Genetic Variations in SMAD7 Are Associated with Colorectal Cancer Risk in the Colon Cancer Family Registry

Abstract

Background

Recent genome-wide studies identified a risk locus for colorectal cancer at 18q21, which maps to the SMAD7 gene. Our objective was to confirm the association between SMAD7 SNPs and colorectal cancer risk in the multi-center Colon Cancer Family Registry.

Materials and Methods

23 tagging SNPs in the SMAD7 gene were genotyped among 1,592 population-based and 253 clinic-based families. The SNP-colorectal cancer associations were assessed in multivariable conditional logistic regression.

Results

Among the population-based families, both SNPs rs12953717 (odds ratio, 1.29; 95% confidence interval, 1.12–1.49), and rs11874392 (odds ratio, 0.80; 95% confidence interval, 0.70–0.92) were associated with risk of colorectal cancer. These associations were similar among the population- and the clinic-based families, though they were significant only among the former. Marginally significant differences in the SNP-colorectal cancer associations were observed by use of nonsteroidal anti-inflammatory drugs, cigarette smoking, body mass index, and history of polyps.

Conclusions

SMAD7 SNPs were associated with colorectal cancer risk in the Colon Cancer Family Registry. There was evidence suggesting that the association between rs12953717 and colorectal cancer risk may be modified by factors such as smoking and use of nonsteroidal anti-inflammatory drugs.

About the Authors:

Xuejuan Jiang

Contributed equally to this work with: Xuejuan Jiang, J. Esteban Castelao

Affiliation: Department of Preventive Medicine, University of Southern California, Los Angeles, California, United States of America

J. Esteban Castelao

Contributed equally to this work with: Xuejuan Jiang, J. Esteban Castelao

Affiliation: Oncology and Genetics Unit, Complejo Hospitalario Universitario de Vigo, Servicio Galego de Saude (SERGAS), Vigo, Spain

David Vandenberg

Affiliation: Department of Preventive Medicine, University of Southern California, Los Angeles, California, United States of America

Angel Carracedo

Affiliation: Genomic Medicine Group, Galician Foundation of Genomic Medicine, Complejo Hospitalario Universitario de Santiago, Servicio Galego de Saude (SERGAS), Instituto de Investigación Sanitaria de Santiago (IDIS), Santiago de Compostela, Spain

Carmen M. Redondo

Affiliation: Oncology and Genetics Unit, Complejo Hospitalario Universitario de Vigo, Servicio Galego de Saude (SERGAS), Vigo, Spain

David V. Conti

Affiliation: Department of Preventive Medicine, University of Southern California, Los Angeles, California, United States of America

Jesus P. Paredes Cotoré

Affiliation: Department of Surgery, University Hospital Santiago de Compostela, Santiago de Compostela, Spain

John D. Potter

Affiliation: Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America

Polly A. Newcomb

Affiliation: Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America

Michael N. Passarelli

Affiliation: Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America

Mark A. Jenkins

Affiliation: Centre for Molecular, Environmental, Genetic and Analytic Epidemiology, School of Population & Global Health, The University of Melbourne, Victoria, Australia

John L. Hopper

Affiliation: Centre for Molecular, Environmental, Genetic and Analytic Epidemiology, School of Population & Global Health, The University of Melbourne, Victoria, Australia

Steven Gallinger

Affiliation: Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada

Loic Le Marchand

Affiliation: University of Hawaii Cancer Research Center, Honolulu, Hawaii, United States of America

María E. Martínez

Affiliation: University of California, San Diego Moores Cancer Center, San Diego, United States of America

Dennis J. Ahnen

Affiliation: Denver VA Medical Center and University of Colorado, Denver, Colorado, United States of America

John A. Baron

Affiliation: Department of Medicine, University of North Carolina School of Medicine, Chapel Hill, North Carolina, United States of America

Noralane M. Lindor

Affiliation: Department of Health Science Research, Mayo Clinic, Scottsdale, Arizona, United States of America

Robert W. Haile

Affiliation: Department of Preventive Medicine, University of Southern California, Los Angeles, California, United States of America

Manuela Gago-Dominguez

* E-mail: [email protected]

Affiliation: Genomic Medicine Group, Galician Foundation of Genomic Medicine, Complejo Hospitalario Universitario de Santiago, Servicio Galego de Saude (SERGAS), Instituto de Investigación Sanitaria de Santiago (IDIS), Santiago de Compostela, Spain

Introduction

It is estimated that inherited susceptibility contributes to ∼35% of all colorectal cancer (CRC) cases [1]. Recent progress through the application of genome-wide association studies (GWASs) have identified a number of common variants involved in the etiology of CRC[2]. Two GWAS [3], [4] identified a risk locus for CRC at 18q21, which maps to SMAD7, a functional candidate gene for CRC. Smad7 plays inhibitory roles in the transforming growth factor beta (TGF-β) signaling pathway [5], [6] which is involved in many cellular processes and has an important role in cancer development and progression [7]. Broderick et al. [3] identified three SNPs (rs4939827, rs12953717, rs4464148) in SMAD7 associated with CRC and the SNP rs4939827 was later replicated as the top-ranking SNP at 18q21 by Tenesa et al. [4]. The association between rs4939827 and CRC risk was also confirmed in a recent meta-analysis [8]; however, significant between-study heterogeneity was observed. In addition, these susceptibility variants were found to be enriched in familial CRC [9], [10]. Furthermore, SMAD7 expression was found to be lower in colorectal cancers than in adenomas regardless of 18q copy number status [11] and the risk allele at rs12953717 was significantly associated with lower SMAD7 expression in lymphoblastoid cell lines [3], suggesting that allele-specific expression of SMAD7 is likely to be the biological mechanism underlying the association between variations in 18q21 and CRC risk.

Given the SMAD7's role in the TGF-β signaling pathway [12] and the significant between-study heterogeneity in the reported association between SMAD7 SNPs and CRC risk, we looked to confirm the association between SMAD7 SNPs and CRC risk in a large family-based case-control study based on the multi-center Colon Cancer Family Registry (Colon CFR), and to examine gene X environmental interactions to identify risk/protective factors that may affect the association between SMAD7 SNPs and CRC risk. Our case-unaffected sibling study design has been shown to be more powerful for controlling potential confounding from population stratification and detecting gene-environment interactions [13].

Materials and Methods

Study population

Data for this study were obtained through the Colon CFR, a National Cancer Institute (NCI)–funded registry of CRC cases, unaffected family members, and population-based controls. The registry is described in detail in Newcomb et al. [14] and Levine et al. [15]. Briefly, the Colon CFR is an international collaborative study initiated in 1997. Participants were recruited from six centers including centers in the University of Southern California Consortium (Arizona, Cleveland Clinic, Colorado, Dartmouth, Minnesota, North Carolina, and University of Southern California), Hawaii (Honolulu), Fred Hutchinson Cancer Research Center (Seattle, WA), Mayo Clinic (Rochester, MN), Cancer Care Ontario (Toronto, Canada), and University of Melbourne (Victoria, Australia) using population-based and clinic-based ascertainment strategies. Cases were recruited in two phases, from 1998 to 2002 (phase 1) and from 2002 to 2007 (phase 2). Phase 2 subjects were enriched in cases more likely to have a family history of CRC. All centers except Fred Hutchinson Cancer Research Center oversampled cases with multiple first-degree relatives reporting CRC or CRC cases diagnosed under age 50 to target families with excess CRC risk. First-degree and some second-degree relatives with CRC were also recruited from families with multiple CRC cases. The clinic-based sample represents multiple-case families at high risk of Hereditary Non-Polyposis Colorectal Cancer or other familial CRC phenotypes.

In the Colon CFR, population-based controls were only obtained from one of the Colon CFR sites (Fred Hutchinson Cancer Research Center), and the total sample size (N = 429) is much smaller than that of the sibling controls (N = 3,115). To make the most use of the available genetic data, in this investigation we used a case/unaffected sibling control design[13] with data from both population-based and clinic-based families in the main effect analyses. Cases were probands and siblings diagnosed with CRC and controls were siblings without CRC at the time of ascertainment. Diagnosis of CRC was based on the following six categories of confirmation [14]: pathologist review of slides; review of pathology report; cancer registry report or medical record(s) indicating treatment for the specific type of cancer; report on a death certificate; self-report; and report by a relative. Therefore, the unaffected status of the siblings was not established through colonoscopy. All cases were interviewed within 5 years of diagnosis (76% within 2 years). There were too few clinic-based case/control pairs for stratified analyses so all stratified analyses used the population-based families only. We excluded monozygous twins and subjects with unknown age or gender, and included only non-Hispanic white subjects. In addition, we also genotyped a random set of unrelated population-based controls (n = 429) from one of the Colon CFR sites (Fred Hutchinson Cancer Research Center). A total of 1,923 cases (1,640 population-based and 283 clinic-based) and their 3,115 unaffected sibling controls (2,621 population-based and 494 clinic-based) were included in the analyses.

Ethics

All subjects signed an informed consent before providing data to the Colon CFR. Ethics approval for the study was obtained from the Institutional Review Boards at each CCFR site: University of Southern California Health Sciences Institutional Review Board, Mount Sinai Hospital Ethics Research Board, University of Hawaii Institutional Review Board, The University of Melbourne Central Human Research Ethics Committee, Fred Hutchinson Cancer Research Center Institutional Review Board, and Mayo Clinic Institutional Review Board.

SNP Selection and genotyping

SMAD7 was genotyped as part of an ongoing study of genes relevant to lipid peroxidation and apoptosis (5R01CA114472-02). Tagging single nucleotide polymorphisms (SNPs) were selected using the program Snagger20 [16] to cover all SNPs with a minor allele frequency (MAF) of ≥0.05 or greater with a pairwise r2 of ≥0.80 in the region covering each gene of interest as well as 20 kb upstream and 10 kb downstream of the gene. The linkage disequilibrium blocks were determined using data from the International HapMap Project White CEPH (Utah residents with ancestry from northern and western Europe) population (HapMap, release 21, July 2006; www.hapmap.org). Finally, three GWAS-identified SNPs, rs4939827, rs12953717, rs4464148, were also included. SNPs were genotyped on the Illumina GoldenGate platform (Illumina, Inc., San Diego, CA) [17] in the University of Southern California, Norris Comprehensive Cancer Center, Molecular Genomics Core facility, using DNA extracted from blood samples [14]. Quality control measures included testing for deviations from Hardy-Weinberg equilibrium (HWE) in non-Hispanic Whites, the inclusion of blinded interplate and intraplate replicates, and mixing cases and controls on genotyping plates. SNPs were excluded from the analysis if there were more than two errors on the replicate genotypes. Marker rs4939827 failed on the Illumina platform and was subsequently excluded. In this analysis, we report results for 23 tSNPs in SMAD7. A total of 133 blinded duplicate pairs were included for genotyping. Concordance for the duplicate samples was >99%.

Microsatellite Instability Testing

All available tumors from the Colon CFR's Jeremy Jass Memorial Pathology Bank were assayed for instability at the following 10 microsatellites: BAT25, BAT26, BAT40, BAT34C4, D5S346, D17S250, D18S55, D10S197, ACTC, and MYCL as described previously [14]. Only subjects with clear results for at least four markers were included. Microsatellite Instability (MSI) data were available for 1,242 (64.4%) of cases (1,106 population-based and 136 clinic-based). Instability at >30% of the tested loci was defined as microsatellite instability high (MSI-H); instability at >10% of loci but <30% of loci was defined as microsatellite instability low (MSI-L); and those with instability at 0 loci were categorized as microsatellite stable (MSS).

Tumor Location

Tumor location was obtained from the pathology report and was available for 1,778 (92.1%) of cases (1,566 population-based and 212 clinic-based). Right colon was defined as the cecum through the splenic flexure; left colon included the descending colon through the sigmoid colon; rectal tumors included the rectosigmoid junction and the rectum.

Statistical Analysis

All statistical analyses were conducted using the R programming language and SAS v9.1 (SAS Institute Inc., Cary, NC).

MAF was estimated from the genotype data from unrelated population-based controls. Pairwise linkage disequilibrium between SNPs was estimated using the square of the correlation coefficient (R2) and D-prime (D′) between markers. We also evaluated Hardy-Weinberg equilibrium for each SNP. No deviations were observed for all SNPs except rs1873190. Among the unrelated population-based controls, statistically significant reduced numbers of heterozygous genotypes of rs1873190 (155 observed vs. 186 expected under HWE; P exact = 0.0006) were observed. The genotyping data of this SNP showed clear genotype separation with a call rate of 98% and concordance rates of 100% among replicates.

In the analysis of main effects, the population- and clinic-based data were analyzed separately. We used multivariable conditional logistic regression with sibship as the matching factor and controlled for age (continuous) and sex. We assessed the SNP-CRC associations assuming a log-additive model. In all analyses, the lower frequency allele was coded as the “risk” allele and individuals were assigned a 0, 1, or 2 representing the number of risk alleles they possessed for that SNP. Test for haplotypic association was performed using SAS/Genetics and haplotype frequencies were estimated by the expectation–maximization algorithm.

Among the population-based families, we also examined the possibility that the SNP-CRC association was modified by other factors including gender, age, use of nonsteroidal anti-inflammatory drugs (NSAIDs), cigarette smoking, alcohol drinking, body mass index (BMI), physical activity (average weekly metabolic equivalent (MET) hours of physical activity throughout adulthood), history of diabetes, polyps, and ulcerative colitis, and family history of CRC in a first-degree relative as reported by the proband. All analyses within the exposure strata were specified in advance based on indications for potential effect modification in the literature. To test this hypothesis, dummy variables representing stratum-specific exposure were created for estimating stratum-specific results in one single conditional logistic regression model. P-values for interactions were estimated from likelihood ratio tests. We also evaluated differences in the association by tumor location (right, left, rectum/rectosigmoid junction) and MSI status (MSS, MSI-L and MSI-H) by stratifying the matched sets on the tumor characteristics of the case. We assigned the sets the MSI or tumor-site category of the case and included interaction terms in the conditional logistic regression models to estimate these stratum-specific odds ratios. Finally, we considered whether inclusion of cases recruited >2 years after diagnosis resulted in biased estimates and results were fundamentally unchanged after exclusion of these cases (data not shown).

Results

Table 1 shows the demographic characteristics of the study population. A total of 1,854 sibships were included in this study. Among the participants, 1,640 cases and 2,621 controls were population-based and 283 cases and 494 controls were clinic-based. Data for tumor site and MSI status were available for 1,778 cases and 1,242 cases respectively.

[Figure omitted. See PDF.]

Table 1. Selected characteristics of cases and sibling controls from participants in the Colon Cancer Family Registry.

https://doi.org/10.1371/journal.pone.0060464.t001

Results of the single SNP analysis are shown in Table 2. SNPs were sorted according to their position on chromosome 18. Assuming a log-additive model, a total of seven SNPs had a P value less than 0.05; after Bonferroni correction for the number of SNPs tested in the SMAD7 gene, only two SNPs, rs11874392 and rs12953717, remained significantly associated with CRC risk among the population-based families. Each minor allele (T) of rs12953717 SNP was associated with a significantly increased risk of CRC (odds ratio-OR, 1.29; 95% confidence interval-CI, 1.12–1.49), whereas each minor allele (T) of rs11874392 was associated with a statistically significantly reduced risk of CRC (OR, 0.80; 95% CI, 0.70–0.92). These associations were similar among the population-based families and the clinic-based families, though they were only significant among the population-based families with its larger sample size. The two SNPs were highly correlated with each other (D′, 0.999; R2, 0.661). In a logistic regression analysis, the inclusion of rs11874392 did not significantly improve the fit of the model over that with rs12953717 alone (P = 0.88). Analysis of rs12953717 and rs11874392 revealed that CT and TA were the most common haplotypes (Table 3) and only TA was associated with higher risk of CRC (P = 9.0×10−5). Haplotype CA was not associated with CRC risk (P = 0.53). No significant associations were observed for the rs4464148 SNP or other SNPs evaluated in SMAD7 after adjustment for multiple testing.

[Figure omitted. See PDF.]

Table 2. SMAD7 tagging SNPs and colorectal cancer risk in the Colon Cancer Family Registry.

https://doi.org/10.1371/journal.pone.0060464.t002

[Figure omitted. See PDF.]

Table 3. Colorectal cancer risk associated with haplotypes formed by SMAD7 SNPs rs12953717 and rs11874392.

https://doi.org/10.1371/journal.pone.0060464.t003

Among the population-based families, we assessed the association between rs12953717 and CRC risk after stratification by related risk/protective factors (Table 4). We observed modest and marginally significant differences in the disease association by use of NSAIDs, cigarette smoking, BMI and history of polyps. The strongest associations were observed among current NSAID users, non-/former smokers, overweight individuals (25≤BMI<30) and individuals with a prior diagnosis of polyps, respectively. There were no statistically significant differences in the SNP-CRC association according to gender, age, alcohol drinking, physical activity, diabetes, and ulcerative colitis. The effect of rs12953717 seemed to be stronger among individuals without a family history of CRC than among those with a family history; however, this difference was not statistically significant (P = 0.39). When the association between rs11874392 and risk of CRC was examined, similar but less pronounced differences by the above factors were observed.

[Figure omitted. See PDF.]

Table 4. Stratified analyses of rs12953717 and rs11874392 by known risk/protective factors among population-based families.

https://doi.org/10.1371/journal.pone.0060464.t004

We also examined the possible heterogeneity of the SNP effect by tumor location and MSI status (Table 5). The strongest associations were observed for distal colon cancer and MSI-L/MSI-H tumors than for other tumors; however, neither difference was statistically significant.

[Figure omitted. See PDF.]

Table 5. SNP rs12953713 and rs11874392 by tumor characteristics among population-based families.

https://doi.org/10.1371/journal.pone.0060464.t005

To address potential survival bias due to the inclusion of cases interviewed up to 5 years after diagnosis, we repeated the analyses of main effects including only probands interviewed within 2 years of diagnosis and their unaffected siblings and results were fundamentally unchanged (data not shown).

Discussion

Consistent with the two GWAS [3], [4], we found a statistically significant association between two SMAD7 polymorphisms, rs12953717 and rs11874392, and CRC in this large family-based case-control study. We observed suggestions that the association of rs12953717 and CRC risk may be modified by use of NSAIDs, cigarette smoking, BMI and history of polyps.

The underlying causal variant for the association between 18q21 variation and CRC risk remains unknown. Pittman et al. [11] reported that a C to G SNP at 44,703,563 bp, a SNP in strong linkage disequilibrium with other genetic variations around 18q21, may be the functional variant responsible for 18q21-associated variations in CRC risk through differential SMAD7 expression and subsequent TGF-β signaling. Compared to the C allele, the risk allele G forms weaker protein–DNA complexes with nuclear extracts and was associated with a reduced expression of SMAD7 in the colorectum. However, it remains unclear how such alterations may promote carcinogenesis. It has been suggested that SMAD7 may induce tumorigenicity by blocking TGF-β-induced growth inhibition and apoptosis [18]. The expression of SMAD7 is very low in epithelial tissues, but is up-regulated in several cancers. Over-expression of SMAD7 has been shown to inhibit TGF-β-mediated induction of endogenous heme oxygenase-1 (HO-1) gene expression [19], an adaptive defense against oxidant stress [20]. In colon cancer cells, stable expression of SMAD7 blocks the TGF-β-mediated activation of NFκB[5], a critical molecule in oxidative-stress-induced apoptosis. In addition, CRC patients with deletion of SMAD7 were found to have a favorable clinical outcome [21]. Conversely, amplification of this gene was associated with a significantly worse prognosis, with a graded effect depending on SMAD7 gene copy number [21].

Given that SMAD7 is involved in intestinal inflammation through its regulation of TGF-β signaling [22], [23] the association between SMAD7 SNPs and CRC risk may be modified by factors affecting inflammation. In human inflammatory bowel tissues, TGF-β1 signaling is disrupted by the up-regulation of SMAD7, leading to an enhanced production of inflammatory cytokines. Inhibition of SMAD7 with a specific antisense oligonucleotide can restore the immunosuppressive TGF-β1 signaling and suppress the inflammatory cytokine production [23]. In our study, there is some indication of differences in the association between rs12953717 and CRC risk by NSAID use, with the effect of the SNP being most pronounced among current users and least among former users. In a population-based case-control study of colon cancer, Slattery et al. [24] also observed slightly stronger associations of rs4939827 and rs12953717 for individuals reporting recent aspirin/NSAID use. Our findings are also consistent with those from a recent study suggesting that CRC-specific survival in postmenopausal women varied according to SMAD7 genotype (rs4939827 and rs4464148) among patients who were regularly using NSAIDs before diagnosis, but not among never-users and former users of NSAIDs [25]. Studies have shown that disruption of TGF-β1 signaling due to high levels of SMAD7 is a feature of colitis and blocking SMAD7 restores TGF- β1 signaling in colitis [26]. Since non-/former smokers have higher risk of ulcerative colitis compared with current smokers [27], our observation of a stronger SMAD7 SNP association among non-/former smokers would lend some support to the inflammation hypothesis. We consistently observed a higher CRC OR for rs12953717 among individuals with ulcerative colitis than those without, although this difference did not reach statistical significance, possibly due to the small number of individuals with ulcerative colitis. On the other hand, smoking increases the risk of the other form of IBD, Crohn's disease, but we were unable to examine the SNP-CRC association stratified by history of Crohn's disease due to the small number of individuals with this disease (N = 32). In addition, the observation of a BMI interaction could also suggest the involvement of inflammation pathways, since obesity is associated with a state of chronic inflammation [28]. However, the observation of the effect of SMAD7 genetic variants only among overweight individuals but not among obese individuals was unexpected and requires confirmation.

SMAD7 may also function through insulin-related pathways. SMAD7 protein expression in the renal cortex is decreased in diabetes [29] and conditional expression of SMAD7 in pancreatic β cells disrupts TGF-β signaling and induces reversible diabetes mellitus [30]. In addition, the SMAD7 SNP rs3764482 (IVS2 –21C>T) has been associated with a reduced risk of type 2 diabetes in mice [31]. We explored this possible mechanism by examining differences in the association of the SNP with CRC risk by diabetes status but did not find strong evidence for an interaction with diabetes (Table 3). The association between rs12953717 and CRC risk remained among non-diabetic individuals.

In addition to confirming the associations between SMAD7 SNPs and CRC risk, we further defined these associations according to other known risk/protective factors for CRC and tumor characteristics. We found a stronger SNP association among individuals with a prior history of polyps. Further exploration by the type, number and size of polyps was not performed because detailed information on these characteristics is lacking. The exact biological mechanism for the interaction between polyps and SMAD7 is unknown; however, a recent study in prostate cancer cells found that SMAD7 can interact with adenomatous polyposis coli (APC) protein in linking the TGF-β type I receptors to the microtubule system to promote cell migration [32], indicating a possible role of SMAD7 in the progression from adenomatous polyps to CRCs. Interpretation of this result needs to be cautious, because the proportion of subjects who had been examined for polyps prior to cancer diagnosis is unknown. There was some indication of differences in the effect of SMAD7 SNPs by tumor location. We observed a significant association of rs12953717 for distal colon tumors, but not for proximal colon tumors and rectal tumors. Such finding is consistent with Curtin et al. [33], who also reported significant associations for distal colon tumors for two SMAD7 SNPs but not for proximal colon and rectal tumors. Similarly, Slattery et al. [24] observed slightly stronger associations for distal than for proximal colon cancers. However, no significant difference in the effect size of rs4939827 by tumor site was reported by Broderick et al. [3]. On the contrary, Tenesa et al. [4] found that rs4939827 was more strongly associated with risk of rectal cancer than for colon cancer. Consistent with most prior studies [4], [24], [34], we did not find any significant differences in the SNP associations by age, gender, family history and MSI status.

Our study has a number of limitations. First, SNP rs4939827 was excluded from our analysis due to genotyping failure; however, given the perfect linkage disequilibrium (R2 = 1.0 in HapMap CEU data Release 28.) between rs4939827 and rs11874392, there is unlikely to have been any loss of information due to this exclusion. Second, the Bonferroni method to adjust for multiple comparisons may be so conservative that some potentially important polymorphisms, such as the variant identified in a previous GWAS of CRC rs4464148 [3], may be overlooked. On the other hand, the observed weaker associations of other SMAD7 SNPs (rs4939832, rs1316447, rs4939837, rs7240215) with CRC risk have not been widely replicated by other studies and therefore could have been chance findings. Third, we used a case-unaffected sibling study design that may reduce the power to detect association between genetic variations and CRC risk because of overmatching on genotypes between cases and their unaffected siblings [13]; however, such design does not lead to biases in estimating genetic relative risks, and is more powerful for detecting gene-environment interactions and controls for potential confounding from population stratification. And lastly, we did not observe statistically significant associations between these common SMAD7 SNPs and CRC risk in the clinic-based families, possibly due to the limited sample size in the clinic-based dataset. Alternatively, the genetic predisposition in the multiple-case, clinic-based families may be largely due to other yet-to-be-identified genetic variations with stronger effects.

In summary, using data from the Colorectal Cancer Family Registry, we confirmed the association between SMAD7 and CRC found in GWAS. Further studies are needed to confirm our results stratified by demographic factors and tumor characteristics and to elucidate the relevant biological mechanisms. With the growing epidemiological evidence linking SMAD7 to CRC susceptibility, studies are needed to investigate potential biological mechanisms by which SMAD7 contributes to the development of CRC.

Acknowledgments

The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Colon Cancer Family Registry, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the Colon Cancer Family Registry. We thank the data collection and management team and all the individuals who participated in the Colon Cancer Family Registry.

Author Contributions

Interpretation of data and revising the article critically for important intellectual content: XJ MG JEC CMR DVC JPPC JDP PAN MNP JLH MEM JAB. Conceived and designed the experiments: MG JEC RWH. Performed the experiments: XJ MG JEC DV. Analyzed the data: XJ MG JEC. Contributed reagents/materials/analysis tools: AC DVC JDP PAN MAJ JLH SG LLM MEM DJA JAB NML RWH. Wrote the paper: XJ MG JEC.

Citation: Jiang X, Castelao JE, Vandenberg D, Carracedo A, Redondo CM, Conti DV, et al. (2013) Genetic Variations in SMAD7 Are Associated with Colorectal Cancer Risk in the Colon Cancer Family Registry. PLoS ONE 8(4): e60464. https://doi.org/10.1371/journal.pone.0060464

References

1. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, et al. (2000) Environmental and heritable factors in the causation of cancer-Analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343: 78–85.

2. Tenesa A, Dunlop MG (2009) New insights into the aetiology of colorectal cancer from genome-wide association studies. Nat Rev Genet 10: 353–358.

3. Broderick P, Carvajal-Carmona L, Pittman AM, Webb E, Howarth K, et al. (2007) A genome-wide association study shows that common alleles of SMAD7 influence colorectal cancer risk. Nat Genet 39: 1315–1317.

4. Tenesa A, Farrington SM, Prendergast JG, Porteous ME, Walker M, et al. (2008) Genome-wide association scan identifies a colorectal cancer susceptibility locus on 11q23 and replicates risk loci at 8q24 and 18q21. Nat Genet 40: 631–637.

5. Grau AM, Datta PK, Zi J, Halder SK, Beauchamp RD (2006) Role of Smad proteins in the regulation of NF-kappaB by TGF-beta in colon cancer cells. Cell Signal 18: 1041–1050.

6. Korchynskyi O, Landstrom M, Stoika R, Funa K, Heldin CH, et al. (1999) Expression of Smad proteins in human colorectal cancer. Int J Cancer 82: 197–202.

7. Massague J (2012) TGFbeta signalling in context. Nat Rev Mol Cell Biol 13: 616–630.

8. Song Q, Zhu B, Hu W, Cheng L, Gong H, et al. (2012) A common SMAD7 variant is associated with risk of colorectal cancer: evidence from a case-control study and a meta-analysis. PLoS ONE 7: e33318.

9. Middeldorp A, Jagmohan-Changur S, van Eijk R, Tops C, Devilee P, et al. (2009) Enrichment of low penetrance susceptibility loci in a Dutch familial colorectal cancer cohort. Cancer Epidemiol Biomarkers Prev 18: 3062–3067.

10. Niittymäki I, Kaasinen E, Tuupanen S, Karhu A, Järvinen H, et al. (2010) Low-penetrance susceptibility variants in familial colorectal cancer. Cancer Epidemiol Biomarkers Prev 19: 1478–1483.

11. Pittman AM, Naranjo S, Webb E, Broderick P, Lips EH, et al. (2009) The colorectal cancer risk at 18q21 is caused by a novel variant altering SMAD7 expression. Genome Res 19: 987–993.

12. Loh YH, Mitrou PN, Wood A, Luben RN, McTaggart A, et al. (2011) SMAD7 and MGMT genotype variants and cancer incidence in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Norfolk Study. Cancer Epidemiol 35: 369–374.

13. Witte JS, Gauderman WJ, Thomas DC (1999) Asymptotic bias and efficiency in case-control studies of candidate genes and gene-environment interactions: basic family designs. Am J Epidemiol 149: 693–705.

14. Newcomb PA, Baron J, Cotterchio M, Gallinger S, Grove J, et al. (2007) Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol Biomarkers Prev 16: 2331–2343.

15. Levine AJ, Figueiredo JC, Lee W, Poynter JN, Conti D, et al. (2010) Genetic variability in the MTHFR gene and colorectal cancer risk using the colorectal cancer family registry. Cancer Epidemiol Biomarkers Prev 19: 89–100.

16. Edlund C, Lee W, Li D, Van Den Berg D, Conti D (2008) Snagger: A user-friendly program for incorporating additional information for tagSNP selection. BMC Bioinformatics 9: 174.

17. Shen R, Fan JB, Campbell D, Chang W, Chen J, et al. (2005) High-throughput SNP genotyping on universal bead arrays. Mutat Res 573: 70–82.

18. Halder SK, Beauchamp RD, Datta PK (2005) Smad7 induces tumorigenicity by blocking TGF-beta-induced growth inhibition and apoptosis. Exp Cell Res 307: 231–246.

19. Hill-Kapturczak N, Truong L, Thamilselvan V, Visner GA, Nick HS, et al. (2000) Smad7-dependent regulation of heme oxygenase-1 by transforming growth factor-beta in human renal epithelial cells. J Biol Chem 275: 40904–40909.

20. Stocker R (1990) Induction of haem oxygenase as a defence against oxidative stress. Free Radic Res Commun 9: 101–112.

21. Boulay JL, Mild G, Lowy A, Reuter J, Lagrange M, et al. (2003) SMAD7 is a prognostic marker in patients with colorectal cancer. Int J Cancer 104: 446–449.

22. Monteleone G, Boirivant M, Pallone F, MacDonald TT (2008) TGF-b1 and Smad7 in the regulation of IBD. Mucosal Immunol 1: S50–S53.

23. Monteleone G, Pallone F, MacDonald TT (2004) Smad7 in TGF-β-mediated negative regulation of gut inflammation. Trends in Immunology 25: 513–517.

24. Slattery ML, Herrick J, Curtin K, Samowitz W, Wolff RK, et al. (2010) Increased risk of colon cancer associated with a genetic polymorphism of SMAD7. Cancer Res 70: 1479–1485.

25. Passarelli MN, Coghill AE, Hutter CM, Zheng Y, Makar KW, et al. (2011) Common colorectal cancer risk variants in SMAD7 are associated with survival among prediagnostic nonsteroidal anti-inflammatory drug users: A population-based study of postmenopausal women. Genes Chromosomes Cancer 50: 875–886.

26. Boirivant M, Pallone F, Di Giacinto C, Fina D, Monteleone I, et al. (2006) Inhibition of Smad7 with a specific antisense oligonucleotide facilitates TGF-beta1-mediated suppression of colitis. Gastroenterology 131: 1786–1798.

27. Lakatos PL, Szamosi T, Lakatos L (2007) Smoking in inflammatory bowel diseases: good, bad or ugly? World J Gastroenterol 13: 6134–6139.

28. Gunter MJ, Leitzmann MF (2006) Obesity and colorectal cancer: epidemiology, mechanisms and candidate genes. J Nutr Biochem 17: 145–156.

29. Dixon A, Maric C (2007) 17β-Estradiol attenuates diabetic kidney disease by regulating extracellular matrix and transforming growth factor-β protein expression and signaling. Am J Physiol Renal Physiol 293: F1678–F1690.

30. Smart NG, Apelqvist ÃA, Gu X, Harmon EB, Topper JN, et al. (2006) Conditional expression of Smad7 in pancreatic beta cells disrupts TGF-beta signaling and induces reversible diabetes mellitus. PLoS Biol 4: e39.

31. Gutierrez-Aguilar R, Benmezroua Y, Balkau B, Marre M, Helbecque N, et al. (2007) Minor contribution of SMAD7 and KLF10 variants to genetic susceptibility of type 2 diabetes. Diabetes Metab 33: 372–378.

32. Ekman M, Mu Y, Lee SY, Edlund S, Kozakai T, et al. (2012) APC and Smad7 link TGFβ type I receptors to the microtubule system to promote cell migration. Mol Biol Cell 23: 2109–2121.

33. Curtin K, Lin WY, George R, Katory M, Shorto J, et al. (2009) Meta association of colorectal cancer confirms risk alleles at 8q24 and 18q21. Cancer Epidemiol Biomarkers Prev 18: 616–621.

34. Thompson CL, Plummer SJ, Acheson LS, Tucker TC, Casey G, et al. (2009) Association of common genetic variants in SMAD7 and risk of colon cancer. Carcinogenesis 30: 982–986.