Human proto-oncogene c-Jun and c-Fos assemble the activator protein-1 complex which is a crucial transcription factor responding to environmental factors and promotes tumorgenesis. We hypothesized that genetic variants in these two genes may alter the carriers' susceptibility to lung cancer. In two independent case–control studies, we genotyped three putative functional polymorphisms (−1318T>G and −673T>C of c-Jun; −60C>T of c-Fos) in southern Chinese and then validated the association in eastern Chinese. We found that compared to −1318TT genotype, the −1318GT/GG variant genotypes had an increased lung cancer risk (OR = 1.46, 95% CI = 1.26–1.69), and the −673CC genotype had an increased lung cancer risk compared to −673TT/CT genotypes (OR = 1.35, 95% CI = 1.17–1.56) in the total 1,559 cases versus 1,679 controls. After combining these two loci, the number of the risk genotypes was associated with increased cancer risk in a dose-response manner (ptrend = 2.21 × 10−11); moreover, the risk genotypes interacted with smoking or drinking status on increasing cancer risk (p values of interaction were 0.009 and 0.007, respectively). Further, we found that those with −1318GT/GG genotypes, −673CC genotypes or both genotypes in c-Jun had higher mRNA and protein expression levels in vivo, and those variants had higher transcription activities in reporter genes in vitro, especially under the stimuli with tobacco extract or alcohol mixture as luciferase assay shown. However, for −60C>T of c-Fos, no significant association was observed for lung cancer risk. Our data suggested that the genetic variants in c-Jun (−1318T>G and −673T>C) increase the carriers' susceptibility to lung cancer via interaction with smoking or drinking on increasing the c-Jun's expression.
Lung cancer is one of the most common causes of mortality in China and worldwide with increasing incidence in the past three decades. It was reported that in China, the deaths due to lung cancer were 493,348 in 2008,1 and in the United States, the deaths were 160,390 in 2010.2 The pathological mechanism of lung cancer is extremely complex involving the stimulus of environmental factors and the effects of internal genetic factors. More and more studies have already recognized that the smoking, air pollution, occupational exposures (i.e., diesel fumes, asbestos, metals, oil mist and welding fumes) and their possible gene–environmental interactions play major roles in cancer development.3–6
Several cellular signaling pathways are involved in the transduction of carcinogenic signals in response to environmental risk factors, which finally activate many nuclear transcription factors and regulate the expressions of related effecter molecules7, 8 and in turn regulate cellular responses and induce tumorgenesis.9–11 The activator protein-1 (AP-1) is such a key signaling transduction molecule and nuclear transcription factor that regulates critical cellular processes, including cell proliferation, differentiation, apoptosis and tumorgenesis.12, 13 As reported, when activated by mitogen-activated protein kinase, wingless-int (WNT) or other pathways,13, 14 the AP-1 will bind to the AP-1 DNA-binding element, like tetradecanoylphorbol 13-acetate (TPA)-responsive elements (TGAC/CTCA)15 in the promoter of many tumor-related genes, such as Ras,16Fas,17Bim,18Bcl-3,19MMP-1 and MMP-3,20Jagged1,21MTS1,22 and increase their expressions,12 which would ultimately affect tumorgenesis.12, 23 Furthermore, overexpression of AP-1 observed in many cancers including lung caner reflected carcinogenic role of AP-1.24–30
The AP-1 is a heterodimeric or homodimeric complex that comprises members of the proto-oncogene c-Jun family proteins (c-Jun, JunB and JunD) and c-Fos family proteins (c-Fos, FosB, Fra-1 and Fra-2).23 In mammalians, AP-1 is the heterodimer comprising c-Jun/c-Fos. The human c-Jun gene is mapped to 1p32-p31, spans over 3.3 kb and encodes c-Jun protein with 331 amino acid residues; the c-Fos gene locates on chromosome 14q24.3 with 3.4 kb in size and encodes c-Fos protein with 380 amino acid residues. Environmental exposures, such as cigarette smoking, radiation and airborne genotoxic carcinogens could induce the expressions of c-Jun/c-Fos or the assembly of AP-1 complex23, 31–34 and further activate the oncogenic signaling pathway through AP-1′s binding to AP-1 DNA-binding element, thus promoting tumorgenesis.25 Recently, one of our previous studies reported that two polymorphisms in c-Jun promoter could increase the gene's expression and were associated with elevated risk of colorectal cancer in Chinese.11 There is a possibility that the genetic variations in these two genes and their interaction with environmental factors alter the susceptibility to human lung cancer. So, we hypothesized that the genetic variations in c-Jun or c-Fos were associated with altered lung cancer risk.
In our study, we conducted two independent hospital-based case–control studies in southern and eastern Chinese populations to investigate the associations between three putative functional polymorphisms (i.e., −1318T>G and −673T>C in c-Jun promoter, −60C>T in c-Fos promoter) and lung cancer risk. Consecutive functional assays were further performed to determine the biological effects of these polymorphisms.
Material and Methods
The two independent case–control datasets have been described previously.35 Briefly, for the discovery set, 1,056 age- (±1 years) and sex frequency-matched lung cancer patients and healthy controls of southern Chinese were used, and for the validation set, 503 patients and 623 age- (±5 years) and sex frequency-matched controls of eastern Chinese were enrolled. All the participants were genetically unrelated ethnic Han Chinese and none had blood transfusion in the last 6 months. Each participant was scheduled for an interview to collect subjects' information on demographic data and selected factors such as smoking status, alcohol use and family history of cancer after a written informed consent was obtained. The definition of smoking status, pack-years smoked, drink status, family history of cancer and family history of lung cancer have been described previously.4 Each subject was asked to donate 5 mL of blood after having given their informed consent. The study was approved by the institutional review boards of Guangzhou Medical University and Soochow University.
We have previously described the single-nucleotide polymorphisms (SNPs) selection (−1318T>G, rs2760501 and −673T>C, rs4646999) in promoter of c-Jun.11 Based on the GeneBank dbSNP Chinese Han Beijing (CHB) database (http://www.ncbi.nlm.nih.gov/), there were no putative functional common SNPs (minor allele frequency >5%) in the exons or 3′-untranslated region (UTR) of this gene. In our study, we resequenced 1-kb promoter region in 60 normal Han Chinese and found that there were three common SNPs in Chinese, that is, rs2760501, rs4646999 and rs4647001. Our sequencing data showed that rs4647001 was in complete linkage disequilibrium (LD) with rs4646999 (D' = 1.00, r2 = 1.00), so, we selected rs2760501 and rs4646999 SNPs in our study. For c-Fos gene, 87 SNPs were reported by GeneBank, and only three loci (rs932577, rs2239615 and rs7101) in promoter are common SNPs, while none is common in 3′-UTR or nonsynonymous polymorphism in exons. After resequencing the promoter region in 60 normal Han Chinese, we found that these three SNPs were in high LD with each other (D' = 1.00, r2 > 0.90), and rs7101 (−60C>T) in c-Fos could represent the genetic information of these three loci. Therefore, we genotyped three putative functional SNPs (rs2760501, rs4646999 of c-Jun and rs7101 of c-Fos) in our study.
The SNPs' genotyping of each subject was performed by Taqman allelic discrimination assay, and ABI PRISM 7900 Sequence Detection Systems (Applied Biosystems, Foster City, CA) was used. Primers and probes were described in Supporting Information Table S1, which were designed by Primer Express 3.0 (Applied Biosystems) and synthesized by Shanghai GeneCore Biotechnologies (Shanghai, China). The genotypes of each SNP were automatically determined by Sequence Detection Systems software 2.0.1 (Applied Biosystems) (Supporting Information Fig. S1). About 10% samples were randomly selected to repeat and 60 randomly selected samples were used for resequence and the results were all 100% concordant (Supporting Information Fig. S2).
c-Jun mRNA expression analysis
As only −1318T>G and −673T>C of c-Jun were found to be significantly associated with lung cancer risk, we then determined whether these polymorphisms had an effect on c-Jun gene expression in vivo. Thirty-two tumor tissues and their adjacent normal tissues were used as previously described.35, 36 Genotypes of these samples were all confirmed by sequencing. Total RNA was extracted by using Trizol Reagent (Invitrogen) and then reverse transcribed to complementary DNA by using oligo primer and SuperscriptII (Invitrogen). Relative mRNA expression level of c-Jun and an internal reference gene β-actin were detected on the ABI Prism 7500 sequence detection system (Applied Biosystems) based on the SYBR-Green method. The primers used for c-Jun were 5′-ACA GAG CAT GAC CCT GAA CC-3′ (forward) and 5′-CCG TTG CTG GAC TGG ATT AT-3′ (reverse) and for β-actin were 5′-GGC GGC ACC ACC ATG TAC CCT-3′ and 5′-AGG GGC CGG ACT CGT CAT ACT-3′. Method of 2Δt was used to demonstrate the level of c-Jun gene's expression (Fig. 1a). All analyses were performed in a blinded fashion with the laboratory persons unaware of genotyping data. Each assay was done in triplicate.
c-Jun protein expression analysis
The association between the two SNPs of c-Jun and gene's protein expression levels was also detected by Western blotting. Proteins were extracted from the previously described 32 lung cancer tissues, and the Western blotting assays were performed as described previously.4, 11 The protein expressions were detected with a Phototope-horseradish peroxidase Western blot detection kit (Cell Signaling Technology, Danvers, MA).The expressions of β-actin were used as internal control. All antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and the relative c-Jun protein expression levels normalized to β-actin were semiquantified with a software Gene Tools (version 4.01, Syngene, Cambridge, UK) assembled in our image machine (G:BOX, Syngene).
Reporter plasmids construction
As a significant association was observed for haplotypes and lung cancer risk, we constructed four haplotype reporter plasmids (i.e., “p-T-T,” “p-T-C,” “p-G-T” and “p-G-C”) derived from −1318T>G and −673T>C (Fig. 2a). The p-T-T reporter was constructed by amplifying the 1956-bp c-Jun promoter region (from −1752 to +204 nt relative to the translation start site) from subjects with −1318TT and −673TT genotypes by using the primers 5′-GG GGT ACC TTA GGG CTG GGG GCA ATG-3′ (forward) and GCC C GAG CTC AGA GAG AAG GTG AAA AGA-3′ (reverse). The amplicons were then subcloned into the pGL3 basic vector (Promega, Madison, WI). The p-T-C, p-G-T and p-G-C reporter constructs were obtained from the p-T-T construct by site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). We also generated two short reporter constructs (i.e., “p-673T,” “p-673C”) with 982 bp (−725 to +257 nt) c-Jun promoter region which only includes the −673T>C polymorphism by using the primers 5′-CT GGT ACC CTA CGA GCAGCCAGACGA-3′ (forward) and 5′-GG GAG CTC GGA CAC TCC CGA AAC ACC-3′ (reverse). The six reporter constructs were all resequenced to confirm the sequence, orientation and integrity of each insert would be right.
The in vitro luciferase assays were performed as previously described.4 Two cell lines, A549 (a human lung adenocarcinoma cell line) and NCI-H520 (a human lung squamous cell line) were seeded onto 24-well plates (100,000 cells per well) and were cotransfected with 1 μg of the reporter plasmid and 10 ng pRL-TK plasmid (Promega) by using FuGENE HD Transfection Reagent (Roche Applied Science, Mannheim, Germany). The luciferase activity was measured with a Dual-Luciferase Reporter Assay System (Promega).
Because smoking and drinking were observed to interact with the risk genotypes on lung cancer risk, we then added tobacco extract or alcohol mixture into the cell medium in luciferase assays. Tobacco extract was prepared as described by Nakamura et al.37 Briefly, the smog of two lit cigarettes was collected by a syringe pump and was sent to 50 mL Dulbecco's modification of eagle's medium (DMEM)-F12. After sealing and complete mixing, 1 M NaOH was used to adjust its pH to 7.4, and filter membrane (pore size: 0.22 μm) was used to remove bacteria. The tobacco extract was added to the medium at a final concentration of 10% at 2 hr before the luciferase activity was detected. Alcohol mixture consists of ethanol and acetaldehyde. The treatment with alcohol mixture was performed as described by Blasiak et al.38 The ethanol was first added into the medium at a final concentration of 10 mM for half an hour, after removal of ethanol, the acetaldehyde at a final concentration of 100 mM was added 1.5 h before the luciferase activity was detected. Independent triplicate experiments were done for each plasmid construct. Differences in the expression levels of different constructs were determined by Student's t test (Fig. 2b).
The χ2 tests were used to assess differences in the distributions of demographic characteristics, alleles and genotypes between cases and controls. The Hardy–Weinberg equilibrium was tested by a goodness-of-fit χ2 test to compare the observed genotype frequencies to the expected ones among the control subjects. The PROC ALLELE statistical procedure in SAS/Genetics (SAS Institute, Cary, NC) software was used to assess the LD of the two SNPs and to infer haplotype frequencies based on their observed genotypes. The associations between case–control status and each SNP/haplotype were estimated using an unconditional logistic regression model with adjustment for age, sex, smoking status, drinking status and family history of cancer. Logistic regression models were also used in the trend test. Bonferroni correction was used to correct for multiple comparisons of SNPs, that is, the Bonferroni adjusted p value = (p value of tested SNP) × k!/(2!(k − 2)!), k is the number of total SNPs. Akaike's information criterion (AIC) was used to select the best genetic-effect model for each SNP (i.e., that in which the AIC value was smallest).39 To increase the study power, we merged the two study populations if the results were homogenous in Breslow–Day test. A multiplicative interaction model was suggested when OR 11> OR 10 × OR 01, in which OR 11is the OR when both factors were present, OR 01 is the OR when only factor 1 was present and OR 10 is the OR when only factor 2 was present to evaluate the possible gene–environment interactions on lung cancer risk.4 The statistical power was calculated by using the PS Software (http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize). The false-positive report probability (FPRP) test was applied to detect the false-positive association findings.40 Student's t test and one-way analysis of variance (ANOVA) test were used to examine the difference in levels of luciferase reporter gene expression between different constructs, and that of c-Jun mRNA expressions in tumor tissues as well as protein levels. Linear regression models with adjustment of age, sex, smoking, alcohol usage and family history of cancer were also used to evaluate the main effect of SNPs or genotypes on the c-Jun expression. All tests were two-sided by using the SAS software (version 9.1; SAS Institute). p < 0.05 was considered statistically significant.
Subject's demographic characteristics
The distribution of demographic characteristics of the discovery set and the validation set are shown in Table 1. Consistent results were observed in the two populations as there was no significant deviation in distributions of age, sex and family lung cancer history between the cases and controls (p > 0.05 for all); there were more current smokers and former smokers in cases than in controls (p = 0.028 for discovery set and p = 1.79 × 10−7 for validation set), especially more male smokers in cases than in controls (p = 5.14 × 10−5 for the discovery set and p = 4.77×10−9 for the validation set). Meanwhile, cases had smoked more pack-years compared to controls (p = 9.40 × 10−12 for the discovery set and p = 0.011 for the validation set); also, there was significant difference of drinking status too (p = 0.042 for the discovery set and p = 0.017 for the validation set). However, other than in the southern Chinese, the family history of cancer was significantly associated with the risk of lung cancer in eastern Chinese population (p = 0.046); the frequency distributions of smoking status and drinking status were not homogeneous (Breslow–Day test p = 0.001 and 0.026, respectively), reflecting different lifestyle between two populations. These variables were adjusted for the multivariate logistic regression model to control possible confounding on the main effects of the studied polymorphisms. We further combined the two populations in stratification analysis to increase the study power. The histological types and clinical stages of the cases were also enumerated in Table 1.
Table 1. Frequency distributions of selected variables in lung cancer patients and cancer-free controls
Associations between the SNPs and lung cancer risk
Table 2 summarizes the genotype and allele distributions of the three polymorphisms among the cases and controls. The observed genotype frequencies of these polymorphisms were all in agreement with the Hardy–Weinberg equilibrium in the control subjects of both sets (p > 0.05 for all), and LD analysis showed a low linkage between the −1318T>G and −673T>C (D' = 0.184, r2 = 0.015); so, we further analyzed the effects of haplotypes and combined genotypes on lung cancer risk recently.
Table 2. Distribution of genotypes of c-Jun and c-Fos gene and associations with the risk of lung cancer
As shown in Table 2, the −1318T>G and −673T>C of c-Jun were significantly associated with increased lung cancer risk, but −60C>T of c-Fos was not. In the discovery set of southern population, the genotypes' frequency of −1318T>G polymorphism differed significantly between cases and controls (group test p = 7.10 × 10−5); the frequency of G allele was significantly higher in cases than in controls (0.231 vs. 0.178, p =1.69 × 10−5). Compared to the most common genotype −1318TT, carriers of −1318GT heterozygote had an increased risk of lung cancer (adjusted odds ratio (OR) = 1.47; 95% confidence interval (CI) =1.22–1.77, p = 6.40 × 10−5); −1318GG homozygote had a further increased risk of lung cancer (adjusted OR = 1.69; 95% CI = 1.12–2.57, p = 0.013), and the adverse genetic effect increased in an allele dose-dependent model (adjusted ptrend=1.53 × 10−5). The genetic effect of the −1318T>G best fit the dominant heredity model as indicated by the AIC value (=2929.86) which was the smallest among all possible assumed models (data not shown), and the −1318GT/GG genotypes confer an increased risk of lung cancer (adjusted OR = 1.49; 95% CI = 1.25–1.79, p = 1.18 × 10−5). For −673T>C of c-Jun, the frequency of the genotypes also differed significantly between cases and controls (group test p = 0.0014) as well as C variant allele (0.589 vs. 0.548, p =0.007). Compared to the most common genotype −673TT, carriers of −673CC variant homozygote had an increased risk of lung cancer (adjusted OR =1.29; 95% CI =1.02–1.64, p = 0.035); but −673CT heterozygote did not (p = 0.465). The effect of −673T>C best fit the recessive heredity model (AIC value is 2929.85); the −673CC variant genotype was significantly associated with an increased lung cancer risk compared to −673TT/CT genotypes (adjusted OR=1.37; 95% CI =1.14–1.64, p = 7.0 × 10−4) as shown in Table 2.
Consistent results were observed in the validation set. The −1318GT/GG variant genotypes confer an increased lung cancer risk (adjusted OR =1.39; 95 % CI =1.08–1.77, p = 0.010), so did the −673CC variant genotype (adjusted OR =1.32; 95% CI =1.03–1.69, p = 0.031) as shown in Table 2.
We further combined the two populations as a merged set for the homogeneity test and showed that the above associations in two populations were homogeneous (Breslow–Day test p = 0.766 for −1318T>G; and Breslow–Day test p = 0.797 for −673T>C). Similarly, the carriers of −1318GT/GG variant genotypes had a 0.46-fold increased risk of lung cancer (adjusted OR = 1.46, 95% CI = 1.26–1.69, p = 2.92 × 10−7; Bonferroni adjusted p = 8.76 × 10−7) and the −673CC variant genotype carriers had a 0.35-fold increased risk of lung cancer (adjusted OR = 1.35, 95% CI = 1.17–1.56, p = 4.80 × 10−5; Bonferroni adjusted p = 1.44 × 10−4).
The haplotypes and combined genotypes of the c-Jun SNPs and lung cancer risk
As shown in Table 3, the difference in the frequency distributions of haplotype alleles between the cases and controls was statistically significant (p = 2.26 × 10−9). When the haplotype “T-T” was used as the reference, the other haplotypes were all significantly associated with an increased risk of lung cancer (“T-C”: OR=1.19, 95% CI =1.07–1.34, p = 0.0018; “G-T”: OR = 1.51, 95% CI = 1.27–1.81, p = 3.20 × 10−6; “G-C”: OR=1.60, 95% CI = 1.33–1.92, p = 2.20 × 10−7, respectively). We further combined the genotypes of these two SNPs based on the number of risk genotypes, that is, −1318GG/GT and −673CC were defined as risk genotype separately, thus the carriers of −1318TT and −673TT/CT have zero risk genotype, the carriers of −1318GG/GT and −673TT/CT or −1318TT and −673CC have one risk genotype and the −1318GG/GT and −673CC carriers have two risk genotypes. We found that the number of risk genotypes was significantly associated with lung cancer risk (p = 4.66 × 10−11). Compared to the zero risk genotype, the carriers of risk genotypes were associated with increased risk of lung cancer in a dose-dependent manner (adjusted OR = 1.52, 95% CI =1.31–1.77 for one and adjusted OR = 1.97, 95% CI = 1.55–2.50 for two risk genotypes; ptrend = 2.21 × 10−11; Bonferroni adjusted p = 6.63 × 10−11).
Table 3. Effect of haplotypes and combined genotypes of c-Jun on lung cancer risk
Stratification and gene–environmental interaction analysis
As shown in Table 4, the significant trends for dose-effect of number of variant genotypes on the risk of lung cancer were observed in all the subgroups except subgroup of female smoker (ptrend = 0.345) and female drinker (ptrend = 0.218) due to the limitation of small sample size; for patients with Stages I and II, borderline significance was observed (ptrend = 0.060 and 0.059, respectively).
Table 4. Stratification analysis of the number of risk genotypes in c-Jun by selected variables in lung cancer patients and controls
Furthermore, as shown in Figure 2c, we observed a positive significant interaction between number of c-Jun risk genotypes and smoking status on increasing lung cancer risk (p = 0.009). We also found a positive significant interaction between number of c-Jun risk genotypes and drinking status (p = 0.007), suggesting a possibility that the effect of c-Jun risk genotypes may be modified (i.e., induced) by smoking or drinking behavior. We further analyzed the FPRP of the interaction, under the assumption of a prior probability of 0.05 and a prior OR of 1.50 as suggested by Wacholder et al.,40 the FPRP for the observed interaction between number of c-Jun combined risk genotypes and smoking or drinking status on the risk of lung cancer yielded values of 0.002 and 0.088, respectively, which are lower than the preset FPRP-level criterion 0.20, suggesting that these findings are noteworthy. Further functional assays with luciferase reporter genes were also performed to test these findings.
Correlation between the c-Jun genotypes and phenotypes of gene expression
As shown in Figure 1a, we found that the mRNA levels of c-Jun in 32 lung cancer tissues were significantly higher than their adjacent normal tissues (p = 0.023). We then found that the mRNA levels of c-Jun were much higher in cases with −1318GT/GG genotypes than −1318TT genotype (p = 0.002), so were in cases with −673CC genotype compared to −673TT/CT genotypes (p = 0.003). Consistently, those cases with one or two risk genotypes had higher mRNA levels than cases with zero risk genotype (ANOVA test: p = 5.0 × 10−4; linear regression test: p < 0.001). The c-Jun protein expression confirmed the mRNA results as shown in Figures 1b and 1c; the levels of c-Jun protein in carriers with−1318GT/GG genotypes were significantly higher than those with −1318TT genotype (p = 0.038), and the c-Jun protein levels in carriers with −673CC genotype were much higher than those with −673TT/CT genotypes (p = 0.022). Similarly, those carriers with one or two risk genotypes had higher protein levels than carriers of zero risk genotype (ANOVA test: p = 0.017; linear regression test: p = 0.005). All these demonstrated that the c-Jun risk genotypes may increase c-Jun gene expression and thus increase susceptibility to lung cancer.
Effects of the c-Jun polymorphisms on transcriptional activity in vitro
Six luciferase reporter gene constructs were generated (Fig. 2a). As shown in Figure 2b, both in A549 and NCI-520, we found that the constructs containing the p-673C driven significantly higher luciferase activity than the constructs containing the p-673T (all p < 0.05), suggesting that the SNP −673T>C had an effect on promoter activity. We also found that the transcriptional activity of the reporter gene driven by the c-Jun promoter containing the p-G-C, p-G-T and p-T-C haplotypes was significantly higher than those driven by the p-T-T haplotype (all p < 0.001), and the highest activity level was observed for the reporter constructs with the p-G-C haplotype in both A549 and NCI-520 cancer cell lines. These indicated that both −1318T>G and −673T>C are functional loci that could regulate the transcriptional activity of the c-Jun. Intriguingly, we further observed that the reporter gene's transcription could be induced by tobacco extract or mixture of ethanol and acetaldehyde (all p < 0.001), especially in the reporter gene containing haplotype p-G-C of c-Jun which has the highest transcriptional activity under the treatments; so, these further supported the previously observed results of interaction between the risk genotypes of c-Jun and smoking or drinking status on cancer risk.
In the present two independent hospital case–control studies, we found that the −1318GT/GG variant genotypes or −673CC variant genotypes in c-Jun increased the risk of lung cancer both in southern and eastern Chinese. We further observed interactions between the c-Jun risk genotypes and smoking or drinking behavior on increasing cancer risk. Functional assays revealed a robust genotype–phenotype correlation that both the risk genotypes conferred higher gene expression invivo; the −1318G or −673C variant allele could significantly increase the transcription activity of the c-Jun promoter in vitro. Furthermore, the treatments with tobacco extract or alcohol mixture could easily induce c-Jun's transcription in vitro, especially for that plasmid containing two variant alleles (i.e., the G-C haplotype). However, for −60C>T SNP of c-Fos, no association was observed for lung cancer risk. To the best of our knowledge, this is the first study to investigate the polymorphisms of AP-1 members' gene and gene–environmental interaction in lung cancer.
Several studies have demonstrated that multitudinous extracellular stimuli, such as tobacco smoking, UV irradiation and peptide growth factors,23, 31–34 could activate c-Jun and then activate tumor-related target genes including Ras,16Fas,41Bim,18Bcl-3,19Jagged1,21MMP-9,42MMP-1 and MMP-320 and thus promote carcinogenesis. Elevated c-Jun protein also can inspirit cell survival by suppressing the function of tumor suppression gene PTEN.43 Consistently, overexpression of c-Jun has been found in many cancers, including lung cancer28–30 and increased c-Jun expression would promote tumor initiation,44 development,29 invasiveness27 and metastasis45 as reported. Here, consistently, we found higher expressions of c-Jun in tumor tissues compared to their adjacent normal tissues. We further demonstrated that the −1318T>G and −673T>C variations in c-Jun could increase the transcriptional activity and thus upregulate the expression of c-Jun and later alter the susceptibility to lung cancer, and it is concordant with the previous biological studies.11
We have previously reported that the polymorphisms −1318T>G and −673T>C of c-Jun were associated with increased risk of colorectal cancer.11 Consistently, here we also observed the significant association between c-Jun variations and lung cancer risk in both southern and eastern Chinese population. In addition, we observed positive significant gene–environment interactions between the c-Jun risk genotypes and smoking or drinking behavior on increasing lung cancer risk. Moreover, we found that either tobacco extract or alcohol mixture can induce the transcriptional activity of c-Jun promoter, especially for reporter genes containing haplotype “G-C” with both two variant alleles of −1318T>G and −673T>C. It is well known that tobacco smoking is the major risk factor of lung cancer,46 and long time alcohol consumption is also a potent cancer risk factor.47 Studies have reported that smoking and drinking can induce AP-1 activity;33, 48, 49 thus, our novel finding of gene–environment interaction between the c-Jun risk genotypes and smoking or drinking behavior on cancer risk is biologically acceptable as smoking or drinking could enlarge the adverse genetic effect of the variants by increasing c-Jun expression.
Being hospital-based case-control studies, restricted Chinese Han populations, some limitations are unavoidable for our study, such as selection bias. However, the genotype frequencies among controls fitted the Hardy–Weinberg disequilibrium law suggesting the randomness of subject selection; the study power is strong too, for −1318T>G, we have achieved a 99.3% study power (two-sided test, α = 0.05) to detect an OR of 1.46 for the −1318GT/GG risk genotypes (which occurred at a frequency of 31.8% in the controls), and for −673T>C, similarly, we have achieved a 98.2% study power to detect an OR of 1.35 for the −673CC risk genotypes (31.4% in the controls). When we combined the risk genotypes, the study power was 99.9%. In addition, those functional assays also confirmed the association between the c-Jun promoter polymorphisms and risk of lung cancer. Taken these together, it appears that our finding that the c-Jun risk genotypes associated with an increased lung cancer risk is unlikely to have been by chance.
In conclusion, in these two independent hospital-based case–control studies of lung cancer, our data suggest that the functional polymorphisms −1318T>G and −673T>C in the promoter region of c-Jun gene confer an increased risk of lung cancer by increasing the promoter activity and that c-Jun variant genotypes may be a biomarker for susceptibility to lung cancer. Studies with larger population in different ethnic groups are warranted to validate our findings.
Our study was supported by the National Natural Scientific Foundation of China grants 30671813, 30872178, 81072366 (J. Lu) and partly by 30872142 (W. Ji), 30972540 (B. Liu), 81001278 (Y. Zhou); the Guangdong Provincial Scientific Research Grants 8251018201000005 (J. Lu), 2008B060600008 (B. Liu); Guangdong Provincial High Level Experts Grants 2010-79 (J. Lu). We thank Dr. Zhanhong Xie, Ms. Wanmin Zeng and Ling Liu for their assistance in recruiting the subjects; Fuman Qiu and Yehua Liu for their assistance.