The Common Variant rs4444235 near BMP4 Confers Genetic Susceptibility of Colorectal Cancer: An Updated Meta-Analysis Based on a Comprehensive Statistical Strategy

Li Liu; Qinji Su; Lixia Li; Xiaohui Lin; Yu Gan; Sidong Chen

doi:10.1371/journal.pone.0100133

Abstract

Objective

We performed an updated meta-analysis, using a comprehensive strategy of a logistic regression and a model-free approach, to evaluate more precisely the role of the rs4444235 variant near the Bone morphogenetic protein-4 (BMP4) gene in susceptibility to colorectal cancer (CRC).

Methods

A total of 19 studies with 28770 cases and 28234 controls were included. Metagen system with logistic regression was applied to choose the most plausible genetic model for rs4444235. Generalized odds ratio (OR_G) metric was used to provide a global test of relationship between rs4444235 and CRC risk.

Results

Metagen analysis suggested the rs4444235 fitted best to an additive model. In assessment of the additive model, heterogeneity was observed (P = 0.059, I² = 36.1), and pooled per-allele OR was 1.08 (95% CI = 1.05–1.11). Based on the model-free approach, pooled OR_G was 1.09 (95% CI = 1.05–1.14) under a random-effect model. Stratified analyses suggested heterogeneity could be in part explained by population ethnicity, study design, sources of controls, and sample size. Sensitivity analysis further supported the robust stability of the current results, by showing similar pooled estimates before and after sequential removal of each study.

Conclusions

This meta-analysis provides a robust estimate of the positive association between the rs4444235 and CRC risk and further emphasizes the importance of the rs4444235 in CRC risk prediction.

Citation: Liu L, Su Q, Li L, Lin X, Gan Y, Chen S (2014) The Common Variant rs4444235 near BMP4 Confers Genetic Susceptibility of Colorectal Cancer: An Updated Meta-Analysis Based on a Comprehensive Statistical Strategy. PLoS ONE 9(6): e100133. https://doi.org/10.1371/journal.pone.0100133

Editor: Xiaoping Miao, MOE Key Laboratory of Environment and Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, China

Received: January 25, 2014; Accepted: May 21, 2014; Published: June 16, 2014

Copyright: © 2014 Liu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by Program for Guangdong Medical Research Foundation (Grant No. B2012176; http://www.gdwst.gov.cn/lixiang/), Foundation for Distinguished Young Talents in Higher Education of Guangdong (Grant No. 2012LYM_0082; http://www.gdhed.edu.cn/main/index.html) and National Natural Science Foundation of China (Grant No. 81302493; http://isisn.nsfc.gov.cn/egrantweb/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Colorectal cancer is a major public health issue in developed countries and is becoming increasingly prevalent in Asia and Africa, with over 1.2 million new cases worldwide each year [1]. As other complex diseases, colorectal cancer is a complex trait driven by diverse etiologies involving in multiple environmental and genetic factors and their interactions [2]. Twin- and familial-based studies have provided clear evidence that approximately 35% of all CRC cases have a genetic component [2]. Of all CRC cases, <5% can be accounted by a combination of some germline mutations with high penetrance, whereas most “sporadic” cases are due to large numbers of common variants with individually small effects [3].

Recently, genome-wide association studies (GWAS) have implicated multiple common single nucleotide polymorphisms (SNPs) in inherited predisposition to CRC [4], [5]. The SNP rs4444235 at chromosome 14q22.2, mapping 9.4 kb upstream region of the gene encoding bone morphogenetic protein 4 (BMP4), was firstly reported by a meta-analysis of GWAS data to be associated with CRC risk, with a combined OR of 1.11 (95% CI = 1.08–1.15, P = 8.1×10⁻¹⁰) [6]. BMP4 is an important member of the BMP signaling pathway, which involves in CRC development through regulation of colorectal stem cell differentiation [7]. This SNP has been proposed to act as a cis-regulator of BMP4 and thus conferred to CRC risk [6]. However, the following replication studies yielded inconsistent results, in part due to “winner curse” in the original report [8], “Proteus phenomenon” in replication data [9], heterogeneous ethnical population, and insufficient statistical power, among other issues.

Meta-analysis, by integrating published data, may be a powerful tool to clarify the inconsistencies across individual studies. Two meta-analyses have been performed to assess rs4444235 in CRC. The meta-analysis by Li et al. [10], including 19893 cases and 22106 controls, assessed multiple genetic models for the rs4444235, which would lead to multiple comparisons or erroneous mode specification without priori biological evidence. The other meta-analysis by Theodoratou et al. [11], including less samples (18607 cases and 19576 controls), utilized a maximum likelihood estimator to decipher plausible model for the rs4444235. However, in this meta-analysis, there was no subgroup analysis undertaken. To overcome the above mentioned shortcomings in the previous meta-analyses, we integrated published data from 28770 cases and 28234 controls, and performed an updated meta-analysis, using a comprehensive statistical strategy. The methodology of logistic regression was applied to estimate the most plausible genetic model in the metagen system [12]. The generalized odds ratio, based on model-free approach, was utilized to provide a global test of genetic association [13]. Stratified analyses were further performed to explore potential sources of heterogeneity. The core aim of this meta-analysis was to provide a more precise and robust evaluation for the role of rs4444235 polymorphism in genetic susceptibility of colorectal cancer.

Materials and Methods

Search Strategy and Identification of Relevant Studies

This meta-analysis were conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement (Checklist S1) [14]. Genetic association studies regarding rs4444235 and colorectal cancer (CRC) risk were searched in the PubMed/MEDLINE and EMBASE databases through October 15, 2013, by using the combinations of the keywords: (“BMP4” or “rs4444235” or “14q22.2”) and (“colorectal cancer” or “Colorectal neoplasmor” or “colon cancer” or “rectal cancer”). The similar search terms was also used for the WANFANG DATA and CNKI databases. The search was supplemented by review of reference lists for all relevant studies and review articles. All relevant reports identified were included without language restriction.

The following inclusion criteria should be fulfilled: (1) either case-control or nested case-control studies; (2) clear definition of colorectal cancer cases; (3) studies evaluating relationship between rs4444235 and CRC risk; (4) providing sufficient data to re-calculate the effect metrics, that was, numbers of genotypes in cases and controls. The authors were contacted via E-mail when eligible articles reported insufficient data. If they were unable to provide detailed data, those articles were excluded. Animal studies, reviews, conference abstracts, editorials and letters were excluded. If more than one ethnical population were in one report, each population was considered separately. Studies overlapping with other studies should be excluded, and the one with the most completed information was included. The first study on the association of rs4444235 by Houlston et al. was excluded [6], due to overlaps with the study by Tomlinson et al. [5]. The latter was chosen because of the larger sample.

Data Extraction

Data were extracted independently and in duplicate by 2 reviewers (L. Liu & Q. Su). The following data was extracted from each article according to a fixed protocol: the first author, publication year, study design, country, ethnicity, source of controls, numbers of cases and controls, mean age of cases, sex ratio, site/type of colorectal cancer, genotyping method, minor allele frequency (MAF), and frequency of genotypes in cases and controls.

Statistical Analysis

Hardy-Weinberg equilibrium in controls was re-analyzed using the goodness-of-fit χ² test (P>0.05). The inverse variance method was applied to estimate the pooled frequency of the risk allele (the C allele) in various ethnical populations. The genetic effect of the rs4444235 in CRC susceptibility was assessed using the approaches described as below:

Metagen system has provided a general framework to decipher the most plausible genetic model for the rs4444235 that treated the genotypes as independent variables in a logistic regression under both fixed and random effects models [12]. Under fixed-effect model, two parameters, θ₂ and θ₃ were estimated using the logistic regression: logit (π_ij) = α_i +θ₂z_i2+θ₃z_i3, where α_i was the indicator of study-specific fixed-effect, OR_TC/TT = exp(θ₂), and OR_CC/TT = exp(θ₃). In order to account for an additive component of heterogeneity, a random-effect logistic regression was performed using the GLLAMM module in STATA software via introducing a study-specific random coefficient: logit (π_ij) = α_i +(θ₂+ν_i2)z_i2+(θ₃₊ν_i3)z_i3. The most plausible genetic model was determined using the following procedure: if θ₂ = θ₃ = 0, no significant genetic-association was suggested; if θ₂ = 0 and θ₃>0, a recessive genetic model was suggested; if θ₂ = θ₃>0, a dominant model was suggested; if θ₃>θ₂>0, a co-dominant model was suggested; if 2θ₂ = θ₃, an additive model was likely. In this meta-analysis, the genetic model of rs4444235 was best fitted with an additive model. Then the per-allele OR of the C allele (additive model ) with corresponding 95% confidence interval (95% CI) was estimated in a logistic regression model, by assigning scores of 0, 1, and 2 to the AA, AC and CC genotypes, respectively. Between-study heterogeneity was assessed by the Cochran’s χ² based Q test and I² metric. If there was no heterogeneity (i.e., if the Q test was significant [P<0.1] or I² was less than 25%), a fixed-effect model was used to pool the estimate; otherwise, a random-effect model was applied. To explore the sources of heterogeneity, stratified analyses were performed, if feasible, according to population ethnicity (Asians, Caucasians, and Africans), sources of controls (population- and hospital-based), study design (GWAS and replication study), and total sample size (≤2000 and >2000).

Additionally, the generalized OR (OR_G), based on a genetic model-free approach, was also introduced in this meta-analysis [13]. The OR_G utilized the complete genotype distribution to provide an estimate of overall gene-disease relationship, given that the mutational load was treated as a graded exposure. Heterogeneity was also assessed for OR_G metric and stratified analysis was also performed.

Sensitivity analysis was performed to assess the influence of single study on pooled estimates. Publication bias was tested by the Egger’s regression test and Begg’s funnel plot. Statistical analyses were conducted in ORGGASMA, metan and metagen modules in STATA software version 13.0. A P value of <0.05 was considered statistically significant, except for estimation of between-study heterogeneity, where a significant level of 0.10 was applied.

Results

The Characteristic of Included Studies

Figure 1 shows a flow diagram of the study selection process. The comprehensive search yielded 56 potentially relevant references. 18 articles were determined to be initially eligible by screening titles and abstracts. After further detailed evaluation, 7 duplicated articles [6], [15], [16], [17], [18], [19], [20] and 3 articles with insufficient data [20], [21], [22], [23] were excluded. 1 article was excluded due to small sample size (92 cases and 96 controls) [24]. 1 study in the article by Tomlinson et al. was excluded due to deviation with Hardy-Weinberg equilibrium [5]. Finally, a total of 7 articles with 19 studies of 28770 cases and 28234 controls were included in this meta-analysis [5], [25], [26], [27], [28], [29], [30]. The characteristics of these studies were summarized in Table 1. Among the included studies, 15 studies were performed in Caucasians, 3 studies in Asians, and 1 study in Africans.

Download:

Figure 1. Flow chart of study selection process.

https://doi.org/10.1371/journal.pone.0100133.g001

Download:

Table 1. Characteristics of included studies in the meta-analysis of rs4444235 and colorectal cancer.

https://doi.org/10.1371/journal.pone.0100133.t001

Pooled Frequency of the Risk Allele (the C Allele) in Controls According to Ethnicity

Significant heterogeneity was seen both in Caucasians and Asians. and thus the random-effect model was applied (all P<0.0001, I² = 82.21 and 92.40, respectively). The pooled frequency of the C allele was 0.463 (95% CI = 0.452–0.474) in Caucasians, similar to that of 0.477 (95% CI = 0.423–0.532) in Asians. Only 1 study was conducted in Africans, and the frequency of the C allele was 0.334.

Overall Meta-analysis of the rs4444235 and Colorectal Cancer Risk

Table 2 summarizes the results of overall meta-analysis. In the metagen analysis, the pooled OR_TC/TT and OR_CC/TT were 1.08 (95% CI = 1.03–1.12) and 1.18 (95% CI = 1.12–1.25), respectively, suggesting an additive model as the most plausible genetic model. Then the additive model for the rs4444235 was assessed using traditional method. In the additive model, heterogeneity was observed (P = 0.059, I² = 36.1), and thus the random-effect model was applied. The variant was significantly associated with increased CRC risk, with a pooled per-allele OR of 1.08 (95% CI = 1.05–1.11; Figure 2). Based on the model-free approach, heterogeneity was also seen (P = 0.063, I² = 35.6). Under the random-effect model, significant result was also produced for the association of rs4444235 and CRC risk, with a pooled OR_G of 1.09 (95% CI = 1.05–1.14).

Download:

Figure 2. The forest plot of the association between rs4444235 and colorectal cancer risk in the additive model.

https://doi.org/10.1371/journal.pone.0100133.g002

Download:

Table 2. Meta-analysis of rs4444235 and colorectal cancer risk.

https://doi.org/10.1371/journal.pone.0100133.t002

Stratification Analysis of the rs4444235 and Colorectal Cancer Risk

When performed stratified analysis by population ethnicity, in Caucasian subgroup of 15 studies, heterogeneity was removed, and the significant association of the rs4444235 still existed for both additive model and OR_G assessment (Table 2). However, in Asians of 3 studies, there was significant heterogeneity (P = 0.040 and 0.041 for additive model and OR_G, respectively), and no significant association was found.

According to the sources of controls, in the population-based subgroup of 13 studies, analysis of the additive model and OR_G both showed significant association of rs4444235 with CRC without evidence of heterogeneity, whereas in the hospital-based subgroup of 8 studies, significant heterogeneity was observed and no significant association was reported.

Regarding to study design, there were 6 GWAS and 13 replication studies. When assessing the additive model and OR_G metric, both subgroups showed the positive genetic association with CRC risk, without evidence of heterogeneity. Interestingly, the pooled estimates in the GWAS (per-allele OR = 1.12; OR_G = 1.14) were slightly larger than those in the subgroup of replication studies (per-allele OR = 1.06; OR_G = 1.07).

The stratified analysis was also conducted according to total sample size (numbers of both cases and controls), into 2 subgroups: the large sample size subgroup (total sample size >2000) with 22064 cases and 20876 controls and the small or moderate size subgroup (total sample size ≤2000) with 6706 cases and 7358 controls. For both additive model and OR_G analyses, heterogeneity was removed in the subgroup with large sample size, whereas in the small or moderate size subgroup, heterogeneity still existed. Both subgroups showed the significant association between the rs4444235 and CRC risk.

Sensitivity Analysis and Publication Bias Assessment

Since between-study heterogeneity was observed in this meta-analysis, we further performed sensitivity analysis under the random-effect model. For the additive model, the sensitivity analysis, by sequentially omitting each study, reported a series of pooled OR with 95% CI exceeding 1.00, and the pooled ORs were similar before and after omitting each study (Table 3). Similar results were suggested for OR_G analysis that no single study significantly altered the pooled OR_G. In the Begg’s and the Egger’s tests, there was no evidence of publication bias for both additive model and OR_G (all P values for Begg’s and Egger’s tests >0.05).

Download:

Table 3. Sensitivity analysis of rs4444235 and colorectal cancer risk.

https://doi.org/10.1371/journal.pone.0100133.t003

Discussion

Currently, traditional meta-analyses of genetic association studies are usually performed by collapsing genotypes in two categories assuming various genetic models. However, these different models are not independent, and a priori biological justification for the choice of a specific model is seldom available [31]. Additionally, interpretation of these results is complicated since a set of different estimates and significance tests are usually provided. In this current meta-analysis of rs4444235 and colorectal cancer risk, we utilized a comprehensive strategy, including the metagen analysis based on logistic regression and OR_G metric based on model-free approach [12], [13], to overcome the drawbacks in traditional meta-analysis of erroneous model specification and multiple model tests with an inflated Type I error rate, and make the interpretation of the current results easier.

In this meta-analysis of 19 case-control studies of 28770 cases and 28234 controls, the metagen analysis indicated that the rs4444235 fitted best to an additive model. Knowledge of the best-fitting model for the rs4444235 may be important in optimizing the use of this SNP in colorectal cancer (CRC) risk prediction. Assessment of additive model indicated that CRC risk was increased by 8% per extra C allele. Based on model-free approach, the generalized OR (OR_G) analysis showed that CRC cases with higher mutational load than healthy individuals have 9% higher risk for CRC susceptibility. Sensitivity analysis further supported the current results, by showing similar ORs before and after sequentially omitting single study. The positive association of the rs4444235 with CRC risk identified by this meta-analysis was also concordant with the findings of previous meta-analyses [10], [11].

rs4444235 is 9.4 kb from the transcription start site of the BMP4. The BMP signaling has vital function in maintenance of Wnt signaling to inhibit differentiation of stem cell near colorectal crypt bases [7]. Heightened expression of BMP pathway members would restrain the Wnt signaling, subsequently activate β-catenin and elevate cells susceptibility to tumor-causing mutations, and ultimately promote colorectal carcinogenesis [7]. Intriguingly, in a recent study, luciferase reporter assay suggested the element to which rs4444235 maps acts as an allele-specific transcriptional enhancer [23]. In CRC cell lines allele-specific expression analysis indicated a significant association of increased BMP4 expression with the C allele [23]. These data have strongly supported the functional role of rs4444235 in CRC development through the cis-acting regulatory influence on BMP4 expression.

Heterogeneity is a pervasive and difficult problem in meta-analysis of genetic association studies. Not surprisingly, heterogeneity existed in this meta-analysis, and thus the findings should be interpreted with caution. Nevertheless, in stratified analysis by ethnicity, heterogeneity was removed in Caucasians and significant association of rs4444235 retained. According to study design, both in GWAS and replication studies, heterogeneity was effectively decreased, and association was also existed. Interestingly, the subgroup of GWAS yielded larger pooled ORs than that in replication data, indicating “winner curse” existed for the rs4444235 in GWAS. In regarding to sample size, only in the subgroup with large sample size heterogeneity was removed, but both subgroups showed significant genetic association. When stratified by sources of controls, heterogeneity was removed in population-based subgroup. These findings suggested the heterogeneity could be in part explained by the distinct natures of population ethnicity, control sources, study design, and sample size across individual studies. Furthermore, no single study had significant influence on the overall estimates in sensitivity analysis, and no publication bias was observed in this meta-analysis, suggesting the robust stability of the current results.

Despite the strength of this study utilizing a comprehensive statistical strategy, some limitations merit serious consideration. In stratified analysis by ethnicity, majority of studies were conducted in Caucasians, only 3 studies and 1 study appraised rs4444235 in Asians and Africans, respectively. No association was seen in Asians and Africans possibly due to small sample size and insufficient power. The relationship of rs4444235 and CRC risk merits more studies in various populations. Only one polymorphism was assessed in this meta-analysis, and this meta-analysis did not give a global view of the genetic variants of BMP4 in CRC susceptibility. Additionally, gene-environment interactions did play more important role in colorectal carcinogenesis as compared with genetic factors [32]. However, only one study so far by Hutter et al. has explored interaction of rs4444235 and environmental factors [33], and thus the interaction could not be appraised in this meta-analysis.

In conclusion, this updated meta-analysis, utilizing a comprehensive strategy, further supports the significant role of rs4444235 in genetic susceptibility of colorectal cancer. Further functional polymorphism-based studies in the whole BMP4 gene are warranted to confirm and extend the current findings in various ethnical populations.

Supporting Information

Checklist S1.

Checklist of Preferred Reporting Items for Systematic Reviews and Meta-analyses statement.

https://doi.org/10.1371/journal.pone.0100133.s001

(DOC)

Author Contributions

Conceived and designed the experiments: LL. Performed the experiments: LL QJS LXL. Analyzed the data: LL QJS. Contributed reagents/materials/analysis tools: LXL XHL YG. Wrote the paper: LL. Checked and modified the manuscript: XHL YG SDC.

References

1. Jemal A, Bray F (2011) Center MM, Ferlay J, Ward E, et al (2011) Global cancer statistics. CA Cancer J Clin 61: 69–90.
- View Article
- Google Scholar
2. Al-Sohaily S, Biankin A, Leong R, Kohonen-Corish M, Warusavitarne J (2012) Molecular pathways in colorectal cancer. J Gastroenterol Hepatol 27: 1423–1431.
- View Article
- Google Scholar
3. Hughes MR, Huang EH (2011) Moleuclar basis of hereditary colorectal cancer. Semin Colon Rectal Surg 22: 65–70.
- View Article
- Google Scholar
4. Peters U, Jiao S, Schumacher FR, Hutter CM, Aragaki AK, et al. (2013) Identification of Genetic Susceptibility Loci for Colorectal Tumors in a Genome-Wide Meta-analysis. Gastroenterology 144: 799–807.e724.
- View Article
- Google Scholar
5. Tomlinson IP, Carvajal-Carmona LG, Dobbins SE, Tenesa A, Jones AM, et al. (2011) Multiple common susceptibility variants near BMP pathway loci GREM1, BMP4, and BMP2 explain part of the missing heritability of colorectal cancer. PLoS Genet 7: e1002105.
- View Article
- Google Scholar
6. Houlston RS, Webb E, Broderick P, Pittman AM, Di Bernardo MC, et al. (2008) Meta-analysis of genome-wide association data identifies four new susceptibility loci for colorectal cancer. Nat Genet 40: 1426–1435.
- View Article
- Google Scholar
7. Hardwick JC, Kodach LL, Offerhaus GJ, van den Brink GR (2008) Bone morphogenetic protein signalling in colorectal cancer. Nat Rev Cancer 8: 806–812.
- View Article
- Google Scholar
8. Zhong H, Prentice RL (2010) Correcting “winner’s curse” in odds ratios from genomewide association findings for major complex human diseases. Genet Epidemiol 34: 78–91.
- View Article
- Google Scholar
9. Ioannidis JP (2005) Why most published research findings are false. PLoS Med 2: e124.
- View Article
- Google Scholar
10. Li J, Sun C, Yuan Y, Liu L, Xiong G, et al. (2012) Bone morphogenetic protein-4 polymorphism and colorectal cancer risk: a meta analysis. Mol Biol Rep 39: 5239–5251.
- View Article
- Google Scholar
11. Theodoratou E, Montazeri Z, Hawken S, Allum GC, Gong J, et al. (2012) Systematic meta-analyses and field synopsis of genetic association studies in colorectal cancer. J Natl Cancer Inst 104: 1433–1457.
- View Article
- Google Scholar
12. Bagos PG, Nikolopoulos GK (2007) A method for meta-analysis of case-control genetic association studies using logistic regression. Stat Appl Genet Mol Biol 6: Article17.
- View Article
- Google Scholar
13. Zintzaras E (2010) The generalized odds ratio as a measure of genetic risk effect in the analysis and meta-analysis of association studies. Stat Appl Genet Mol Biol 9: Article21.
- View Article
- Google Scholar
14. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6: e1000097.
- View Article
- Google Scholar
15. Lubbe SJ, Whiffin N, Chandler I, Broderick P, Houlston RS (2012) Relationship between 16 susceptibility loci and colorectal cancer phenotype in 3146 patients. Carcinogenesis 33: 108–112.
- View Article
- Google Scholar
16. Niittymaki I, Kaasinen E, Tuupanen S, Karhu A, Jarvinen H, et al. (2010) Low-penetrance susceptibility variants in familial colorectal cancer. Cancer Epidemiol Biomarkers Prev 19: 1478–1483.
- View Article
- Google Scholar
17. Niittymaki I, Tuupanen S, Li Y, Jarvinen H, Mecklin JP, et al. (2011) Systematic search for enhancer elements and somatic allelic imbalance at seven low-penetrance colorectal cancer predisposition loci. BMC Med Genet 12: 23.
- View Article
- Google Scholar
18. Lubbe SJ, Di Bernardo MC, Broderick P, Chandler I, Houlston RS (2012) Comprehensive evaluation of the impact of 14 genetic variants on colorectal cancer phenotype and risk. Am J Epidemiol 175: 1–10.
- View Article
- Google Scholar
19. Win AK, Hopper JL, Buchanan DD, Young JP, Tenesa A, et al. (2013) Are the common genetic variants associated with colorectal cancer risk for DNA mismatch repair gene mutation carriers? Eur J Cancer 49: 1578–1587.
- View Article
- Google Scholar
20. Fernandez-Rozadilla C, Palles C, Carvajal-Carmona L, Peterlongo P, Nici C, et al. (2013) BMP2/BMP4 colorectal cancer susceptibility loci in northern and southern european populations. Carcinogenesis 34: 314–318.
- View Article
- Google Scholar
21. Thean LF, Li HH, Teo YY, Koh WP, Yuan JM, et al. (2012) Association of Caucasian-identified variants with colorectal cancer risk in Singapore Chinese. PLoS One 7: e42407.
- View Article
- Google Scholar
22. He J, Wilkens LR, Stram DO, Kolonel LN, Henderson BE, et al. (2011) Generalizability and epidemiologic characterization of eleven colorectal cancer GWAS hits in multiple populations. Cancer Epidemiol Biomarkers Prev 20: 70–81.
- View Article
- Google Scholar
23. Lubbe SJ, Pittman AM, Olver B, Lloyd A, Vijayakrishnan J, et al. (2012) The 14q22.2 colorectal cancer variant rs4444235 shows cis-acting regulation of BMP4. Oncogene 31: 3777–3784.
- View Article
- Google Scholar
24. Mates IN, Csiki I, Mates D, Constantinescu V, Badea P, et al. (2010) Association of common genetic variants with colorectal cancer risk in a Romanian sample. Chirurgia (Bucur) 105: 749–757.
- View Article
- Google Scholar
25. Fernandez-Rozadilla C, de Castro L, Clofent J, Brea-Fernandez A, Bessa X, et al. (2010) Single nucleotide polymorphisms in the Wnt and BMP pathways and colorectal cancer risk in a Spanish cohort. PLoS One 5: e12673.
- View Article
- Google Scholar
26. von Holst S, Picelli S, Edler D, Lenander C, Dalen J, et al. (2010) Association studies on 11 published colorectal cancer risk loci. Br J Cancer 103: 575–580.
- View Article
- Google Scholar
27. Xiong F, Wu C, Bi X, Yu D, Huang L, et al. (2010) Risk of genome-wide association study-identified genetic variants for colorectal cancer in a Chinese population. Cancer Epidemiol Biomarkers Prev 19: 1855–1861.
- View Article
- Google Scholar
28. Ho JW, Choi SC, Lee YF, Hui TC, Cherny SS, et al. (2011) Replication study of SNP associations for colorectal cancer in Hong Kong Chinese. Br J Cancer 104: 369–375.
- View Article
- Google Scholar
29. Kupfer SS, Anderson JR, Hooker S, Skol A, Kittles RA, et al.. (2010) Genetic heterogeneity in colorectal cancer associations between African and European americans. Gastroenterology 139: 1677–1685, 1685 e1671–1678.
30. Li FX, Yang XX, Hu NY, Du HY, Ma Q, et al. (2012) Single-nucleotide polymorphism associations for colorectal cancer in southern chinese population. Chin J Cancer Res 24: 29–35.
- View Article
- Google Scholar
31. Kavvoura FK, Ioannidis JP (2008) Methods for meta-analysis in genetic association studies: a review of their potential and pitfalls. Hum Genet 123: 1–14.
- View Article
- Google Scholar
32. Zhong R, Liu L, Zou L, Sheng W, Zhu B, et al. (2013) Genetic variations in the TGFbeta signaling pathway, smoking and risk of colorectal cancer in a Chinese population. Carcinogenesis 34: 936–942.
- View Article
- Google Scholar
33. Hutter CM, Chang-Claude J, Slattery ML, Pflugeisen BM, Lin Y, et al. (2012) Characterization of gene-environment interactions for colorectal cancer susceptibility loci. Cancer Res 72: 2036–2044.
- View Article
- Google Scholar

[ref1] 1. Jemal A, Bray F (2011) Center MM, Ferlay J, Ward E, et al (2011) Global cancer statistics. CA Cancer J Clin 61: 69–90.
View Article
Google Scholar

[2] View Article

[3] Google Scholar

[ref2] 2. Al-Sohaily S, Biankin A, Leong R, Kohonen-Corish M, Warusavitarne J (2012) Molecular pathways in colorectal cancer. J Gastroenterol Hepatol 27: 1423–1431.
View Article
Google Scholar

[5] View Article

[6] Google Scholar

[ref3] 3. Hughes MR, Huang EH (2011) Moleuclar basis of hereditary colorectal cancer. Semin Colon Rectal Surg 22: 65–70.
View Article
Google Scholar

[8] View Article

[9] Google Scholar

[ref4] 4. Peters U, Jiao S, Schumacher FR, Hutter CM, Aragaki AK, et al. (2013) Identification of Genetic Susceptibility Loci for Colorectal Tumors in a Genome-Wide Meta-analysis. Gastroenterology 144: 799–807.e724.
View Article
Google Scholar

[11] View Article

[12] Google Scholar

[ref5] 5. Tomlinson IP, Carvajal-Carmona LG, Dobbins SE, Tenesa A, Jones AM, et al. (2011) Multiple common susceptibility variants near BMP pathway loci GREM1, BMP4, and BMP2 explain part of the missing heritability of colorectal cancer. PLoS Genet 7: e1002105.
View Article
Google Scholar

[14] View Article

[15] Google Scholar

[ref6] 6. Houlston RS, Webb E, Broderick P, Pittman AM, Di Bernardo MC, et al. (2008) Meta-analysis of genome-wide association data identifies four new susceptibility loci for colorectal cancer. Nat Genet 40: 1426–1435.
View Article
Google Scholar

[17] View Article

[18] Google Scholar

[ref7] 7. Hardwick JC, Kodach LL, Offerhaus GJ, van den Brink GR (2008) Bone morphogenetic protein signalling in colorectal cancer. Nat Rev Cancer 8: 806–812.
View Article
Google Scholar

[20] View Article

[21] Google Scholar

[ref8] 8. Zhong H, Prentice RL (2010) Correcting “winner’s curse” in odds ratios from genomewide association findings for major complex human diseases. Genet Epidemiol 34: 78–91.
View Article
Google Scholar

[23] View Article

[24] Google Scholar

[ref9] 9. Ioannidis JP (2005) Why most published research findings are false. PLoS Med 2: e124.
View Article
Google Scholar

[26] View Article

[27] Google Scholar

[ref10] 10. Li J, Sun C, Yuan Y, Liu L, Xiong G, et al. (2012) Bone morphogenetic protein-4 polymorphism and colorectal cancer risk: a meta analysis. Mol Biol Rep 39: 5239–5251.
View Article
Google Scholar

[29] View Article

[30] Google Scholar

[ref11] 11. Theodoratou E, Montazeri Z, Hawken S, Allum GC, Gong J, et al. (2012) Systematic meta-analyses and field synopsis of genetic association studies in colorectal cancer. J Natl Cancer Inst 104: 1433–1457.
View Article
Google Scholar

[32] View Article

[33] Google Scholar

[ref12] 12. Bagos PG, Nikolopoulos GK (2007) A method for meta-analysis of case-control genetic association studies using logistic regression. Stat Appl Genet Mol Biol 6: Article17.
View Article
Google Scholar

[35] View Article

[36] Google Scholar

[ref13] 13. Zintzaras E (2010) The generalized odds ratio as a measure of genetic risk effect in the analysis and meta-analysis of association studies. Stat Appl Genet Mol Biol 9: Article21.
View Article
Google Scholar

[38] View Article

[39] Google Scholar

[ref14] 14. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6: e1000097.
View Article
Google Scholar

[41] View Article

[42] Google Scholar

[ref15] 15. Lubbe SJ, Whiffin N, Chandler I, Broderick P, Houlston RS (2012) Relationship between 16 susceptibility loci and colorectal cancer phenotype in 3146 patients. Carcinogenesis 33: 108–112.
View Article
Google Scholar

[44] View Article

[45] Google Scholar

[ref16] 16. Niittymaki I, Kaasinen E, Tuupanen S, Karhu A, Jarvinen H, et al. (2010) Low-penetrance susceptibility variants in familial colorectal cancer. Cancer Epidemiol Biomarkers Prev 19: 1478–1483.
View Article
Google Scholar

[47] View Article

[48] Google Scholar

[ref17] 17. Niittymaki I, Tuupanen S, Li Y, Jarvinen H, Mecklin JP, et al. (2011) Systematic search for enhancer elements and somatic allelic imbalance at seven low-penetrance colorectal cancer predisposition loci. BMC Med Genet 12: 23.
View Article
Google Scholar

[50] View Article

[51] Google Scholar

[ref18] 18. Lubbe SJ, Di Bernardo MC, Broderick P, Chandler I, Houlston RS (2012) Comprehensive evaluation of the impact of 14 genetic variants on colorectal cancer phenotype and risk. Am J Epidemiol 175: 1–10.
View Article
Google Scholar

[53] View Article

[54] Google Scholar

[ref19] 19. Win AK, Hopper JL, Buchanan DD, Young JP, Tenesa A, et al. (2013) Are the common genetic variants associated with colorectal cancer risk for DNA mismatch repair gene mutation carriers? Eur J Cancer 49: 1578–1587.
View Article
Google Scholar

[56] View Article

[57] Google Scholar

[ref20] 20. Fernandez-Rozadilla C, Palles C, Carvajal-Carmona L, Peterlongo P, Nici C, et al. (2013) BMP2/BMP4 colorectal cancer susceptibility loci in northern and southern european populations. Carcinogenesis 34: 314–318.
View Article
Google Scholar

[59] View Article

[60] Google Scholar

[ref21] 21. Thean LF, Li HH, Teo YY, Koh WP, Yuan JM, et al. (2012) Association of Caucasian-identified variants with colorectal cancer risk in Singapore Chinese. PLoS One 7: e42407.
View Article
Google Scholar

[62] View Article

[63] Google Scholar

[ref22] 22. He J, Wilkens LR, Stram DO, Kolonel LN, Henderson BE, et al. (2011) Generalizability and epidemiologic characterization of eleven colorectal cancer GWAS hits in multiple populations. Cancer Epidemiol Biomarkers Prev 20: 70–81.
View Article
Google Scholar

[65] View Article

[66] Google Scholar

[ref23] 23. Lubbe SJ, Pittman AM, Olver B, Lloyd A, Vijayakrishnan J, et al. (2012) The 14q22.2 colorectal cancer variant rs4444235 shows cis-acting regulation of BMP4. Oncogene 31: 3777–3784.
View Article
Google Scholar

[68] View Article

[69] Google Scholar

[ref24] 24. Mates IN, Csiki I, Mates D, Constantinescu V, Badea P, et al. (2010) Association of common genetic variants with colorectal cancer risk in a Romanian sample. Chirurgia (Bucur) 105: 749–757.
View Article
Google Scholar

[71] View Article

[72] Google Scholar

[ref25] 25. Fernandez-Rozadilla C, de Castro L, Clofent J, Brea-Fernandez A, Bessa X, et al. (2010) Single nucleotide polymorphisms in the Wnt and BMP pathways and colorectal cancer risk in a Spanish cohort. PLoS One 5: e12673.
View Article
Google Scholar

[74] View Article

[75] Google Scholar

[ref26] 26. von Holst S, Picelli S, Edler D, Lenander C, Dalen J, et al. (2010) Association studies on 11 published colorectal cancer risk loci. Br J Cancer 103: 575–580.
View Article
Google Scholar

[77] View Article

[78] Google Scholar

[ref27] 27. Xiong F, Wu C, Bi X, Yu D, Huang L, et al. (2010) Risk of genome-wide association study-identified genetic variants for colorectal cancer in a Chinese population. Cancer Epidemiol Biomarkers Prev 19: 1855–1861.
View Article
Google Scholar

[80] View Article

[81] Google Scholar

[ref28] 28. Ho JW, Choi SC, Lee YF, Hui TC, Cherny SS, et al. (2011) Replication study of SNP associations for colorectal cancer in Hong Kong Chinese. Br J Cancer 104: 369–375.
View Article
Google Scholar

[83] View Article

[84] Google Scholar

[ref29] 29. Kupfer SS, Anderson JR, Hooker S, Skol A, Kittles RA, et al.. (2010) Genetic heterogeneity in colorectal cancer associations between African and European americans. Gastroenterology 139: 1677–1685, 1685 e1671–1678.

[ref30] 30. Li FX, Yang XX, Hu NY, Du HY, Ma Q, et al. (2012) Single-nucleotide polymorphism associations for colorectal cancer in southern chinese population. Chin J Cancer Res 24: 29–35.
View Article
Google Scholar

[87] View Article

[88] Google Scholar

[ref31] 31. Kavvoura FK, Ioannidis JP (2008) Methods for meta-analysis in genetic association studies: a review of their potential and pitfalls. Hum Genet 123: 1–14.
View Article
Google Scholar

[90] View Article

[91] Google Scholar

[ref32] 32. Zhong R, Liu L, Zou L, Sheng W, Zhu B, et al. (2013) Genetic variations in the TGFbeta signaling pathway, smoking and risk of colorectal cancer in a Chinese population. Carcinogenesis 34: 936–942.
View Article
Google Scholar

[93] View Article

[94] Google Scholar

[ref33] 33. Hutter CM, Chang-Claude J, Slattery ML, Pflugeisen BM, Lin Y, et al. (2012) Characterization of gene-environment interactions for colorectal cancer susceptibility loci. Cancer Res 72: 2036–2044.
View Article
Google Scholar

[96] View Article

[97] Google Scholar

Figures

Abstract

Objective

Methods

Results

Conclusions

Introduction

Materials and Methods

Search Strategy and Identification of Relevant Studies

Data Extraction

Statistical Analysis

Results

The Characteristic of Included Studies

Pooled Frequency of the Risk Allele (the C Allele) in Controls According to Ethnicity

Overall Meta-analysis of the rs4444235 and Colorectal Cancer Risk

Stratification Analysis of the rs4444235 and Colorectal Cancer Risk

Sensitivity Analysis and Publication Bias Assessment

Discussion

Supporting Information

Checklist S1.

Author Contributions

References