NFκB1 and NFκBIA Polymorphisms Are Associated with Increased Risk for Sporadic Colorectal Cancer in a Southern Chinese Population

Background Nuclear factor κB (NFκB) plays a key role in the regulation of apoptosis. The function of NFκB is inhibited by binding to NFκB inhibitor (IκB), and disruption of the balance of NFκB and IκB is related to the development of many diseases, including tumors. Therefore, we hypothesized that the NFκB1 (-94del/insATTG) and NFκBIA (2758 A>G) polymorphisms were associated with colorectal cancer (CRC) susceptibility. Methods In a hospital-based case–control study of 1001 CRC patients and 1005 cancer-free controls frequency matched by age and sex, we genotyped polymorphisms using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method and performed luciferase assays and Western blotting analysis to identify whether genetic variants in NFκBIA alter its gene expressions and functions and thus cancer risk. Results We found that both NFκB1-94 ins/delATTG and NFκBIA 2758 A>G polymorphisms were correlated with CRC risk (OR = 1.47; 95%CI = 1.14–1.86, and OR = 1.38; 95% CI = 1.14–1.66, respectively). Furthermore, when evaluated these two polymorphisms together, the combined genotypes with 2 variant (risk) alleles (2758GG and -94ins/ins+del/ins) were associated with an increased risk of CRC (OR = 1.71; 95% CI = 1.23–2.38) compared to 0 variant, and the significant trend for 2 variant (risk) alleles were more pronounced among subgroups of aged <60 years, women, never drinkers, never smokers, persons with a normal BMI and those with a family history of cancer(Ptrend<0.01). Moreover, luciferase assays showed that the G allele in the 3′UTR significantly decreased NFκBIA mRNA stability and the A allele regulation by miRNA449a in vitro and that the NFκBIA protein expression levels of the AA+AG variant carriers were significantly higher in peritumoral tissues than those of the 2758GG genotype. Conclusion NFκB1 and NFκBIA polymorphisms appear to jointly contribute to risk of CRC. These two variants may be a genetic modifier for CRC susceptibility in this southern Chinese population.


Introduction
Colorectal cancer (CRC) is the third most common cancer in men and the second most common cancer in women around the globe, and it is estimated that there were approximately 1.2 million newly diagnosed CRC cases and 608,700 related deaths in 2008 [1]. Records from the municipal death registry of the city of Guangzhou, Guangdong, China, indicate that CRC is the fifth most common cancer. The mortality rate was dramatically increased from 4.33/10 5 persons in 1970's to 12.13/10 5 persons in 2000's [2]. The majority of CRC cases (approximately 80%) are sporadic [3], but a hereditary predisposition is present in 20-35% of patients, suggesting that both genetic and environmental factors contribute to CRC development [4]. Alcohol drinking and tobacco use [5,6], dietary and lifestyle factors [7], and inflammatory bowel disease such as ulcerative colitis [8,9] have shown to be associated with CRC risk. Although many people are exposed to these risk factors, only some of the exposed individuals develop CRC, indicating that genetic variation partly determines individual susceptibility to colorectal tumorigenesis.
Apoptosis, a highly regulated cellular process, participates in development, tissue homeostasis maintenance and elimination of unwanted cells [10]. Dysregulation in this process is likely to contribute to tumorigenesis [11]. The biochemical pathways of apoptosis are complicated and depend on not only the cells but also the inducers of apoptosis. Substantial evidence suggests that the occurrence and development of cancer is associated with both extended cell survival and suspended apoptosis in precancerous lesions and, consequently, aberrant apoptosis may allow for unchecked cell growth [12].
Nuclear factor kappa B (NFkB) is a major transcription regulator of the immune response, cell adhesion, differentiation, proliferation, and apoptosis [13]. Five members(p50/p105, p65/ RelA, c-Rel, RelB, and p52/p100) in the NFkB family have been identified, and the dimeric form of NFkB1 p50/RelA is the major form [14]. In the resting cell, NFkB is inactivated in the cytoplasm through association with a sequestering inhibitory protein, IkBa, b or c, and the most common protein of this family is the NFkB inhibitor a (NFkBIA) [15]. In the classical activation pathway, the phosphorylation and degradation of the inhibitory proteins lead to NFkB dissociation from the NFkB complex and translocation to the nucleus, where it can activate the transcription of a large number of genes [16]. As an important transcription factor, NFkB mediates the survival response by inhibiting p53-dependent apoptosis and up-regulating anti-apoptotic members of the Bcl-2 family and caspase inhibitors [17,18]. In contrast, NFkB is also activated by both the extrinsic and intrinsic apoptotic stimuli and mediates upregulation of pro-apoptotic genes such as TRAIL R2/ DR5, Fas, and Fas ligand [19,20]. An inappropriate activation of NFkB could disturb tissue homeostasis and lead to dysregulated apoptosis. Furthermore, activity of NFkB has been observed in several types of cancers including CRC [21,22], indicating it may play an important role in tumorigenesis [23,24].
There has been no previous report on the association between NFkB1 and NFkBIA polymorphisms and CRC risk. As the NFkB/ IkB system plays an important regulatory role in the apoptotic pathway and dysregulated expression of the NFkB1 and NFkBIA has been observed in CRC, we hypothesized that combined NFkB1 and NFkBIA polymorphisms may be associated with increased risk of CRC. To test this hypothesis, we genotyped the NFkB1-94 insertion/deletionATTG and NFkBIA 2758A.G polymorphisms in our ongoing hospital-based case-control study of CRC in a southern Chinese population, and further performed luciferase assays and Western blotting analysis to identify whether genetic variants in NFkBIA alter its gene expressions and functions and thus cancer risk.

Ethics statement
The study protocol was approved by the institutional review boards of Sun Yat-Sen University. Written informed consent was obtained from each participant after a full explanation of the study. In the interim, a total of 1005 cancer-free controls were randomly selected from a subject pool of more than 10,000 individuals who participated in health check-up programs at the community health stations in Guangzhou, China.

Study subjects and sample collection
The study participants were interviewed and data on smoking status, alcohol use and other factors including family history of cancer were obtained using a structured questionnaire. Smoking status, alcohol use and family history of cancer were defined as described previously [38]. Subjects whose body mass index (BMI) was ,18.5 kg/m 2 were categorized as being underweight, subjects whose BMI was from 18.5 to 24.0 kg/m 2 were normal body weight, those who have a BMI .24.0 kg/m 2 were overweight [39]. Cases belonging to familial adenomatous polyposis and those cases that fulfilled the criteria of Amsterdam for hereditary nonpolyposis CRC were excluded.

Genotyping
Five mL blood was collected from each participant and genomic DNA was extracted using the DNA Blood Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Genotyping was performed by the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method. For determination of the NFkB1 promoter (rs28362491) polymorphism, the SNP-containing fragment was amplified using the following primers: 59-TGGGCACAAGTCGTTTATGA-39 (forward) and 59-CTGGAGCCGGTAGGGAAG-39 (reverse). PCR was run at 94uC for 3 min followed by 30 cycles of 94uC for 30 s, 56uC for 30 s, and 72uC for 60 s with a final extension at 72uC for 10 min. The PCR products (281/285 bp in size) were digested with PfIMI (Fermentas, Vilnius, Lithuania) at 37uC overnight followed by 2% agarose gel electrophoresis. For determination of the NFkBIA (rs696) polymorphism, the SNPcontaining fragment was amplified using the following primers: 59-GGCTGAAAGAACATGGACTTG-39 (forward) and 59-GTACACCATTTACAGGAGGG -39 (reverse). The PCR was run at 94uC for 5 min followed by 32 cycles of 94uC for 30 s, 54.3uC for 45 s and 72uC for 60 s with a final extension at 72uC for 10 min. The amplified fragments were digested with HaeIII (Fermentas, Vilnius, Lithuania) overnight at 37uC followed by 2% agarose gel electrophoresis. Genotype analysis was done by two experimenters independently who were blinded to the status of the subjects as patient or control. To further validate the genotyping results, we randomly selected 10% samples for each of the 2 SNPs to perform repeat assays, and the results were 100% concordant. Additionally, 5% of the PCR products for each target genotype were purified and confirmed by direct sequencing ( Figure S1 and Figure S2).

Cell culture
Three human colorectal adenocarcinoma cell lines, HCT116, HT29 and SW480 were purchased from Culture Collection of Chinese Academy of Science (Shanghai, China) and routinely cultured in Dulbecco's Modified Eagle's Medium or RPMI 1640 medium supplemented with 100 units/ml of penicillin, 100 mg/ml of streptomycin, and 10% fetal bovine serum (FBS). The cells were grown at 37uC with 5% CO2 in a humidified incubator.

Western blotting analysis
To analyze the correlation between NFkBIA polymorphisms in its 39UTR (2758A.G) and NFkBIA protein expression levels in tissues, 32 paired tumor and peritumoral tissues were obtained from sporadic CRC patients resected at the Sixth Affiliated Hospital of Sun Yat-Sen University and archived in the tumor bank. All tissues samples were histologically confirmed. Immunoblotting assays were performed as previously described [38] and antibodies against NFkBIA and b-actin (Precision Task Group, Chicago, USA) were used for the procedure. Protein densitometry was performed using Gel-Pro Analyzer 4.0 software (Media Cybernetics, Inc., Silver spring, MD) and NFkBIA expression was normalized against b-actin.

Statistical analysis
Two-sided chi-square tests were used to assess differences in the distributions of age, sex, smoking status, alcohol use, BMI, menstruation history, and family history of cancer between cases and controls. The Hardy-Weinberg equilibrium (HWE) was tested by a goodness-of-fit chi-square test to compare the expected genotype frequencies with observed genotype frequencies (p2+2pq+q2 = 1) in cancer-free controls. The association between case-control status and each SNP, measured by the odds ratio (OR) and its corresponding 95% confidence interval (CI), was estimated using an unconditional logistic regression model, with and without adjustment for age, sex, smoking status, alcohol drinking status, BMI, tumor site, and family history of cancer.
Recent studies indicate that an analysis of the combined genotypes might be more scientifically significant than an analysis of a single polymorphism in predicting the disease associations. Therefore the combined genotype data were further stratified by age, sex, and smoking status, alcohol drinking status, BMI, tumor site, and family history of cancer. Logistic regression modeling was also used for the trend test. Student's t test was done to examine the difference in levels of luciferase reporter gene expression between different constructs. Kruskal-Wallis one-way ANOVA tests were used for analyzing NFkBIA protein expression in peritumoral tissues of different genotypes. All tests were two-sided by using the SAS software (version 9.1; SAS Institute, Cary, NC). P,0.05 was considered statistically significant.

Characteristics of the study population
The distribution of demographic characteristics of the 1001 sporadic CRC cases and 1005 cancer-free controls are shown in Table 1. Overall, no statistically significant difference was observed in age, sex, and smoking status between the cases and controls (P.0.05 in all). Compared with controls, there were more ever drinkers (cases vs. controls, 45.5% vs. 23.6%) (P,0.01) in the cohort of CRC cases. Moreover, compared with controls, CRC cases were more likely to have a family history of cancer or a higher BMI (P,0.05 in both). Consequently, these variables were further adjusted for in the multivariate logistic regression model to avoid possible confounding on the main effects of the SNPs under study. They were also used in the subsequent stratification and gene-environment interaction analysis.
PCR-RFLP study of the NFkB1 promoter region and the NFkBIA 39UTR region We carried out analysis of the NFkB1 promoter region by the PCR-RFLP method. Two ATTG repeats are present at the NFkB1 promoter region, and one allele that has an ATTG insertion (ins) was cleaved into a 45-bp and a 240-bp fragment after digestion with PfIMI while the other deletion allele (del) that has only one ATTG at its promoter was not cleaved (Figure 1.a). Furthermore, after digestion with HaeIII, the 2758GG genotype produced a 316-bp and a 108-bp band whereas the 2758AA genotype produced a single 424-bp band, and heterozygotes displayed all 3 bands (Figure 1.b).
The NFkB1-94ins/delATTG (rs28362491) and NFkBIA 2758A.G (rs696) polymorphisms are associated with the risk of sporadic CRC The genotype and allele distributions of the NFkB1-94ins/ delATTG (rs28362491) and NFkBIA 2758A.G (rs696) polymorphisms among the cases and controls are summarized in Table 2. The observed genotype frequencies of these two polymorphisms were all in agreement with the Hardy-Weinberg equilibrium in the control subjects (P.0.05). As shown in Table 2, for rs28362491, the distribution of the co-dominant genetic model (del/del vs. ins/ins vs. del/ins), and the dominant model (ins/ ins+del/ins vs. del/del) differed significantly between CRC cases and controls (ins/ins: OR = 1.70; 95% CI = 1.29-2.25; P,0.01; del/ins: OR = 1.33; 95% CI = 1.03-1.74; P,0.01; ins/ins+del/ ins: OR = 1.47; 95% CI = 1.14-1.86; P,0.01, respectively). For rs696, there was a significant difference in the distribution of the rs696 genotypes between CRC cases and controls (P,0.01). Significant differences were also noted in the distribution of the recessive model of rs696 (GG versus AA+AG) between CRC cases and controls (OR = 1.38; 95% CI = 1.14-1.66; P,0.05). Consistently, significant association was found between the two SNPs and the risk of sporadic CRC.

Stratification analysis of the association of combined
NFkB1 and NFkBIA polymorphisms with risk of CRC We further analyzed the combined genotypes of the NFkB1 and NFkBIA polymorphisms by examining age, sex, smoking, drinking, BMI, tumor site, and family history of cancer by logistic regression. We combined the NFkB1 and NFkBIA polymorphisms based on the number of variant (risk) alleles (i.e., 2758GG and -94ins/ins+del/ins). As shown in Table 3, compared with the NFkB1-94 del/del and NFkB1IA 2758AA+AG genotype, one variant combined genotype carriers who were ,60 years of age had a higher risk of CRC (OR = 1.57; 95% CI = 1.04-2.38) (P,0.05).   MiR-449a mimics suppressed the activities of the NFkBIA 2758 A.G polymorphism We further analyzed the 39UTR of the NFkBIA 2758 A.G polymorphism using a computer algorithm and found that the polymorphism could affect miRNA binding. MiR-449a and miR-34b, which are involved in a wide variety of biological functions, were found to have a binding site within the 39-UTR of the NFkBIA 2758 A.G polymorphism (Figure 2a). To determine the allele-specific effect of the NFkBIA 2758 A.G variant on the activity of the 39UTR and whether this polymorphism affected miRNA binding, we transfected CRC cell lines SW480, HT29, and HCT116 with reporter plasmids carrying the 2758G or 2758A allele and miR mimics or inhibitors (Figure 2a). We found that, in the absence of miR mimics or inhibitors, compared with the 2758A allele, the 2758G allele exhibited decreased luciferase activities ( P,0.05) (Figure 2b). MiR-449a mimics could reduce the relative luciferase activities of the 2758A allele while miR-449a inhibitors significantly up-regulate the relative luciferase activities of the 2758A allele in all the three cell lines (P,0.01 for HT29; P,0.05 for SW480 and HCT116) (Figure 2b). On the other hand, miR-34b mimics failed to exert any noticeable effects on the relative luciferase activities of the 2758G allele in these cell lines (data not shown). Taken together, these results demonstrated that the 2758G allele in the 39UTR of the NFkBIA reduced normalized luciferase activity compared to the 2758A allele, likely corresponding to reduced mRNA stability or translational efficacy and that miR-449a have the ability to bind and partially repress luciferase expression via the NFkBIA 39UTR segment when carrying the 2758A allele of NFkBIA 2758 A.G polymorphism.

Association of the NFkBIA39UTR polymorphisms with the NFkBIA protein expression
We were also interested in investigating whether the NFkBIA39UTR polymorphisms (2758A.G) were associated with increased or reduced expression of NFkBIA protein. We collected 32 paired tumor and peritumoral tissues from the untreated CRC patients of different genotypes. Immunoblotting analysis revealed that the levels of NFkBIA (NFkBIA/b-actin protein ratio) in the peritumoral tissues of 12 cases of the 2758AA genotype was 0.8060.09 and those of 8 cases of the 2758AG genotype was 0.6360.11, both of which were significantly higher than those of the other 12 cases of the 2758GG genotype (0.4760.06) (analysis of variance test, P,0.05) (Figure 3). However, in the tumor tissues, NFkBIA expression levels did not differ significantly among the cases of different genotypes (data not shown). These results indicated that NFkBIA levels were significantly higher in peritumoral tissues from patients of the 2758AA+AG genotypes than those of the 2758GG genotypes.

Discussion
In the present study, we investigated the associations of the NFkB1-94del/insATTG and NFkB1IA 2758A.G polymorphisms Table 2. Distribution of genotypes in NFkB1 and NFkBIA, and results of logistic regression analysis for associations with risk of colorectal cancer.

Genotypes
Patients n (%) Controls a n (%) P b Crude OR (95% CI) Adjusted OR (95% CI) c with risk of CRC in a southern Han Chinese population. We found that both the NFkB1-94del/ins ATTG and NFkBIA 2758A.G polymorphisms were associated with increased risk of CRC. For the NFkB1-94del/insATTG polymorphism, with the -94del/del as the reference, we found that the -94(ins/ins+del/ins) genotype was associated with a statistically significantly increased risk of CRC. For the NFKBIA 2758A.G polymorphism, with the 2758(AA+GA) as the reference, we also found that the 2758GG genotype was associated with a statistically significantly increased risk of CRC. Furthermore, when we evaluated NFkB1 and NFKBIA polymorphisms in combination, we found that the combined 2758GG and -94ins/ins+del/ins genotype was associated with a significantly increased risk of CRC compared with those without the 2758GG and -94ins/ins+del/ins genotype, and this increased risk was more pronounced among younger than 60 years, women, never drinkers, never smokers, persons with a normal BMI and those with family history of cancer. In the in vitro assays, we also found that, compared with the 2758A allele, the 2758G variant allele showed significantly decreased mRNA stability and/or translational efficacy. Furthermore, we found that NFkBIA levels significantly higher in peritumoral tissues from patients of the 2758AA+AG genotypes than those of the 2758GG genotypes, suggesting that the 2758A.G polymorphism is potentially functional and the polymorphism 2758A.G at 39UTR of NFkBIA could affect gene expression. To the best of our knowledge, this is the first study to investigate whether NFkB1 and NFkBIA polymorphism and their combined polymorphism were associated with risk of CRC.
There are several lines of evidence supporting our findings. The NFkB1 pathways seem to play a critical role in multiple human pathologies by regulating the transcription of genes involved in the immune response, cell proliferation, and apoptosis [13]. It has been shown that alterations of NFkB1 expression plays an important role in the protection of cells from apoptosis [40]. NFkB1 activity has been observed in various types of cancer [41], including breast cancer [42] and CRC [21], to contribute to tumor angiogenesis and progression [43]. There are also many experimental data suggests that NFkB1/IkB pathway may participate in tumor cell invasion as well [44]. Therefore, the variants of the NFkB1 and NFkBIA genes, if functional, could be expected to have an effect on cell death, and thus, carcinogenesis. Several association studies have reported that the NFkB1 and NFkBIA polymorphisms is related to the development of inflammatory and other diseases including ulcerative colitis, Graves' disease, and diabetes mellitus, and susceptibility to tumors including melanoma, bladder cancer and CRC in different ethnic groups [32,36,45,46,47,48]. Gao et al.'s study reported a lack of an association between the NFkBIA2758 A.G polymorphism in northern Chinese population, but showed an association with the Swedish population and CRC risk [36]. Previous studies have provided evidence that the del allele may result in relatively decreased NFkB1 transcript levels and hence decreased p50/p105 NFkB1 protein production [32]. In our study, we further found that the change of the 2758 A to G allele in the 39UTR of NFkBIA decreased luciferase activities as assessed by luciferase assays. Our functional in vitro experiments suggested that NFkBIA 2758 A.G variants may affect mRNA stability. However, that the NFkBIA 2758 A.G variants affects translational efficacy or conduces to differential nuclear RNA processing or export also cannot be completely excluded [49]. Furthermore, it is well known that miRNAs can also cause mRNA cleavage or translational repression by forming imperfect base pairing with the 39-UTR of target genes. In silico analysis of the NFkBIA mRNA sequence predicted the 2758 A.G variant of NFkBIA generates a potential seed site for miR-449a and our ex vivo luciferase data indicated that miR-449a reduced the relative luciferase activities via the NFkBIA 39UTR target site created by the 2758A allele. The results indicated that the 2758A allele strengthens the binding of miR-449a with 39UTR of NFkBIA, which in turn inhibits the expression of NFkBIA. Recent evidence indicate that miRNAs can bind to the 39 UTRs of mRNAs and affect their translation, thus regulating cell proliferation, apoptosis and tumorigenesis [50]. Experimental studies concluded that SNPs located in miRNAbinding sites affect miRNA target expression and function, which is potentially associated with cancers [51]. This reconcile with our results that the NFkBIA 2758A-allele may enhance binding of miR-449a and affect NFkBIA gene expression and change of the 2758 A to G allele in the 39UTR of NFkBIA may affect mRNA stability or translational efficacy subsequent diminished protein levels. Consistently, we observed that NFkBIA protein expression was higher in the peritumoral tissues but not in tumor tissues from patients of the 2758AA+GG genotype than those of the 2758GG genotype. These data suggest that wildtype A allele may result in overexpression of NFkBIA in peritumoral normal tissues and hence increased NFkBIA protein production, resulting in inhibition of NFkB activity. However, carrying a mutant 2758G allele result in relatively decreased NFkBIA mRNA stability and hence diminished NFkBIA protein production. As a major inhibitor of NFkB, decreased expression and dysfunction of NFkBIA may be a direct result of the activation of NFkB and thus cancer [16]. This is also consistent with our initial expectations, since CRC has been associated with increased levels of NFkBIA G allele. These findings indicate that the NFkB1 -94del/ ins ATTG and NFkBIA 2758 A.G polymorphisms are functional. Our data from this relatively large sample size study further support the notion that the NFkB1 and NFkBIA polymorphisms are potentially implicated in cancer risk.
In the present study, we also observed that the combined effect of the NFkB1 and NFkBIA polymorphisms on risk of CRC was more pronounced among younger than 60 years, women, never drinkers, never smokers, persons with a normal BMI and those with a family history of cancer, suggesting that, in these subpopulations, gene-environment interaction may be very weak and the combined effect was an independent risk factor for these subpopulations. Compared with the published data, our results indicate that the genotype distributions of the NFkB1 and NFkBIA polymorphisms vary with ethnicity. For example, the frequencies of the del/del, del/ins, and ins/ins genotypes of the NFkB1-94del/ insATTG among our 1005 southern Han Chinese control subjects were 18.5%, 51.9%, and 29.6%, respectively, compared with 15.6%, 45.9%, and 38.4%, respectively, of 307 Germans in the study by Kathrin et al. [37]. Similarly, the frequencies of the AA, AG and GG genotypes of the NFkBIA 2758A.G in our controls were 21.2%, 52.8%, and 26.1%, respectively, compared with 6%, 45%, and 49%, respectively, of 109 cancer-free Australian controls in the study by Curran et al. [52]. However, the frequencies of the del/del, del/ins, and ins/ins genotypes of the NFkB1-94del/insATTG (18.5%, 51.9%, and 29.6%, respectively) were very similar to the published data by Sun et al. (17%, 58%, and 24%, respectively) [53]. In our southern Han Chinese controls, the frequency of the NFkBIA genotype distributions is also similar to that reported for that of the northern Han Chinese [36]. In this study, we also found that the insertion of NFkB1-94del/insATTG polymorphism increased CRC risk, which was contrary to the results reported by Andersen et al [54]. It is likely that the discrepancy results from the genetic difference in ethnicity (the frequencies of the del/del, del/ins, and ins/ins genotypes of the NFkB1-94del/insATTG among our 1005 southern Han Chinese control subjects were 18.5%, 51.9%, and 29.6%, respectively, compared with 13.5%, 45.9%, and 40.6%, respectively, of 756 Danes in the study). In addition, environmental effects such as dietary and lifestyle may also contribute to the discrepancy [7]. However, this hypothesis warrants further investigation.
In conclusion, our results suggested that both NFkB1 and NFkBIA polymorphisms have effect on risk of CRC. These findings suggest that the NFkB1 and NFkBIA polymorphisms may jointly contribute to the risk of CRC in a southern Chinese population, which were consistent with the functional assays we performed. Our study indicated that the NFkB1-94(ins/ins+del/ ins) and NFkBIA GG polymorphism may be a genetic marker for susceptibility to CRC in Chinese populations. However, additional studies with more detailed data on environmental exposure and survival data are required to verify these findings. Therefore, future population-based studies are needed to verify the findings.