impact of type 2 diabetes susceptibility variants on...

1

Impact of type 2 diabetes susceptibility variants on quantitative glycemic traits reveals

mechanistic heterogeneity

Running title: Physiological characterization of known type 2 diabetes loci

Antigone S. Dimas1,2

*, Vasiliki Lagou1,3

*, Adam Barker4*, Joshua W. Knowles

5*, Reedik

Mägi1,3,6

, Marie-France Hivert7,8

, Andrea Benazzo9, Denis Rybin

10, Anne U. Jackson

11,

Heather M. Stringham11

, Ci Song12,13

, Antje Fischer-Rosinsky14

, Trine Welløv Boesgaard15

,

Niels Grarup16

, Fahim A. Abbasi5, Themistocles L. Assimes

5, Ke Hao

17, Xia Yang

18, Cécile

Lecoeur19

, Inês Barroso20,21

, Lori L. Bonnycastle22

, Yvonne Böttcher23

, Suzannah

Bumpstead20

, Peter S. Chines22

, Michael R. Erdos22

, Jurgen Graessler24

, Peter Kovacs25

,

Mario A. Morken22

, Narisu Narisu22

, Felicity Payne20

, Alena Stancakova26

, Amy J. Swift22

,

Anke Tönjes23,27

, Stefan R. Bornstein24

, Stéphane Cauchi19

, Philippe Froguel19,28

, David

Meyre19,29

, Peter E.H. Schwarz24

, Hans-Ulrich Häring30

, Ulf Smith31

, Michael Boehnke11

,

Richard N. Bergman32

, Francis S. Collins22

, Karen L. Mohlke33

, Jaakko Tuomilehto34-36

,

Thomas Quertemous5, Lars Lind

37, Torben Hansen

16,38, Oluf Pedersen

16,39-41, Mark Walker

42,

Andreas F.H. Pfeiffer14,43

, Joachim Spranger14

, Michael Stumvoll23,27

, James B. Meigs8,44

,

Nicholas J. Wareham4, Johanna Kuusisto

26, Markku Laakso

26, Claudia Langenberg

4, Josée

Dupuis45,46

, Richard M. Watanabe*47

, Jose C. Florez*44,48,49

, Erik Ingelsson*1,12

, Mark I.

McCarthy*1,3,50

, Inga Prokopenko*1,3,28

on behalf of the MAGIC investigators

* These authors contributed equally to this work.

Correspondence should be addressed to:

Inga Prokopenko, MSc, PhD

Department of Genomics of Common Disease

School of Public Health

Imperial College London

Burlington Danes Building,

Hammersmith Hospital,

Du Cane Road, London,

W12 0NN, UK

Phone: +4420 759 46501

E-mail: [email protected]

Prof. Mark I. McCarthy

OCDEM, Churchill Hospital, University of Oxford

Old Road, Headington, OX3 7LJ

Email: [email protected]

Phone: +44 1865 857298

Erik Ingelsson, MD, PhD, FAHA

Professor of Molecular Epidemiology

Department of Medical Sciences

Page 3 of 73 Diabetes

Diabetes Publish Ahead of Print, published online December 2, 2013

2

Molecular Epidemiology and Science for Life Laboratory

UCR/MTC

Dag Hammarskjölds väg 14B

Uppsala Science Park

SE-752 37 Uppsala

Sweden

Phone: +46-70-7569422

Fax: +46-18-515570

E-mail: [email protected]

Jose C. Florez, MD, PhD

Diabetes Unit / Center for Human Genetic Research

Simches Research Building - CPZN 5.250

Massachusetts General Hospital

185 Cambridge Street

Boston, MA 02114

Phone: 617.643.3308

Fax: 617.643-6630

Richard M. Watanabe, PhD

Departments of Preventive Medicine and Physiology & Biophysics

Diabetes and Obesity Research Institute of USC

Keck School of Medicine of USC

2250 Alcazar Street, CSC-204

Los Angeles, CA 90089-9073

Phone: 1-323-442-2053

Affiliations

1. Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, OX3

7BN, UK.

2. Alexander Fleming, Biomedical Sciences Research Center, 34 Fleming Street, Vari,

16672 Athens, Greece.

3. Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford,

UK, OX3 7LJ.

4. MRC Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital,

Cambridge, UK.

5. Department of Medicine and Cardiovascular Institute, Stanford University School of

Medicine, Stanford, CA 94305, USA.

6. Estonian Genome Center, University of Tartu, Tartu, 51010, Estonia.

7. Department of Medicine, Université de Sherbrooke, Sherbrooke (Quebec), Canada.

8. General Medicine Division, Massachusetts General Hospital, Boston,Massachusetts,

USA.

9. Department of Biology and Evolution, University of Ferrara, Ferrara, Italy.

10. Boston University Data Coordinating Center, Boston, Massachusetts, MA 02118,

USA.

11. Department of Biostatistics and Center for Statistical Genetics, University of

Michigan School of Public Health, Ann Arbor, Michigan 48109, USA.

Page 4 of 73Diabetes

3

12. Department of Medical Sciences, Molecular Epidemiology and Science for Life

Laboratory, Uppsala University, Uppsala, Sweden.

13. Department of Medical Epidemiology and Biostatistics, Karolinska Institutet,

Stockholm, Sweden.

14. Charité-Universitätsmedizin Berlin, Department of Endocrinology and Metabolism.

15. Steno Diabetes Center, Gentofte, Denmark.

16. The Novo Nordisk Foundation Center for Basic Metabolic Research, Faculty of

Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.

17. Department of Genetics and Genomic Sciences, Institute of Genomics and Multiscale

Biology, Mount Sinai School of Medicine, New York, NY 10029-6574, USA.

18. Department of Integrative Biology and Physiology, University of California, 610

Charles E. Young Dr. East, Los Angeles, CA 90095, USA.

19. CNRS UMR8199-Institute of Biology, Pasteur Institute, Lille 2-Droit et Santé

University, Lille, France.

20. Wellcome Trust Sanger Institute, Hinxton, CB10 1SA, UK.

21. University of Cambridge Metabolic Research Laboratories and NIHR Cambridge

Biomedical Research Centre, Level 4, Institute of Metabolic Science, Box 289,

Addenbrooke’s Hospital, Cambridge CB2 OQQ, UK.

22. Genome Technology Branch, National Human Genome Research Institute, Bethesda,

MD.

23. Leipzig University Medical Center, IFB AdiposityDiseases, Liebigstr. 21, 04103

Leipzig, Germany.

24. Department of Medicine III, Division of Prevention and Care of Diabetes, University

of Dresden, 01307 Dresden, Germany.

25. Interdisciplinary Center for Clinical Research (IZKF) Leipzig, Liebigstr. 21, 04103

Leipzig, Germany.

26. Department of Medicine, University of Eastern Finland and Kuopio University

Hospital, 70210 Kuopio, Finland.

27. Department of Medicine, University of Leipzig, Liebigstr. 18, 04103 Leipzig,

Germany.

28. Department of Genomics of common diseases, Imperial College London, London,

UK.

29. Department of Clinical Epidemiology & Biostatistics, McMaster University,

Hamilton, Canada.

30. Department of Internal Medicine, Division of Endocrinology, Diabetology, Vascular

Medicine, Nephrology and Clinical Chemistry, University of Tübingen, Tübingen,

Germany.

31. Lundberg Laboratory for Diabetes Research, Center of Excellence for Metabolic and

Cardiovascular Research, Department of Molecular and Clinical Medicine,

Sahlgrenska Academy, University of Gothenburg, SE-413 45 Gothenburg, Sweden.

32. Department of Physiology & Biophysics, Keck School of Medicine, University of

Southern California, Los Angeles, California 90033, USA.

33. Department of Genetics, University of North Carolina Chapel Hill, North Carolina

27599, USA.

34. Diabetes Prevention Unit, National Institute for Health and Welfare, 00271 Helsinki,

Finland.

35. Centre for Vascular Prevention, Danube-University Krems, 3500 Krems, Austria.

36. King Abdulaziz University, Jeddah 21589, Saudi Arabia.


4

37. Department of Medical Sciences, Uppsala University, Akademiska sjukhuset,

Uppsala, Sweden.

38. Faculty of Health Sciences, University of Southern Denmark, Odense, Denmark.

39. Hagedorn Research Institute, Copenhagen, Denmark.

40. Institute of Biomedical Science, Faculty of Health Sciences, University of

Copenhagen, Copenhagen, Denmark.

41. Faculty of Health Sciences, University of Aarhus, Aarhus, Denmark.

42. Institute of Cellular Medicine, Newcastle University, UK.

43. German Institute of Human Nutrition, Department of Clinical Nutrition.

44. Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115,

USA.

45. Department of Biostatistics, Boston University School of Public Health, Boston,

Massachusetts, MA 02118, USA.

46. The National Heart, Lung, and Blood Institute’s Framingham Heart Study,

Framingham, Massachusetts, USA.

47. Departments of Preventive Medicine and Physiology & Biophysics, Keck School of

Medicine of USC, Los Angeles, CA 90033, USA.

48. Center for Human Genetic Research and Diabetes Research Center (Diabetes Unit),

Massachusetts General Hospital, Boston, Massachusetts 02114, USA.

49. Program in Medical and Population Genetics, Broad Institute, Cambridge,

Massachusetts 02142, USA.

50. Oxford NIHR Biomedical Research Centre, Churchill Hospital, Oxford OX3 7LJ,

UK.


5

ABSTRACT

Patients with established type 2 diabetes display both beta-cell dysfunction and insulin

resistance. To define fundamental processes leading to the diabetic state, we examined

the relationship between type 2 diabetes risk variants at 37 established susceptibility

loci and indices of proinsulin processing, insulin secretion and insulin sensitivity. We

included data from up to 58,614 non-diabetic subjects with basal measures, and 17,327

with dynamic measures. We employed additive genetic models with adjustment for sex,

age and BMI, followed by fixed-effects inverse variance meta-analyses. Cluster analyses

grouped risk loci into five major categories based on their relationship to these

continuous glycemic phenotypes. The first cluster (PPARG, KLF14, IRS1, GCKR) was

characterized by primary effects on insulin sensitivity. The second (MTNR1B, GCK)

featured risk alleles associated with reduced insulin secretion and fasting

hyperglycemia. ARAP1 constituted a third cluster characterized by defects in insulin

processing. A fourth cluster (including TCF7L2, SLC30A8, HHEX/IDE, CDKAL1,

CDKN2A/2B) was defined by loci influencing insulin processing and secretion without

detectable change in fasting glucose. The final group contained twenty risk loci with no

clear-cut associations to continuous glycemic traits. By assembling extensive data on

continuous glycemic traits, we have exposed the diverse mechanisms whereby type 2

diabetes risk variants impact disease predisposition.


6

MAIN TEXT

Type 2 diabetes is a metabolic disorder characterized by impaired insulin secretion and

reduced sensitivity to the peripheral actions of insulin. Both genetic and environmental

factors contribute to the development of type 2 diabetes [1] but the fundamental mechanistic

defects contributing to the evolution of disease remain far from clear. Recently, genome-wide

association efforts have extended the number of loci robustly implicated in type 2 diabetes

risk to more than 60 [2-5]. Each of these loci contains sequence variants that are causal for

disease risk and elucidation of the mechanisms through which these operate has the potential

to reveal processes fundamental to disease pathogenesis.

To date, systematic review of the effects of disease risk variants on processes contributing to

the diabetic state has mostly been restricted to the examination of basal indices of beta-cell

function or insulin sensitivity [2, 3]. These studies have demonstrated that most, but not all,

of these loci exert their primary effects on disease risk through deficient insulin secretion

rather than insulin resistance [2, 4, 6].

Further dissection of these mechanisms requires more intensive phenotyping in risk allele

carriers and controls. In principle, such studies, particularly if performed in non-diabetic

individuals, can provide readouts of the status of various aspects of intermediary metabolism.

For example, disproportionately raised levels of circulating fasting proinsulin (PI), as

compared to those of fasting insulin (FI), reflect beta-cell stress and impairment in early

insulin processing. Fasting 32,33 split PI provides a more detailed assessment of the impact

of genetic variants on insulin processing [7]. Dynamic tests of insulin secretion following oral

and/or intravenous (IV) glucose administration can provide insights into early beta-cell

dysfunction and loss of early insulin release [8]. Studies of selected diabetes risk loci

conducted using more refined dynamic measures of insulin secretion and/or sensitivity have,


7

in the main, corroborated the broad inferences obtained from basal measures [9-13], but have

often been underpowered with respect to more detailed mechanistic insights. To date, the

most substantial effort to characterize the physiology of risk allele carriers involved a survey

of 19 loci defined by their primary associations with quantitative glycemic traits, such as

fasting glucose (FG) and FI, eight of which were also type 2 diabetes risk loci [14]. This

study underscored the heterogeneity of genetic effects on glucose homeostasis.

Here we expand upon that study to examine the effect of a total of 37 type 2 diabetes risk loci

across a broad range of quantitative measures of glycemic metabolism. To do this, within the

Meta-Analyses of Glucose and Insulin-related traits Consortium (MAGIC), we have gathered

basal and dynamic testing data from 15 studies providing detailed measures of insulin

processing (fasting PI; 32,33 split PI), insulin secretion (insulinogenic index; acute insulin

response [AIR]), insulin sensitivity (indices derived from oral and IV tests of glucose-

stimulated insulin secretion [15]) and insulin clearance (C-peptide), and combined these with

existing published data for FG, FI and homeostasis model assessments (HOMA) data from

the MAGIC meta-analysis of fasting traits [16]. In doing so, we are able to demonstrate that

the mechanisms of action of these various disease risk loci can be grouped into a number of

specific categories.

Research Design and Methods

Contributing studies. Three partially overlapping collections of samples were used (Table 1,

Supplementary Table 1): (i) nine studies with detailed physiologic basal and/or dynamic

measures after oral glucose stimulation, from a total of 23,443 individuals; (ii) twenty-nine

studies, including up to 58,614 individuals, with fasting trait data available from the MAGIC

genome-wide meta-analysis [16]; and (iii) seven studies, including 4,180 subjects, with IV-


8

derived measures of insulin sensitivity. All participants were adult, non-diabetic (diabetes

defined by clinical diagnosis, diabetes treatment or fasting plasma glucose levels ≥7.0

mmol/l), and of European ancestry. Individuals with impaired FG or impaired glucose

tolerance were maintained in the analyses. Subjects provided informed consent, and all

studies were approved by local ethics committees.

Phenotypes. We collected available data from participating studies for any of 14

intermediate quantitative glycemic phenotypes, grouped as follows: (i) fasting glycemic traits

including FG (Nmax=58,614), FI (Nmax=52,379), and derived HOMA indices of beta-cell

function (HOMA-B) and insulin resistance (HOMA-IR) (Nmax=50,908) [16]; (ii) oral glucose

tolerance test (OGTT)-derived measures including insulinogenic index (pmol/mmol,

N=11,268), and/or insulin sensitivity indices (ISIs, Supplementary Table 2) including

Belfiore ISI [17] (Nmax=10,348), Stumvoll ISI [18] (Nmax=10,239), Matsuda ISI [19]

(Nmax=10,364) and Gutt ISI [20] (Nmax=13,158); (iii) circulating levels of fasting intact PI

(pmol/l) adjusted for concomitant FI (pmol/l), measured in plasma or serum (Nmax=13,912);

(iv) IV measures (up to 4,180 individuals from seven studies) including M/I derived from

euglycemic hyperinsulinemic clamp (Nmax=2,626), SI from the frequently sampled IV glucose

tolerance test (FSIGT, Nmax=1,173), and steady state plasma glucose (SSPG) from the insulin

suppression test (Nmax=381); (v) AIR analysed by the incremental area under the insulin

curve from 0-10 minutes, (Nmax=1,135); (vi) C-peptide (Nmax=5,059) and 32,33 split PI

(pmol/l) adjusted for concomitant FI (pmol/l, Nmax=2,568), ((ii)-(vi) summarized in Table 1).

Given the wide range of sample sizes available for different traits, we divided these 14

phenotypes into two groups based on a sample size cut-off of 10,000. This resulted in ten

“principal” traits with data from >10,000 individuals ((i)-(iii) above) and four traits with data

on fewer individuals ((iv)-(vi)). We focused initial analyses on principal traits.


9

Trait normalisation. Insulinogenic index, ISIs, PI, AIR, C-peptide, and 32,33 split PI were

natural log transformed. IV insulin sensitivity measures (M/I, SI and SSPG) were z-score

transformed to enable meta-analysis as a single IV trait.

SNP definition and proxies. We included 37 of the 38 type 2 diabetes-associated loci

documented in the DIAGRAM consortium meta-analysis [2]. This subset of known risk

variants was selected given their relatively large effects on type 2 diabetes risk (earliest

discovered variants), identification in individuals of European descent, and good

representation on genome-wide association study (GWAS) or custom genotype panels within

the contributing studies. Variants at FTO were excluded given the well-documented primary

association with BMI, which mediates the effect of FTO on type 2 diabetes risk [21, 22].

None of the other 37 loci has evidence for primary BMI associations. At these 37 loci, we

included data for the lead SNP and a total of 126 alternative proxy SNPs. In studies where

data for the lead SNP were not available, we chose the best proxy SNP for each locus on a

study-specific basis, using r2 measures from CEU HapMap [23] (Supplementary Table 3).

QC and exclusion criteria were as previously described [14] (Supplementary Note,

Supplementary Table 3). Not all samples had called genotypes on the sex chromosomes, and

as a result the maximum sample size for chromosome X locus DUSP9 was 7,642 individuals

(i.e. <10,000). We thus excluded this locus from the main analyses.

Statistical analysis. Linear regression was performed to test for association, under an

additive genetic model, between SNPs and quantitative glycemic traits adjusting for age, sex,

and BMI within each cohort. Cohort-specific effect estimates and standard errors derived

from the regression models were then combined in an inverse variance-weighted fixed effects

meta-analysis using GWAMA [24] or METAL [25]. Association P values are reported

without correction for multiple testing. Two-sided P values<0.05 were considered significant


10

given high prior probabilities for association of established type 2 diabetes risk loci (reported

previously at genome-wide significance level P<5x10-8

) with glycemic traits [2, 4, 26, 27].

To investigate the impact of risk variants on physiologic traits, for selected trait pairs, we

plotted the standardised beta-coefficient estimates of the effects to account for differences in

trait transformations and power of individual meta-analyses.

Cluster analysis of physiologic traits and type 2 diabetes loci. To explore the

physiological basis of type 2 diabetes associations, we performed a primary cluster analysis

using principal traits only, and a subsidiary analysis that included all 14 traits. We also

performed cluster analysis of the 36 loci (excluding DUSP9) which grouped their effects on

principal traits. We used meta-analysis z-scores to perform complete linkage hierarchical

clustering and aligned all effects to the disease risk-increasing allele. In this type of cluster

analysis, the distance between two clusters is computed as the maximum distance between a

pair of traits/SNPs that map in separate clusters [28]. Locus clusters were defined by L2, a

Euclidean distance dissimilarity measure. The uncertainty of hierarchical clustering was

evaluated via multiscale bootstrap resampling [29]. Ten thousand bootstrap replicates were

generated to compute a probability for the strength of support for each dendrogram node, and

to evaluate topology sensitivity to sample size for each phenotype. We subsequently

performed a centroid-based clustering analysis to identify the most supported number of

clusters, where the full set of SNPs could be structured. In this clustering method orthogonal

transformation results in a reduced set of observations for each locus, translating ten

phenotypes to two linearly uncorrelated principal components. Dendrograms were created

where markers were forced into k groups (from two to eight), and the Calinski index [30] was

computed as a measure of clustering support. We then performed principal component

analysis (PCA) to centroid-based clustering results to visualise graphically the assignment of

SNPs to inferred clusters.


11

Results

Association meta-analysis. After excluding DUSP9, we first examined the pattern of trait

associations across the 36 remaining loci. First, we highlight those specific trait-SNP

associations with the strongest statistical support (Table 2, Supplementary Tables 4 and 5).

The strongest association was seen between type 2 diabetes risk alleles at the HHEX/IDE

locus and reduced insulinogenic index (P=3.6×10-21

, Supplementary Fig. 1A). We also

confirmed to genome-wide levels of significance, associations between: (i) 32,33 split PI and

ARAP1; and (ii) insulinogenic index and MTNR1B, in line with previous findings in partially-

overlapping data sets [7, 14, 31-35] and in addition to the established overlapping

associations with basal FG, FI and PI. Furthermore, we uncovered strong associations (which

here we define as P<5×10-5

) between: (i) insulinogenic index and CDKAL1 (Supplementary

Fig. 1B); (ii) 32,33 split PI and HNF1A (Supplementary Fig. 1C); and (iii) AIR and

MTNR1B, KCNQ1 and CDKN2A/B (Supplementary Fig. 1D-F). For all variants showing at

least nominal association (P<0.05) with AIR, the diabetes risk allele reduced AIR. In the joint

analysis of samples with IV-derived indices of insulin sensitivity, nominally significant

associations were observed between the disease risk allele and reduced insulin sensitivity for

the IRS1 (P=7×10-4

) and ADCY5 (P=6×10-3

) loci (Table 2, Supplementary Fig. 2A and B).

Cluster analyses. Complete linkage cluster analysis focusing on the structure of trait

relationships (Supplementary Fig. 3 principal traits; Supplementary Fig. 4 all traits)

confirmed that traits grouped in a meaningful manner consistent with expected physiological

relationships. We observed, for example, grouping of the four ISI measures, and of HOMA-B

and insulinogenic index. The Calinski index, which determines best partitioning and optimal

number of these trait clusters, was maximal at eight, indicating no large sub-clusters

(Supplementary Fig. 3B).


12

We next re-clustered data, using principal traits only, to define relationships amongst the 36

type 2 diabetes loci (Fig. 1A). The unimodal distribution of the Calinski index was maximal

at five, with the same number of groups emerging from an alternative clustering analysis

based on centroids (Supplementary Fig. 5). Group assignment for the optimal Nclusters

determined using the Calinski index was coherent with results from linkage clustering. On

this basis, we defined four locus clusters with distinctive phenotypic features (Fig. 1B) which

we term: insulin resistance (IR); hyperglycemic (HG); proinsulin processing (PI); and beta

cell (BC). The twenty remaining loci not included in these phenotypic clusters we describe as

forming an unclassified group (UC). In bootstrap resampling (Fig. 1A), the baseline

branching nodes were highly supported (strength P≥0.84, with best support for a node

defined as Pmax=1.00 and absence of support Pmin=0) with the only exception of the PI group

where there was slightly less evidence of separation from the BC-UC clade (strength

P=0.64).

Four clusters of type 2 diabetes risk loci with distinctive phenotypic features. Here we

describe the features of each of the four clusters. To visualise some of these groupings, Fig. 2

and Supplementary Fig. 6 present selected scatterplots of physiological trait pairs.

Association P values reported in the text are selected for illustrative purposes with full data

detailed in Table 2 and Supplementary Tables 4 and 5.

The IR cluster contains four loci (IRS1, GCKR, PPARG, KLF14) and is characterized by the

association between type 2 diabetes risk alleles and reduced insulin sensitivity, as evidenced

by higher HOMA-IR (IRS1 [P=7×10-10

], GCKR [P=4×10-16

], PPARG [P=4×10-6

], KLF14

[P=7×10-7

]) and by reduced ISIs, with the strongest associations seen for decreased ISI-

Matsuda (IRS1 [P=2×10-7

], GCKR [P=0.008], PPARG [P=0.004]). The IR loci also tended to

show elevation of FG, FI, PI (IRS1 [P=0.006], GCKR [P=0.050]) and C-peptide levels (IRS1


13

[P=0.01], GCKR [P=0.008]). We observed some evidence that carriers of risk alleles at these

loci had higher values for HOMA-B (IRS1 [P=3×10-7

]) and insulinogenic index (IRS1

[P=4×10-4

], Fig. 2A, Supplementary Fig. 6A), which we interpret as likely to reflect, at least

in truncated ascertainment amongst non-diabetic individuals (see Discussion). Given access

to data from the largest set of individuals yet characterized for IV indices of insulin

sensitivity (Nmax=4,180), we specifically evaluated the effects of IR loci on these measures.

We observed an association between the disease risk variant at IRS1 and reduced IV

measures of insulin sensitivity (P=7×10-4

, Table 2) but no nominally significant associations

were seen at GCKR, PPARG or KLF14. Dropping either FI or HOMA-IR measures from the

cluster analyses (these traits are highly correlated) had only subtle effects on cluster

definitions and relationships: for example in analyses that retain FI but exclude HOMA-IR,

three loci, KCNQ1 [rs231362], JAZF1 and HMGA2 moved from the UC group to the IR

cluster. Adjustment for BMI improved estimates of both basal and dynamic measures of

insulin sensitivity (data not shown).

The HG cluster comprises the risk loci mapping near MTNR1B and GCK. Type 2 diabetes

risk alleles at these loci are associated with markedly reduced insulin secretion, combined

with fasting hyperglycemia, even in non-diabetic individuals (Fig. 2B). The insulin secretory

defect is manifest in reduced HOMA-B (MTNR1B [P=4×10-26], GCK [P=1×10

-11]),

insulinogenic index (MTNR1B [P=1×10-14

], GCK [P=0.03]) (Fig. 2B), and AIR (MTNR1B

[P=4×10-6

], GCK [P=0.051]). There was some evidence for inverse associations with insulin

sensitivity, with type 2 diabetes risk alleles associated with higher HOMA-IR (MTNR1B

[P=0.002], GCK [P=4×10-4

]) (Fig. 2C), but no consistent associations were found with ISIs

or IV measures of insulin sensitivity.


14

The PI cluster actually consists of a single locus, ARAP1 (also named CENTD2). The key

feature of this locus is that the type 2 diabetes risk allele is associated with lower levels of

intact PI (P=2×10-50

) and 32,33 split PI (P=1x10-8

) (Supplementary Fig. 6B and C),

combined with reduced insulin secretion (insulinogenic index [P=2×10-5

], HOMA-B

[P=3x10-6

]) (Supplementary Fig. 6B and D) [7]. Insulin sensitivity was also slightly reduced,

as shown by lower ISIs, Matsuda and Gutt ISIs being nominally significant (P=0.033 and

0.050, respectively). Although excluded from the cluster analysis due to smaller sample size,

risk alleles at DUSP9 were also associated with lower PI values (P=0.035, Table 2) and

reduced insulinogenic index (P=3×10-4

), indicating that this locus may share some

mechanistic overlap with ARAP1. However, in contrast to ARAP1, the risk allele at DUSP9

was associated with increased insulin sensitivity (e.g. Matsuda ISI, P=9×10-4

).

The nine loci in the BC cluster (TCF7L2, SLC30A8, HHEX/IDE, CDKAL1, CDKN2A/2B,

THADA, DGKB, PROX1, ADCY5) are characterised by reduced insulin secretion in the

presence of increased PI and, in comparison with the HG cluster, only modest increases in

fasting glycemia. For all nine loci, we observed associations with lower HOMA-B (Fig. 2C),

higher FG, and lower FI (Table 2). The strongest associations were seen at TCF7L2, but all

loci also showed similar patterns of reduced insulinogenic index (Fig. 2B) and AIR

(Supplementary Fig. 6E), as well as increased PI (Supplementary Fig. 6C). There was little

systematic evidence for effects on HOMA or IV measures of insulin sensitivity

(Supplementary Fig. 6F), though the type 2 diabetes risk allele at ADCY5 displayed reduced

sensitivity on IV measures (P=6×10-3

, Supplementary Fig. 6F).

The UC group includes 20 loci for which, despite large sample sizes, detailed phenotyping

and established effects on risk for type 2 diabetes, no systematic evidence of association with

basal or dynamic glycemic phenotypes was detected. The only nominal associations observed


15

were at HNF1B (TCF2) (lower FI [P=8×10-4

], Table 2) and HNF1A (lower 32,33 split PI

[P=8×10-6

], insulinogenic index [P=0.005], Table 2).

We observed no material difference in the overall picture of trait associations when we

compared results from this study, with those previously described by Ingelsson et al. [14] in

partially overlapping samples, at the eight loci examined in both.

Discussion

We present here the most comprehensive report to date concerning the role of type 2 diabetes

risk loci on physiologic trait variability in non-diabetic individuals. In addition, this study

includes the largest number of individuals phenotyped using “gold standard” IV-derived

measures of insulin sensitivity. Through cluster analysis, we have shown that, in terms of

diabetes risk allele effect size on intermediate glycemic traits, type 2 diabetes risk loci fall

into five broad categories.

The IR cluster loci, including those mapping near PPARG, KLF14, IRS1 and GCKR, are

characterized by consistent effects across multiple indices of insulin sensitivity, in both the

basal and stimulated state. Three of the four IR loci (PPARG excluded) ranked in the lower

half of type 2 diabetes risk loci in terms of effect size: this may indicate that some of the loci

in the UC group act through weak effects on insulin sensitivity (as has indeed been suggested

for HMGA2) [2]. The observation that disease risk alleles at IR loci are associated with some

degree of improved beta-cell function (such as measured by HOMA-B) is likely to reflect a

combination of complementary factors. An increase in insulin secretion as a result of higher

insulin resistance is a known physiological phenomenon, which occurs until individuals

decompensate and shift towards development of diabetes. Therefore, alleles that primarily

raise insulin resistance may be secondarily associated with improved insulin secretion


16

through a compensatory mechanism. In addition, the apparent association with insulin

sensitivity may in part reflect the fact that analysis was restricted to non-diabetic subjects:

this can lead to a truncation effect, whereby individuals with genotypes that favor both high

insulin resistance and poor insulin secretion are preferentially depleted from the analysis due

to their diabetes predisposition [6].

The PI cluster is limited to the type 2 diabetes risk variant at ARAP1, the cardinal feature

being a marked reduction in fasting PI levels in carriers of the disease risk allele. The

combination of diabetes risk reduced basal and stimulated insulin secretion and reduced PI

levels runs counter to the usual epidemiological associations [7]. Our data therefore support

previous assertions that the ARAP1 risk variant increases risk of type 2 diabetes through

defects in early steps of insulin production [7].

The HG cluster includes loci characterised by a prominent reduction in basal and stimulated

beta-cell function [35-37], resulting in a marked increase in fasting glucose levels. This

glycemic effect is consistent with the fact that the type 2 diabetes risk alleles at these loci are

associated with modest reductions in basal insulin sensitivity (as measured by HOMA-IR).

These associations with insulin action phenotypes did not however extend to the stimulated

measures, and may simply reflect limitations in the HOMA model. This cluster includes both

GCK and MTNR1B, and whilst they share similar features in terms of their effects on FG,

HOMA-B, and insulin/glucagon ratios [11], external data suggest that these genes act through

different mechanisms. In individuals with MODY2 [38], severe loss-of-function mutations in

GCK lead to impaired glucose sensing, and a higher homeostatic set point for glucose. At

MTNR1B, several studies have shown that lower insulin secretion is accompanied by

markedly reduced insulin sensitivity [11] and increased insulin response to GLP-1 [39]. The


17

apparent lack of effects on PI measures for MTNR1B, as for GCK, points to an effect on the

secretory function of beta-cells, rather than on PI processing [14].

The BC cluster includes a number of loci associated with defective insulin secretion but

without the marked glycemic effects seen in the HG cluster. These include several of the loci

with the strongest allelic effects on diabetes risk, such as those at TCF7L2, SLC30A8,

CDKAL1 and CDKN2A/2B, suggesting that the distinction from the loci in HG cluster is not

simply a consequence of the degree of allelic impact. Although the BC loci map to a single

cluster, there is a degree of heterogeneity with respect to other traits. For example, disease

risk alleles at TCF7L2 and SLC30A8 are associated with increased PI, reduced insulinogenic

index and lower FI levels, underscoring defects in insulin processing and secretion [7]. In

contrast, the risk alleles at HHEX/IDE, DGKB, CDKAL1, PROX1 and CDKN2A/2B display

reduced insulinogenic index and AIR, with no effect on fasting PI, suggesting defects during

first phase insulin response and early insulin secretion. These phenotypic associations within

the BC cluster highlight the potential for further subdivision of loci according to

pathophysiological patterns.

The UC group combines all remaining loci, which although genome-wide significant for type

2 diabetes, display no discernible impact on glycemic measures across the large sample sets

assembled here. Most of these loci rank in the lower half of signals in terms of type 2

diabetes effect size and modest functional impact may offer a partial explanation for this

observation. However, it is notable that the UC group includes many of the loci where the

common variant effects are highly likely to be mediated by transcripts implicated in

monogenic and syndromic forms of diabetes (including HNF1A, HNF1B [40], KCNJ11 [41]

and WFS1 [42]). Thus, even in the setting where rare coding mutations result in severe

abrogation of islet function, and common variants acting through those same genes influence


18

type 2 diabetes risk to stringent levels of significance, those same common variants can also

be compatible with normal glucose homeostasis [2].

This study combined information from a large number of individuals phenotyped using IV

measures of insulin sensitivity, including over 2,600 subjects examined using the “gold

standard” euglycemic hyperinsulinemic clamp. Overall, the data do not support the view that

the combination of IV measures used in this study offers sufficient boost in precision to

compensate for the reduction in sample size when compared to the numbers available for the

basal or OGTT-derived measures. Except at IRS1 and ADCY5, neither the IV measures nor

the OGTT-derived ISIs generated powerful signals of association with deficient insulin

action, even at loci where the existing evidence for an effect on insulin sensitivity is

compelling (e.g. PPARG, KLF14 [2, 43]).

The study presented here is the largest investigation of its kind published to date: we set out,

to maximise the sample size for each trait included. One consequence of this strategy is that

the characteristics of the particular samples informative for each trait may differ, raising the

possibility of artefacts when comparing genotypic associations across traits. However, the

extent to which our findings are both internally consistent, and in broad agreement with

published data describing more detailed phenotypic studies of individual variants, provides

considerable reassurance. An alternative strategy, which restricted analyses to a small core of

individuals with data available for all traits, would have resulted in a substantial loss of

power.

We recognise some other limitations of this study, and of others of its kind. First, the

complexity of the metabolic phenotypes examined, the longitudinal dimension of diabetes

development and progression, and the relative imprecision of the experimental tools at our

disposal, limit the inferences that are possible. This is most obviously seen in the large


19

numbers of loci, which though genome-wide significant for type 2 diabetes, show no

evidence of relationship with intermediate metabolic traits. We also note, for example, that

the measures of early insulin secretion employed (insulinogenic index, AIR) may, in part,

reflect insulin clearance, a trait that was not evaluated directly. Second, in the interests of

avoiding the secondary effects of diabetes and its treatment, analyses were restricted to non-

diabetic individuals: this is likely to have introduced some truncation effects such as the

apparent enhancement of beta-cell function for disease risk alleles disrupting insulin

sensitivity. These have the potential to lead to incorrect inference, if not recognised as such

[6, 44]. Third, the variants we have examined are not, in most cases, known to be causal for

the association signals detected, leaving open the possibility that, at some loci, the causal

alleles, once identified, may have more pronounced effects on intermediate traits than the

variants studied here. Finally additional recently published disease associations [3] of variants

with lesser genetic effects were not considered. Future efforts to extend our study to

incorporate these variants will likely require substantially larger sample sizes.

In summary, in this examination of the intermediate metabolic phenotypic associations of

proven type 2 diabetes risk alleles, we have demonstrated that the loci fall into a limited

number of broad mechanistic categories. These highlight the diverse mechanisms

contributing to individual risk of disease. These data will guide future experimental studies

which use more specific, tailored tests to further dissect these key pathogenetic processes.


20

Table 1. Studies and sample sizes of physiologic glycemic traits studied.

Study

Number of

loci

covered

Number of individuals included in analysis by study

Insulinogenic

Index

ISI

Belfiore

ISI

Stumvoll

ISI

Matsuda

ISI

Gutt Proinsulin M/I SI SSPG AIR

C-

peptide

32,33 Split

Proinsulin

ELY 31 1,435 1,441 1,425 1,311 1,476 1,600 - - - - 1,602 1,598

EUGENE2 36 - - - - - - 595 - - - - -

FHS 35 - - - - 2,599 5,752 - - - - - -

FUSION 10 - - - - - - - 538 - 557 925 -

MESYBEPO 30 883 887 894 885 1,060 - - - - - - -

METSIM 30 6,847 6,898 6,794 6,898 6,933 5,076 - - - - - -

NHANES 12 - - - - - - - - - - 1,221 -

PIVUS 29 - - - - - 911 - - - - - -

QTL FAMILIES 34 262 265 267 274 268 - - 261 - 214 274 -

RISC 36 1,015 - - - - - 1,042 - - - - -

PARTNERS/ROCHE 7 583 608 600 619 612 - - - - - - -

SORBS 36 756 758 765 758 763 - - - - - 718 -

STANFORD 36 - - - - - - - - 381 - - -

ULSAM 29 978 987 976 984 989 979 989 - - - - 979

UNG92 35 - - - - - - - 374 - 373 374 -

Total N * 11,268 10,348 10,239 10,364 13,158 13,912 2,626 1,173 381 1,135 5,059 2,568

All data have been adjusted for age, sex and BMI.

* Total number of individuals with phenotype, non-missing covariates and genetic data available.


21

Table 2. Effects of 37 SNPs previously associated with type 2 diabetes on physiologic glycemic measures.

Principal traits

Effectsignificance

Other traits†

Effectsignificance

Dynamic traits Fasting traits

Nearest gene (SNP) Alleles

(effect/

other)

Proinsulin

(n=13,912)

Insulinoge

nic Index

(n=11,268)

ISI

Belfiore

(n=10,348)

ISI

Stumvoll

(n=10,239)

ISI

Matsuda

(n=10,364)

ISI

Gutt

(n=13,158)

Fasting

Glucose^

(n=58,614)

Fasting

Insulin^

(n=52,379)

HOMA-

B^

(n=50,908)

HOMA-

IR^

(n=50,908)

M/I, SI, SSPG

, (n=4,180)

Insulin Resistance (IR)

PPARG (rs13081389) A/G 0.008 0.009 -0.019* -0.021* -0.035* -0.015* 0.010 0.025*** 0.017** 0.025*** -0.025

IRS1 (rs7578326) A/G 0.014* 0.037** -0.023*** -0.013* -0.040*** -0.020*** 0.005 0.019**** 0.014*** 0.019**** -0.070***

GCKR (rs780094) C/T 0.010* 0.009 -0.008 -0.015* -0.019* 0.007 0.033**** 0.019**** 0.006* 0.023**** -0.019

KLF14 (rs972283) G/A 0.004 -0.002 -0.003 -0.008 -0.001 -0.008 0.007* 0.013*** 0.009** 0.014*** -0.002

Hyperglycemic (HG)

MTNR1B (rs10830963) G/C 0.009 -0.083**** 0.006 -0.007 -0.004 -0.006 0.079**** 4x10-4 -0.033**** 0.011* 0.012

GCK (rs4607517) A/G -7x10-4 -0.036* -0.005 -0.015 -0.008 -0.016* 0.065**** 0.006 -0.024**** 0.014** 0.015

Proinsulin (PI)

DUSP9 (rs5945326) A/G -0.013* -0.037* 0.012* 0.010 0.022** 0.006 0.011 -0.013 -2.890 * -0.066 0.010

ARAP1 (rs1552224) A/C -0.093**** -0.054** -4x10-4 -0.013* -0.013 -0.010* 0.024**** -0.007* -0.015*** -0.005 -0.010

Beta Cell (BC)

TCF7L2 (rs7903146) T/C 0.044**** -0.063*** 0.009 -0.009 0.010 -0.017** 0.026**** -0.014*** -0.019**** -0.010* 0.014

CDKAL1 (rs10440833) A/T 0.008 -0.102*** 0.005 -0.011 -0.010 -0.001 0.014** -0.006* -0.009* -0.004 0.047*

CDKN2A/B (rs10965250) G/A -4x10-4 -0.045* 0.013* -7x10-4 0.012 0.005 0.017** -0.005 -0.012** -0.002 0.047

THADA (rs11899863) C/T 0.015 0.008 0.010 -0.004 0.016 -0.007 0.020** -0.010* -0.020** -0.008 -0.053

HHEX/IDE (rs5015480) C/T 0.006 -0.102**** 0.020*** -0.010 0.017* -0.005 0.010* -0.001 -0.005* 0.001 -0.004

SLC30A8 (rs3802177) G/A 0.038**** -0.041** 0.001 -0.004 -0.001 -0.011* 0.033**** -0.003 -0.016*** 2x10-4 -0.014

ADCY5 (rs11708067) A/G 0.016* -0.009 -0.008 0.001 -0.003 -0.014* 0.022*** -0.006 -0.015*** -0.001 -0.063*

PROX1 (rs340874) C/T 0.009 -0.028* 0.008* 0.008 0.009 -0.001 0.017*** -0.002 -0.009** -3x10-4 -0.002

DGKB/TMEM195 (rs2191349) T/G -0.001 -0.028* 0.011* 0.007 0.027** 0.006 0.032**** -0.002 -0.017**** 0.002 -5x10-4

Unclassified (UC)

CHCHD9 (rs13292136) C/T -5x10-4 -0.025 0.012 -0.001 0.026* 0.006 -0.003 -0.005 -0.002 -0.005 0.034

HMGA2 (rs1531343) C/G 0.010 0.030 -0.013 0.002 -0.027* 0.001 0.013* 0.007 -0.001 0.011* -0.034

HNF1B (TCF2) (rs4430796) G/A -5x10-4 -0.044 0.007 -0.011 0.008 -0.013 -0.001 -0.012** -0.011** -0.012** 0.025


22

IGF2BP2 (rs1470579) C/A -0.011 -0.038* 0.013 0.010 0.010 -8x10-4 0.016*** -0.001 -0.006* 1x10-4 0.013

JAZF1 (rs849134) A/G 0.003 0.013 -0.005 -0.005 -0.014 -0.006 0.007* -1x10-4 -0.001 -2x10-4 0.015

WFS1 (rs1801214) T/C -0.019** -0.017 0.002 -8x10-4 0.008 -0.008 0.008* -0.001 -0.005* 3x10-4 -0.016

TSPAN8/LGR5 (rs4760790) A/G 0.006 -0.016 0.001 0.010 -0.003 -4x10-4 0.004 -5x10-4 -0.002 5x10-4 0.016

ZBED3 (rs4457053) G/A -0.008 -0.014 4x10-4 0.002 4x10-4 -0.004 0.019*** 0.002 -0.005 0.003 -0.008

BCL11A (rs243021) A/G 0.007 0.018 0.004 0.008 0.013 0.009* 0.004 0.001 3x10-4 0.000 -0.006

KCNQ1 (rs163184) G/T 0.016 -0.011 0.002 0.009 0.009 -0.004 0.012** 0.002 -0.004 0.004 0.015

HNF1A (rs7957197) T/A -0.011 -0.036* 0.009 -0.006 -0.003 5x10-4 -0.001 -0.004 -0.003 -0.004 0.005

NOTCH2 (rs10923931) T/G -1x10-5 -0.012 0.011 0.003 0.019 0.001 0.002 -0.003 -0.004 -0.001 0.038

CDC123/CAMK1D (rs12779790) G/A -0.008 -0.029* 0.001 -0.002 -0.006 0.002 0.014** -0.002 -0.007* -0.001 0.020

KCNQ1 (rs231362) G/A -0.004 0.005 -8x10-5 -0.004 -0.003 -0.005 0.017*** 0.007* -0.001 0.007* -0.021

ADAMTS9 (rs6795735) C/T 6x10-4 0.009 0.006 0.003 0.011 0.001 0.008* 0.002 -0.001 0.003 -0.004

KCNJ11 (rs5215) C/T 0.001 -0.024 0.002 -0.015 9x10-4 0.004 -0.004 -0.001 0.003 -0.002 -0.018

PRC1 (rs8042680) A/C -0.008 -0.016 0.003 0.007 0.011 -0.002 0.011* -0.002 -0.004 -3x10-4 0.003

TP53INP1 (rs896854) T/C -0.003 -0.003 -0.001 -0.006 -0.002 3x10-4 0.007* -0.002 -0.004 -0.001 0.007

ZFAND6 (rs11634397) G/A 0.008 -0.010 -0.001 -0.005 -0.007 -0.005 0.001 0.003 0.002 0.003 0.032

HCCA2 (rs2334499) T/C -0.017* -0.018 0.010 0.016 0.015 -0.006 0.001 -0.004 -0.004 -0.004 0.031

Within each cluster, loci were ordered by the previously described size of their effects on T2D risk. These risk estimates were derived from stage

2 metabochip data from Morris et al [3], with the exception of DUPS9 for which we used the OR reported in Voight et al [2] (DUSP9 was not

included in the Metabochip design). Effects represent beta-coefficients. Stars next to per allele effect correspond to P values equal or less than: *

0.05, ** 10-3

, *** 10-5

, **** 10-8

. ^ Fasting trait results have been previously reported in Manning et al. [16]. DUSP9 was grouped together with

ARAP1 as risk alleles at both loci were associated with lower PI values and reduced insulinogenic index. However, in contrast to ARAP1, the risk

allele at DUSP9 was associated with increased insulin sensitivity.


23

AUTHOR CONTRIBUTIONS

A.Ba., A.Be., A.S.D., J.F., M-F.H., E.I., J.K., V.L., C.La., M.I.M, R.M., I.P., R.W. wrote the

manuscript.

A.Ba., A.Be., A.S.D., J.D., V.L., C.La., M.I.M, R.M., I.P., D.R. carried out meta-analysis,

clustering, and higher level analysis of the data.

(EUGENE2) H-U.H., M.L., U.S.; (Ely) N.J.W.; (Framingham Heart Study) J.B.M.; (FUSION)

M.B., R.N.B., F.S.C., K.L.M., J.T., R.M.W.; (MeSyBePo) J.S., A.F.H.P.; (METSIM) J.K., M.L.;

(NHANES III)/(Partners/Roche) J.B.M.; (PIVUS) E.I., L.L.; (QTL/UNG92) T.H., O.P.; (RISC)

M.W.; (Sorbs) M.S.; (Stanford IST) T.Q.; (ULSAM) E.I. designed, managed and coordinated the

project.

(EUGENE2) H-U.H., M.L., U.S.; (Ely) N.J.W.; (Framingham Heart Study) J.B.M., J.C.F.;

(FUSION) R.N.B., J.T.; (MeSyBePo) J.S., A.F-R.; (METSIM) A.S., J.K., M.L.; (NHANES

III)/(Partners/Roche) J.B.M.; (PIVUS) E.I., L.L.; (QTL/UNG92) T.W.B., T.H.; (RISC) M.W.;

(Sorbs) A.T.; (Stanford IST) T.Q., F.A.A.; (ULSAM) E.I. conducted phenotyping.

(EUGENE2) H-U.H., M.L., U.S.; (Ely) I.B., S.B., F.P.; (Framingham Heart Study) J.D., J.B.M.;

(FUSION) L.L.B, M.R.E., A.J.S., N.N., M.A.M., P.S.C.; (M-A) M.I.M.; (METSIM) L.L.B, M.R.E.,

A.J.S., N.N., M.A.M., P.S.C.; (NHANES III)/(Partners/Roche) J.B.M.; (PIVUS) E.I.;

(QTL/UNG92) N.G.; (RISC) M.W.; (Sorbs) P.K., Y.B.; (Stanford IST) T.L.A., J.W.K, K.H., X.Y.;

(ULSAM) E.I. carried out genotyping.

(Ely) C.La., A.Ba.; (Framingham Heart Study) J.D., D.R., M-F.H., J.C.F.; (FUSION) A.U.J.,

H.M.S.; (M-A) A.S.D.; (MeSyBePo) J.S., A.F-R.; (METSIM) A.U.J., H.M.S.; (NHANES

III)/(Partners/Roche) J.B.M.; (PIVUS) E.I., C.S.; (QTL/UNG92) N.G., T.W.B.; (Sorbs) A.S.D.,

I.P., R.M., V.L.; (Stanford IST) T.L.A., J.W.K, F.A.A., K.H., X.Y.; (ULSAM) E.I., C.S. carried out

initial data analysis.

All authors have read and approved the manuscript.

A.S.D. and I.P are the guarantors of this work and, as such, had full access to all the data in the

study and take responsibility for the integrity of the data and the accuracy of the data analysis


24

ACKNOWLEDGEMENTS

The major funding for this meta-analysis was provided through the European Community's Seventh

Framework Programme (FP7/2007-2013), ENGAGE project, grant agreement HEALTH-F4-2007-

201413.

Ely: I.B. acknowledges funding from the Wellcome Trust grant WT098051, United Kingdom NIHR

Cambridge Biomedical Research Centre and the MRC Centre for Obesity and Related Metabolic

Diseases.

EUGENE 2: EUGENE2 was supported by a grant from the European Community’s FP6 EUGENE2

no. LSHM-CT-2004-512013.

Framingham Heart Study: This research was conducted in part using data and resources from the

Framingham Heart Study of the National Heart Lung and Blood Institute of the National Institutes

of Health and Boston University School of Medicine. The analyses reflect intellectual input and

resource development from the Framingham Heart Study investigators participating in the SNP

Health Association Resource (SHARe) project. This work was partially supported by the National

Heart, Lung and Blood Institute's Framingham Heart Study (Contract No. N01-HC-25195) and its

contract with Affymetrix. Inc. for genotyping services (Contract No. N02-HL-6-4278). A portion of

this research utilized the Linux Cluster for Genetic Analysis (LinGA-II) funded by the Robert

Dawson Evans Endowment of the Department of Medicine at Boston University School of

Medicine and Boston Medical Center. Also supported by National Institute for Diabetes and

Digestive and Kidney Diseases (NIDDK) R01 DK078616 to J.B.M., J.D. and J.C.F., NIDDK K24

DK080140 to J.B.M., and a Doris Duke Charitable Foundation Clinical Scientist Development

Award to J.C.F.

FUSION: We would like to thank the many Finnish volunteers who generously participated in the

FUSION, D2D, Health 2000, Finrisk 1987, Finrisk 2002, and Savitaipale studies from which we

chose our FUSION GWA and replication cohorts (no overlap with Health 2000 replication cohort).

We also thank Terry Gliedt and Peggy White for informatics and analysis support. The Center for

Inherited Disease Research performed the GWA genotyping. Support for this study was provided

by the following: NIH grants DK069922 (R.M.W.), U54 DA021519 (R.M.W.), DK062370 (M.B.),

and DK072193 (K.L.M.). Additional support comes from the National Human Genome Research

Institute intramural project number 1 Z01 HG000024 (F.S.C.).

MESYBEPO: We are indebted to all subjects who participated to this study. This work was

supported by grants from "Agence Nationale de la Recherche" and "Conseil Regional Nord-Pas de

Calais/Fonds européen de développement économique et régional". We thank Marianne Deweider

and Fredéric Allegaert for the DNA bank management, Philippe Gallina, Philippe Delfosse and

Sylvie Poulain for the recruitment of most obese adults, "Department of Nutrition of Paris Hotel

Dieu Hospital" for the contribution to the recruitment effort and Stéphane Lobbens for genotyping.

METSIM: Support was provided by grant 124243 from the Academy of Finland.

MeSyBePo: J.S. was supported by a Clinical Research Group (KFO218/1) and a Heisenberg-

Professorship (SP716/2-1) of the Deutsche Forschungsgemeinschaft (DFG). J.S. was also funded by

a research group on molecular nutrition of the Bundesministerium für Bildung und Forschung

(BMBF).


25

NHANES: We thank Dr. Sekar Kathiresan at the Massachusetts General Hospital and Broad

Institute of Harvard and MIT for his collaboration in assembling the NHANES DNA samples, and

Jody E. McLean at the Division of Health and Nutrition Examination Surveys, National Center for

Health Statistics, Centers for Disease Control and Prevention, for her assistance with the analysis of

NHANES genetic data.

Partners/Roche: We thank Roche Pharmaceuticals for its support and collaboration in assembling

the Partners/Roche cohort. Supported by the Mallinckrodt General Clinical Research Program,

National Center for Research Resources, NIH RR01066.

PIVUS and ULSAM: Genotyping was performed by the SNP&SEQ Technology Platform in

Uppsala (www.genotyping.se). We thank Tomas Axelsson, Ann-Christine Wiman and Caisa

Pöntinen for their excellent assistance with genotyping. The SNP Technology Platform is supported

by Uppsala University, Uppsala University Hospital and the Swedish Research Council for

Infrastructures. E.I. was supported by grants from the Swedish Research Council, the Swedish

Heart-Lung Foundation, the Swedish Foundation for Strategic Research, and the Royal Swedish

Academy of Science.

QTL families and UNG92: Lundbeck Foundation Centre of Applied Medical Genomics for

Personalized Disease Prediction, Prevention and Care (LuCAMP), The Danish Diabetes

Association and The Danish Research Council.

RISC: The RISC Study is supported by European Union grant QLG1-CT-2001-01252 and

AstraZeneca.

Sorbs: This work was supported by grants from the German Research Council (SFB- 1052 “Obesity

mechanisms”), from the German Diabetes Association and from the DHFD (Diabetes Hilfs- und

Forschungsfonds Deutschland). P.K. is funded by the Boehringer Ingelheim Foundation. IFB

Adiposity Diseases is supported by the Federal Ministry of Education and Research (BMBF),

Germany, FKZ: 01EO1001.

We would like to thank Knut Krohn (Microarray Core Facility of the Interdisciplinary Centre for

Clinical Research, University of Leipzig) for the genotyping/analytical support and Joachim Thiery

(Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics, University of

Leipzig) for clinical chemistry services. We thank Nigel W. Rayner (WTCHG, University of

Oxford, UK) for the excellent bioinformatics support. I.P. and V.L. were partial funded through the

European Community's Seventh Framework Programme (FP7/2007-2013), ENGAGE project, grant

agreement HEALTH-F4-2007-201413. R.M. is funded by European Commission under the Marie

Curie Intra-European Fellowship and an EFSD New Horizons grant. M.I.M. is supported by

Wellcome Trust grant 098381.

Stanford: Sample collection for the Stanford IST cohort was carried out in part in the Clinical and

Translational Research Unit, Stanford University, with funds provided by the National Center for

Research Resources, 2 M01 RR000070, U.S. Public Health Service. J.W.K. is supported by an

AHA Fellow to Faculty Transition award, 10FTF3360005.


26

CONFLICTS OF INTEREST

I.B. and spouse own stock in GlaxoSmithKline and Incyte. J.B.M. currently serves on a consultancy

board for Interleukin Genetics. R.M.W. has received consulting honoraria from Merck & Co.,

receives research funding from Merck & Co., and receives research material support from Takeda

Pharmaceuticals North America. J.C.F. has received consulting honoraria from Daiichi-Sankyo and

AstraZeneca. A.S.D., V.L., A.Ba., J.W.K., R.M., M-F.H., A.Be., D.R., A.U.J., H.M.S., C.S.,A.F-R.,

T.W.B., N.G., F.A.A., T.L.A., K.H., X.Y., C.Le., L.L.B., Y.B., S.B., P.S.C., M.R.E., J.G., P.K.,

M.A.M., N.N., F.P., A.S., A.J.S., A.T., S.R.B., S.C., P.F., D.M., P.E.H.S., H-U.H., U.S., M.B.,

R.N.B., F.S.C., K.L.M., J.T., T.Q., L.L., T.H., O.P., M.W., A.F.H.P., J.S., M.S., N.J.W., J.K., M.L.,

C.La., J.D., E.I., M.I.M., I.P. declare no conflict of interest.


27

REFERENCES

1. Stumvoll, M., B.J. Goldstein, and T.W. van Haeften, Type 2 diabetes: principles of

pathogenesis and therapy. Lancet, 2005. 365(9467): p. 1333-46.

2. Voight, B.F., et al., Twelve type 2 diabetes susceptibility loci identified through large-scale

association analysis. Nat Genet, 2010. 42(7): p. 579-89.

3. Morris, A.P., et al., Large-scale association analysis provides insights into the genetic

architecture and pathophysiology of type 2 diabetes. Nat Genet, 2012. 44(9): p. 981-990.

4. Dupuis, J., et al., New genetic loci implicated in fasting glucose homeostasis and their

impact on type 2 diabetes risk. Nat Genet, 2010. 42(2): p. 105-16.

5. Billings, L.K. and J.C. Florez, The genetics of type 2 diabetes: what have we learned from

GWAS? Ann N Y Acad Sci, 2010. 1212: p. 59-77.

6. Florez, J.C., Newly identified loci highlight beta cell dysfunction as a key cause of type 2

diabetes: where are the insulin resistance genes? Diabetologia, 2008. 51(7): p. 1100-10.

7. Strawbridge, R.D., J; Prokopenko, I; et al., Genome-wide association identifies nine

common variants associated with fasting proinsulin levels and provides new insights into

the pathophysiology of type 2 diabetes. (in press). Diabetes, 2011.

8. Ashcroft, F.M. and P. Rorsman, Diabetes mellitus and the beta cell: the last ten years. Cell,

2012. 148(6): p. 1160-71.

9. Banasik, K., et al., The effect of FOXA2 rs1209523 on glucose-related phenotypes and risk

of type 2 diabetes in Danish individuals. BMC Med Genet, 2012. 13: p. 10.

10. Grarup, N., et al., The diabetogenic VPS13C/C2CD4A/C2CD4B rs7172432 variant impairs

glucose-stimulated insulin response in 5,722 non-diabetic Danish individuals. Diabetologia,

2011. 54(4): p. 789-94.

11. Jonsson, A., et al., Effect of Common Genetic Variants Associated with Type 2 Diabetes and

Glycemic Traits on alpha- and beta-cell Function and Insulin Action in Man. Diabetes,

2013.

12. Nielsen, T., et al., Type 2 diabetes risk allele near CENTD2 is associated with decreased

glucose-stimulated insulin release. Diabetologia, 2011. 54(5): p. 1052-6.

13. Wagner, R., et al., Glucose-raising genetic variants in MADD and ADCY5 impair

conversion of proinsulin to insulin. PLoS One, 2011. 6(8): p. e23639.

14. Ingelsson, E., et al., Detailed physiologic characterization reveals diverse mechanisms for

novel genetic Loci regulating glucose and insulin metabolism in humans. Diabetes, 2010.

59(5): p. 1266-75.

15. Ferrannini, E. and A. Mari, How to measure insulin sensitivity. Journal of hypertension,

1998. 16(7): p. 895-906.

16. Manning, A.K., et al., A genome-wide approach accounting for body mass index identifies

genetic variants influencing fasting glycemic traits and insulin resistance. Nat Genet, 2012.

44(6): p. 659-669.

17. Belfiore, F., S. Iannello, and G. Volpicelli, Insulin sensitivity indices calculated from basal

and OGTT-induced insulin, glucose, and FFA levels. Molecular genetics and metabolism,

1998. 63(2): p. 134-41.

18. Stumvoll, M., et al., Use of the oral glucose tolerance test to assess insulin release and

insulin sensitivity. Diabetes care, 2000. 23(3): p. 295-301.

19. Matsuda, M. and R.A. DeFronzo, Insulin sensitivity indices obtained from oral glucose

tolerance testing: comparison with the euglycemic insulin clamp. Diabetes care, 1999.

22(9): p. 1462-70.

20. Gutt, M., et al., Validation of the insulin sensitivity index (ISI(0,120)): comparison with

other measures. Diabetes research and clinical practice, 2000. 47(3): p. 177-84.


28

21. Fall, T., et al., The role of adiposity in cardiometabolic traits: a mendelian randomization

analysis. PLoS Med, 2013. 10(6): p. e1001474.

22. Frayling, T.M., et al., A common variant in the FTO gene is associated with body mass

index and predisposes to childhood and adult obesity. Science, 2007. 316(5826): p. 889-94.

23. International HapMap, C., A haplotype map of the human genome. Nature, 2005. 437(7063):

p. 1299-320.

24. Magi, R. and A.P. Morris, GWAMA: software for genome-wide association meta-analysis.

BMC Bioinformatics, 2010. 11: p. 288.

25. Willer, C.J., Y. Li, and G.R. Abecasis, METAL: fast and efficient meta-analysis of

genomewide association scans. Bioinformatics, 2010. 26(17): p. 2190-1.

26. Prokopenko, I., et al., Variants in MTNR1B influence fasting glucose levels. Nat Genet,

2009. 41(1): p. 77-81.

27. Zeggini, E., et al., Meta-analysis of genome-wide association data and large-scale

replication identifies additional susceptibility loci for type 2 diabetes. Nat Genet, 2008.

40(5): p. 638-45.

28. Everitt, B.S., S. Landau, and M. Leese, Cluster Analysis (Fourth ed.), ed. Arnold. 2001,

London.

29. Shimodaira, H., Approximately unbiased tests of regions using multistep-multiscale

bootstrap resampling. Annals of Statistics, 2004. 32: p. 2616-2641.

30. Caliński, T. and J. Harabasz, A dendrite method for cluster analysis. Communications in

Statistics, 1974. 3(1): p. 1-27.

31. Gonzalez-Sanchez, J.L., et al., Association of variants of the TCF7L2 gene with increases in

the risk of type 2 diabetes and the proinsulin:insulin ratio in the Spanish population.

Diabetologia, 2008. 51(11): p. 1993-7.

32. Kirchhoff, K., et al., Polymorphisms in the TCF7L2, CDKAL1 and SLC30A8 genes are

associated with impaired proinsulin conversion. Diabetologia, 2008. 51(4): p. 597-601.

33. Loos, R.J., et al., TCF7L2 polymorphisms modulate proinsulin levels and beta-cell function

in a British Europid population. Diabetes, 2007. 56(7): p. 1943-7.

34. Stolerman, E.S., et al., TCF7L2 variants are associated with increased proinsulin/insulin

ratios but not obesity traits in the Framingham Heart Study. Diabetologia, 2009. 52(4): p.

614-20.

35. Lyssenko, V., et al., Common variant in MTNR1B associated with increased risk of type 2

diabetes and impaired early insulin secretion. Nature genetics, 2009. 41(1): p. 82-8.

36. Langenberg, C., et al., Common genetic variation in the melatonin receptor 1B gene

(MTNR1B) is associated with decreased early-phase insulin response. Diabetologia, 2009.

52(8): p. 1537-42.

37. Matschinsky, F.M., Regulation of pancreatic beta-cell glucokinase: from basics to

therapeutics. Diabetes, 2002. 51 Suppl 3: p. S394-404.

38. Fajans, S.S., G.I. Bell, and K.S. Polonsky, Molecular mechanisms and clinical

pathophysiology of maturity-onset diabetes of the young. N Engl J Med, 2001. 345(13): p.

971-80.

39. Simonis-Bik, A.M., et al., Gene variants in the novel type 2 diabetes loci

CDC123/CAMK1D, THADA, ADAMTS9, BCL11A, and MTNR1B affect different aspects of

pancreatic beta-cell function. Diabetes, 2010. 59(1): p. 293-301.

40. Bell, G.I. and K.S. Polonsky, Diabetes mellitus and genetically programmed defects in beta-

cell function. Nature, 2001. 414(6865): p. 788-91.

41. Gloyn, A.L., et al., Activating mutations in the gene encoding the ATP-sensitive potassium-

channel subunit Kir6.2 and permanent neonatal diabetes. N Engl J Med, 2004. 350(18): p.

1838-49.


29

42. Awata, T., et al., Missense variations of the gene responsible for Wolfram syndrome

(WFS1/wolframin) in Japanese: possible contribution of the Arg456His mutation to type 1

diabetes as a nonautoimmune genetic basis. Biochem Biophys Res Commun, 2000. 268(2):

p. 612-6.

43. Small, K.S., et al., Identification of an imprinted master trans regulator at the KLF14 locus

related to multiple metabolic phenotypes. Nat Genet, 2011. 43(6): p. 561-4.

44. Helgason, A., et al., Refining the impact of TCF7L2 gene variants on type 2 diabetes and

adaptive evolution. Nat Genet, 2007. 39(2): p. 218-25.


30

Figure legends

Figure 1. Cluster analysis of effects of 36 type 2 diabetes loci on principal physiologic traits.

Clustering of traits with meta-analysis results from at least 10,000 individuals (principal traits). The

existence of five clusters was revealed using two clustering approaches: (A) Complete linkage

dendrogram of type 2 diabetes SNPs with P values (%) indicating the robustness of each branching

event (shown in red).We named the clusters as hyperglycemic (HG) loci linked to reduced beta-cell

function following glucose stimulation, insulin resistance (IR) loci with a primary effect on IR at

basal measurements, proinsulin (PI) locus linked to decreased fasting proinsulin, beta cell (BC) loci

associated with defective beta cell function, and unclassified (UC) loci with no apparent impact on

glycemic measures. Strong support exists for the baseline branching notes (strength P≥0.84)

whereas branching of IR from the BC-UC clade shows lesser evidence for support (strength

P=0.64). (B) Calinski index computed on the centroid-based clustering of type 2 diabetes SNPs

provides further evidence for the existence of five locus groups.

Figure 2. Scatter plots of allelic effect size estimates for selected trait pairs.

In each scatter plot loci were assigned to the groups defined from the cluster analysis of principal

traits (groups highlighted by different colors). (A) Insulinogenic index vs. FI: this plot highlights

the effects of loci linked to IR (PPARG, KLF14, IRS1, GCKR) with respect to FI and insulinogenic

index. (B) Insulinogenic index vs. FG: the plot reveals the largest impact of HG loci (MTNR1B and

GCK) on FG driven by reduced beta-cell function. Large negative effects on insulinogenic index are

also seen for CDKAL1 and HHEX/IDE, but with very modest effects on FG. (C) HOMA-B vs.

HOMA-IR: the plot shows the separation of the BC, HG and IR clusters. Cluster group colors are

HG: orange, IR: green, PI: pink, BC: red, UC: blue. Loci named in the box are coded numerically

within the respective scatter plot.


Figure 1. Cluster analysis of effects of 36 type 2 diabetes loci on principal physiologic traits. Clustering of traits with meta-analysis results from at least 10,000 individuals (principal traits). The

existence of five clusters was revealed using two clustering approaches: (A) Complete linkage dendrogram of type 2 diabetes SNPs with P values (%) indicating the robustness of each branching event (shown in

red).We named the clusters as hyperglycemic (HG) loci linked to reduced beta-cell function following glucose stimulation, insulin resistance (IR) loci with a primary effect on IR at basal measurements, proinsulin (PI) locus linked to decreased fasting proinsulin, beta cell (BC) loci associated with defective beta cell function,

and unclassified (UC) loci with no apparent impact on glycemic measures. Strong support exists for the

baseline branching notes (strength P≥0.84) whereas branching of IR from the BC-UC clade shows lesser evidence for support (strength P=0.64). (B) Calinski index computed on the centroid-based clustering of

type 2 diabetes SNPs provides as further evidence for the existence of five locus groups. 414x435mm (96 x 96 DPI)


305x523mm (96 x 96 DPI)


Figure 2. Scatter plots of allelic effect size estimates for selected trait pairs. In each scatter plot loci were assigned to the groups defined from the cluster analysis of principal traits

(groups highlighted by different colors). (A) Insulinogenic index vs. FI: this plot highlights the effects of loci linked to IR (PPARG, KLF14, IRS1, GCKR) with respect to FI and insulinogenic index. (B) Insulinogenic index vs. FG: the plot reveals the largest impact of HG loci (MTNR1B and GCK) on FG driven by reduced beta-cell function. Large negative effects on insulinogenic index are also seen for CDKAL1 and HHEX/IDE, but with very modest effects on FG. (C) HOMA-B vs. HOMA-IR: the plot shows the separation of the BC, HG and IR clusters. Cluster group colors are HG: orange, IR: green, PI: pink, BC: red, UC: blue. Loci named in the box

are coded numerically within the respective scatter plot. 557x398mm (96 x 96 DPI)


542x398mm (96 x 96 DPI)


572x398mm (96 x 96 DPI)


1

Impact of type 2 diabetes susceptibility variants on quantitative glycemic traits reveals

mechanistic heterogeneity

Running title: Physiological characterization of known type 2 diabetes loci

Antigone S. Dimas1,2*, Vasiliki Lagou1,3*, Adam Barker4*, Joshua W. Knowles5*, Reedik

Mägi1,3,6, Marie-France Hivert7,8, Andrea Benazzo9, Denis Rybin10, Anne U. Jackson11, Heather M. Stringham11, Ci Song12,13, Antje Fischer-Rosinsky14, Trine Welløv Boesgaard15, Niels

Grarup16, Fahim A. Abbasi5, Themistocles L. Assimes5, Ke Hao17, Xia Yang18, Cécile Lecoeur19, Inês Barroso20,21, Lori L. Bonnycastle22, Yvonne Böttcher23, Suzannah Bumpstead20, Peter S. Chines22, Michael R. Erdos22, Jurgen Graessler24, Peter Kovacs25, Mario A. Morken22, Narisu Narisu22, Felicity Payne20, Alena Stancakova26, Amy J. Swift22, Anke Tönjes23,27, Stefan

R. Bornstein24, Stéphane Cauchi19, Philippe Froguel19,28, David Meyre19,29, Peter E.H. Schwarz24, Hans-Ulrich Häring30, Ulf Smith31, Michael Boehnke11, Richard N. Bergman32,

Francis S. Collins22, Karen L. Mohlke33, Jaakko Tuomilehto34-36, Thomas Quertemous5, Lars Lind37, Torben Hansen16,38, Oluf Pedersen16,39-41, Mark Walker42, Andreas F.H. Pfeiffer14,43,

Joachim Spranger14, Michael Stumvoll23,27, James B. Meigs8,44, Nicholas J. Wareham4, Johanna Kuusisto26, Markku Laakso26, Claudia Langenberg4, Josée Dupuis45,46, Richard M.

Watanabe*47, Jose C. Florez*44,48,49, Erik Ingelsson*1,12, Mark I. McCarthy*1,3,50, Inga Prokopenko*1,3,28 on behalf of the MAGIC investigators

* These authors contributed equally to this work.

Correspondence should be addressed to:

Inga Prokopenko, MSc, PhD Department of Genomics of Common Disease School of Public Health Imperial College London Burlington Danes Building, Hammersmith Hospital, Du Cane Road, London, W12 0NN, UK Phone: +4420 759 46501 E-mail: [email protected] Prof. Mark I. McCarthy OCDEM, Churchill Hospital, University of Oxford Old Road, Headington, OX3 7LJ Email: [email protected] Phone: +44 1865 857298 Erik Ingelsson, MD, PhD, FAHA Professor of Molecular Epidemiology Department of Medical Sciences Molecular Epidemiology and Science for Life Laboratory UCR/MTC Dag Hammarskjölds väg 14B Uppsala Science Park SE-752 37 Uppsala Sweden


2

Phone: +46-70-7569422 Fax: +46-18-515570 E-mail: [email protected] Jose C. Florez, MD, PhD Diabetes Unit / Center for Human Genetic Research Simches Research Building - CPZN 5.250 Massachusetts General Hospital 185 Cambridge Street Boston, MA 02114 Phone: 617.643.3308 Fax: 617.643-6630 Richard M. Watanabe, PhD Departments of Preventive Medicine and Physiology & Biophysics Diabetes and Obesity Research Institute of USC Keck School of Medicine of USC 2250 Alcazar Street, CSC-204 Los Angeles, CA 90089-9073 Phone: 1-323-442-2053

Affiliations

1. Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, OX3 7BN, UK.

2. Alexander Fleming, Biomedical Sciences Research Center, 34 Fleming Street, Vari, 16672 Athens, Greece.

3. Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, UK, OX3 7LJ.

4. MRC Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital, Cambridge, UK.

5. Department of Medicine and Cardiovascular Institute, Stanford University School of Medicine, Stanford, CA 94305, USA.

6. Estonian Genome Center, University of Tartu, Tartu, 51010, Estonia. 7. Department of Medicine, Université de Sherbrooke, Sherbrooke (Quebec), Canada. 8. General Medicine Division, Massachusetts General Hospital, Boston,Massachusetts,

USA. 9. Department of Biology and Evolution, University of Ferrara, Ferrara, Italy. 10. Boston University Data Coordinating Center, Boston, Massachusetts, MA 02118, USA. 11. Department of Biostatistics and Center for Statistical Genetics, University of Michigan

School of Public Health, Ann Arbor, Michigan 48109, USA. 12. Department of Medical Sciences, Molecular Epidemiology and Science for Life

Laboratory, Uppsala University, Uppsala, Sweden. 13. Department of Medical Epidemiology and Biostatistics, Karolinska Institutet,

Stockholm, Sweden. 14. Charité-Universitätsmedizin Berlin, Department of Endocrinology and Metabolism. 15. Steno Diabetes Center, Gentofte, Denmark. 16. The Novo Nordisk Foundation Center for Basic Metabolic Research, Faculty of Health

and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.


3

17. Department of Genetics and Genomic Sciences, Institute of Genomics and Multiscale Biology, Mount Sinai School of Medicine, New York, NY 10029-6574, USA.

18. Department of Integrative Biology and Physiology, University of California, 610 Charles E. Young Dr. East, Los Angeles, CA 90095, USA.

19. CNRS UMR8199-Institute of Biology, Pasteur Institute, Lille 2-Droit et Santé University, Lille, France.

20. Wellcome Trust Sanger Institute, Hinxton, CB10 1SA, UK. 21. University of Cambridge Metabolic Research Laboratories and NIHR Cambridge

Biomedical Research Centre, Level 4, Institute of Metabolic Science, Box 289, Addenbrooke’s Hospital, Cambridge CB2 OQQ, UK.

22. Genome Technology Branch, National Human Genome Research Institute, Bethesda, MD.

23. Leipzig University Medical Center, IFB AdiposityDiseases, Liebigstr. 21, 04103 Leipzig, Germany.

24. Department of Medicine III, Division of Prevention and Care of Diabetes, University of Dresden, 01307 Dresden, Germany.

25. Interdisciplinary Center for Clinical Research (IZKF) Leipzig, Liebigstr. 21, 04103 Leipzig, Germany.

26. Department of Medicine, University of Eastern Finland and Kuopio University Hospital, 70210 Kuopio, Finland.

27. Department of Medicine, University of Leipzig, Liebigstr. 18, 04103 Leipzig, Germany. 28. Department of Genomics of common diseases, Imperial College London, London, UK. 29. Department of Clinical Epidemiology & Biostatistics, McMaster University, Hamilton,

Canada. 30. Department of Internal Medicine, Division of Endocrinology, Diabetology, Vascular

Medicine, Nephrology and Clinical Chemistry, University of Tübingen, Tübingen, Germany.

31. Lundberg Laboratory for Diabetes Research, Center of Excellence for Metabolic and Cardiovascular Research, Department of Molecular and Clinical Medicine, Sahlgrenska Academy, University of Gothenburg, SE-413 45 Gothenburg, Sweden.

32. Department of Physiology & Biophysics, Keck School of Medicine, University of Southern California, Los Angeles, California 90033, USA.

33. Department of Genetics, University of North Carolina Chapel Hill, North Carolina 27599, USA.

34. Diabetes Prevention Unit, National Institute for Health and Welfare, 00271 Helsinki, Finland.

35. Centre for Vascular Prevention, Danube-University Krems, 3500 Krems, Austria. 36. King Abdulaziz University, Jeddah 21589, Saudi Arabia. 37. Department of Medical Sciences, Uppsala University, Akademiska sjukhuset, Uppsala,

Sweden. 38. Faculty of Health Sciences, University of Southern Denmark, Odense, Denmark. 39. Hagedorn Research Institute, Copenhagen, Denmark. 40. Institute of Biomedical Science, Faculty of Health Sciences, University of Copenhagen,

Copenhagen, Denmark. 41. Faculty of Health Sciences, University of Aarhus, Aarhus, Denmark. 42. Institute of Cellular Medicine, Newcastle University, UK. 43. German Institute of Human Nutrition, Department of Clinical Nutrition. 44. Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA. 45. Department of Biostatistics, Boston University School of Public Health, Boston,

Massachusetts, MA 02118, USA.


4

46. The National Heart, Lung, and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts, USA.

47. Departments of Preventive Medicine and Physiology & Biophysics, Keck School of Medicine of USC, Los Angeles, CA 90033, USA.

48. Center for Human Genetic Research and Diabetes Research Center (Diabetes Unit), Massachusetts General Hospital, Boston, Massachusetts 02114, USA.

49. Program in Medical and Population Genetics, Broad Institute, Cambridge, Massachusetts 02142, USA.

50. Oxford NIHR Biomedical Research Centre, Churchill Hospital, Oxford OX3 7LJ, UK.

SUPPLEMENTARY NOTES AND FIGURES

Supplementary Note. Genotyping and quality control.

Supplementary Table 1. Cohort descriptions and baseline characteristics (Excel file attached).

Supplementary Table 2. Definitions of physiological measurements.

Supplementary Table 3. Genetic markers investigated.

Supplementary Table 4. Effects of 37 SNPs previously associated with type 2 diabetes on

physiologic glycemic measures with genetic data available in >10,000 non-diabetic individuals

(principal traits).

Supplementary Table 5. Effects of 37 SNPs previously associated with type 2 diabetes on

physiologic glycemic measures with genetic data available in <10,000 non-diabetic individuals.

Supplementary Figure 1. Forest plots for meta-analyses of insulinogenic index, split

proinsulin and acute insulin response (AIR). (A) insulinogenic index at HHEX/IDE, (B)

insulinogenic index at CDKAL1, (C) split proinsulin at HNF1A, (D) AIR at MTNR1B, (E) AIR

at KCNQ1, (F) AIR at CDKN2A/B.

Supplementary Figure 2. Forest plots for meta-analyses of intravenous (IV) insulin sensitivity

traits. (A) IRS1, (B) ADCY5.

Supplementary Figure 3. Results of cluster analysis using principal traits.

Supplementary Figure 4. Results of cluster analysis using all traits.

Supplementary Figure 5. Principal component analysis (PCA) of type 2 diabetes SNPs.

Supplementary Figure 6. Scatter plots of allelic effect size estimates for given trait pairs.


5

Supplementary note

Genotyping and quality control

Genotyping data was obtained de novo for eleven studies and in silico (HapMap imputed) for four studies. Initial association testing for each cohort was performed using Merlin, STATA 10.1, R (LMEKIN package, R 2.10.0 and R 2.9.2), SNPTEST v1.1.4, SPSS 18.0, and SAS (9.1.3 and 9.1.9). Prior to meta-analysis, low quality SNPs were filtered in individual cohorts using the following criteria: (1) call rate for directly genotyped SNPs <0.70, (2) SNP Hardy-Weinberg equilibrium P<10-6, (3) SNPs minor allele count (MAC) ≥10 (calculated as total number of observed alleles [twice the sample size] at each SNP multiplied by minor allele frequency [MAF]). In samples where initial genotyping of an index SNP failed, a proxy SNP in strong linkage disequilibrium (LD) with the original SNP was genotyped and reported wherever possible.

Supplementary Table 2. Definitions of physiological measurements.

Index Equation Reference

Insulinogenic

Index

[(30-min insulin-fasting insulin)]/[(30-min glucose-fasting

glucose)]

ISI Belfiore

2/[({[0.5*fasting PG (mmol/L)] + OGTT 1-h PG + (0.5*OGTT

2-h PG)} / 11.36) * ({[0.5*fasting PI (pmol/L)] + OGTT 1-h PI

+ (0.5*OGTT 2-h PI)} / 638) + 1]

(1)

ISI Gutt

{75000 + [fasting PG (mg/L) - OGTT 2-h PG] * 0.19 * body

weight (kg)/120} / [(fasting PG + OGTT 2-h PG) / 2] /

log10{[fasting PI (mU/L) + OGTT 2-h PI] / 2}

(2)

ISI Matsuda 10 000√{fasting PG (mg/dL) * fasting PI (µU/mL) * [mean

OGTT PG * mean OGTT PI]} (3)

ISI Stumvoll 0.226 - [0.0032 * BMI (kg/m2)] - [0.0000645 * OGTT 2-h PI

(pmol/L)] - [0.00375 * OGTT 1.5-h PG (mmol/L)] (4)

Abbreviations: PG: plasma glucose, OGTT: oral glucose tolerance test, PI: plasma insulin.


6

Supplementary Table 3. Genetic markers investigated.

Lead SNP Chr

Position

(b36)

Nearest

gene

Effect

allele*/

Other MAF Proxy used

r2

with

lead

SNP†

rs10830963 11 92348358 MNTR1B G/C 0.30 rs7112766 0.861 rs11020124 0.725 rs1387153 0.703 rs10830956 0.703

rs4607517 7 44202193 GCK A/G 0.20 rs730497 1 rs2908289 1 rs1799884 1 rs6975024 1

rs13081389 3 12264800 PPARG A/G 0.04 rs1596417 1 rs17036130 1 rs1562040 1 rs1801282 0.536

rs972283 7 130117394 KLF14 G/A 0.45 rs4731702 1 rs3996352 1 rs11765979 1 rs10954284 1

rs7578326 2 226728897 IRS1 A/G 0.36 rs2943651 0.964 rs2943632 0.964 rs13423088 0.929 rs2943641 0.74

rs780094 2 27594741 GCKR C/T 0.38 rs780093 1 rs1260326 0.932 rs2911711 0.87 rs1260333 0.87

rs1552224 11 72110746 ARAP1 A/C 0.13 rs11603334 1 rs11605166 0.853 rs613937 0.736

rs7903146 10 114748339 TCF7L2 T/C 0.25 rs4506565 0.917 rs4132670 0.917

rs3802177 8 118254206 SLC30A8 G/A 0.25 rs13266634 1 rs11558471 0.957

rs5015480 10 94455539 HHEX/

IDE C/T 0.45 rs10882102 1 rs1111875 1 rs12778642 0.904

rs2191349 7 15030834 DGKB/

TMEM195 T/G 0.47 rs2191348 1 rs4719433 1 rs6947830 1 rs2215383 1 rs10244051 1

rs340874 1 212225879 PROX1 C/T 0.49 rs340835 0.765


7

rs340883 0.548

rs10965250 9 22123284 CDKN2A/

B G/A 0.19 rs10965250 1 rs10811661 0.95 rs2383208 0.95

rs10440833 6 20796100 CDKAL1 A/T 0.25 rs9368222 1 rs7766070 1 rs7756992 1 rs7754840 0.677

rs11899863 2 43472323 THADA C/T 0.07 rs7607777 1 rs6743071 1 rs13408002 1 rs17406646 1 rs7578597 0.786

rs11708067 3 124548468 ADCY5 A/G 0.22 rs11717195 0.86 rs2124500 0.819 rs6798189 0.795 rs7613951 0.753

rs4430796 17 33172153 HNF1B

(TCF2) G/A 0.47 rs2005705 0.632 rs757210 0.615

rs2334499 11 1653425 HCCA2 T/C 0.44 rs4752780 0.964 rs4417225 0.934 rs4752781 0.931 rs10839220 0.583

rs7957197 12 119945069 HNF1A T/A 0.15 rs12427353 0.832 rs7965349 0.818 rs7305618 0.55

rs10923931 1 120319482 NOTCH2 T/G 0.12 rs2641317 1 rs2934381 1 rs1493694 1 rs2453040 1 rs2793831 1

rs13292136 9 81141948 CHCHD9 C/T 0.07 rs17791555 1 rs13297387 1 rs1328405 1 rs1328405 1

rs1801214 4 6353923 WFS1 T/C 0.27 rs6446480 0.917 rs10937720 0.917 rs13108780 0.917 rs13128674 0.917 rs10010131 0.917

rs4457053 5 76460705 ZBED3 G/A 0.26 rs6878122 0.864 rs7708285 0.862

rs8042680 15 89322341 PRC1 A/C 0.22 rs6496743 1 rs6496742 0.811 rs11853073 0.811 rs8028856 0.81

rs1470579 3 187011774 IGF2BP2 C/A 0.29 rs1470579 1


8

rs6767484 1 rs6769511 1 rs11711477 1 rs9859406 1

rs896854 8 96029687 TP53INP1 T/C 0.48 rs13255935 1 rs13257021 0.967 rs7845219 0.846 rs12548874 0.743

rs12779790 10 12368016 CDC123/

CAMK1D G/A 0.23 rs11257655 0.802 rs7069060 0.617 rs10906111 0.562

rs11634397 15 78219277 ZFAND6 G/A 0.41 rs870185 0.897 rs12441266 0.897 rs1357335 0.897 rs4778582 0.896

rs4760790 12 69921061 TSPAN8/

LGR5 A/G 0.23 rs4760915 1 rs7306184 1 rs1353362 0.955 rs7961581 0.909

rs5215 11 17365206 KCNJ11 C/T 0.41 rs1557765 0.902 rs757110 0.899 rs7124355 0.745

rs243021 2 60438323 BCL11A A/G 0.46 rs243020 1 rs243019 1 rs243018 0.967

rs849134 7 28162747 JAZF1 A/G 0.47 rs849135 1 rs849133 1 rs1635852 1 rs864745 0.967

rs1531343 12 64461161 HMGA2 C/G 0.10 rs2612067 1 rs2583922 1 rs2358949 1 rs1122590 1 rs2884592 1

rs231362 11 2648047 KCNQ1 G/A 0.49 rs163184 11 2803645 KCNQ1 G/T 0.44

rs6795735 3 64680405 ADAMTS9 C/T 0.45 rs7428936 1 rs4611812 1 rs17727064 1 rs4607103 0.28

rs5945326 X 152553116 DUSP9 A/G 0.21 rs3020789 0.939 rs2301142 0.548

*-Effect allele defined as the type 2 diabetes risk increasing allele. †-r2 values based on HapMap CEU, release 22


9

Supplementary Table 4. Effects of 37 SNPs previously associated with type 2 diabetes on physiologic glycemic measures with genetic data

available in more than 10,000 non-diabetic individuals.

Dynamic traits

Fasting traits Insulino- genic index (n=11,268)

ISI

Belfiore

(n=10,34)

ISI

Stumvoll

(n=10,23

9)

ISI

Matsuda

(n=10,364)

ISI Gutt

(n=13,158)

Proinsulin

(n=13,912)

Fasting

Glucose*

(n=58,614)

Fasting

Insulin*

(n=52,379) HOMA-B*

(n=50,908)

HOMA-

IR*

(n=50,908)

SNP Nearest

gene

Alleles

(effect

/

other)

Effect (SE) P

Effect

(SE) P

Effect

(SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

Insulin Resistance (IR)

rs13081389 PPARG A/G 0.0092 (0.017)

-0.019 (0.0069)

-0.021 (0.0082)

-0.035 (0.012)

-0.015 (0.0071)

0.0082 (0.0088)

0.0097 (0.0062)

0.025 (0.0052)

0.017 (0.0048)

0.025 (0.0055)

0.59 0.0055 0.012 0.0037 0.039 0.35 0.12 1.6×10-6 4.96×10-4 4.37×10-6

rs972283 KLF14 G/A -0.0020 (0.011)

-0.0030 (0.0042)

-0.0083 (0.0052)

-0.0013 (0.0077)

-0.0079 (0.0041)

0.0045 (0.0052)

0.0066 (0.0032)

0.013 (0.0027)

0.0091 (0.0026)

0.014 (0.0029)

0.85 0.47 0.11 0.86 0.055 0.39 0.040 4.41×10-6 4.16×10-4 6.84×10-7

rs7578326 IRS1 A/G 0.037 (0.010)

-0.023 (0.0043)

-0.013 (0.0050)

-0.040 (0.0075)

-0.020 (0.0042)

0.014 (0.0052)

0.0045 (0.0034)

0.019 (0.0029)

0.014 (0.0027)

0.019 (0.0031)

0.00038 1.19×10-7 0.0090 1.46×10-7 2.47×10-6 0.0063 0.19 2.68×10-11 2.65×10-7 6.95×10-10

rs780094 GCKR C/T 0.0090 (0.010)

-0.0077 (0.0041)

-0.015 (0.0048)

-0.019 (0.0073)

0.0072 (0.0039)

0.010 (0.0053)

0.033 (0.0032)

0.019 (0.0027)

0.0058 (0.0026)

0.023 (0.0029)

0.38 0.061 0.0020 0.0077 0.061 0.050 3.30×10-24 3.36×10-12 0.022 4.68×10-16 Hyperglycemic (HG)

rs10830963 MTNR1B G/C -0.083 (0.011)

0.0057 (0.0043)

-0.0073 (0.0051)

-0.0043 (0.0076)

-0.0059 (0.0043)

0.0093 (0.0058)

0.079 (0.0038)

0.0004 (0.0033)

-0.033 (0.0031)

0.011 (0.0035)

1.23×10-14 0.18 0.15 0.57 0.17 0.11 8.72×10-95 0.91 4.23×10-26 0.0018

rs4607517 GCK A/G -0.036 (0.017)

-0.0048 (0.0070)

-0.015 (0.0083)

-0.0078 (0.012)

-0.016 (0.0060)

-0.00068 (0.0080)

0.065 (0.0043)

0.0064 (0.0037)

-0.024 (0.0035)

0.014 (0.0039)

0.031 0.49 0.077 0.52 0.0069 0.93 1.39×10-51 0.081 1.32×10-11 3.89×10-4 Proinsulin (PI)


10

rs1552224 ARAP1 A/C -0.054 (0.012)

-0.00038 (0.0048)

-0.013 (0.0062)

-0.013 (0.0089)

-0.010 (0.0052)

-0.093 (0.0062)

0.024 (0.0042)

-0.0075 (0.0034)

-0.015 (0.0032)

-0.0051 (0.0037)

1.54×10-5 0.94 0.033 0.15 0.050 2.16×10-50 9.40×10-9 0.029 3.16×10-6 0.16 Beta Cell (BC)

rs7903146 TCF7L2 T/C -0.063 (0.012)

0.0094 (0.0051)

-0.0085 (0.0058)

0.010 (0.0085)

-0.017 (0.0046)

0.044 (0.0061)

0.026 (0.0036)

-0.014 (0.0031)

-0.019 (0.0029)

-0.010 (0.0033)

1.40×10-7 0.064 0.14 0.24 0.00021 4.20×10-13 9.55×10-13 4.81×10-6 4.25×10-11 0.0022

rs3802177 SLC30A8 G/A -0.041 (0.011)

0.00066 (0.0043)

-0.0039 (0.0055)

-0.0012 (0.0081)

-0.011 (0.0050)

0.038 (0.0062)

0.033 (0.0035)

-0.0035 (0.0031)

-0.016 (0.0028)

0.0002 (0.0033)

0.00024 0.88 0.47 0.88 0.034 1.25×10-9 4.32×10-21 0.26 3.41×10-8 0.96

rs5015480 HHEX/ID

E C/T -0.10 (0.011) 0.020 (0.0042)

-0.010 (0.0053)

0.017 (0.0078)

-0.0052 (0.0041)

0.0063 (0.0053)

0.010 (0.0032)

-0.0008 (0.0027)

-0.0051 (0.0025)

0.0007 (0.0028)

3.63×10-21 1.46×10-6 0.054 0.030 0.21 0.23 0.0011 0.77 0.042 0.81

rs2191349 DGKB/TMEM195 T/G

-0.028 (0.010)

0.011 (0.0040)

0.0071 (0.0047)

0.027 (0.0072)

0.0060 (0.0038)

-0.0011 (0.0049)

0.032 (0.0032)

-0.0017 (0.0027)

-0.017 (0.0025)

0.0018 (0.0028)

0.0058 0.0059 0.13 0.00018 0.12 0.83 1.15×10-24 0.51 2.07×10-11 0.52

rs340874 PROX1 C/T -0.028 (0.010)

0.0082 (0.0041)

0.0079 (0.0047)

0.0090 (0.0072)

-0.00066 (0.0040)

0.0086 (0.0055)

0.017 (0.0032)

-0.0019 (0.0028)

-0.0094 (0.0026)

-0.0003 (0.003)

0.0060 0.044 0.096 0.21 0.87 0.12 6.80×10-8 0.50 2.45×10-4 0.91

rs10965250 CDKN2A/

B G/A -0.045 (0.014)

0.013 (0.0057)

-0.00072 (0.0070)

0.012 (0.010)

0.0046 (0.0054)

-0.00039 (0.0071)

0.017 (0.0042)

-0.0049 (0.0037)

-0.012 (0.0035)

-0.0023 (0.0039)

0.0013 0.023 0.92 0.26 0.39 0.96 5.17×10-5 0.18 4.27×10-4 0.55

rs10440833 CDKAL1 A/T -0.10 (0.023) 0.0047 (0.0090)

-0.011 (0.010)

-0.0095 (0.017)

-0.0013 (0.0066)

0.0077 (0.0085)

0.014 (0.0035)

-0.0061 (0.0029)

-0.0088 (0.0027)

-0.0043 (0.0031)

6.84×10-6 0.60 0.30 0.58 0.84 0.37 5.19×10-5 0.036 0.0012 0.17

rs11899863 THADA C/T 0.0078 (0.020)

0.0096 (0.0080)

-0.0036 (0.0095)

0.016 (0.015)

-0.0071 (0.0072)

0.015 (0.0097)

0.020 (0.0054)

-0.010 (0.0048)

-0.020 (0.0046)

-0.0080 (0.0051)

0.70 0.23 0.70 0.28 0.32 0.12 2.28×10-4 0.036 1.30×10-5 0.12

rs11708067 ADCY5 A/G -0.0089 (0.013)

-0.0079 (0.0051)

0.0012 (0.0062)

-0.0025 (0.0093)

-0.014 (0.0047)

0.016 (0.0064)

0.022 (0.0039)

-0.0060 (0.0034)

-0.015 (0.0032)

-0.0011 (0.0036)

0.49 0.13 0.85 0.79 0.0024 0.012 1.89×10-8 0.076 5.39×10-6 0.76 Unclassified (UC)

rs4430796 HNF1B

(TCF2) G/A -0.044 (0.023)

0.0072 (0.0091)

-0.011 (0.011)

0.0082 (0.018)

-0.013 (0.0095)

-0.00053 (0.016)

-0.0007 (0.0039)

-0.012 (0.0035)

-0.011 (0.0031)

-0.012 (0.0037)

0.061 0.43 0.32 0.64 0.17 0.97 0.85 8.25×10-4 5.14×10-4 9.08×10-4


11

rs2334499 HCCA2 T/C -0.018 (0.023)

0.0098 (0.0089)

0.016 (0.0099)

0.015 (0.017)

-0.0057 (0.0068)

-0.017 (0.0086)

0.0011 (0.0032)

-0.0044 (0.0027)

-0.0036 (0.0026)

-0.0037 (0.0029)

0.43 0.27 0.11 0.38 0.40 0.043 0.73 0.11 0.16 0.21

rs7957197 HNF1A T/A -0.036 (0.018)

0.0086 (0.0075)

-0.0062 (0.0091)

-0.0029 (0.014)

0.00049 (0.0061)

-0.011 (0.0087)

-0.0014 (0.004)

-0.0043 (0.0033)

-0.0028 (0.0031)

-0.0045 (0.0035)

0.049 0.25 0.49 0.83 0.94 0.22 0.73 0.20 0.36 0.20

rs10923931 NOTCH2 T/G -0.012 (0.016)

0.011 (0.0065)

0.0026 (0.0078)

0.019 (0.012)

0.0011 (0.0065)

-0.000015 (0.0084)

0.0025 (0.005)

-0.0034 (0.0042)

-0.0043 (0.0039)

-0.0014 (0.0044)

0.45 0.085 0.74 0.11 0.87 1.00 0.63 0.41 0.28 0.76

rs13292136 CHCHD9 C/T -0.025 (0.016)

0.012 (0.0069)

-0.0013 (0.0084)

0.026 (0.012)

0.0063 (0.0069)

-0.00054 (0.0086)

-0.0028 (0.0057)

-0.0054 (0.0046)

-0.0022 (0.0042)

-0.0050 (0.0049)

0.13 0.078 0.87 0.028 0.36 0.95 0.63 0.24 0.61 0.30

rs1801214 WFS1 T/C -0.017 (0.011)

0.0016 (0.0045)

-0.00084 (0.0052)

0.0082 (0.0080)

-0.0076 (0.0042)

-0.019 (0.0053)

0.0081 (0.0033)

-0.0011 (0.0028)

-0.0054 (0.0027)

0.0003 (0.003)

0.12 0.72 0.87 0.30 0.071 0.00023 0.016 0.67 0.042 0.92

rs4457053 ZBED3 G/A -0.014 (0.012)

0.00038 (0.0049)

0.0021 (0.0058)

0.00039 (0.0083)

-0.0040 (0.0045)

-0.0078 (0.0055)

0.019 (0.004)

0.0016 (0.0033)

-0.0053 (0.0031)

0.0031 (0.0035)

0.21 0.94 0.72 0.96 0.36 0.16 2.38×10-6 0.61 0.084 0.38

rs8042680 PRC1 A/C -0.016 (0.011)

0.0030 (0.0046)

0.0074 (0.0053)

0.011 (0.0081)

-0.0021 (0.0043)

-0.0082 (0.0057)

0.011 (0.0034)

-0.0016 (0.0028)

-0.0036 (0.0027)

-0.0003 (0.003)

0.15 0.51 0.16 0.19 0.62 0.15 0.0015 0.56 0.17 0.93

rs1470579 IGF2BP2 C/A -0.038 (0.018)

0.013 (0.0068)

0.0096 (0.0079)

0.0096 (0.013)

-0.00083 (0.0054)

-0.011 (0.0073)

0.016 (0.0034)

-0.0012 (0.0028)

-0.0062 (0.0027)

0.0001 (0.003)

0.030 0.062 0.23 0.46 0.88 0.15 2.34×10-6 0.66 0.022 0.96

rs896854 TP53INP

1 T/C -0.0035 (0.010)

-0.0013 (0.0038)

-0.0056 (0.0049)

-0.0025 (0.0069)

0.00026 (0.0040)

-0.0031 (0.0049)

0.0069 (0.0031)

-0.0023 (0.0026)

-0.0045 (0.0025)

-0.0013 (0.0028)

0.73 0.74 0.26 0.72 0.95 0.53 0.028 0.38 0.070 0.65

rs12779790 CDC123/

CAMK1D G/A -0.029 (0.013)

0.0014 (0.0055)

-0.0021 (0.0064)

-0.0060 (0.0096)

0.0016 (0.0053)

-0.0082 (0.0068)

0.014 (0.0042)

-0.0018 (0.0036)

-0.0068 (0.0034)

-0.0009 (0.0038)

0.029 0.80 0.74 0.53 0.76 0.23 5.61×10-4 0.62 0.046 0.81

rs11634397 ZFAND6 G/A -0.0098 (0.010)

-0.00098 (0.0043)

-0.0055 (0.0052)

-0.0068 (0.0071)

-0.0054 (0.0042)

0.0081 (0.0051)

0.0014 (0.0035)

0.0030 (0.0029)

0.0023 (0.0028)

0.0029 (0.0031)

0.34 0.82 0.29 0.34 0.19 0.11 0.68 0.31 0.41 0.35

rs4760790 TSPAN8/

LGR5 A/G -0.016 (0.013)

0.00096 (0.0050)

0.0096 (0.0061)

-0.0035 (0.0092)

-0.00039 (0.0046)

0.0061 (0.0060)

0.0040 (0.0036)

-0.0005 (0.003)

-0.0021 (0.0029)

0.0005 (0.0032)


12

0.21 0.85 0.11 0.70 0.93 0.31 0.27 0.88 0.47 0.89

rs5215 KCNJ11 C/T -0.024 (0.017)

0.0017 (0.0065)

-0.015 (0.0076)

0.00091 (0.012)

0.0035 (0.0052)

0.0010 (0.0071)

-0.0040 (0.0032)

-0.0010 (0.0027)

0.0027 (0.0025)

-0.0020 (0.0029)

0.15 0.79 0.057 0.94 0.50 0.89 0.22 0.72 0.28 0.48

rs243021 BCL11A A/G 0.018 (0.011)

0.0045 (0.0040)

0.0084 (0.0051)

0.013 (0.0074)

0.0090 (0.0041)

0.0070 (0.0052)

0.0038 (0.0032)

0.0009 (0.0027)

0.0003 (0.0025)

0.00 (0.0029)

0.093 0.26 0.10 0.067 0.029 0.18 0.23 0.72 0.89 0.99

rs849134 JAZF1 A/G 0.013 (0.011)

-0.0052 (0.0042)

-0.0047 (0.0052)

-0.014 (0.0077)

-0.0061 (0.0041)

0.0035 (0.0052)

0.0074 (0.0032)

-0.0001 (0.0027)

-0.0014 (0.0025)

-0.0002 (0.0029)

0.23 0.21 0.37 0.068 0.14 0.50 0.019 0.97 0.57 0.943

rs1531343 HMGA2 C/G 0.030 (0.019)

-0.013 (0.0076)

0.0017 (0.0087)

-0.027 (0.014)

0.0012 (0.0067)

0.010 (0.0090)

0.013 (0.0054)

0.0074 (0.0045)

-0.0013 (0.0043)

0.011 (0.0048)

0.11 0.084 0.84 0.050 0.86 0.26 0.018 0.10 0.76 0.029

rs231362 KCNQ1 G/A 0.0045 (0.011)

-0.000085 (0.0045)

-0.0039 (0.0053)

-0.0033 (0.0079)

-0.0051 (0.0044)

-0.0037 (0.0052)

0.017 (0.0035)

0.0069 (0.003)

-0.0010 (0.0028)

0.0066 (0.0032)

0.68 0.98 0.46 0.67 0.24 0.48 2.7×10-6 0.020 0.72 0.038

rs163184 KCNQ1 G/T -0.011 (0.022)

0.0024 (0.0088)

0.0092 (0.0096)

0.0087 (0.017)

-0.0037 (0.0079)

0.016 (0.010)

0.012 (0.0032)

0.0019 (0.0028)

-0.0045 (0.0027)

0.0039 (0.003)

0.63 0.79 0.34 0.60 0.64 0.11 1.90×10-4 0.51 0.091 0.20

rs6795735 ADAMTS

9 C/T 0.0088 (0.010)

0.0060 (0.0043)

0.0027 (0.0051)

0.011 (0.0076)

0.0012 (0.0041)

0.00056 (0.0050)

0.0076 (0.0032)

0.0025 (0.0027)

-0.0014 (0.0025)

0.0032 (0.0029)

0.40 0.16 0.60 0.14 0.77 0.91 0.017 0.35 0.57 0.26 Other

rs5945326 DUSP9 A/G -0.037 (0.010)

0.012 (0.0038)

0.010 (0.0048)

0.022 (0.0067)

0.0059 (0.0047)

-0.013 (0.0060)

0.011 (0.013)

-0.013 (0.0138)

-2.89 (1.3345)

-0.066 (0.041)

0.00032 0.0017 0.033 0.00085 0.21 0.030 0.40 0.35 0.030 0.11


13

Supplementary Table 5. Effects of 37 SNPs previously associated with type 2 diabetes on physiologic glycemic measures with genetic data

available in fewer than 10,000 non-diabetic individuals. IR: insulin resistance, HG: hyperglycemic, PI: proinsulin, BC: beta cell, UC: unclassified.

SNP Nearest gene Alleles

(effect/other)

M/I, SI, SSPG

(n=4,180) Acute Insulin

Response

(n=1,135)

C-Peptide

(n=5,059) 32,33 Split

Proinsulin

(n=2,568)

Effect (SE) P

Effect (SE) P

Effect (SE) P

Effect (SE) P

IR

rs13081389 PPARG A/G -0.025 (0.049) -0.025 (0.066) 0.015 (0.012) 0.0050 (0.032)

0.60 0.70 0.22 0.88 rs972283 KLF14 G/A -0.0023 (0.020) 0.019 (0.043) 0.0068 (0.0084) 0.0024 (0.017)

0.91 0.65 0.42 0.88 rs7578326 IRS1 A/G -0.070 (0.021) -0.0061 (0.043) 0.021 (0.0083) 0.021 (0.018)

0.00065 0.89 0.013 0.24 rs780094 GCKR C/T -0.019 (0.019) 0.034 (0.033) 0.020 (0.0074) 0.027 (0.017)

0.32 0.31 0.0079 0.12

HG

rs10830963 MTNR1B G/C 0.012 (0.021) -0.16 (0.035) -0.0069 (0.0080) -0.0034 (0.019)

0.55 4.81×10-6 0.39 0.86 rs4607517 GCK A/G 0.015 (0.025) -0.097 (0.050) -0.014 (0.014) -0.0097 (0.038)

0.55 0.051 0.32 0.80

PI

rs1552224

ARAP1

A/C -0.010 (0.026) 0.69

-0.0064 (0.054)

0.91

0.019 (0.012)

0.11

-0.13 (0.023)

1.39×10-8

BC

rs7903146 TCF7L2 T/C 0.014 (0.021) -0.054 (0.039) -0.021 (0.011) 0.012 (0.030) 0.50 0.17 0.051 0.68

rs3802177 SLC30A8 G/A -0.014 (0.020) -0.11 (0.033) 0.0096 (0.0077) 0.066 (0.018)


14

0.47 0.0016 0.21 0.00019 rs5015480 HHEX/IDE C/T -0.0040 (0.020) -0.11 (0.044) 0.0011 (0.0087) 0.017 (0.017)

0.84 0.013 0.90 0.33 rs2191349 DGKB/TMEM195 T/G -0.00054 (0.018) -0.013 (0.032) -0.0055 (0.0072) -0.00057 (0.017)

0.98 0.67 0.44 0.97 rs340874 PROX1 C/T -0.0020 (0.018) -0.043 (0.032) -0.00058 (0.0072) 0.00035 (0.016)

0.91 0.18 0.94 0.98 rs10965250 CDKN2A/B G/A 0.047 (0.024) -0.18 (0.044) -0.0022 (0.011) 0.011 (0.022)

0.051 3.57×10-5 0.84 0.60

rs10440833 CDKAL1 A/T 0.047 (0.021) -0.15 (0.048) 0.00083 (0.013) -0.042 (0.029) 0.026 0.0020 0.95 0.15

rs11899863 THADA C/T -0.053 (0.032) 0.013 (0.066) 0.012 (0.013) 0.038 (0.028)

0.10 0.84 0.38 0.17 rs11708067 ADCY5 A/G -0.063 (0.023) -0.013 (0.040) -0.0018 (0.0088) 0.062 (0.019)

0.0061 0.75 0.84 0.0012

UC

rs4430796 HNF1B (TCF2) G/A 0.025 (0.032) -0.021 (0.042) 0.00040 (0.0087) 0.030 (0.022)

0.45 0.61 0.96 0.18 rs2334499 HCCA2 T/C 0.031 (0.021) - 0.0043 (0.018) -0.016 (0.026)

0.14 - 0.81 0.53 rs7957197 HNF1A T/A 0.0053 (0.024) -0.079 (0.054) 0.0094 (0.011) -0.094 (0.021)

0.82 0.14 0.39 8.08×10-6

rs10923931 NOTCH2 T/G 0.038 (0.030) -0.083 (0.066) -0.017 (0.014) -0.0073 (0.027) 0.21 0.21 0.22 0.78

rs13292136 CHCHD9 C/T 0.034 (0.042) 0.040 (0.068) -0.0091 (0.017) -0.010 (0.044)

0.43 0.56 0.58 0.82 rs1801214 WFS1 T/C -0.016 (0.023) -0.063 (0.034) -0.0083 (0.0079) -0.0048 (0.021)

0.48 0.063 0.29 0.82


15

rs4457053 ZBED3 G/A -0.0084 (0.023) -0.013 (0.046) -0.017 (0.0098) -0.014 (0.018)

0.72 0.77 0.084 0.43 rs8042680 PRC1 A/C 0.0034 (0.025) 0.00053 (0.043) -0.014 (0.0092) -0.013 (0.024)

0.89 0.99 0.12 0.57 rs1470579 IGF2BP2 C/A 0.013 (0.021) -0.15 (0.046) -0.0026 (0.0091) -0.0040 (0.018)

0.54 0.0014 0.77 0.82 rs896854 TP53INP1 T/C 0.0071 (0.019) 0.030 (0.041) -0.012 (0.0086) -0.022 (0.017)

0.70 0.46 0.16 0.19 rs12779790 CDC123/CAMK1D G/A 0.020 (0.024) -0.015 (0.053) -0.028 (0.011) -0.016 (0.021)

0.40 0.78 0.015 0.46 rs11634397 ZFAND6 G/A 0.032 (0.021) 0.075 (0.045) -0.0070 (0.0093) -0.022 (0.017)

0.12 0.096 0.45 0.21 rs4760790 TSPAN8/LGR5 A/G 0.016 (0.022) 0.019 (0.046) -0.012 (0.0087) -0.026 (0.018)

0.45 0.68 0.18 0.15 rs5215 KCNJ11 C/T -0.018 (0.020) -0.099 (0.042) -0.0053 (0.0088) 0.035 (0.017)

0.37 0.018 0.54 0.042 rs243021 BCL11A A/G -0.0056 (0.023) 0.0037 (0.041) -0.0053 (0.0086) 0.027 (0.022)

0.81 0.93 0.54 0.22 rs849134 JAZF1 A/G 0.015 (0.019) -0.0017 (0.041) 0.0092 (0.0084) 0.021 (0.017)

0.44 0.97 0.27 0.20 rs1531343 HMGA2 C/G -0.034 (0.036) -0.12 (0.075) 0.0062 (0.014) -0.0042 (0.034)

0.35 0.11 0.65 0.90 rs231362 KCNQ1 G/A -0.021 (0.020) -0.18 (0.041) 0.0065 (0.0091) 0.0096 (0.017)

0.28 1.63×10-5 0.47 0.58

rs163184 KCNQ1 G/T 0.015 (0.021) - 0.0076 (0.017) 0.041 (0.027)

0.48 - 0.66 0.12 rs6795735 ADAMTS9 C/T -0.0045 (0.020) 0.031 (0.051) -0.0017 (0.0095) 0.012 (0.018)

0.82 0.54 0.86 0.52


16

Other rs5945326 DUSP9 A/G 0.069 (0.047) -0.019 (0.038) -0.018 (0.015) - 0.15 0.63 0.23 -


17

(A) Insulinogenic index at HHEX/IDE

(B) Insulinogenic index atCDKAL1


18

(C) 32,33 split proinsulin at HNF1A

(D) AIR at MTNR1B


19

(E) AIR at KCNQ1

(F) AIR at CDKN2A/B


20

Supplementary Figure 1. Forest plots for meta-analyses of insulinogenic index, split

proinsulin and acute insulin response (AIR). (A) insulinogenic index at HHEX/IDE, (B) insulinogenic index at CDKAL1, (C) 32,33 split proinsulin at HNF1A, (D) AIR at MTNR1B, (E) AIR at KCNQ1, (F) AIR at CDKN2A/B.


21

(A) IV insulin sensitivity traits at IRS1

(B) IV insulin sensitivity traits at ADCY5

Supplementary Figure 2. Forest plots for meta-analyses of intravenous (IV) insulin sensitivity traits (A) IRS1, (B) ADCY5.


22

(A) (B)

Supplementary Figure 3. Results of cluster analysis using principal traits. (A) Clustering of ten traits with meta-analysis results from at least 10,000 individuals (principal traits). Traits include: insulinogenic index, proinsulin, ISI Belfiore, ISI Gutt, ISI Matsuda, ISI Stumvoll and four fasting traits: fasting glucose (FG), fasting insulin (FI), HOMA-B, HOMA-IR. Closest links are observed between the four ISIs, which are connected to the


23

HOMA-IR/FI branch and constitute a larger clade of traits linked to insulin sensitivity and resistance. This clade is linked to proinsulin, making up a cluster bearing insulin secretion and sensitivity traits. In turn, this is connected to the HOMA-B/insulinogenic index clade representing effects of type 2 diabetes loci on beta-cell function and early insulin secretion. Finally, the effects of type 2 diabetes loci on FG cluster separately from all other traits. (B) Calinski index computed as a measure of clustering support. The most robust set contained eight trait groups underscoring a clear distinction between most of the investigated physiologic phenotypes.


24

Supplementary Figure 4. Results of cluster analysis using all 14 traits. (A) Complete linkage dendrogram of the 14 traits. Cluster analysis confirmed the trait connectivity observed for principal traits and revealed a direct link between AIR and insulinogenic index. C-peptide and split proinsulin were connected directly and linked subsequently to the clade of insulin secretion measurements. IV measures cluster with the ISIs (Matsuda/Belfiore clade). The C-peptide/split proinsulin cluster was connected directly to ISIs/ IV measures and this whole clade was linked subsequently to the group of insulin secretion measurements. (B) Calinski index computed as a measure of clustering support. The most robust set contained twelve trait groups underscoring a clear distinction between the investigated physiologic phenotypes.

(A) (B)


25

Supplementary Figure 5. Principal component analysis (PCA) of type 2 diabetes SNPs. Loci are grouped using centroid-based clustering. Each group is represented using a specific color and group name is shown next to each ellipse (see main text for loci belonging to each group). The two components explain the 64.29% of the total phenotypic variance. Numbers on ellipses correspond to locus groups: 1: hyperglycemic (HG), 2: insulin resistance (IR), 3: proinsulin (PI), 4: beta cell (BC), 5: unclassified (UC).


26

(A)


27

(B)


28

(C)


29

(D)


30

(E)


31

(F)

Supplementary Figure 6. Scatter plots of allelic effect size estimates for selected trait pairs. For each scatter plot, loci are assigned into the groups defined from the cluster analysis of principal traits (groups highlighted by different colours). (A) Insulinogenic index vs. M/I SI SSPG reveals elevated insulinogenic index at IRS1 and a negative correlation of insulinogenic index with the compensatory effects IV measures. (B) Proinsulin vs. HOMA-B, this plot highlights the effect of ARAP1 on proinsulin. (C) Proinsulin vs. 32,33 split proinsulin shows the diminished proinsulin levels at ARAP1. (D)


32

Insulinogenic index vs. 32,33 split proinsulin reveals a similar effect for ARAP1 on insulinogenic index. (E) Acute insulin response vs. M/I SI SSPG (combined IV measures): this plot illustrates a weak negative correlation of acute insulin response with compensatory effects IV measures. (F) ISI Belfiore vs. M/I SI SSPG plot reflects good correlation between the two measures of insulin sensitivity. Cluster group colors are HG: orange, IR: green, PI: pink, BC: red, UC: blue.


33

REFERENCES

1. Belfiore F, Iannello S, Volpicelli G: Insulin sensitivity indices calculated from basal and OGTT-induced insulin, glucose, and FFA levels. Mol Genet Metab 1998;63:134-141

2. Gutt M, Davis CL, Spitzer SB, Llabre MM, Kumar M, Czarnecki EM, Schneiderman N, Skyler JS, Marks JB: Validation of the insulin sensitivity index (ISI(0,120)): comparison with other measures. Diabetes Res Clin Pract 2000;47:177-184

3. Matsuda M, DeFronzo RA: Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes care 1999;22:1462-1470

4. Stumvoll M, Mitrakou A, Pimenta W, Jenssen T, Yki-Jarvinen H, Van Haeften T, Renn W, Gerich J: Use of the oral glucose tolerance test to assess insulin release and insulin sensitivity. Diabetes care 2000;23:295-301


COHORT ELY EUGENE2 FHS (proinsulin) FHS (Gutt) FUSION MESYBEPO METSIM NHANES PIVUS QTL FAMILIES RISCPARTNERS/ROCH

ESORBS STANFORD ULSAM UNG92

CONTACT PERSON/ANALYST CL ML JF JF RW JS AUJ JBM EI NG CL JF AT/IP

Ke Hao, Fahim

Abbasi, Tim

Assimes, Josh

Knowles

EI NG

Ethnicity European

Northern

European, German

(Finland, Sweden,

Denmark,

germany)

White WhiteEuropean

descentEuropean

Northern

European

(white,

Caucasian)

European descent

Northern

European (white,

Caucasian)

Northern

European (white,

Caucasian)

European

Northern

European (white,

Caucasian)

Sorbs (Slavonic

origin)White, Caucasian

Northern

European (white,

Caucasian)

Northern European

(white, Caucasian)

Country (Sample source) UK USA USA Finland Germany Finland USA Sweden Denmark Multi-Centre USA Germany

United States,

San Francisco

Bay Area

Sweden Denmark

Type (Population-based or case-

control)Population-based Offspring of T2DM

Population-

based

Population-

based

Family-based:

nondiabetic

offspring and

spouses of T2D

affected sibs

population-

based

Population-

basedPopulation-based

Population-

based

Type 2 diabetes

offspringPopulation-based Case-Control Population-based Population based Population-based Population-based

OGTT MEASUREMENTS

Sample (Plasma? Blood?)

Fresh venous

plasma in lithium

heparin or

sodium fluoride

with heparin

(glucose); frozen

at -80C until

measurement

(insulin)

Plasma NA NA NA Plasma Plasma NA NAGlucose: plasma;

insulin: serum

Fresh venous

plasma in lithium

heparin or sodium

fluoride with

heparin

Fasting Plasma Serum NA PlasmaGlucose: plasma;

insulin: serum

Time points for collection0, 30, 60, 120

mins0, 30, 60, 90, 120 NA NA NA 0,30,60,90,120 0, 30, 120 min NA NA

0, 10, 20, 30, 40,

50, 60, 75, 90,

105, 120 mins

0, 30, 60, 90, 120

mins0, 30, 120 mins 0, 30, 120 min NA

0, 30, 60, 90, 120

minsNA

Glucose assay Hexokinase assayGlucose oxidase

colorimetric assayNA NA NA

Hexokinase,

ABX Pentra

Reagents

Glucose

hexokinase

(Konelab

System

Reagents,

Glucose (HK),

Thermo Fisher

Scientific,

Vantaa, Finland)

NA NA

hexokinase/G6P-

DH technique

(Boehringer

Mannheim,

Germany).

Glucose oxidase

method (Cobas

Integra, Roche)

Hexokinase

Hexokinase

method

(Automated

analyser Modular,

Roche Diagnostics,

Mannheim,

Germany)

NA

Glucose

dehydrogenase

method (Gluc-DH,

Merck, Darmstadt,

Germany)

hexokinase/G6P-

DH technique

(Boehringer

Mannheim,

Germany).

Insulin assayImmunometric

assay

Immunoassay

(ADVIA Centaur

Insulin IRI,

Siemens Medical

Solutions

Diagnostics)

NA NA NA ELISA, Mercodia

Immunoassay

(ADVIA Centaur

Insulin IRI,

Siemens

Medical

Solutions

Diagnostics)

NA NA

ELISA (Dako

Diagnostics, Ely,

UK)

Specific time-

resolved

fluroimmunoassa

y (AutoDELFIA

Insulin kit; Wallac

Oy, Turku,

Finland)

Human specific

insulin RIA, Linco

Research Inc., St

Louis MO, USA.

The cross-

reactivity of

Pharmacia insulin

antibody with

proinsulin is 0%

AutoDELFIA Insulin

assay (PerkinElmer

Life and Analytical

Sciences, Turku,

Finland)

NA

Immunoreactive

insulin: Enzymatic-

immunological

assay (Enzymun,

Boehringer

Mannheim)

ELISA (Dako

Diagnostics, Ely,

UK)


Reference

Addenbrooke's

Hospital,

Cambridge

Laakso M, et al:

Insulin sensitivity,

insulin release and

glucagon-like

peptide-1 levels in

persons with

impaired …

Diabetologia

51:502-11, 2008

NA NA NA NA NA NA NA NA

Hills SA, Balkau B,

Coppack SW,

Dekker JM, Mari

A, Natali A,

Walker M,

Ferrannini E; EGIR-

RISC Study Group.

The EGIR-RISC

STUDY (The

European group

for the study of

insulin resistance:

relationship

between insulin

sensitivity and

cardiovascular

disease risk): I.

Methodology and

objectives.

Diabetologia.

2004

Mar;47(3):566-70.

NA NA

Ingelsson E,

Sundström J,

Ärnlöv J, Zethelius

B, Lind L. Insulin

Resistance and

Risk of Congestive

Heart Failure.

JAMA 2005;

294(3):334-341

NA

PROINSULIN MEASUREMENTS

Sample (Fasting? Plasma?

Blood?)Fasting plasma NA Fasting plasma NA NA Fasting plasma NA Fasting plasma NA Fasting plasma NA NA NA Fasting plasma NA

Assay

Monoclonal

antibody-based

two-site

immunoradiomet

ric assay

NALinco humna

proinsulin RIA

Linco humna

proinsulin RIANA NA

RIA (human

proinsulin Ria

kit, Linco

Research, St.

Charles, MO,

USA)

NA

Two-site

immunometric

assay technique

NA

Monoclonal

antibody-based

two-site

immunoradiometr

ic assay

NA NA NA

Two-site

immunometric

assay technique

NA

Reference 2669734 NA

http://www.mill

ipore.com/catal

ogue/item/hpi-

15k

http://www.mill

ipore.com/catal

ogue/item/hpi-

15k

NA NA NA NA

Sobey WJ, Beer

SF, Carrington

CA, Clark PMS,

Hales CN.

Sensitive and

specific two-site

immunoradiome

tric assays for

human insulin,

proinsulin, 65-

66split and 32-

33 split

proinsulins.

Biochem J 1989;

260: 535-541

NA PMID: 17846745 NA NA NA

Sobey WJ, Beer SF,

Carrington CA,

Clark PMS, Hales

CN. Sensitive and

specific two-site

immunoradiometri

c assays for human

insulin, proinsulin,

65-66split and 32-

33 split

proinsulins.

Biochem J 1989;

260: 535-541

NA

C-PEPTIDE MEASUREMENTS NA NA

Sample (Fasting? Blood?

Serum?)Fasting plasma NA NA NA Fasting plasma NA NA Fasting serum NA fasting serum Fasting plasma NA Fasting serum NA NA fasting serum

Assay

A two-step time

resolved

fluorometric

assay with

samples assayed

in singleton on a

1235 AutoDELFIA

automatic

immunoassay

system (Perkin

Elmer, UK)

NA NA NA Novo RIA NA NA RIA NA

Monoclonal

antibody-based

two-site

immunoradiometr

ic assay

NA LIAISON (DiaSorin) NA NA RIA


Reference

Christopher P,

Hattersley AT,

Sutton PJ. C-

peptide

measurement:

clinical

applications and

sample storage

using a new wide

range assay.

Proc. ACB Natl.

Meeting, p.69

(1991)

NA NA NA NA NA NA

http://www.cdc.go

v/nchs/data/nhan

es/nhanes3/cdrom

/nchs/manuals/lab

man.pdf

NA NA PMID: 14968294 NA NA NA NA PMID: 1239396

EXCLUSIONS AND NUMBERS OF

SAMPLES PER PHENOTYPE

Exclusions

Fasting plasma

glucose>=7.0

mmol/l

Diabetes (known

diabetes or fasting

plasma glucose >=

7.0 mmol/l or 120

min. glucose

during OGTT >=

11.1 mmol/l)

Non-fasting

individuals,

other diabetic

treatment,

fasting glucose

≥ 7 mmol/L

Non-fasting

individuals,

other diabetic

treatment,

fasting glucose

≥ 7 mmol/L

Type 2 diabetes,

on metab.

medication

Diabetics

(fasting plasma

glucose <=7.0 or

diabetes

treatment)

Diabetics

(fasting plasma

glucose>=7.0 or

diabetes

treatment) or 2-

h glucose at

least 11.1

mmol/l

Diabetics (fasting

plasma

glucose>=7.0 or

diabetes

treatment)

Diabetics (fasting

plasma

glucose>=7.0 or

diabetes

treatment)

Diabetes (known

diabetes or

fasting plasma

glucose > 7.0

mmol/l or 120

min. glucose

during OGTT >

11.1 mmol/l)

Fasting plasma

glucose>=7.0

mmol/l

Diabetics

Diabetics (fasting

plasma

glucose>=7.0 or

diabetes

treatment)

Diabetics (known

diabetes or

fasting plasma

glucose > 7.0

mmol/l ) or

diabetes

treatment

Diabetics (fasting

plasma

glucose>=7.0 or

diabetes

treatment)

Fasting glucose > 7

Samples with M-value: N all

(%males / %females)0 874 (41/59) 0 0 0 0 0 0 0 0 1319 (44.8/55.2) 0 0 0 989(100/0) 0

Samples with SSPG: N all

(%males / %females)0 0 0 0 0 0 0 0 0 0 0 0 0

381 (42.3% male/

57.7% female)0 0

Samples with FSIGT: N all

(%males / %females)0 0 0 0

545 (46.97 /

53.03)0 0 0 0 267 (44.2/55.8) 0 0 0 0 0 376 (49.2/50.8)

Samples with ISI-Stumvoll: N all

(%males / %females)1454 (45.9/54.1) 861 (41/59) 0 0 0

919 (30.69 /

69.31)6888 (100/0) 0 0 272 (43.8/56.2) 1322 (44.8/55.2) 619 (45.2 / 54.8) 820 (40.4/59.6) 0 976 (100/0) 0

Samples with insulinogenic

index: N all (%males / %females)1440 (46.3/53.7) 861 (41/59) 0 0 0

905 (30.28 /

69.72)6941 (100/0) 0 0 268 (44/56) 970 (45.4/54.6) 621 (45.1 / 54.9) 810 (40.4/59.6) 0 979 (100/0) 0

Samples with proinsulin: N all

(%males / %females)1604 (46.3/53.4) 0

5759

(46.7/53.3)0 0 0 5128 (100/0) 0 911 (48.9/51.1) NA 900 (43.8/56.2) 0 0 0 989 (100/0) 0

Samples with C-peptide: N all

(%males / %females)1606 (44.9/55.0) 0 0 0

1005

(44.7/55.3)0 0 1217 (38.4 / 61.6) 0 281 (43.1/56.9) 1465 (44.7/55.3) 0 747 (41.1/58.9) 0 0 376 (49.2/50.8)

Samples with ISI-Belfiore: N all

(%males / %females)1446 (45.9/54.1) 0 0 0 0

912 (30.48 /

69.52)6996 (100/0) 0 0 271 (43.9/56.1) 1278 (44.5/55.5) 616 (45.1 / 54.9) 812 (40.4/59.6) 0 989(100/0) 0

Samples with ISI-Matsuda: N all

(%males / %females)1315 (46.0/54.0) 853 (41/59) 0 0 0

902 (30.27 /

69.73)6996 (100/0) 0 0 281 (42.7/57.3) 1158 (44.9/55.1) 628 (44.9 / 55.1) 812 (40.4/59.6) 0 986 (100/0) 0

Samples with ISI-Gutt: N all

(%males / %females)1480 (45.5/54.5) 860 (41/59) 0

2604

(45.8/54.2)0

1090 (32.39 /

67.61)7028 (100/0) 0 0 273 (43.6/56.4) 1289 (44.6/55.4) 612 (44.9 / 55.1) 816 (40.3/59.7) 0 991 (100/0) 0

PHENOTYPE DISTRIBU TIONS(ALLVALUESINORIGINALUNITSBEFORETRANSFORMATIONS)

Age [Mean (sd) males / Mean

(sd) females], years

61.23 (9.18) /

60.55 (9.08)

40.4 (10.5) / 39.9

(10.3)

48.5 (13.4) /

48.9 (13.8)

54.0 (9.9) / 54.1

(9.8)

44.37 (13.37) /

47.25 (13.46)

51.28 (14.64) /

49.75 (13.75)

57.07 (6.98) /

NA

51.2 (20.6) / 51.5

(20.3)

70.13 (0.17) /

70.26 (0.15)

41.4 (11.8) / 44.3

(12.6)

43.41 (8.48) /

44.60 (8.19)

52.68 (12.96) /

52.53 (12.80)

47.8 (16.6) / 48.1

(16.1)52 (9) / 49 (10) 70.99 (0.63) / NA

25.5 (3.45) / 25

(3.55)

BMI [Mean (sd) males / Mean

(sd) females], kg/m2

27.13 (3.86)

/27.06 (5.36)

26.9 (4.26) / 26.4

(5.35)

28.2 (4.3) / 26.8

(5.8)

27.8 (3.9) / 26.1

(4.9)

27.39 (4.19) /

27.28 (5.50)

28.40 (5.50) /

29.44 (6.51)

26.79 (3.76) /

NA

26.9 (4.8) / 26.4

(5.8)

26.82 (3.64) /

26.85 (4.66)

26.3 (4.1) / 26.3

(5.07)

26.52 (3.53) /

24.93 (4.40)

27.76 (5.19) /

27.08 (7.23)

26.5 (4.8) / 27.2

(5.1)

30.3 (5.1) / 30.1

(5.6)25.99(3.20) / NA

24.1 (3.38) / 23

(3.91)

Fasting plasma glucose [Mean

(sd) males / Mean (sd) females],

mmol/l

5.10 (0.53) / 4.88

(0.53)

5.29 (0.52) / 4.95

(0.50)

5.46 (0.48) /

5.15 (0.50)

5.36 (0.47) /

5.12 (0.50)

5.36 (0.54) /

5.10 (0.53)

5.75 (0.57) /

5.68 (0.58)5.70 (0.48) / NA

5.39 (0.49) / 5.15

(0.55)

5.02 (0.55) / 4.95

(0.56)

5.26 (0.501) /

5.07 (0.538)

5.24 (0.53) / 4.95

(0.59)

4.97 (0.47) / 4.77

(0.49)5.4 (0.5) / 5.4 (1.1) 100 (16) / 96 (12) 5.37 (0.56) / NA

5.15 (0.466) / 4.82

(0.368)

Fasting insulin [Mean (sd) males

/ Mean (sd) females], pmol/l

58.75 (34.04) /

54.21 (34.78)

52.0 (57.5) / 49.9

(54.8)

25.7 (19.2) /

22.5 (16.9)

30.3 (11.2) /

27.2 (8.7)

76.83 (46.30) /

72.48 (53.06)

54.13 (39.16) /

49.31 (30.55)

49.09 (33.9) /

NA

62.4 (44.0) / 57.9

(36.9)

52.73 (32.66) /

50.24 (30.61)

42.5 (30.2) / 42.5

(33.3)

38.21 (20.83) /

34.26 (19.42)

88.49 (62.33) /

81.12 (77.59)

42.3 (32.7) / 42.3

(29.1)NA 73.25 (40.25) / NA

35 (21.1) / 39.3

(23)

M-value [Mean (sd) males /

Mean (sd) females]NA

39.5 (15.8) / 43.2

(16.8)NA NA NA NA NA NA NA NA

124.1 (64.2) /

155.1 (74.4)NA NA NA 5.46 (1.94) / NA NA

Original units NA umol/kg/min NA NA NA NA NA NA NA NAmmol*min-1*kg

ffm-1*nM-1NA NA NA mg/kg bw/min NA

SSPG [Mean (sd) males / Mean

(sd) females]NA NA NA NA NA NA NA NA NA NA NA NA NA

163 (71) / 155

(74)NA NA

Original units NA NA NA NA NA NA NA NA NA NA NA NA NA mg/dl NA NA

Si from FSIGT [Mean (sd) males /

Mean (sd) females]NA NA NA NA

7.04 (4.61) /

7.52 (4.34)NA NA NA NA

9.64 (6.18) / 10.7

(6.1)NA NA NA NA NA

15.3 (8.84) / 15.2

(9.71)

Original units NA NA NA NA x10-5 min-1/pM NA NA NA NA x10-5 min-1/pM NA NA NA NA NA x10-5 min-1/pM

ISI-Stumvoll [Mean (sd) males /

Mean (sd) females]

0.085 (0.029) /

0.089 (0.033)

0.092 (0.027) /

0.091 (0.028)NA NA NA

0.0781 (0.0338)

/ 0.0749

(0.0324)

0.089 (0.024) /

NANA NA

0.0941 (0.0258)

/ 0.0963

(0.0285)

0.096 (0.024) /

0.103 (0.026)

0.085 (0.031) /

0.088 (0.040)

0.1 (0.02) / 0.1

(0.03)NA 0.082 (0.028) / NA NA


Insulinogenic index [Mean (sd)

males / Mean (sd) females],

pmol/mmol

111.4 (88.3) /

139.4 (205.3)

117.0 (114.7) /

132.7 (484.3)NA NA NA

126.33 (209.45)

/ 102.23

(128.18)

136.5 (190.2) /

NANA NA

86.3 (55.1) / 113

(98.4)

7.92 (10.42) /

9.81 (14.75)

114.8 (386.9) /

163.3 (383.8)

80.2 (67.9) / 145.6

(357.4)NA 88.88 (70.26) / NA NA

Proinsulin [Mean (sd) males /

Mean (sd) females], pmol/l

4.82 (3.52) / 3.77

(2.75)NA

12.3 (10.9) / 9.2

(7.5)NA NA NA

13.31 (6.16) /

NANA

10.33 (7.82) /

8.48 (7.03)NA

9.76 (11.30) /

7.27 (9.64)NA NA NA 7.16 (5.61) / NA NA

C-peptide [Mean (sd) males /

Mean (sd) females], pmol/l

705.0 (282.0) /

639.5 (282.0)NA NA NA

1.77 (0.83) /

1.65 (0.77)NA NA

740.0 (406.6) /

676.6 (420.1)NA

521 (196) / 523

(200)

636.1 (299.1) /

561.3 (257.7)NA

6.5 (0.36) / 6.48

(0.35)NA NA

454 (156) / 493

(160)

ISI-Belfiore [Mean (sd) males /

Mean (sd) females]

1.03 (0.28) / 1.09

(0.28)NA NA NA NA

0.9692 (0.3051)

/ 0.9885

(0.2670)

1.02 (0.30) /

NANA NA

1.12 (0.29) / 1.15

(0.273)

1.55 (0.25) / 1.55

(0.23)

0.98 (0.30) / 1.03

(0.30)

1.2 (.27) / 1.18

(.26)NA 0.96 (0.27) / NA NA

ISI-Matsuda [Mean (sd) males /

Mean (sd) females]

5.78 (3.38) / 6.58

(4.08)

2.64 (1.98) / 2.80

(1.79)NA NA NA

6.015 (4.26) /

6.012 (3.53)

6.96 (4.13) /

NANA NA

7.01 (4.03) / 7.6

(4.86)

8.21 (5.14) / 9.25

(5.20)

5.13 (3.30) / 5.78

(3.16)

10.18 (6.9) / 9.3

(5.31)NA 4.21 (2.60) / NA NA

ISI-Gutt [Mean (sd) males /

Mean (sd) females]

14.37 (4.41) /

14.76 (4.30)

81.2 (222.8) / 86.1

(29.4)NA

26.8 (7.2) / 26.2

(6.7)NA

53.77 (27.10) /

46.67 (24.12)

50.13 (22.26) /

MANA NA

62.3 (48.1) / 51.3

(18.1)

16.30 (5.37) /

15.95 (4.36)

26.96 (8.38) /

26.99 (7.33)

80.01 (50.88) /

60.76 (31.82)NA 39.31 (13.82 / NA NA

GENOTYPING

In silico/De novo De novo In silico In silico In silico De novo De novo De novo De novo De novo De novo De novo De novo In silico In silico De novo De novo

Genotyping platform & SNP

panelSequenom iPLEX

Illumina Infinium

HumanHap550

BeadChip

Affymetrix 500K

and MIPS 50K

Affymetrix 500K

and MIPS 50K

iPLEX

Sequenom

MassARRAY

Oxford

iPLEX

Sequenom

MassARRAY

Sequenom iPLEXIllumina Golden

Gate assay

KASPar SNP

genotypingKASPar Chemistry Sequenom iPLEX

500K Affymetrix

GeneChip (250K

Sty and 250K Nsp

arrays, Affymetrix,

Inc) and Affymetrix

Genome-Wide

Human SNP Array

6.0

Affymetrix SNP

Array 6.0

Illumina Golden

Gate assay

KASPar SNP

genotyping

Genotyping centreWTSI/ MRC

Epidemiology

Finnish Genome

CenterAffymetrix Affymetrix NHGRI Oxford NHGRI Broad Institute

Uppsala SNP

Technology

Platform

Kbiosciences KBioscience Broad Institute

Microarray Core

Facility of the

Interdisciplinary

Centre for Clinical

Research,

University of

Leipzig, Germany

and ATLAS Biolabs

GmbH, Berlin,

Germany

Rosetta

Uppsala SNP

Technology

Platform

Kbiosciences

Genotyping calling algorithm Sequenom

Illumina

Beadstation

Genotyping

Solution

BRLMM BRLMM

Sequenom

MassArray

Typer 3.4

NA

Sequenom

MassArray

Typer 3.4

SequenomGenCall

(Illumina)KASPar KASPar Sequenom

BRLMM algorithm

(Affymetrix, Inc)

for 500K and

Birdseed Algorithm

for Genome-Wide

Human SNP Array

6.0

BirdSuite GenCall (Illumina) KASPar

SAMPLE QC

Call rate [filter detail / N

individuals excluded]94%-99.2% > 0.98 ≥ 97% / 79 ≥ 97% / 77 NA 94.3% - 97.3% > 96-97% 98.7% (60% / 86) ≥ 90% / 0 ≥95% / 0 99.7% (96% / 14) >0.94 > 0.98 ≥ 90% / 7

Ethnic outliers excluded - Yes - - NA NA NA - - - - - Yes Yes - -

Other exclusions -Duplicates, gender

mismatch

Heterozygosity

5 SD from mean

(< 25.758% or >

29.958%) or

excess

mendelian

inconsistencies

/ n=10

Heterozygosity

5 SD from mean

(< 25.758% or >

29.958%) or

excess

mendelian

inconsistencies

/ n=13

NA NA NA - - - - -Duplicates, gender

mismatch

Duplicates,

gender mismatch- -

Individuals for analysis 1612 595 5759 2604 1546 1130 7043 912 911 286 1428 648 921 381 1030 376

SNP QC (prior to imputation)

MAF [filter detail / N SNPs

excluded]1% / 0 < 2% 1% / 68 953 1% / 68 953 1% / 0 1% / 0 1% / 0 1% / 0 1% / 0 1% / 0 1% / 0 1% / 0 1% < 2% 1% / 0 1% / 0

HWE [filter detail / N SNPs

excluded]0.01 / 0 1 x 10-6 10-6 / 20 999 10-6 / 20 999 0.01 / 1 0.01 / 1 0.01 / 2 0.01 / 0 0.001 / 0 0.001 / 0 0.01 / 0 0.01 / 0 10-4 1 x 10-6 0.001 / 0 0.001 / 0

Call rate [filter detail / N SNPs

excluded]90% / 0 < 90% 95% / 23 312 95% / 23 312 95% / 0 92.5% / 0 90% / 0 90% / 0 95% / 0 95% / 0 90% / 0 90% / 0 0.95 < 90% 95% / 0 95% / 0

Other NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA

SNP number in QC'd dataset NA 378,163 378,163 NA NA NA NA NA NA NA NA 378,513 NA NA

IMPUTATION STATS

Imputation software NA MACH MACH MACH NA NA NA NA NA NA NA NA IMPUTE MACH NA NA


Imputation quality metrics NASNPs of MACH r^2

< 0.3 are excludedr2hat > 0.3 r2hat > 0.3 NA NA NA NA NA NA NA NA Proper-info > 0.4

SNPs of MACH

r^2 < 0.3 are

excluded

NA NA

Other SNP QC filters applied? NA

SNPs of MAF < 5%

and HWE<1E-6 are

excluded

MAF > 1% MAF > 1% NA NA NA NA NA NA NA NAMAF>1%, HWE<10-

4

SNPs of MAF <

5% and HWE<1E-

6 are excluded

NA NA

DATA ANALYSIS

Number of SNPs in analysis / N

imputed16 / 0 38 /0 19 / 17 19 / 17 16 / 0 32 / 0 18 / 0 18 / 0 61 / 0 36 / 0 16 / 0 18 / 0 19 / 17 38 / 0 61 / 0 36 / 0

Trait transformation M-value NAZ score

transformationNA NA NA NA NA NA NA NA

Z-score

transformationNA NA NA

Z-score

transformationNA

Trait transformation SSPG NA NA NA NA NA NA NA NA NA NA NA NA NAZ-score

transformationNA NA

Trait transformation FSIGT NA NA NA NAZ-score

transformationNA NA NA NA Natural log NA NA NA NA NA Natural log

Trait transformation ISI-Stumvoll Natural log NA NA NA NA Natural log Natural log NA NA Natural log Natural log natural log Natural log NA Natural log NA

Trait transformation

insulinogenic indexNatural log NA NA NA NA Natural log Natural log NA NA Natural log Natural log natural log Natural log NA Natural log NA

Trait transformation proinsulin Natural log NA Natural log Natural log NA Natural log Natural log Natural log Natural log NA NA NA NA Natural log NA

Trait transformation C-peptide Natural log NA NA NA Natural log Natural log NA NA NA Natural log Natural log NA Natural log NA NA Natural log

Trait transformation ISI-Belfiore Natural log NA NA NA NA Natural log Natural log NA NA Natural log Natural log natural log Natural log NA Natural log NA

Trait transformation ISI-Matsuda Natural log NA NA NA NA Natural log Natural log NA NA Natural log Natural log natural log Natural log NA Natural log NA

Trait transformation ISI-Gutt Natural log NA NA NA NA Natural log Natural log NA NA Natural log Natural log natural log Natural log NA Natural log NA

Adjustments in first model Age, sexage, sex and age,

sex, BMI

Age, sex, natural

log of fasting

insulin

Age, sex, natural

log of fasting

insulin

Age, sex Age, sex Age Age, sex Age, sex Age, sex, family Age, sex, centre Age + Sex Age, sexage, sex and age,

sex, BMIAge Age, sex

Analysis method Linear regression Linear regressionLinear mixed

effect model

Linear mixed

effect model

Linear

regression

Linear

regression

Linear

regressionGLM Linear regression Mixed model Linear regression GLM Linear regression Linear regression Linear regression Linear regression

Software for analysis STATA 10.1 R 2.9.2R LMEKIN

package

R LMEKIN

packageMerlin SPSS 18.0 Merlin SAS 9.1.3 STATA 10.1 SPSS STATA 10.1 SAS 9.1.9 GWAMA R 2.9.2 STATA 10.1 R 2.10.0

REFERENCES

Study sample reference PMID: 17257284

Laakso M, et al:

Insulin sensitivity,

insulin release and

glucagon-like

peptide-1 levels in

persons with

impaired …

Diabetologia

51:502-11, 2008

– -

Fischer A, Fisher

E, Mˆhlig M,

Schulze M,

Hoffmann K,

Weickert MO,

Schueler R,

Osterhoff M,

Pfeiffer AFH,

Boeing H,

Spranger J:

KCNJ11 E23K

Affects Diabetes

Risk and Is

Associated With

the Disposition

Index. Diabetes

Care 31:87-89,

2008

Stancakova A,

Javorsky M,

Kuulasmaa T,

Haffner SM,

Kuusisto J,

Laakso M;

Changes in

insulin

sensitivity and

insulin release

in relation to

glycemia and

glucose

tolerance in

6414 Finnish

men. DIabetes

2009

May;58(5):1212-

21. Epub 2009

Feb 17

Ingelsson E,

Hulthe J, Lind L.

Inflammatory

markers in

relation to

insulin resistance

and the

metabolic

syndrome. Eur J

Clin Invest

2008;38(7):502-

9

Pubmed ID:

14968294

Ai M et al Clin

Chim Acta. 2009

May 22. [Epub

ahead of print]

Tönjes et al., EJHG,

2009 (in press -

PMID: 19584900),

Tönjes et al.,

HumMolGenet,

2009 (in press)

Ryan MC, Fenster

Farin HM, Abbasi

F, Reaven GM:

Comparison of

waist

circumference

versus body mass

index in

diagnosing

metabolic

syndrome and

identifying

apparently

healthy subjects

at increased risk

of cardiovascular

disease. Am J

Cardiol 102:40-

46, 2008

Zethelius B, Byberg

L, Hales CN, Lithell

H, Berne C.

Proinsulin and

acute insulin

response

independently

predict Type 2

diabetes mellitus

in men--report

from 27 years of

follow-up study.

Diabetologia

2003;46(1):20-6

PMID: 8787683

Website

http://www.mrc-

epid.cam.ac.uk/S

tudies/Ely/

NA

http://www.ncb

i.nlm.nih.gov/pr

ojects/gap/cgi-

bin/study.cgi?st

udy_id=phs000

007.v4.p2

http://www.ncb

i.nlm.nih.gov/pr

ojects/gap/cgi-

bin/study.cgi?st

udy_id=phs000

007.v4.p2

- NA NA

http://www.cdc.go

v/nchs/nhanes.ht

m

http://www.med

sci.uu.se/pivus/p

ivus.htm

NAhttp://www.egir.o

rg/egirrisc/-

http://innere.unikli

nikum-

leipzig.de/_forschu

ng/schwerpunkte/

sorbs.html

NAhttp://www.pubca

re.uu.se/ULSAMNA


impact of type 2 diabetes susceptibility variants on...

Documents