Figures
Abstract
Background
Insulin receptor substrate-2 (IRS-2), a signaling adaptor protein, was involved in two cancer-related pathways (the phosphatidylinositol 3′-kinase (PI3K) and the extracellular signal-regulated kinase (ERK) pathways). Several studies have evaluated the association between IRS2 rs1805097 (G>A) polymorphisms and the risk of colorectal and breast cancer. However, the results were inconsistent.
Methodology/Principal Findings
A meta-analysis of seven published case-control studies (4 studies with 4798 cases and 5478 controls for colorectal cancer and 3 studies with 2108 cases and 2507 controls for breast cancer) were conducted to assess the strength of association using crude odd ratios (ORs) with 95% confidence intervals (CIs). For colorectal cancer, no obvious associations were found for all genetic models (homozygote comparison OR = 0.96, 95%CI 0.85–1.08, Pheterogeneity = 0.97; heterozygote comparison: OR = 0.91, 95%CI 0.73–1.13, Pheterogeneity<0.01; dominant model: OR = 0.92, 95%CI 0.80–1.06, Pheterogeneity = 0.05; recessive model: OR = 1.02, 95%CI 0.91–1.14, Pheterogeneity = 0.60). In the subgroup analysis by ethnicity, control source and consistency of frequency with Hardy-Weinberg equilibrium (HWE), still no significant associations were observed. For breast cancer, also no obvious associations were found for all genetic models (homozygote comparison: OR = 0.95, 95%CI 0.71–1.26, Pheterogeneity = 0.10; heterozygote comparison: OR = 1.00, 95%CI 0.89–1.14, Pheterogeneity = 0.71; dominant model: OR = 0.98, 95%CI 0.87–1.10, Pheterogeneity = 0.55; recessive model: OR = 0.95, 95%CI 0.72–1.25, Pheterogeneity = 0.07). We performed subgroup analyses by sample size and did not find an association.
Citation: Hu Y, Zhou M, Zhang K, Kong X, Hu X, Li K, et al. (2014) Lack of Association between Insulin Receptor Substrate2 rs1805097 Polymorphism and the Risk of Colorectal and Breast Cancer: A Meta-Analysis. PLoS ONE 9(1): e86911. https://doi.org/10.1371/journal.pone.0086911
Editor: Xun Ai, Loyola University Chicago, United States of America
Received: August 10, 2013; Accepted: December 16, 2013; Published: January 30, 2014
Copyright: © 2014 Hu 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 research was supported by the National Natural Science Foundation of China (No. 81171979). 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
Insulin receptor substrates (IRs) are signaling adaptor proteins consisting of six members (IRS-1-6) [1], [2]. Among the six family members, insulin receptor substrates 1 and 2(IRS-1and IRS-2) are the most thoroughly characterized members, owing to their wide tissue expression in rodents and humans [3]. IRS -2 shares significant structure with IRS -1, in that both proteins contain a N-terminal pleckstrin homology (PH)domain, phosphotyrosine binding (PTB) domains as well as a C-terminal tail consisting of numerous tyrosine and serine phosphorylation sites [4], [5]. The crucial role played by IRS-1 and IRS-2 in the regulation of insulin signaling was widely demonstrated by studies on knockout animal models. IRS-1 null mice showed growth retardation and mild resistance to insulin, but did not develop diabetes. IRS-2 null mice displayed metabolic defects in liver, muscle, and adipose tissues and they developed diabetes owing to pancreatic β-cell failure [6]. Recently, studies have shown they had a redundant role in mediating insulin action in hepatocytes. It was demonstrated that the deletion of both IRS1 and IRS2 genes in the liver of mice (L-DKO mice) prevented activation of hepatic Akt-Foxo1 phosphorylation and resulted in the development of diabetes [7], [8].
In addition, researchers have found out IRS-1 and IRS-2 mediate mitogenic and antiapoptotic signaling via binding to receptor tyrosine kinases (RKTs) such as ligand-phosphorylated insulin-like growth factor I receptor (IGF-IR) or insulin receptor (IR) [2], [9]. Moreover, considerable studies have revealed that this two signaling adaptors have themselves been shown to be transforming oncogenes [10]. IRS1 plays a central role in cancer cell proliferation, in contrast, IRS2 is associated with cancer cell motility and metastasis [9]. In addition, they are able to translocate into the nucleus and regulate transcription of genes involved in different stages of cancer progression [11]. Elizabeth et al reported that IRS2 may be a driver oncogene in colorectal cancer and over-expressed IRS-2 activated the PI3 kinase pathway and increase cell adhesion [12]. Porter et al. and Chan et al. both figured out that a role for IRS-2 in cell migration rather than proliferation was shown in breast cancer [11], [13]. Also, Mathieu et al showed that deregulated expression of IRS-2 may contribute to liver tumor progression [14].
Until now, about 3644 single nucleotide polymorphisms (SNP) in the IRS1 gene and 1704 SNP in the IRS2 gene have been reported (http://www.ncbi.nlm.nih.gov/SNP), some of which have been shown as susceptible loci for several kinds of diseases, such as cancer and Type 2 diabetes mellitus (T2DM) [15]–[19]. For example, an important IRS1 polymorphism rs1801278 (G>A) has been extensively investigated as a determinant of insulin resistance and a meta-analysis demonstrated that the A carriers significantly increased the risk of T2DM in those subjects whose mean age at diagnosis was less than 45 years [15], [17]. Furthermore, many epidemiological studies suggested this polymorphism affected the risk of many cancer types, including breast, colorectal, ovarian, prostate cancer and multiple myeloma [18], [20]–[23]. The IRS2 gene is located on chromosome 13q34 [12]. The IRS2 polymorphism rs2289046(A>G), which is a 3′UTR SNP, has already been reported it is closely related to the onset of pancreatic, breast and colorectal cancer [24]–[26]. Another IRS2 rs1805097(G>A) polymorphism, a nonsynonymous SNP that was predicted to affect splicing, transcriptional regulation, and post-translational modification, is common [minor allele frequency (MAF) = 0.30] (http://www.ncbi.nlm.nih.gov/SNP) and most frequently studied for the association with cancer risk, especially colorectal and breast cancer [18], [20], [21], [27]–[30]. In Chinese and northern Europe populations this polymorphism did not show associations with insulin sensitivity, insulin secretary function or T2DM, but in Italian and Asian Indian populations the variant allele may increase susceptibility to T2DM in obese people [16].
Although reported studies have focused on the association between the IRS2 rs1805097(G>A) polymorphism and the risk of colorectal and breast cancer in diverse populations, the results remain inconclusive, partially due to the relatively small sample size in each of the published studies [30]. To confirm the effect of IRS2 rs1805097 polymorphism on colorectal and breast cancer risk, a meta-analysis was performed.
Methods
Search Strategy
A literature search of Pubmed and EMbase was performed independently by two authors (updated to August 9, 2013), using the key words: ‘Insulin receptor substrate2 or IRS2’, ‘polymorphism or variation’, ‘cancer or carcinoma or tumor or adenocarcinoma or neoplasm’. We evaluated all associated publications to retrieve the most eligible literature. Their reference lists were searched manually to identify additional eligible studies. When overlapping data of the same patient population were included in more than one publication, only the most recent or complete study was included. The results were limited to papers published in English.
Inclusion Criteria and Exclusion Criteria
The following criteria were used for inclusion of the identified articles: (1) case-control studies for human, (2) investigation of the IRS2 rs1805097 polymorphism and colorectal and breast cancer risk, (3) sufficient published data for estimating the OR and their corresponding 95%CI. Exclusion criteria included: (1) comment and review, (2) duplication of the previous publications, (3) no usable data reported.
Data Extraction
Two investigators independently extracted data and reached consensus on all the items. For each eligible study, the following information was extracted: first author’s name, year of publication, ethnicity, cancer types, matching variables (age and sex), numbers of cases and controls, and genotype frequencies for cases and controls. Eligible studies were defined as population-based and hospital-based according to the control source. Different ethnicity descents were classified as Caucasian, Asian, and Mixed that included more than one ethnic descent. And when studies included subjects of more than one ethnicity and were able to separate, data were extracted separately for each ethnic group. We also classified a study as a large study if its total sample size was more than 1000, or it would be categorized as a small one.
Statistical Analysis
We used the PRISMA checklist as protocol of the meta-analysis and followed the guideline (Checklist S1). HWE was firstly tested by Chi-square in each control group (P<0.05 was considered representative of statistical significance). The minor allele frequency (MAF) was also calculated for the controls. We calculated the OR and its 95%CI to assess the association between IRS2 rs1805097 polymorphism and the risk of colorectal and breast cancer. Pooled ORs were computed for four genetic models of comparison: homozygote (AA vs GG), heterozygote (AG vs GG ), dominant (AA/AG vs GG), and recessive (AA vs AG/GG). A statistical test for heterogeneity was performed based on the I2 test and Q test [31], [32]. If I2>50% or P≤0.10which indicated heterogeneity in the comparison models among studies, so the estimated pooled ORs for each study were calculated using a random effects model (the DerSimonian and Laird method) [33]. Otherwise, the fixed-effects model was suitable (the Mantel–Haenszel method) [34]. Subgroup analyses were performed by ethnicity, source of control and consistency of frequency with HWE for colorectal cancer. For breast cancer, subgroup analyses were performed by sample size. Sensitivity analyses were also performed to identify the stability of the meta-analysis results.
Evaluation of Publication Bias
Funnel plotting, in which the standard error of the log OR in each study was plotted against its log OR, was used to assess potential publication bias. An asymmetric plot suggested a possible publication bias. Funnel plot asymmetry was further assessed by the method of Egger’s linear regression test [35]. The significance of the intercept was determined by the t-test (P<0.05 was considered representative of statistically significant publication bias). The intercept provides a measure of asymmetry, and the larger its deviation from zero the more pronounced the asymmetry. We did not use Egger’s linear regression to test publication bias in a subgroup less than three studies.
All statistical tests were performed with the STATA software, version 12.0 (Stata Corporation, College Station, TX).
Results
Study Characteristics
There were 94 articles identified by literature search from PubMed and EMBASE. The flow chart in Figure 1 summarized the selection process. In total, 10 articles were retrieved for further detailed evaluation [20], [21], [28]–[30], [36]–[40]. As shown in Figure 1, among them, 3 studies on colon and rectal cancer [36]–[38] and 2 studies on colon [39], [40] were excluded because of the overlapping data by the same authors. As 1 article presented the association in non-Hispanic white (NHW) and Hispanic populations, each study in the literature was considered separately [20]. 1 study investigated colon cancer and rectal cancer with the different controls, so we also treated them as separate studies [28]. At last, a total of 7 studies, including 4798 cases and 5478 controls for colorectal cancer and 2108 cases and 2507 controls for breast cancer, were used in the meta-analysis. All 7 studies were written in English. The 7 separate studies consisted of 4 Caucasian and 3 mixed ethnicity. There were 4 studies on colorectal cancer and 3 studies on breast cancer. Noticeably, deviation from HWE of genotype frequencies among the controls was detected in 2 studies [28]. Characteristics, genotypic frequencies (G/G, G/A, and A/A) and the MAF calculated for the controls of all individual studies (all were greater than 0.05) were listed in Table 1.
Meta-analysis Results
Table 2 showed the main results of this meta-analysis in details. Overall, no significant association between IRS2 rs1805097 polymorphism and colorectal cancer risk was observed (homozygote comparison: OR = 0.96, 95%CI 0.85–1.08, Pheterogeneity = 0.97; heterozygote comparison: OR = 0.91, 95%CI 0.73–1.13, Pheterogeneity<0.01; dominant model: OR = 0.92, 95%CI 0.80–1.06, Pheterogeneity = 0.05; recessive model: OR = 1.02, 95%CI 0.91–1.14, Pheterogeneity = 0.60) (Figure 2). Obvious heterogeneity was observed in the heterozygote and dominant comparisons. To determine the cause of heterogeneity among the studies and to obtain more accurate results, we conducted further meta-analyses stratified according to ethnicity, source of control and HWE. However, the heterogeneity still did not decrease. Similarly, the results of the reanalysis revealed that this polymorphism was not significantly associated with colorectal cancer risk. For breast cancer, also no obvious associations were found for all genetic models (homozygote comparison: OR = 0.95, 95%CI 0.71–1.26, Pheterogeneity = 0.10; heterozygote comparison: OR = 1.00, 95%CI 0.89–1.14, Pheterogeneity = 0.71; dominant model: OR = 0.98, 95%CI 0.87–1.10, Pheterogeneity = 0.55; recessive model: OR = 0.95, 95%CI 0.72–1.25, Pheterogeneity = 0.07) (Figure 3). In the homozygote and recessive models, we found there was obvious heterogeneity between studies. Then, we performed subgroup analysis by sample size to assess the source of the heterogeneity, but the results still did not change.
Sensitivity Analysis and Bias Diagnostics
Funnel plots and Egger’s test were performed to assess publication bias of the literature. The results of Egger’s test are shown in Table 3. In our overall analysis, no evidence of publication bias was detected for both colorectal and breast cancer. Similar results were revealed by the shapes of the funnel plots (data not shown). A single study involved in the meta-analysis was deleted each time to reflect the influence of the individual data-set to the pooled ORs, and the corresponding pooled ORs were not materially altered except for heterozygote and dominant models in the subgroup of PB or mixed for colorectal cancer (data not shown).
Discussion
Accumulating number of genetic association and genome-wide association studies (GWASs) have focused on the association between gene polymorphisms and cancer risk [41], [42]. However, the findings are generally inconsistent, probably due to some limitation in these studies such as small sample size. Meta-analysis is considered a powerful tool for summarizing the contradicting results from different studies with more statistical power, so that it can obtain more reliable results than a single study [43]. The findings suggested that the IRS2 gene polymorphism was not significantly associated with both colorectal and breast cancer risk. In subgroup analysis, no significant association was observed. To the best of our knowledge, the present meta-analysis is the first article to assess the role of IRS2 rs1805097 polymorphism in colorectal and breast cancer.
As one of the typical signaling adaptors, IRS-2 was involved in the phosphatidylinositol 3′ -kinase (PI3K) and the extracellular signal-regulated kinase (ERK) pathways, despite the fact it did not have intrinsic kinase activity and require upstream activators [10]. Additionally, IRS-2 could also interact integrins, hormones and cytokines such as IL-4 in a non-canonical manner by binding to cytoplasmic kinases(e.g.JAK) [11]. Owing to its presence in these important cancer-related pathways, IRS-2 was considered to be one of factors accelerating tumor progression and metastasis [3]. Jackson et al. reported that IRS-2 dependent signaling promoted cell motility and metastasis in human breast cancer cell [44].
An amino acid substitution of Gly to Asp at codon 1057 in IRS2 gene by transversion of G to A was located close to two putative tyrosine phosphorylation sites at positions 1042 and 1072, and might change the tertiary structure and function of the protein [30], [45]. The relationship of the IRS2 rs1805097 polymorphism to phenotypic features of insulin resistance, T2MD, polycystic ovary syndrome and several kinds of cancer has been intensively studied with controversial results. Several case-control studies found significant associations between IRS2 rs1805097 polymorphism and gastric, colon and endometrial cancer [3], [28], [46]. However, these results were not confirmed by other studies [18], [29]. In this present meta-analysis, we retrieved 4 studies with 4798 cases and 5478 controls for colorectal cancer and 3 studies with 2108 cases and 2507 controls for breast cancer, no significant association was observed in all genetic models. Several factors contributing to the discrepancy between published studies might be different sample size, disease mechanisms or carcinogen exposure in different populations. There was a possibility that the IRS2 variant genotypes may be tissue-specific. D’Alfonso et al. also found that the common IRS2 rs1805097 variant did not appear to affect the level of IRS-2 expression and the ability to bind the P85 regulatory subunit of PI3-kinase [47]. Maybe this polymorphism was not functional but in linkage disequilibrium with a currently unrecognized functional polymorphism [48].
Certain potential limitations existed in our meta-analysis. Firstly, the number of published studies in our meta-analysis was insufficient, and the studies with small samples might decrease statistical power to properly evaluate the association. Secondly, when regarding with the ethnicity, most of the included studies conducted on Caucasians. Thus, it would be important to have more studies and samples from other ethnicity so that more accurate conclusions about the relationship between the IRS2 rs1805097 polymorphism and colorectal and breast cancer risk might be determined. Thirdly, a more precise analysis could have been conducted, if individual data such as sex, body mass index, smoking and drinking status were available. Fourthly, besides a relatively high level of heterogeneity was detected in some comparisons, the influences of the individual data setting on the pooled ORs were significant in the subgroup of PB or mixed for colorectal cancer, which indicated the instability of the result. Therefore, the results should be interpreted with caution and confirmed from an additional analysis with more published studies in the furture. Fifthly, not all of studies are population-based. Owing to some types of unhealthy life styles or certain genotypes, controls in a hospital-based study might be susceptible to cancer. Therefore, controls selected from the healthy population were more representative for the general population, and might contribute more reliable results to a meta-analysis. Finally, gene-environment interactions should be considered in further studies if individual data of environmental exposure were available.
In conclusion, this meta-analysis suggests that the IRS2 rs1805097 polymorphism may not contribute to the colorectal and breast cancer risk. However, large-sample studies of different ethnic groups with well matched cases and controls are necessary to further clarify the role of IRS2 rs1805097 polymorphisms and these two types of cancer susceptibility in the future.
Supporting Information
Checklist S1.
The guidelines of the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement.
https://doi.org/10.1371/journal.pone.0086911.s001
(DOC)
Author Contributions
Conceived and designed the experiments: YH MZ XQK LL. Performed the experiments: YH MZ KZ. Analyzed the data: YH MZ KL XYH. Contributed reagents/materials/analysis tools: YH MZ. Wrote the paper: YH MZ XQK LL.
References
- 1. Cai D, Dhe-Paganon S, Melendez PA, Lee J, Shoelson SE (2003) Two new substrates in insulin signaling, IRS5/DOK4 and IRS6/DOK5. J Biol Chem 278: 25323–25330.
- 2. Lee YH, White MF (2004) Insulin receptor substrate proteins and diabetes. Arch Pharm Res 27: 361–370.
- 3. Zhao XM, Chen J, Yang L, Luo X, Xu LL, et al. (2012) Association between IRS-2 G1057D polymorphism and risk of gastric cancer. World J Gastrointest Oncol 4: 9–15.
- 4. Sun XJ, Wang LM, Zhang Y, Yenush L, Myers MG Jr, et al. (1995) Role of IRS-2 in insulin and cytokine signalling. Nature 377: 173–177.
- 5. Stohr O, Hahn J, Moll L, Leeser U, Freude S, et al. (2011) Insulin receptor substrate-1 and -2 mediate resistance to glucose-induced caspase-3 activation in human neuroblastoma cells. Biochim Biophys Acta 1812: 573–580.
- 6. Withers DJ, Gutierrez JS, Towery H, Burks DJ, Ren JM, et al. (1998) Disruption of IRS-2 causes type 2 diabetes in mice. Nature 391: 900–904.
- 7. Guo S, Copps KD, Dong X, Park S, Cheng Z, et al. (2009) The Irs1 branch of the insulin signaling cascade plays a dominant role in hepatic nutrient homeostasis. Mol Cell Biol 29: 5070–5083.
- 8. Guo S (2013) Molecular Basis of Insulin Resistance: The Role of IRS and Foxo1 in the Control of Diabetes Mellitus and Its Complications. Drug Discov Today Dis Mech 10: e27–e33.
- 9. Reuveni H, Flashner-Abramson E, Steiner L, Makedonski K, Song R, et al. (2013) Therapeutic destruction of insulin receptor substrates for cancer treatment. Cancer Res 73: 4383–4394.
- 10. Dearth RK, Cui X, Kim HJ, Hadsell DL, Lee AV (2007) Oncogenic transformation by the signaling adaptor proteins insulin receptor substrate (IRS)-1 and IRS-2. Cell Cycle 6: 705–713.
- 11. Chan BT, Lee AV (2008) Insulin receptor substrates (IRSs) and breast tumorigenesis. J Mammary Gland Biol Neoplasia 13: 415–422.
- 12. Day E, Poulogiannis G, McCaughan F, Mulholland S, Arends MJ, et al. (2013) IRS2 is a candidate driver oncogene on 13q34 in colorectal cancer. Int J Exp Pathol 94: 203–211.
- 13.
Porter HA, Perry A, Kingsley C, Tran NL, Keegan AD (2013) IRS1 is highly expressed in localized breast tumors and regulates the sensitivity of breast cancer cells to chemotherapy, while IRS2 is highly expressed in invasive breast tumors. Cancer Lett.
- 14. Boissan M, Beurel E, Wendum D, Rey C, Lecluse Y, et al. (2005) Overexpression of insulin receptor substrate-2 in human and murine hepatocellular carcinoma. Am J Pathol 167: 869–877.
- 15. Prudente S, Morini E, Trischitta V (2009) Insulin signaling regulating genes: effect on T2DM and cardiovascular risk. Nat Rev Endocrinol 5: 682–693.
- 16. Bodhini D, Radha V, Deepa R, Ghosh S, Majumder PP, et al. (2007) The G1057D polymorphism of IRS-2 gene and its relationship with obesity in conferring susceptibility to type 2 diabetes in Asian Indians. Int J Obes (Lond) 31: 97–102.
- 17. Morini E, Prudente S, Succurro E, Chandalia M, Zhang YY, et al. (2009) IRS1 G972R polymorphism and type 2 diabetes: a paradigm for the difficult ascertainment of the contribution to disease susceptibility of ‘low-frequency-low-risk’ variants. Diabetologia 52: 1852–1857.
- 18. Neuhausen SL, Slattery ML, Garner CP, Ding YC, Hoffman M, et al. (2005) Prostate cancer risk and IRS1, IRS2, IGF1, and INS polymorphisms: strong association of IRS1 G972R variant and cancer risk. Prostate 64: 168–174.
- 19. Çayan F, Gen R, Ateş NA, Ayaz L, Akbay E, et al. (2011) Association of Insulin Receptor Substrate-2 Gene Polymorphism with Ovarian Cancer. BALKAN MEDICAL JOURNAL 28: 14–17.
- 20. Slattery ML, Sweeney C, Wolff R, Herrick J, Baumgartner K, et al. (2007) Genetic variation in IGF1, IGFBP3, IRS1, IRS2 and risk of breast cancer in women living in Southwestern United States. Breast Cancer Res Treat 104: 197–209.
- 21. Samowitz WS, Wolff RK, Ma KN, Andersen K, Caan B, et al. (2006) Polymorphisms in insulin-related genes predispose to specific KRAS2 and TP53 mutations in colon cancer. Mutat Res 595: 117–124.
- 22. Ding YC, McGuffog L, Healey S, Friedman E, Laitman Y, et al. (2012) A nonsynonymous polymorphism in IRS1 modifies risk of developing breast and ovarian cancers in BRCA1 and ovarian cancer in BRCA2 mutation carriers. Cancer Epidemiol Biomarkers Prev 21: 1362–1370.
- 23. Birmann BM, Tamimi RM, Giovannucci E, Rosner B, Hunter DJ, et al. (2009) Insulin-like growth factor-1- and interleukin-6-related gene variation and risk of multiple myeloma. Cancer Epidemiol Biomarkers Prev 18: 282–288.
- 24. Dong X, Li Y, Tang H, Chang P, Hess KR, et al. (2012) Insulin-like growth factor axis gene polymorphisms modify risk of pancreatic cancer. Cancer Epidemiol 36: 206–211.
- 25. Karimi K, Mahmoudi T, Karimi N, Dolatmoradi H, Arkani M, et al. (2013) Is there an Association between Variants in Candidate Insulin Pathway Genes IGF-I, IGFBP-3, INSR, and IRS2 and Risk of Colorectal Cancer in the Iranian Population? Asian Pac J Cancer Prev 14: 5011–5016.
- 26. Feigelson HS, Teras LR, Diver WR, Tang W, Patel AV, et al. (2008) Genetic variation in candidate obesity genes ADRB2, ADRB3, GHRL, HSD11B1, IRS1, IRS2, and SHC1 and risk for breast cancer in the Cancer Prevention Study II. Breast Cancer Res 10: R57.
- 27. Lee PH, Shatkay H (2008) F-SNP: computationally predicted functional SNPs for disease association studies. Nucleic Acids Res 36: D820–824.
- 28. Slattery ML, Murtaugh M, Caan B, Ma KN, Neuhausen S, et al. (2005) Energy balance, insulin-related genes and risk of colon and rectal cancer. Int J Cancer 115: 148–154.
- 29. Pechlivanis S, Pardini B, Bermejo JL, Wagner K, Naccarati A, et al. (2007) Insulin pathway related genes and risk of colorectal cancer: INSR promoter polymorphism shows a protective effect. Endocr Relat Cancer 14: 733–740.
- 30. Wagner K, Hemminki K, Grzybowska E, Klaes R, Butkiewicz D, et al. (2004) The insulin-like growth factor-1 pathway mediator genes: SHC1 Met300Val shows a protective effect in breast cancer. Carcinogenesis 25: 2473–2478.
- 31. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21: 1539–1558.
- 32. Cochran WG (1954) The combination of estimates from different experiments. Biometrics 10: 101–129.
- 33. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7: 177–188.
- 34. Mantel N, Haenszel W (1959) Statistical aspects of the analysis of data from retrospective studies of disease. The Challenge of Epidemiology: Issues and Selected Readings 1: 533–553.
- 35. Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315: 629–634.
- 36. Slattery ML, Samowitz W, Curtin K, Ma KN, Hoffman M, et al. (2004) Associations among IRS1, IRS2, IGF1, and IGFBP3 genetic polymorphisms and colorectal cancer. Cancer Epidemiol Biomarkers Prev 13: 1206–1214.
- 37. Slattery ML, Samowitz W, Hoffman M, Ma KN, Levin TR, et al. (2004) Aspirin, NSAIDs, and colorectal cancer: possible involvement in an insulin-related pathway. Cancer Epidemiol Biomarkers Prev 13: 538–545.
- 38. Slattery ML, Curtin K, Wolff R, Ma KN, Sweeney C, et al. (2006) PPARgamma and colon and rectal cancer: associations with specific tumor mutations, aspirin, ibuprofen and insulin-related genes (United States). Cancer Causes Control 17: 239–249.
- 39. Slattery ML, Wolff RK, Herrick J, Caan BJ, Potter JD (2008) Leptin and leptin receptor genotypes and colon cancer: gene-gene and gene-lifestyle interactions. Int J Cancer 122: 1611–1617.
- 40. Slattery ML, Wolff RK, Curtin K, Fitzpatrick F, Herrick J, et al. (2009) Colon tumor mutations and epigenetic changes associated with genetic polymorphism: insight into disease pathways. Mutat Res 660: 12–21.
- 41.
Chang CQ, Yesupriya A, Rowell JL, Pimentel CB, Clyne M, et al.. (2013) A systematic review of cancer GWAS and candidate gene meta-analyses reveals limited overlap but similar effect sizes. Eur J Hum Genet.
- 42. Wu B, Liu HL, Zhang S, Dong XR, Wu G (2012) Lack of an association between two BER gene polymorphisms and breast cancer risk: a meta-analysis. PLoS One 7: e50857.
- 43. Munafo MR, Flint J (2004) Meta-analysis of genetic association studies. Trends Genet 20: 439–444.
- 44. Jackson JG, Zhang X, Yoneda T, Yee D (2001) Regulation of breast cancer cell motility by insulin receptor substrate-2 (IRS-2) in metastatic variants of human breast cancer cell lines. Oncogene 20: 7318–7325.
- 45. Mammarella S, Romano F, Di Valerio A, Creati B, Esposito DL, et al. (2000) Interaction between the G1057D variant of IRS-2 and overweight in the pathogenesis of type 2 diabetes. Hum Mol Genet 9: 2517–2521.
- 46. Cayan F, Tok E, Aras-Ates N, Ayaz L, Akbay E, et al. (2010) Insulin receptor substrate-2 gene polymorphism: is it associated with endometrial cancer? Gynecol Endocrinol 26: 378–382.
- 47. D’Alfonso R, Marini MA, Frittitta L, Sorge R, Frontoni S, et al. (2003) Polymorphisms of the insulin receptor substrate-2 in patients with type 2 diabetes. J Clin Endocrinol Metab 88: 317–322.
- 48. Alonso A, Marsal S, Tortosa R, Canela-Xandri O, Julia A (2013) GStream: improving SNP and CNV coverage on genome-wide association studies. PLoS One 8: e68822.