Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

A novel TCF7L2 type 2 diabetes SNP identified from fine mapping in African American women

  • Stephen A. Haddad ,

    sahaddad@bu.edu

    Affiliation Slone Epidemiology Center at Boston University, Boston, MA, United States of America

  • Julie R. Palmer,

    Affiliation Slone Epidemiology Center at Boston University, Boston, MA, United States of America

  • Kathryn L. Lunetta,

    Affiliation Department of Biostatistics, Boston University School of Public Health, Boston, MA, United States of America

  • Maggie C. Y. Ng,

    Affiliation Center for Genomics and Personalized Medicine Research, Center for Diabetes Research, Wake Forest School of Medicine, Winston-Salem, NC, United States of America

  • MEDIA Consortium ,

    Membership of the MEDIA Consortium is provided in the Acknowledgments.

  • Edward A. Ruiz-Narváez

    Affiliation Slone Epidemiology Center at Boston University, Boston, MA, United States of America

A novel TCF7L2 type 2 diabetes SNP identified from fine mapping in African American women

  • Stephen A. Haddad, 
  • Julie R. Palmer, 
  • Kathryn L. Lunetta, 
  • Maggie C. Y. Ng, 
  • MEDIA Consortium, 
  • Edward A. Ruiz-Narváez
PLOS
x

Abstract

SNP rs7903146 in the Wnt pathway’s TCF7L2 gene is the variant most significantly associated with type 2 diabetes to date, with associations observed across diverse populations. We sought to determine whether variants in other Wnt pathway genes are also associated with this disease. We evaluated 69 genes involved in the Wnt pathway, including TCF7L2, for associations with type 2 diabetes in 2632 African American cases and 2596 controls from the Black Women’s Health Study. Tag SNPs for each gene region were genotyped on a custom Affymetrix Axiom Array, and imputation was performed to 1000 Genomes Phase 3 data. Gene-based analyses were conducted using the adaptive rank truncated product (ARTP) statistic. The PSMD2 gene was significantly associated with type 2 diabetes after correction for multiple testing (corrected p = 0.016), based on the nine most significant single variants in the +/- 20 kb region surrounding the gene, which includes nearby genes EIF4G1, ECE2, and EIF2B5. Association data on four of the nine variants were available from an independent sample of 8284 African American cases and 15,543 controls; associations were in the same direction, but weak and not statistically significant. TCF7L2 was the only other gene associated with type 2 diabetes at nominal p <0.01 in our data. One of the three variants in the best gene-based model for TCF7L2, rs114770437, was not correlated with the GWAS index SNP rs7903146 and may represent an independent association signal seen only in African ancestry populations. Data on this SNP were not available in the replication sample.

Introduction

African American women experience a greater burden from type 2 diabetes compared to U.S. women of European ancestry. Incidence in African American women is more than twice that in U.S. white women, with >50% of this excess rate remaining after adjustment for known type 2 diabetes risk factors including body mass index (BMI) [1]. In addition, African Americans with diabetes have poorer glycemic control [2] and an increased risk of diabetic complications and mortality [3] compared to whites. Given these racial disparities, it is critical that more studies be conducted to investigate the etiology of type 2 diabetes in African American women.

More than 75 genetic loci for type 2 diabetes have been discovered in European, Asian, and Mexican ancestry populations [47], while only three novel variants have been discovered in genome-wide association studies (GWAS) of African ancestry (AA) populations [8,9]. Attempts to replicate type 2 diabetes associations from European samples in AA populations suggest that a majority of the variants show associations in the same direction in AA samples [8,1013]. However, only a few loci have achieved statistical significance in replication attempts. Most notable is SNP rs7903146 in the TCF7L2 gene, the variant most significantly associated with type 2 diabetes to date.

TCF7L2 encodes a transcription factor that plays an important role in the Wnt signaling pathway, and its risk alleles appear to be associated with impaired insulin secretion / beta-cell function [14,15]. The Wnt pathway is one of the cell’s most important developmental and growth regulatory mechanisms [16], critical in determining cell fate, proliferation, polarity, and cell death during embryonic development, and also in adult tissue homeostasis. Abnormalities in Wnt signaling have been implicated in a variety of human diseases [17].

The Wnt signaling pathway is actually a group of signal transduction pathways: the canonical Wnt pathway leads to the regulation of gene transcription, and multiple non-canonical Wnt pathways regulate the cell’s cytoskeleton and calcium stores [17]. All Wnt signaling pathways are initiated by the binding of a Wnt ligand to a Frizzled family transmembrane receptor. In the case of the canonical pathway, the resulting intracellular signaling cascade leads to the inactivation of a β-catenin destruction complex [18]. β-catenin thus avoids destruction and translocates from the cytoplasm to the nucleus where it interacts with TCF7L2 and other transcription factors, replacing transcriptional repressors and recruiting coactivators [17,19].

Genes involved in the β-catenin destruction complex may influence susceptibility to type 2 diabetes, given the critical role this complex plays in Wnt signal transduction with the resulting downstream effects on diabetes locus TCF7L2. AA as well as European individuals may be affected, considering that the TCF7L2 / diabetes association is seen across racial groups. Under one scenario, gene mutations might render the β-catenin destruction complex inactive at all times. In this situation, β-catenin would avoid destruction even in the absence of Wnt ligands, thereby accumulating in the cytoplasm and nucleus and binding to TCF7L2 and other transcription factors. These transcription factors would then act mostly as transcriptional activators, and overexpression of some of their target genes may lead to diabetes pathology. With this type of scenario in mind, the present study was initiated to investigate genes involved in the β-catenin destruction complex for evidence of variants that may impact risk of type 2 diabetes in AA women. Given the small effect sizes generally seen for common susceptibility variants, the present analyses utilized gene-based testing in an attempt to identify important genes with multiple risk variants that might otherwise be missed in a SNP-based approach.

Methods

Study population

The data source for the current analyses was the Black Women’s Health Study (BWHS) [20], a prospective cohort study of health and illness among U.S. black women that began in 1995 when 59,000 African American women 21–69 years of age from across the U.S. completed a 14-page postal health questionnaire. Biennial follow-up questionnaires ascertain new cases of type 2 diabetes and other health outcomes and update covariate data. Through 2013, follow-up had been completed for 88% of the potential years of follow-up for the baseline cohort. The BWHS was granted approval by the Institutional Review Board of Boston University, and all study subjects provided written informed consent.

The accuracy of self-reported diabetes in the BWHS was previously assessed using medical records from a sample of 227 women who reported this diagnosis [21]. Type 2 diabetes was confirmed in 96% of these women, and another 2% were found to have other types of diabetes. The prevalence of undiagnosed diabetes in the BWHS was also previously assessed [22], using data from collected blood samples. Of the 1873 cohort members who provided a blood sample in the first year of blood collection and had never reported diabetes, 120 (6.4%) had HbA1c levels of 6.5% (47.5 mmol/mol) or higher, meeting criteria for diabetes [23].

About 50% of BWHS study participants provided DNA samples for analysis, and these subjects were found to be highly representative of all BWHS participants across a number of factors including geographic region, education, and BMI. A case-control sample was drawn from among participants with DNA samples for genotyping and analysis: incident cases of type 2 diabetes were selected, and one control was matched to each case on birth year (+/- 2 years) and geographic region of residence.

We sought replication of the top associations from the BWHS in up to 8284 African American cases and 15,543 controls from the MEDIA (Meta-analysis of type 2 diabetes in African Americans) Consortium, which has been previously described [9]. MEDIA includes 17 African American type 2 diabetes GWAS.

SNP selection

The Reactome database [24,25] (http://www.reactome.org/) was used to identify 68 genes that code for proteins involved in the Wnt pathway’s β-catenin destruction complex. Tag SNPs were then selected for each of these 68 genes (+/- 20 kb regions surrounding them) in order to capture (at r2 ≥ 0.9) all SNPs with minor allele frequency (MAF) ≥ 5%, based on the African populations in 1000 Genomes [26] (http://www.1000genomes.org/). In addition, tag SNPs were selected for the +/- 100 kb region surrounding the TCF7L2 index SNP rs7903146.

Genotyping and QC

Genotyping of the selected Wnt pathway SNPs was performed in two batches totaling 6080 samples (including duplicates), as part of a custom Affymetrix Axiom array that contained 45,747 SNPs chosen for several type 2 diabetes projects. The Axiom array data underwent extensive QC procedures carried out by Affymetrix and Slone Epidemiology Center. About 13% of samples were removed due to high missing call rates (defined as >5%), poor reproducibility, or Dish-QC values <0.6. About 17% of SNPs were removed due to poor cluster properties, high missing call rates (defined as >10%), deviation from Hardy-Weinberg equilibrium (p <10−5 in controls), or high rates of discordant calls across duplicate samples. Only SNPs that passed QC in both sample batches were retained for analyses. After the application of these QC filters and the consolidation of 63 expected and confirmed duplicate sample pairs, the full type 2 diabetes data set contained 5228 subjects (2632 cases and 2596 controls) and 38,008 SNPs, including 3430 SNPs selected for the current analyses of Wnt pathway genes.

Imputation

After prephasing the study data with SHAPEIT version 2 [27], imputation was performed using the IMPUTE2 software [28] and the 1000 Genomes Phase 3 data as the reference panel (5/2/2013 1000 Genomes data, October 2014 haplotype release). Imputation resulted in a total of 32,165 Wnt pathway SNPs with MAF ≥ 0.5% and imputation info score ≥ 0.5 for analysis. The imputation info score used for SNP filtering was the imputation metric produced by IMPUTE2 [29].

Association analysis

We first computed genotype principal components using the smartpca program in the EIGENSOFT package [30], based on 18,825 genotyped and pruned common (MAF >5%) SNPs in the full type 2 diabetes data set. The principal components of genotype were tested for association with case status after accounting for the study covariates: age at baseline, geographical region, and genotyping batch. For all association analyses, we included principal components that had p <0.1 in this multivariable model.

Gene-based association analyses were conducted using the adaptive rank truncated product (ARTP) statistic [31], as implemented in the R package ARTP2 [32]. The ARTP method was selected for its ability to optimize the number of single SNP p-values combined in each gene-based test. According to the options we set, the ARTP2 program selected an optimal test for each gene using between one and 10 SNPs per gene. All genotyped and imputed Wnt pathway SNPs were input into ARTP2 for analysis. Based on the program parameters chosen, ARTP2 removed 10,445 SNPs with MAF <2% in order to eliminate low frequency, imputed SNPs. Next, it identified pairs of SNPs with linkage disequilibrium (LD) r2 >0.8 within each gene and removed the SNP with the lower MAF from each pair, resulting in removal of 14,918 SNPs. After implementation of the MAF and LD filters, 6802 SNPs remained for gene-based analysis.

Single SNP association tests, required as input for gene-based testing, were performed using logistic regression analyses of the imputed dosage genotype data. All statistical models were adjusted for age at baseline, geographical region, genotyping batch, and genotype principal components.

Results

The results of the gene-based analyses are shown in Table 1. One gene, PSMD2, was significantly associated with the risk of type 2 diabetes after a Bonferroni correction for the 69 genes tested (nominal p = 2.2 x 10−4, corrected p = 0.016). One other gene, GWAS locus TCF7L2, was associated with a nominal p <0.01 (p = 1.5 x 10−3), but this result did not survive a correction for multiple testing.

thumbnail
Table 1. Associations of Wnt pathway genes with risk of type 2 diabetes in the BWHS.

https://doi.org/10.1371/journal.pone.0172577.t001

Table 2 shows the genetic variants that were included in the best models selected for genes PSMD2 and TCF7L2. The best model selected for the PSMD2 region included nine genetic variants. The best model selected for TCF7L2 included three genetic variants. The most significantly associated variant in the TCF7L2 region was the GWAS index SNP rs7903146 (p = 1.0 x 10−5), which was associated with a ~20% increased risk of type 2 diabetes (OR 1.21, 95% CI 1.11, 1.32).

thumbnail
Table 2. Genetic variants comprising the optimal models for PSMD2 and TCF7L2: associations with risk of type 2 diabetes in the BWHS.

https://doi.org/10.1371/journal.pone.0172577.t002

Although we had removed correlated SNPs prior to gene-based testing, the r2 threshold used was 0.8, and there was moderate LD (0.45 < r2 < 0.8) in the study sample among five of the top six variants in the best model for PSMD2. Nevertheless, the nine variants included in the best model for PSMD2 comprised four distinct LD groups, using r2 = 0.35 as the cutoff for LD grouping. For TCF7L2, there was moderate LD between the top two variants in the best model (rs7903146 and rs34872471 had r2 = 0.77), but rs114770437 was not correlated (r2 <0.05 with each of the top two SNPs).

We next reviewed SNPs in the PSMD2 and TCF7L2 regions that had been removed by ARTP2 during pruning, in case any of the excluded SNPs were of interest due to potential functionality. In the +/- 20 kb region surrounding PSMD2, a genotyped missense SNP rs2178403 (A/G, Met/Val), located in gene EIF4G1 and excluded from gene-based analyses due to its high LD (r2 = 0.93) with SNP rs1879244 (Table 2), was associated with diabetes risk, with a p-value smaller than that of the nine SNPs from the best model (p = 8.0 x 10−5). The A allele of rs2178403 had a frequency of 6.6% in the study controls and was associated with a 30% decreased risk of type 2 diabetes (OR 0.70, 95% CI 0.58, 0.83).

Given that the most significant SNP in the PSMD2 region was potentially functional and the top SNP in TCF7L2 was the GWAS index SNP, we assessed how much of the association signal in each region was driven by these top SNPs. We reran single variant analyses in these two regions, conditioning on those SNPs. The results of the conditional analyses are shown in Table 3. When we conditioned on rs2178403, three of the nine variants in the best model for PSMD2 remained nominally significant (p <0.05). When we conditioned on rs7903146, SNP rs114770437 in TCF7L2 remained nominally significant (p = 1.3 x 10−3). Thus, both regions may contain multiple independent signals.

thumbnail
Table 3. Genetic variants comprising the optimal models for PSMD2 and TCF7L2: analyses conditioning on the top SNP in each region.

https://doi.org/10.1371/journal.pone.0172577.t003

A haplotype analysis of rs7903146 and rs114770437 in TCF7L2 showed the presence of only three of the four possible haplotypes including common haplotype rs7903146-C / rs114770437-G (63%), and haplotypes T/G (30%) and C/A (7%). An omnibus test assessing the joint effect of all haplotypes on the risk of type 2 diabetes was significant with p = 2.5 x 10−7. Compared to the C/G haplotype, the T/G haplotype was associated with an 18% increased risk of type 2 diabetes (OR 1.18, 95% CI 1.09, 1.29), and the C/A haplotype was associated with a 23% reduction in risk (OR 0.77, 95% CI 0.66, 0.90) (Table 4).

thumbnail
Table 4. Haplotype analysis of SNPs rs7903146 and rs114770437 in TCF7L2.

https://doi.org/10.1371/journal.pone.0172577.t004

We sought replication in the MEDIA Consortium for the top SNPs in the PSMD2 region and for the potentially novel risk variant rs114770437 in TCF7L2. Replication data were not available for rs114770437, but data were available for four of the nine variants in the best model for PSMD2 (rs939317, rs9846954, rs2376524, and rs1687230). In MEDIA, these four variants had effect estimates pointing in the same direction as BWHS, but the odds ratios were quite small (≤ 1.08 for the risk alleles), and none of the associations were statistically significant (p >0.05). Results of a meta-analysis combining BWHS and MEDIA for these four SNPs are shown in S1 Table.

Discussion

Gene-based analyses of common variants in the vicinity of β-catenin destruction complex genes identified an association between the PSMD2 gene region and type 2 diabetes in 2632 AA cases and 2596 AA controls. Eight of the nine variants in the best model for PSMD2 were not located within PSMD2 itself but were instead located within other surrounding genes on chromosome 3q27.l (EIF4G1, ECE2, and EIF2B5) (Table 2). The most significant variant in the PSMD2 region, missense SNP rs2178403, is located within a plausible diabetes candidate gene, EIF4G1. EIF4G1 encodes a component of the multi-subunit protein complex EIF4F. The EIF4F complex facilitates recruitment of mRNA to the ribosome, which is the rate-limiting step in protein synthesis. There is evidence that compromised insulin signaling in pancreatic beta cells downregulates EIF4G1, leading to the inhibition of carboxypeptidase E (CPE) expression, with a subsequent reduction of proinsulin processing and a corresponding increase in the levels of circulating proinsulin [34].

While EIF4G1 is a potential susceptibility gene, the results of our analyses conditioning on rs2178403 suggest that the association signal in the PSMD2 region, if valid, may not be fully captured by variants in EIF4G1 alone. Furthermore, other genes in this region, including PSMD2 itself, could be linked to diabetes pathology. The PSMD2 gene was included in this study because of its involvement in the Wnt pathway’s β-catenin destruction complex: PSMD2 encodes a regulatory subunit of the 26S proteasome, and it is the 26S proteasome that carries out the actual destruction of β-catenin (as well as other ubiquitinated proteins) [35,36]. It has been shown that a high fat diet downregulates hepatic transcription of PSMD2 in mice that are resistant to the development of insulin resistance and non-alcoholic fatty liver disease (NAFLD), while upregulating transcription in mice with susceptibility to developing insulin resistance and NAFLD [37].

ECE2 is another possible susceptibility gene near PSMD2. The enzyme encoded by ECE2 converts big endothelin-1 to the vasoconstrictor endothelin-1, and is involved in the processing of several neuroendocrine peptides. This enzyme may also act as a methyltransferase. A mouse study reported an association between hyperglycemia at an early stage of autoimmune diabetes and downregulation of ECE2 transcription in the kidneys [38]. In our study, the most significant variant in the top model for the PSMD2 region, rs55808452, was located within an intron of ECE2, although it should be noted that this variant was in moderate LD with several EIF4G1 variants including missense SNP rs2178403 (r2 = 0.54).

It should be acknowledged that the association we observed for the PSMD2 region may very well be a false positive result given that four of the top variants in this region failed to replicate in the large AA sample from the MEDIA Consortium. In addition, four of the five variants that were not available in MEDIA were in moderate LD (0.7 < r2 < 0.8) with at least one of the SNPs that failed replication. If SNPs in this region are truly associated, they likely have small effects as represented by the MEDIA estimates. Although the MEDIA estimates were close to the null (odds ratios between 0.92 and 0.98), they were all in the same direction as our study. Thus, the possibility of true, small effects does exist.

Apart from the PSMD2 region, the other interesting finding from the present study concerned the GWAS gene TCF7L2. SNP rs114770437 (BWHS MAF = 7.8%) was one of three variants included in the best gene-based model for TCF7L2 and was not correlated with the GWAS index SNP rs7903146 (BWHS MAF = 28.2%). The minor A allele at rs114770437 was associated with a 27% reduction in the risk of type 2 diabetes. The association with this SNP remained nominally significant after control for rs7903146 (conditional OR = 0.77; p = 1.3 x 10−3). Thus, rs114770437 may represent an independent association signal in TCF7L2 in AA populations. SNP rs114770437 is monomorphic in 1000 Genomes European samples, and this may explain the results of a Bayesian fine mapping analysis by the Wellcome Trust Case Control Consortium (WTCCC), which suggested that no such secondary signal exists in TCF7L2 in Europeans [39]. In the WTCCC study, the posterior probability that rs7903146 was driving the TCF7L2 association signal was 75%. An additional 13% of the posterior probability was accounted for by correlated SNP rs34872471, the second most significant SNP in the best model for TCF7L2 in our study. No other SNP accounted for more than 3% of the posterior probability.

Despite a respectable sample size of 2632 AA cases and 2596 AA controls, the present study had limited power to detect individual SNP associations. Still, we replicated the association of the TCF7L2 GWAS index SNP rs7903146. The failure of TCF7L2 to achieve significance in our gene-based analyses is likely due to the inherent power limitations of the ARTP gene-based approach in situations where much of a gene’s association is driven by a single SNP. In our application of the ARTP method, each gene test had to correct for having considered up to 10 SNPs. Another limitation of our study was the use of imputed genotypes for many SNPs. However, SNPs with an imputation info score <0.5 or MAF <2% were excluded from the association analyses in order to improve the accuracy of the data used. Lastly, non-differential misclassification of diabetes in our sample, though likely to be small, may have resulted in underestimation of the associations.

In summary, we observed a significant association between the PSMD2 gene region and type 2 diabetes in women of African ancestry in a gene-based analysis. This finding opens the possibility that PSMD2, a gene involved in the Wnt pathway’s β-catenin destruction complex, or another nearby gene such as EIF4G1 or ECE2, may be a susceptibility locus for type 2 diabetes. It is also possible that the observed association is a false positive result, given the failed replication of a subset of the top SNPs in this region. Our analyses also suggested a possible association signal in TCF7L2 that is independent of the GWAS index SNP rs7903146 and may be present only in AA populations. Replication is needed in additional AA samples in order to validate our findings.

Supporting information

S1 Table. Meta-analysis of BWHS and MEDIA for the PSMD2 region.

https://doi.org/10.1371/journal.pone.0172577.s001

(DOCX)

Acknowledgments

We thank the Black Women’s Health Study participants for their continuing participation in this research effort.

The MEDIA consortium

Lead author: Maggie CY Ng; Email: mng@wakehealth.edu

Members:

Maggie C.Y. Ng1,2, Daniel Shriner3, Brian H. Chen4,5, Jiang Li2, Wei-Min Chen6,7, Xiuqing Guo8, Jiankang Liu9, Suzette J. Bielinski10, Lisa R. Yanek11, Michael A. Nalls12, Mary E. Comeau13,14, Laura J. Rasmussen-Torvik15, Richard A. Jensen16,17, Daniel S. Evans18, Yan V. Sun19, Ping An20, Sanjay R. Patel21, Yingchang Lu22,23, Jirong Long24, Loren L. Armstrong25, Lynne Wagenknecht26, Lingyao Yang14, Beverly M. Snively14, Nicholette D. Palmer1,2,27, Poorva Mudgal2, Carl D. Langefeld13,14, Keith L. Keene28, Barry I. Freedman29, Josyf C. Mychaleckyj6,7, Uma Nayak6,7, Leslie J. Raffel30, Mark O. Goodarzi30, Y-D Ida Chen8, Herman A. Taylor Jr31,32, Adolfo Correa31, Mario Sims31, David Couper33, James S. Pankow34, Eric Boerwinkle35, Adebowale Adeyemo3, Ayo Doumatey3, Guanjie Chen3, Rasika A. Mathias11,36, Dhananjay Vaidya11,37, Andrew B. Singleton12, Alan B. Zonderman38, Robert P. Igo Jr39, John R. Sedor40,41, Edmond K. Kabagambe42, David S. Siscovick16,17,43, Barbara McKnight16,44, Kenneth Rice16,44, Yongmei Liu45, Wen-Chi Hsueh46, Wei Zhao47, Lawrence F. Bielak47, Aldi Kraja20, Michael A. Province20, Erwin P. Bottinger22, Omri Gottesman22, Qiuyin Cai24, Wei Zheng24, William J. Blot48, William L. Lowe25, Jennifer A. Pacheco49, Dana C. Crawford50, Stephen S. Rich6, M. Geoffrey Hayes25, Xiao-Ou Shu24, Ruth J.F. Loos22,23,51, Ingrid B. Borecki20, Patricia A. Peyser47, Steven R. Cummings18, Bruce M. Psaty16,17,43,52, Myriam Fornage35, Sudha K. Iyengar39, Michele K. Evans53, Diane M. Becker11,54, W.H. Linda Kao37, James G. Wilson55, Jerome I. Rotter8, Michèle M. Sale6,56,57, Simin Liu4,58,59, Charles N. Rotimi3, Donald W. Bowden1,2,27

1 Center for Genomics and Personalized Medicine Research, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 2 Center for Diabetes Research, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 3 Center for Research on Genomics and Global Health, National Human Genome Research Institute, Bethesda, Maryland, United States of America, 4 Program on Genomics and Nutrition, School of Public Health, University of California Los Angeles, Los Angeles, California, United States of America, 5 Center for Metabolic Disease Prevention, School of Public Health, University of California Los Angeles, Los Angeles, California, United States of America, 6 Center for Public Health Genomics, University of Virginia, Charlottesville, Virginia, United States of America, 7 Department of Public Health Sciences, University of Virginia, Charlottesville, Virginia, United States of America, 8 Institute for Translational Genomics and Population Sciences, Los Angeles BioMedical Research Institute at Harbor-UCLA Medical Center, Torrance, California, United States of America, 9 Department of Medicine, University of Mississippi Medical Center, Jackson, Mississippi, United States of America, 10 Division of Epidemiology, Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota, United States of America, 11 The GeneSTAR Research Program, Division of General Internal Medicine, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland, United States of America, 12 Laboratory of Neurogenetics, National Institute on Aging, National Institutes of Health, Bethesda, Maryland, United States of America, 13 Center for Public Health Genomics, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 14 Department of Biostatistical Sciences, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 15 Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States of America, 16 Cardiovascular Health Research Unit, University of Washington, Seattle, Washington, United States of America, 17 Department of Medicine, University of Washington, Seattle, Washington, United States of America, 18 San Francisco Coordinating Center, California Pacific Medical Center Research Institute, San Francisco, California, United States of America, 19 Department of Epidemiology and Biomedical Informatics, Emory University, Atlanta, Georgia, United States of America, 20 Division of Statistical Genomics, Washington University School of Medicine, St. Louis, Missouri, United States of America, 21 Division of Sleep Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, United States of America, 22 The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, United States of America, 23 The Genetics of Obesity and Related Metabolic Traits Program, Icahn School of Medicine at Mount Sinai, New York, New York, United States of America, 24 Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, Tennessee, United States of America, 25 Division of Endocrinology, Metabolism and Molecular Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States of America, 26 Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 27 Department of Biochemistry, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 28 Department of Biology, Center for Health Disparities, East Carolina University, Greenville, North Carolina, United States of America, 29 Department of Internal Medicine, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 30 Medical Genetics Research Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States of America, 31 Department of Medicine, University of Mississippi Medical Center, Jackson, Mississippi, United States of America, 32 Jackson State University, Tougaloo College, Jackson, Mississippi, United States of America, 33 Collaborative Studies Coordinating Center, Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, 34 Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota, United States of America, 35 Human Genetics Center, University of Texas Health Science Center at Houston, Houston, Texas, United States of America, 36 Division of Allergy and Clinical Immunology, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland, United States of America, 37 Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States of America, 38 Laboratory of Personality and Cognition, National Institute on Aging, National Institutes of Health, Baltimore, Maryland, United States of America, 39 Department of Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, Ohio, United States of America, 40 Department of Medicine, Case Western Reserve University, MetroHealth System campus, Cleveland, Ohio, United States of America, 41 Department of Physiology and Biophysics, Case Western Reserve University, Cleveland, Ohio, United States of America, 42 Division of Epidemiology, Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America, 43 Department of Epidemiology, University of Washington, Seattle, Washington, United States of America, 44 Department of Biostatistics, University of Washington, Seattle, Washington, United States of America, 45 Department of Epidemiology and Prevention, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States of America, 46 Department of Medicine, University of California, San Francisco, California, United States of America, 47 Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan, United States of America, 48 Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, Tennessee; International Epidemiology Institute, Rockville, Maryland, United States of America, 49 Center for Genetic Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States of America, 50 Center for Human Genetics Research and Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, Tennessee, United States of America, 51 Child Health and Development Institute, Icahn School of Medicine at Mount Sinai, New York, New York, United States of America, 52 Department of Health Services, University of Washington, Seattle, Washington, United States of America, 53 Health Disparities Unit, National Institute on Aging, National Institutes of Health, Baltimore Maryland, United States of America, 54 Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States of America, 55 Department of Physiology and Biophysics, University of Mississippi Medical Center, Jackson, Mississippi, United States of America, 56 Department of Medicine, University of Virginia, Charlottesville, Virginia, United States of America, 57 Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, Virginia, United States of America, 58 Department of Epidemiology, University of California Los Angeles, Los Angeles, California, United States of America, 59 Departments of Epidemiology and Medicine, Brown University, Providence, Rhode Island, United States of America

Author Contributions

  1. Conceptualization: SAH JRP EAR.
  2. Data curation: SAH EAR.
  3. Formal analysis: SAH.
  4. Funding acquisition: JRP EAR.
  5. Investigation: JRP EAR.
  6. Methodology: SAH JRP KLL EAR.
  7. Project administration: JRP EAR.
  8. Resources: JRP MCYN EAR.
  9. Software: SAH.
  10. Supervision: JRP EAR.
  11. Validation: SAH MCYN EAR.
  12. Visualization: SAH.
  13. Writing – original draft: SAH.
  14. Writing – review & editing: SAH JRP KLL MCYN EAR.

References

  1. 1. Brancati FL, Kao WH, Folsom AR, Watson RL, Szklo M. Incident type 2 diabetes mellitus in African American and white adults: the Atherosclerosis Risk in Communities Study. JAMA J Am Med Assoc. 2000;283: 2253–2259.
  2. 2. Kirk JK, D’Agostino RB, Bell RA, Passmore LV, Bonds DE, Karter AJ, et al. Disparities in HbA1c Levels Between African-American and Non-Hispanic White Adults With Diabetes: A meta-analysis. Diabetes Care. 2006;29: 2130–2136. pmid:16936167
  3. 3. Lanting LC, Joung IM, Mackenbach JP, Lamberts SW, Bootsma AH. Ethnic differences in mortality, End-stage complications, and quality of care among diabetic patients a review. Diabetes Care. 2005;28: 2280–2288. pmid:16123507
  4. 4. Qi Q, Hu FB. Genetics of type 2 diabetes in European populations: T2D genetics in Europeans. J Diabetes. 2012;4: 203–212. pmid:22781158
  5. 5. Morris AP, Voight BF, Teslovich TM, Ferreira T, Segrè AV, Steinthorsdottir V, et al. Large-scale association analysis provides insights into the genetic architecture and pathophysiology of type 2 diabetes. Nat Genet. 2012;44: 981–990. pmid:22885922
  6. 6. Hara K, Fujita H, Johnson TA, Yamauchi T, Yasuda K, Horikoshi M, et al. Genome-wide association study identifies three novel loci for type 2 diabetes. Hum Mol Genet. 2014;23: 239–246. pmid:23945395
  7. 7. Mahajan A, Go MJ, Zhang W, Below JE, Gaulton KJ, Ferreira T, et al. Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet. 2014;46: 234–244. pmid:24509480
  8. 8. Palmer ND, McDonough CW, Hicks PJ, Roh BH, Wing MR, An SS, et al. A Genome-Wide Association Search for Type 2 Diabetes Genes in African Americans. Kronenberg F, editor. PLoS ONE. 2012;7: e29202. pmid:22238593
  9. 9. Ng MCY, Shriner D, Chen BH, Li J, Chen W-M, Guo X, et al. Meta-analysis of genome-wide association studies in african americans provides insights into the genetic architecture of type 2 diabetes. PLoS Genet. 2014;10: e1004517. pmid:25102180
  10. 10. Waters KM, Stram DO, Hassanein MT, Le Marchand L, Wilkens LR, Maskarinec G, et al. Consistent Association of Type 2 Diabetes Risk Variants Found in Europeans in Diverse Racial and Ethnic Groups. McCarthy MI, editor. PLoS Genet. 2010;6: e1001078. pmid:20865176
  11. 11. Saxena R, Elbers CC, Guo Y, Peter I, Gaunt TR, Mega JL, et al. Large-scale gene-centric meta-analysis across 39 studies identifies type 2 diabetes loci. Am J Hum Genet. 2012;90: 410–425. pmid:22325160
  12. 12. Cooke JN, Ng MCY, Palmer ND, An SS, Hester JM, Freedman BI, et al. Genetic Risk Assessment of Type 2 Diabetes-Associated Polymorphisms in African Americans. Diabetes Care. 2012;35: 287–292. pmid:22275441
  13. 13. Haiman CA, Fesinmeyer MD, Spencer KL, Buzkova P, Voruganti VS, Wan P, et al. Consistent Directions of Effect for Established Type 2 Diabetes Risk Variants Across Populations: The Population Architecture using Genomics and Epidemiology (PAGE) Consortium. Diabetes. 2012;61: 1642–1647. pmid:22474029
  14. 14. Lyssenko V, Lupi R, Marchetti P, Del Guerra S, Orho-Melander M, Almgren P, et al. Mechanisms by which common variants in the TCF7L2 gene increase risk of type 2 diabetes. J Clin Invest. 2007;117: 2155–2163. pmid:17671651
  15. 15. Voight BF, Scott LJ, Steinthorsdottir V, Morris AP, Dina C, Welch RP, et al. Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis. Nat Genet. 2010;42: 579–589. pmid:20581827
  16. 16. Grant SFA, Thorleifsson G, Reynisdottir I, Benediktsson R, Manolescu A, Sainz J, et al. Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet. 2006;38: 320–323. pmid:16415884
  17. 17. Saito-Diaz K, Chen TW, Wang X, Thorne CA, Wallace HA, Page-McCaw A, et al. The way Wnt works: Components and mechanism. Growth Factors. 2013;31: 1–31. pmid:23256519
  18. 18. Baarsma HA, Königshoff M, Gosens R. The WNT signaling pathway from ligand secretion to gene transcription: Molecular mechanisms and pharmacological targets. Pharmacol Ther. 2013;138: 66–83. pmid:23328704
  19. 19. Rao TP, Kuhl M. An Updated Overview on Wnt Signaling Pathways: A Prelude for More. Circ Res. 2010;106: 1798–1806. pmid:20576942
  20. 20. Rosenberg L, Adams-Campbell L, Palmer JR. The Black Women’s Health Study: a follow-up study for causes and preventions of illness. J Am Med Womens Assoc 1972. 1995;50: 56–58.
  21. 21. Krishnan S, Cozier YC, Rosenberg L, Palmer JR. Socioeconomic status and incidence of type 2 diabetes: results from the Black Women’s Health Study. Am J Epidemiol. 2010;171: 564–570. pmid:20133518
  22. 22. Vimalananda VG, Palmer JR, Gerlovin H, Wise LA, Rosenzweig JL, Rosenberg L, et al. Night-shift work and incident diabetes among African-American women. Diabetologia. 2015;58: 699–706. pmid:25586362
  23. 23. Gillett MJ. International Expert Committee report on the role of the A1c assay in the diagnosis of diabetes: Diabetes Care 2009; 32(7): 1327–1334. Clin Biochem Rev Aust Assoc Clin Biochem. 2009;30: 197–200. pmid:19502545
  24. 24. Croft D, Mundo AF, Haw R, Milacic M, Weiser J, Wu G, et al. The Reactome pathway knowledgebase. Nucleic Acids Res. 2014;42: D472–477. pmid:24243840
  25. 25. Milacic M, Haw R, Rothfels K, Wu G, Croft D, Hermjakob H, et al. Annotating cancer variants and anti-cancer therapeutics in reactome. Cancers. 2012;4: 1180–1211. pmid:24213504
  26. 26. McVean GA, Altshuler DM, Durbin RM, Abecasis GR, Bentley DR, Chakravarti A, et al. An integrated map of genetic variation from 1,092 human genomes. Nature. 2012;491: 56–65. pmid:23128226
  27. 27. Delaneau O, Marchini J, McVean GA, Donnelly P, Lunter G, Marchini JL, et al. Integrating sequence and array data to create an improved 1000 Genomes Project haplotype reference panel. Nat Commun. 2014;5: 3934. pmid:25653097
  28. 28. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 2009;5: e1000529. pmid:19543373
  29. 29. Marchini J, Howie B. Genotype imputation for genome-wide association studies. Nat Rev Genet. 2010;11: 499–511. pmid:20517342
  30. 30. Patterson N, Price AL, Reich D. Population Structure and Eigenanalysis. PLoS Genet. 2006;2: e190. pmid:17194218
  31. 31. Yu K, Li Q, Bergen AW, Pfeiffer RM, Rosenberg PS, Caporaso N, et al. Pathway analysis by adaptive combination of P-values. Genet Epidemiol. 2009;33: 700–709. pmid:19333968
  32. 32. Zhang H, Wheeler B, Yu K, Yang Y. ARTP2: Pathway and Gene-Level Association Test [Internet]. 2016. Available: https://cran.r-project.org/web/packages/ARTP2/index.html
  33. 33. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience. 2015;4: 7. pmid:25722852
  34. 34. Liew CW, Assmann A, Templin AT, Raum JC, Lipson KL, Rajan S, et al. Insulin regulates carboxypeptidase E by modulating translation initiation scaffolding protein eIF4G1 in pancreatic cells. Proc Natl Acad Sci. 2014;111: E2319–E2328. pmid:24843127
  35. 35. Hwang S-G, Yu S-S, Ryu J-H, Jeon H-B, Yoo Y-J, Eom S-H, et al. Regulation of -Catenin Signaling and Maintenance of Chondrocyte Differentiation by Ubiquitin-independent Proteasomal Degradation of -Catenin. J Biol Chem. 2005;280: 12758–12765. pmid:15695815
  36. 36. Stamos JL, Weis WI. The -Catenin Destruction Complex. Cold Spring Harb Perspect Biol. 2013;5: a007898–a007898. pmid:23169527
  37. 37. Waller-Evans H, Hue C, Fearnside J, Rothwell AR, Lockstone HE, Caldérari S, et al. Nutrigenomics of High Fat Diet Induced Obesity in Mice Suggests Relationships between Susceptibility to Fatty Liver Disease and the Proteasome. Guillou H, editor. PLoS ONE. 2013;8: e82825. pmid:24324835
  38. 38. Ortmann J, Nett PC, Celeiro J, Hofmann-Lehmann R, Tornillo L, Terracciano LM, et al. Downregulation of renal endothelin-converting enzyme 2 expression in early autoimmune diabetes. Exp Biol Med Maywood NJ. 2006;231: 1030–1033.
  39. 39. Maller JB, McVean G, Byrnes J, Vukcevic D, Palin K, Su Z, et al. Bayesian refinement of association signals for 14 loci in 3 common diseases. Nat Genet. 2012;44: 1294–1301. pmid:23104008