D.L. and H.S., 2006AA020706, 2004CB518605, 06XD14015 and 05DZ22201 to W.H.); National Outstanding Youth Science Foundation of China (30425001 to H.S., 30625019 to W.H.); National ‘211’ Environmental Genomics to D.L.


   Footnotes
 
{dagger} The first two authors contributed equally to this work. Back


   Acknowledgments
 
The authors would like to thank Shuhua Xu, Zhengwen Jiang, Yi Wang, Minhua Shao, Rui Li and Wenqing Fu for their helpful comments and discussion.

Conflict of Interest Statement: None declared.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 Funding
 References
 

  1. Parkin DM, et al. Global cancer statistics, 2002. CA Cancer J. Clin. (2005) 55:74–108.[Abstract/Free Full Text]
  2. American Cancer Society, I. In: Cancer facts and figures 2003 (2003) Atlanta, GA: American Cancer Society.
  3. Tsou JA, et al. DNA methylation analysis: a powerful new tool for lung cancer diagnosis. Oncogene (2002) 21:5450–5461.[CrossRef][Web of Science][Medline]
  4. Nadel JA. Role of epidermal growth factor receptor activation in regulating mucin synthesis. Respir. Res. (2001) 2:85–89.[CrossRef][Web of Science][Medline]
  5. Adler KB, et al. Interactions between respiratory epithelial cells and cytokines: relationships to lung inflammation. Ann. NY Acad. Sci. (1994) 725:128–145.[Web of Science][Medline]
  6. Takeyama K, et al. Activation of epidermal growth factor receptors is responsible for mucin synthesis induced by cigarette smoke. Am. J. Physiol. Lung Cell. Mol. Physiol. (2001) 280:L165–L172.[Abstract/Free Full Text]
  7. Jones R, et al. Goblet cell glycoprotein and tracheal gland hypertrophy in rat airways: the effect of tobacco smoke with or without the anti-inflammatory agent phenylmethyloxadiazole. Br. J. Exp. Pathol. (1973) 54:229–239.[Web of Science][Medline]
  8. Balkwill F, et al. Inflammation and cancer: back to Virchow? Lancet (2001) 357:539–545.[CrossRef][Web of Science][Medline]
  9. Fitzpatrick FA. Inflammation, carcinogenesis and cancer. Int. Immunopharmacol. (2001) 1:1651–1667.[CrossRef][Web of Science][Medline]
  10. Godschalk R, et al. Comparison of multiple DNA adduct types in tumor adjacent human lung tissue: effect of cigarette smoking. Carcinogenesis (2002) 23:2081–2086.[Abstract/Free Full Text]
  11. Ames BN, et al. The causes and prevention of cancer. Proc. Natl Acad. Sci. USA (1995) 92:5258–5265.[Abstract/Free Full Text]
  12. Boffett P, et al. Lung cancer risk in a population-based cohort of patients hospitalized for asthma in Sweden. Eur. Respir. J. (2002) 19:127–133.[Abstract/Free Full Text]
  13. Sasco AJ, et al. A case-control study of lung cancer in Casablanca, Morocco. Cancer Causes Control (2002) 13:609–616.[CrossRef][Web of Science][Medline]
  14. Cohen BH, et al. A common familial component in lung cancer and chronic obstructive pulmonary disease. Lancet (1977) 2:523–526.[Web of Science][Medline]
  15. Standiford TJ, et al. Peroxisome proliferator-activated receptor-{gamma} as a regulator of lung inflammation and repair. Proc. Am. Thorac. Soc. (2005) 2:226–231.[Abstract/Free Full Text]
  16. Kota BP, et al. An overview on biological mechanisms of PPARs. Pharmacol. Res. (2005) 51:85–94.[CrossRef][Web of Science][Medline]
  17. Debril MB, et al. The pleiotropic functions of peroxisome proliferator-activated receptor gamma. J. Mol. Med. (2001) 79:30–47.[CrossRef][Web of Science][Medline]
  18. Kersten S, et al. Roles of PPARs in health and disease. Nature (2000) 405:421–424.[CrossRef][Medline]
  19. Keshamouni VG, et al. Peroxisome proliferator-activated receptor-gamma activation inhibits tumor progression in non-small-cell lung cancer. Oncogene (2004) 23:100–108.[CrossRef][Web of Science][Medline]
  20. Inoue K, et al. Expression of peroxisome proliferator-activated receptor (PPAR)-gamma in human lung cancer. Anticancer Res. (2001) 21:2471–2476.[Web of Science][Medline]
  21. Tsubouchi Y, et al. Inhibition of human lung cancer cell growth by the peroxisome proliferator-activated receptor-gamma agonists through induction of apoptosis. Biochem. Biophys. Res. Commun. (2000) 270:400–405.[CrossRef][Web of Science][Medline]
  22. Li M, et al. Activation of peroxisome proliferator-activated receptor-gamma by troglitazone (TGZ) inhibits human lung cell growth. J. Cell. Biochem. (2005) 96:760–774.[CrossRef][Web of Science][Medline]
  23. Bren-Mattison Y, et al. Peroxisome proliferator-activated receptor-gamma (PPAR(gamma)) inhibits tumorigenesis by reversing the undifferentiated phenotype of metastatic non-small-cell lung cancer cells (NSCLC). Oncogene (2005) 24:1412–1422.[CrossRef][Medline]
  24. Han S, et al. Suppression of prostaglandin E2 receptor subtype EP2 by PPARgamma ligands inhibits human lung carcinoma cell growth. Biochem. Biophys. Res. Commun. (2004) 314:1093–1099.[CrossRef][Web of Science][Medline]
  25. Satoh T, et al. Activation of peroxisome proliferator-activated receptor-gamma stimulates the growth arrest and DNA-damage inducible 153 gene in non-small cell lung carcinoma cells. Oncogene (2002) 21:2171–2180.[CrossRef][Web of Science][Medline]
  26. Hu Z, et al. Genetic variants in MGMT and risk of lung cancer in Southeastern Chinese: a haplotype-based analysis. Hum. Mutat. (2007) 28:431–440.[CrossRef][Medline]
  27. Fajas L, et al. The organization, promoter analysis, and expression of the human PPARgamma gene. J. Biol. Chem. (1997) 272:18779–18789.[Abstract/Free Full Text]
  28. Fajas L, et al. PPARgamma3 mRNA: a distinct PPARgamma mRNA subtype transcribed from an independent promoter. FEBS Lett. (1998) 438:55–60.[CrossRef][Web of Science][Medline]
  29. Akaike H. A new look at the statistical model identification. IEEE Trans. Automat. Contr (1974) 19:716–723.[CrossRef]
  30. Lewontin RC. On measures of gametic disequilibrium. Genetics (1988) 120:849–852.[Abstract/Free Full Text]
  31. Gabriel SB, et al. The structure of haplotype blocks in the human genome. Science (2002) 296:2225–2229.[Abstract/Free Full Text]
  32. Stephens M, et al. A comparison of Bayesian methods for haplotype reconstruction from population genotype data. Am. J. Hum. Genet. (2003) 73:1162–1169.[CrossRef][Web of Science][Medline]
  33. Abbott W, et al. Polymorphism in intron 1 of the interferon-gamma gene influences both serum immunoglobulin E levels and the risk for chronic hepatitis B virus infection in Polynesians. Immunogenetics (2007) 59:187–195.[CrossRef][Medline]
  34. Zhou L, et al. Intron 1 sequences are required for pancreatic expression of the human proglucagon gene. Am. J. Physiol. Regul. Integr. Comp. Physiol. (2006) 290:R634–R641.[Abstract/Free Full Text]
  35. Kawada N, et al. Role of intron 1 in smooth muscle alpha-actin transcriptional regulation in activated mesangial cells in vivo. Kidney Int. (1999) 55:2338–2348.[CrossRef][Web of Science][Medline]
  36. Estany J, et al. Association of a CA repeat polymorphism at intron 1 of the IGF1 gene with circulating insulin-like growth factor 1 concentration, growth and fatness in swine. Physiol. Genomics (2007) 31:236–243.[Abstract/Free Full Text]
  37. Sasaki H, et al. Decreased perioxisome proliferator-activated receptor gamma gene expression was correlated with poor prognosis in patients with lung cancer. Lung Cancer (2002) 36:71–76.[CrossRef][Web of Science][Medline]
  38. Meirhaeghe A, et al. A functional polymorphism in a STAT5B site of the human PPAR{gamma}3 gene promoter affects height and lipid metabolism in a French population. Arterioscler. Thromb. Vasc. Biol. (2003) 23:289–294.[Abstract/Free Full Text]
  39. Stumvoll M, et al. The peroxisome proliferator-activated receptor-gamma2 Pro12Ala polymorphism. Diabetes (2002) 51:2341–2347.[Abstract/Free Full Text]
  40. Huang JT, et al. Interleukin-4-dependent production of PPAR-gamma ligands in macrophages by 12/15-lipoxygenase. Nature (1999) 400:378–382.[CrossRef][Medline]
  41. Wang AC, et al. Peroxisome proliferator-activated receptor-gamma regulates airway epithelial cell activation. Am. J. Respir. Cell Mol Biol. (2001) 24:688–693.[Abstract/Free Full Text]
  42. Ma H, et al. Tagging single nucleotide polymorphisms in excision repair cross-complementing group 1 (ERCC1) and risk of primary lung cancer in a Chinese population. Pharmacogenet. Genomics (2007) 17:417–423.[Medline]
  43. Ohashi J, et al. The power of genome-wide association studies of complex disease genes: statistical limitations of indirect approaches using SNP markers. J. Hum. Genet. (2001) 46:478–482.[CrossRef][Web of Science][Medline]
  44. Pritchard JK, et al. The allelic architecture of human disease genes: common disease-common variant...or not? Hum. Mol. Genet. (2002) 11:2417–2423.[Abstract/Free Full Text]
  45. Nyholt DR. A simple correction for multiple testing for single-nucleotide polymorphisms in linkage disequilibrium with each other. Am. J. Hum. Genet. (2004) 74:765–769.[CrossRef][Web of Science][Medline]
  46. Churchill GA, et al. Empirical threshold values for quantitative trait mapping. Genetics (1994) 138:963–971.[Abstract]
  47. Dudbridge F, et al. Detecting multiple associations in genome-wide studies. Hum. Genomics. (2006) 2:310–317.[Medline]
  48. Dudbridge F, et al. Efficient computation of significance levels for multiple associations in large studies of correlated data, including genomewide association studies. Am. J. Hum. Genet. (2004) 75:424–435.[CrossRef][Web of Science][Medline]
  49. Dudbridge F. A note on permutation tests in multistage association scans. Am. J. Hum. Genet. (2006) 78:1094–1095.[CrossRef][Medline]
Received September 8, 2007; revised November 14, 2007; accepted December 1, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Vet PatholHome page
A. K. Bauer and E. A. Rondini
REVIEW PAPER: The Role of Inflammation in Mouse Pulmonary Neoplasia
Vet. Pathol., May 1, 2009; 46(3): 369 - 390.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
29/2/342    most recent
bgm285v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, D.
Right arrow Articles by Lu, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chen, D.
Right arrow Articles by Lu, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?