OBJECTIVE

We evaluated the clinical usefulness of a genetic risk score (GRS) based on 14 well-established variants for type 2 diabetes.

RESEARCH DESIGN AND METHODS

We analyzed 14 SNPs at HHEX, CDKAL1, CDKN2B, SLC30A8, KCNJ11, IGF2BP2, PPARG, TCF7L2, FTO, KCNQ1, IRS-1, GCKR, UBE2E2, and C2CD4A/B in 1,487 Japanese individuals (724 patients with type 2 diabetes and 763 control subjects). A GRS was calculated according to the number of risk alleles by counting all 14 SNPs (T-GRS) as well as 11 SNPs related to β-cell function (β-GRS) and then assessing the association between each GRS and the clinical features.

RESULTS

Among the 14 SNPs, 4 SNPs were significantly associated with type 2 diabetes in the present Japanese sample (P < 0.0036). The T-GRS was significantly associated with type 2 diabetes (P = 5.9 × 10−21). Among the subjects with type 2 diabetes, the β-GRS was associated with individuals receiving insulin therapy (β = 0.0131, SE = 0.006, P = 0.0431), age at diagnosis (β = −0.608, SE = 0.204, P = 0.0029), fasting serum C-peptide level (β = −0.032, SE = 0.0140, P = 0.022), and C-peptide index (β = −0.031, SE = 0.012, P = 0.0125).

CONCLUSIONS

Our data suggest that the β-GRS is associated with reduced β-cell functions and may be useful for selecting patients who should receive more aggressive β-cell–preserving therapy.

Type 2 diabetes affects nearly 300 million individuals worldwide, and its prevalence continues to increase in many countries, including Japan (1). Although the precise mechanisms underlying the development and progression of type 2 diabetes have not been elucidated, a combination of multiple genetic and/or environmental factors contribute to the pathogenesis of the disease (2,3). Impaired insulin secretion and insulin resistance, the two main pathophysiological mechanisms leading to type 2 diabetes, have a significant genetic component (4).

Recent studies have confirmed ~40 genetic loci associated with type 2 diabetes (5); most of these loci were discovered in genome-wide association studies (616), with the exception of PPARG (17), KCNJ11 (18), and WFS1 (19), which were identified using candidate gene approaches, and TCF7L2, which was discovered using a linkage-positional cloning strategy (20). Among them, many loci (at least 10), such as MTNR1B, SLC30A8, THADA, TCF7L2, KCNQ1, CAMK1D, CDKAL1, IGF2BP2, HNF1B, and CENTD2, have been shown to be associated with impaired β-cell functions, whereas only a few loci such as PPARG, IRS1, and FTO have been associated with insulin resistance (13).

Although the molecular mechanisms responsible for the susceptibility effect can be well assigned for some loci, such as those at KCNJ11 and SLC30A8, the mechanisms by which most genetic loci contribute to the development of type 2 diabetes are not understood.

Recently, the construction of a genetic risk score (GRS) using information on these diabetes susceptibility loci has been shown to be useful for evaluating the risk of the development of type 2 diabetes in individuals (2126). However, the currently available genetic information is obviously insufficient for predicting the development of type 2 diabetes, and little is known about the detailed relationship between the GRS and the clinical features of type 2 diabetes. In the current study, we selected 14 well-replicated and well-established genetic variants associated with type 2 diabetes in the Japanese population (25,2732) and constructed a GRS, which may predict mechanism (β-cell function and insulin resistance) of diabetes development, to evaluate the possibility that currently available genetic information can be translated into clinical practice.

All patients with type 2 diabetes who regularly attended the outpatient clinics in five hospitals—University of Toyama Hospital (Toyama, Japan), Shakaihoken Takaoka Hospital, Saiseikai Takaoka Hospital, Nanto City Hospital (Nanto, Japan), and Asahi General Hospital (Asahi-machi, Japan)—were asked to participate in this study. Among them, informed consent was obtained from 724 patients between January 2008 and December 2009, and these 724 patients were enrolled in the current study as case subjects (62.3% male, mean ± SD age 64.9 ± 11.1 years, and A1C 7.5 ± 1.3%) (Table 1). We also enrolled control individuals (n = 763) selected from subjects who had undergone an annual health check-up at the Itoigawa General Hospital (Itoigawa, Japan), Aoi Hospital (Tonami, Japan), Amenithy Tsukioka Hospital (Toyama, Japan), Hida City Hospital (Hida, Japan), Sakurai Hospital (Kurobe, Japan), Hokuriku chuo Hospital (Oyabe, Japan), and the above five hospitals. The inclusion criteria for the nondiabetic control subjects were as follows: 1) >50 years of age, 2) A1C values <6.0%, 3) no family history of type 2 diabetes in first- and second-degree relatives, and 4) no past history of a diagnosis of diabetes. Diabetes was diagnosed based on the 1998 American Diabetes Association criteria (33). The exclusion criteria for the case subjects with diabetes were diabetes caused by 1) liver dysfunction, 2) steroids and other drugs that might increase glucose levels, 3) malignancy, 4) monogenic disorders known to cause diabetes, and 5) individuals who tested positive for anti-GAD antibody. Characteristics of the participants are presented in Table 1.

Table 1

Clinical characteristics of the participants

Clinical characteristics of the participants
Clinical characteristics of the participants

We also performed an examination of another cohort for the association of GRS with type 2 diabetes (homeostasis model assessment [HOMA] of β-cell function or HOMA of insulin resistance [HOMA-IR]), which was conducted in Tokyo University, Tokyo, Japan (30) (type 2 diabetes cases, n = 1,182, 59.6% male, age 65.3 ± 9.5 years, and A1C 7.7 ± 1.6%; nondiabetic subjects, n = 859, 44.4% male, age 69.5 ± 6.8 years, and A1C 5.6 ± 0.2%) (Supplementary Table 1). The inclusion criteria for the nondiabetic control subjects and the exclusion criteria for the case subjects with diabetes were identical between the two studies, except for the age of control individuals >60 years in the Tokyo University study.

Collection of clinical information

We obtained clinical information including the current BMI, maximum BMI, family history of diabetes, age at diagnosis, blood chemistry (including plasma glucose, insulin level, serum C-peptide, and serum creatinine) at fasting state, diabetes complications, and use of antidiabetes drugs. Patients who were required to inject >10 units of insulin a day continuously were regarded as undergoing insulin therapy.

Diabetic nephropathy was defined as having a urinary albumin-to-creatinine ratio ≥30 mg/gCr, determined in at least two consecutive overnight samples collected over a 3- to 6-month period. Patients diagnosed as having a urinary tract infection, other glomerular diseases, or gross hematuria were excluded.

All patients underwent ophthalmologic examinations, including funduscopic examination. We defined nonproliferative diabetic retinopathy, proliferative diabetic retinopathy, and a history of photocoagulation or vitrectomy as indicating the presence of diabetic retinopathy. All the study procedures were approved by the ethics committee of the University of Toyama, and informed consent was obtained from all of the participants.

Genotyping assay

Genomic DNA was extracted from peripheral blood (QIAamp DNA blood kit; QIAGEN, Hilden, Germany). We selected 14 single nucleotide polymorphisms (SNPs) at genetic loci that had been previously shown to be robustly associated with type 2 diabetes in seven recent studies performed in Japanese populations (25,2732). The following SNPs were examined: in KCNJ11 (rs5219), in HHEX (rs1111875), in CDKAL1 (rs7756992), near CDKN2B (rs10811661), in SLC30A8 (rs13266634), in IGF2BP2 (rs4402960), in PPARG (rs1801282), in TCF7L2 (rs7903146), in FTO (rs8050136), near IRS-1 (rs2943641), in GCKR (rs780094), in UBE2E2 (rs761-2463), in C2CD4A-C2CD4B (rs7172432), and in KCNQ1 (rs2237892). The genotyping of these SNPs was performed using TaqMan SNP Genotyping assays (Applied Biosystems, Foster City, CA) or a multiplex-PCR-invader assay as described previously (34,35).

The success rates for these assays were >95%, and the concordance rate, based on duplicate comparisons in 763 control participants and 724 type 2 diabetic patients, was 99.4%. A tagging approach to detect all variations completely covering each genomic region has not been used. Although no apparent deviations in the genotype distributions from Hardy-Weinberg equilibrium (HWE) were observed for all of the SNPs (P ≥ 0.001) (6), some of them had borderline results for the HWE test (rs13266634 in control; rs2237892 in control) (Supplementary Table 2).

Construction of GRS

We combined the information on the 14 SNPs using an allele count model (21). To construct the GRS, we summed the number of risk alleles of all 14 SNPs included in this study in each individual, assuming an equal and additive effect of each allele (T-GRS). The T-GRS was distributed normally in both the control and the diabetic subjects.

We further classified these 14 genetic variants into two categories: 1) 11 β-cell function–related SNPs (rs1111875 in HHEX, rs7756992 in CDKAL1, rs10811661 in CDKN2B, rs13266634 in SLC30A8, rs4402960 in IGF2BP2, rs7903146 in TCF7L2, rs780094 in GCKR, rs7612463 in UBE2E2, rs7172432 in C2CD4A/B, rs2237892 in KCNQ1, and rs5219 in KCNJ11) and 2) three insulin resistance/obesity-related variants (rs1801282 in PPARG, rs8050136 in FTO, and rs2943641 in IRS-1), based on previously reported information (13). We then calculated the GRS of the β-cell function–related SNPs (β-GRS) and the insulin resistance and obesity-related SNPs (R-GRS). The β-GRS and R-GRS were also distributed normally in both the control and diabetic groups.

Statistical analysis

Differences in clinical features, such as the insulin secretory capacity and age at the time of the diagnosis of diabetes, between the risk allele groups were determined using ANOVA and multiple regression analysis after adjustments for related covariables. Results with P values <0.05 were considered statistically significant.

We performed HWE tests according to the method described by Nielsen et al. (36). The proportions of genotypes for each SNP were compared between the type 2 diabetic case and the nondiabetic control subjects using a multiple logistic regression analysis with or without adjustments for age, sex, and BMI. The allele-specific odds ratios (ORs) were calculated using logistic regression with or without adjustments for age, sex, and BMI. Variables with skewed distributions were logarithmically (natural) transformed for further analyses. Quantitative trait analyses were performed using a multiple linear regression analysis with or without adjustments for related covariables. Bonferroni correction was applied to correct for multiple testing errors, and P < 0.0036 (0.05 divided by 14: the total number of SNPs studied) was considered significant.

The effects of the GRS on the clinical features and quantitative metabolic traits were examined by calculating the β values for the risk allele score using linear generalized estimating equations. P values <0.05 were considered statistically significant for this analysis.

The statistical analyses were performed using JMP for Windows version 8.00 software (SAS Institute, Cary, NC). The power of the sample size for the current study to identify the association of previously reported SNP loci with type 2 diabetes was calculated using “CaTS power calculator for genetic studies” software (http://www.sph.umich.edu/csg/abecasis/CaTS/).

Associations of each of the 14 SNPs with type 2 diabetes and quantitative metabolic traits

Among the 14 SNPs from 14 loci, 4 SNPs (rs7756992 in CDKAL1, rs10811661 near CDKN2B, rs13266634 in SLC30A8, and rs2237892 in KCNQ1) were found to be significantly associated with type 2 diabetes (Supplementary Table 3) (P = 1.7 × 10−5, 7.5 × 10−6, 2.8 × 10−3, and 1.4 × 10−7, respectively) after adjustments for age, sex, and BMI; the association of rs2237892 in KCNQ1 was the strongest in the present Japanese sample, as reported previously (16,28). rs4402960 in IGF2BP2, rs2943641 near IRS-1, rs780094 in GCKR, and rs5219 in KCNJ11 showed a nominal association with type 2 diabetes (P = 0.010, P = 0.028, P = 0.013, and P = 0.033, respectively), and rs7172432 in C2CD4A/B tended to be associated with type 2 diabetes (P = 0.073). As for rs7903146 in TCF7L2, rs1111875 in HHEX, rs1801282 in PPARG, rs8050136 in FTO, and rs7612463 in UBE2E2, we were unable to detect any SNPs that were significantly associated with type 2 diabetes in the present Japanese sample (P = 0.659, 0.773, 0.997, 0.187, and 0.207, respectively). The effect directions of the above-mentioned SNPs, with the exception of rs1111875 in HHEX, were consistent with those in previous reports (OR >1, P = 9.2 × 10−4, binomial test) (16,28).

We next studied the associations of the T-GRS (equivalent to the sum of the risk alleles of the 14 SNPs studied here), the β-GRS (equivalent to the sum of the 11 β-cell function–related genes), and the R-GRS (equivalent to the sum of the three obesity and insulin resistance-related genes) with the development of type 2 diabetes. The T-GRS and β-GRS were significantly associated with the development of type 2 diabetes (T-GRS OR 1.26 [95% CI 1.20–1.33], P = 5.9 × 10−21 [Supplementary Fig. 1]; β-GRS 1.26 [1.20–1.33], P = 1.1 × 10−19 [Supplementary Table 3]; and R-GRS, nominally associated with the development of type 2 diabetes, 1.18 [1.02–1.37], P = 0.024 [Supplementary Table 3]). We further determined that when all of the participants were stratified according to the β-GRS (high-risk genetic group [H]-β-GRS ≥12; intermediate risk [I], 12 > β-GRS ≥ 10; and low risk [L]-β-GRS <10) or the R-GRS (H-R-GRS ≥5; I, 5 > R-GRS ≥ 4; and L-R-GRS <4 (Supplementary Table 4), the risk of developing diabetes in the H-β-GRS and the H-R-GRS groups (n = 108) was 6.2-fold higher than in the L-β-GRS and the L-R-GRS groups (n = 78) (Supplementary Fig. 2). Interestingly, an effect of the R-GRS was only seen in the L-β-GRS group (OR 1.43 [95% CI 1.06–1.95], P = 0.02) and not in the H-β-GRS groups (1.17 [0.85–1.61], P = 0.34) (Supplementary Fig. 2), suggesting that the β-GRS has a predominant effect on conferring susceptibility to type 2 diabetes over the R-GRS. To statistically evaluate the interaction between β-GRS and R-GRS, we performed a stepwise logistic regression analysis using strategies of both forward selection (addition of each parameter) and backward selection (starting from all parameters). The results indicated that significant interaction was observed when we added β-GRS to R-GRS (P < 0.001), whereas the effect of addition of R-GRS to β-GRS was modest (P = 0.03).

We next examined the associations of each genetic variant with quantitative metabolic traits related to type 2 diabetes. None of the SNPs had a significant effect on the HOMA-β or HOMA-IR by themselves, but the β-GRS and R-GRS showed stronger association with the HOMA-β (P = 0.025) and HOMA-IR (P = 0.0004), respectively, than single SNP alone, in control individuals and patients with type 2 diabetes who were not treated with medications (Table 2). We further examined the association of the three types of GRS with type 2 diabetes and quantitative traits in a previously published independent cohort, which was conducted in Tokyo University. In this cohort, the association between T-GRS and type 2 diabetes (OR 1.18 [95% CI 1.13–1.24], P = 2.08 × 10−12) and β-GRS and HOMA-β (β of ln-HOMA-β = −0.0377, SE = 0.0103, P = 0.0003) was statistically significant, whereas the association of the R-GRS with HOMA-IR did not reach a statistically significant level (β of ln-HOMA-IR = 0.0294, SE = 0.0290, P = 0.3120) (Supplementary Table 5).

Table 2

Association of the 14 SNPs with quantitative traits related to glucose metabolism in control subjects and diabetic subjects

Association of the 14 SNPs with quantitative traits related to glucose metabolism in control subjects and diabetic subjects
Association of the 14 SNPs with quantitative traits related to glucose metabolism in control subjects and diabetic subjects

Investigation of combined effects of GRS on the clinical features of type 2 diabetes

We next examined the association of the T-GRS with clinical features, such as the maximum BMI, the age at the time of diagnosis, and the individuals presently receiving insulin therapy (Supplementary Table 6). Significant inverse correlations were observed between the T-GRS and the maximum BMI (β of maximum BMI −0.225 [95% CI −0.367 to −0.083], P = 0.002) and the age at diagnosis (β of age at diagnosis −0.663 [−1.048 to −0.278], P = 0.0008). We also found that the individuals receiving insulin therapy were positively associated with the T-GRS (β of insulin therapy 0.249 [0.025–0.473], P = 0.029).

We then divided all the participants into three approximately equally sized strata according to the T-GRS: L-T-GRS, I-T-GRS, and H-T-GRS genetic groups, as described in Supplementary Table 4. The characteristics of the three groups are shown in Table 3. In the H-T-GRS group, the duration of diabetes was significantly longer (P < 0.01) and the current BMI was lower (P < 0.05) than those in the L-T-GRS group (Table 3). We next studied the association of the T-GRS with clinical features such as the maximum BMI, the age at the time of diagnosis, and the percentage of individuals receiving insulin therapy (Table 3). We found that the maximum BMI in the H-T-GRS group (27.1 ± 4.2) was significantly lower than that in the L-T-GRS group (28.5 ± 4.6) (P < 0.01). In addition, the age at the time of the diagnosis of diabetes in the H-T-GRS group (49.8 ± 12.4 years) was significantly younger than that in the L-T-GRS group (52.5 ± 11.4 years) (P < 0.001) after adjustments for sex and the maximum BMI (Table 3). The percentage of individuals receiving insulin therapy in the H-T-GRS group (34.9%) was greater than that in the L-T-GRS group (22.7%) (P < 0.05) after adjustments for age, sex, current BMI, duration of diabetes, class of antihyperglycemic drugs, and present HbA1c level.

Table 3

Clinical characteristics of the three groups according to the T-GRS of 14 SNPs in patients with type 2 diabetes

Clinical characteristics of the three groups according to the T-GRS of 14 SNPs in patients with type 2 diabetes
Clinical characteristics of the three groups according to the T-GRS of 14 SNPs in patients with type 2 diabetes

We next examined the associations of the genetic risk score of β-cell function–related SNPs (β-GRS) with the clinical features (Table 4). We found that the β-GRS was associated with individuals receiving insulin therapy (β of insulin therapy 0.0131 [95% CI 0.0004–0.0259], P = 0.0431) and a younger age at diagnosis (β of age at diagnosis −0.608 [−1.008 to −0.208], P = 0.0029). Furthermore, we found a significant inverse correlation between the β-GRS and β-cell function–related parameters including the fasting serum C-peptide (F-CPR) (β of serum C-peptide −0.036 [−0.065 to −0.007], P = 0.0140) and the C-peptide index (CPI) (β of CPI −0.031 [−0.056 to −0.005], P = 0.0179) after adjustments for age, sex, BMI, duration of diabetes, class of antihyperglycemic drugs, fasting plasma glucose, the presence of diabetic nephropathy, and the presence of diabetic retinopathy. We also examined the association of T-GRS with these parameters, but as expected the β-GRS had stronger effects on basal insulin secretion than the T-GRS (Supplementary Table 6). The R-GRS was not associated with any parameters (Table 4).

Table 4

Association of β-GRS and R-GRS with quantitative metabolic traits and clinical information in diabetic subjects

Association of β-GRS and R-GRS with quantitative metabolic traits and clinical information in diabetic subjects
Association of β-GRS and R-GRS with quantitative metabolic traits and clinical information in diabetic subjects

We further tried, as much as possible, to include all information of European study–derived type 2 diabetes variants in the GRS. Overall, the 36 SNP GRS constructed with the 14 SNPs and additional 22 SNPs, however, did not show stronger association with each metabolic trait than the original T-, β-, and R-GRS in this study (Supplementary Tables 7 and 8).

In the current study, we examined 14 SNP loci, which were robustly shown to be susceptibility loci for type 2 diabetes in the Japanese population, and constructed a GRS to evaluate the usefulness of this genetic information in clinical practice. We found that most SNPs (13 of 14) showed a directionally consistent association with the results of previous reports (616), and constructed GRS (T-GRS) showed a much stronger association with type 2 diabetes than any of the single SNPs alone. The T-GRS was also associated with age at the time of the diagnosis of diabetes. Additionally, we found that a β-GRS, consisting of eleven β-cell function–related SNPs, was associated with requirement of insulin therapy and a reduced basal insulin secretion level in Japanese patients with type 2 diabetes.

Currently, >40 loci have been confirmed as susceptibility loci for type 2 diabetes in populations of European origin (5), but the integration of this information can only explain ∼10% of type 2 diabetes heritability; therefore, currently available genetic information for type 2 diabetes is likely insufficient for predicting disease progression and has failed to provide a significant impact on human health care in the general population to date.

Previously, several groups investigated the impact of a GRS for loci related to β-cell function and reported that the GRS was significantly associated with glucose-stimulated insulin secretion (GSIS) in nondiabetic subjects or the subjects with impaired glucose tolerance (4,26,37,38). However, the effects of the GRS on clinical features have not been evaluated in patients with type 2 diabetes. In the current study, we demonstrated for the first time that a β-GRS was associated with individuals receiving insulin therapy and possessing reduced basal insulin secretion, as evaluated using the F-CPR or CPI, among type 2 diabetic subjects. Thus, the β-GRS may be useful for predicting future reductions in basal insulin secretion, resulting in the need for insulin injections to control plasma glucose levels. Interestingly, the association of β-GRS with the reduction in basal insulin secretion long after the onset (e.g., >10 years) was independent of confounding factors, such as age, sex, BMI, duration of diabetes, presence of microvascular complications, and the use of insulin secretagogues. Of note, since the presence of microvascular complications reflects chronic hyperglycemia, the declining β-cell function in individuals with higher β-GRS may not be a consequence of the relatively longer terms of hyperglycemia. Thus, we think that the evaluation of β-GRS at an earlier stage of the disease may be useful, and patients with a higher β-GRS should be strongly encouraged to receive specialized therapies, such as intensive lifestyle modifications and/or the earlier introduction of β-cell–preserving therapy, such as the use of glucagon-like peptide 1 receptor agonists or medications that ameliorate insulin resistance.

In the current study, we were able to replicate the previously reported associations of 8 of the 14 loci in a Japanese population (4 significantly [P < 0.0036] and 4 modestly [P < 0.05]) (25,2732). As for rs7903146 in TCF7L2, rs1111875 in HHEX, rs1801282 in PPARG, rs8050136 in FTO, and rs7612463 in UBE2E2, which were reported to be associated with type 2 diabetes in previous Japanese reports (25,28,30), we were unable to detect any SNPs that were significantly associated with type 2 diabetes in the present Japanese sample (P = 0.659, 0.773, 0.997, 0.187, and 0.207, respectively). However, since the effect directions of most of the SNP loci (13 of 14) were consistent with the results of previous reports and the estimated study power was 15−81% for the 6 unreplicated SNPs (Supplementary Table 2), the lack of replication might be explained by the insufficient power of the current study. In the quantitative trait analyses using control individuals and type 2 diabetic patients with no medications, we did not observe any significant association between each of the single SNPs and glycemic traits, but the β-GRS and R-GRS showed stronger association with the HOMA-β (P = 0.025) and HOMA-IR (P = 0.0004), respectively, indicating that the constructed β-GRS and R-GRS in the current study were appropriate and useful for evaluating the genetic effects on susceptibility to the disease or on related quantitative traits, even among a relatively small study population. The association of the three types of GRS with the quantitative traits could also be consistently observed in an independent cohort, which was conducted in Tokyo University (28,30), further validating the usefulness of the GRS.

Since HOMA indices have some limitations as indicators of β-cell functions or peripheral insulin sensitivity, evaluation of other independent measures of insulin secretion or resistance, such as 2-h glucose and insulin measurements, is required to confirm our findings. We demonstrated that the β-GRS was associated with a reduced basal insulin secretion in diabetic subjects with an average disease duration of 13.6 years. We also observed that the β-GRS was inversely associated with the GSIS determined by disposition index (β of ln–disposition index −0.102 [95% CI −0.006 to −0.194], P = 0.038, after adjustments for age, sex, FPG, and BMI) at the onset of diabetes (n = 134 [unpublished results]); therefore, the β-GRS may be involved in the GSIS at the onset of diabetes, and a further reduction in basal insulin secretion in patients with a higher β-GRS long after the onset of diabetes (>10 years) may contribute to the need for insulin injections. A cohort study involving a larger number of subjects is needed to clarify this point.

In conclusion, we have shown that the β-GRS, as determined using eleven β-cell function–related loci, is associated with a lower basal insulin secretion and the percentage of individuals requiring insulin therapy among Japanese subjects with type 2 diabetes. These results suggest that the evaluation of β-GRS at an earlier stage of the disease may be useful, and patients with a higher β-GRS should receive specialized therapy, including guidance regarding intensive lifestyle modifications and β-cell–preserving therapy.

A slide set summarizing this article is available online.

This work was partly supported by a grant from the Ministry of Education, Culture, Sports, Science and Technology, Japan, to S.M.

No potential conflicts of interest relevant to this article were reported.

M.I. wrote the manuscript and researched data. S.M. researched data, wrote the manuscript, and edited the manuscript. Y.K. researched data. A.Takan., H.K., S.M., and K.H. contributed to discussion. A.Takah. researched data. H.F. researched data. K.H. researched data and reviewed the manuscript. T.K. reviewed the manuscript. K.T. wrote the manuscript and reviewed and edited the manuscript. K.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 70th Scientific Sessions of the American Diabetes Association, Orlando, Florida, 25–29 June 2010.

The authors are grateful to all the subjects who took part in this study. The authors thank Dr. Sachie Asamizu of Shakaihoken Takaoka Hospital; Dr. Rie Oka and Dr. Susumu Miyamoto of Hokuriku Cho Hospital; Dr. Kunimasa Yagi of the University of Kanazawa; Dr. Isao Usui, Dr. Manabu Ishiki, Dr. Toshiyasu Sasaoka, Dr. Chikaaki Kobashi, Dr. Katsuya Yamazaki, Dr. Masaharu Urakaze, Dr. Shiho Fujisaka, Dr. Satoko Senda, Dr. Hikari Suzuki, and Dr. Yu Yamazaki of Toyama University; Dr. Rie Temaru and Dr. Mariko Ikubo of Nanto City Hospital; Dr. Naoji Akagawa and Dr. Yasuo Fukushima of Asahi General Hospital; Dr. Kazuko Taki of Amenithy Tsukioka Hospital; Dr. Yasuhumi Igarashi of Aoi Hospital; Dr. Shigeki Sawasaki of Hida City Hospital; Dr. Hisae Honoki of Saiseikai Takaoka Hospital; and Dr. Hirohumi Oda of Sakurai Hospital for recruiting the study subjects. The authors thank Dr. Momoko Horikoshi of the University of Tokyo for researching data. The authors also thank the technical staff at the Laboratory for Endocrinology and Metabolism, RIKEN Center for Genomic Medicine, for technical assistance.

1
Wild
S
,
Roglic
G
,
Green
A
,
Sicree
R
,
King
H
.
Global prevalence of diabetes: estimates for the year 2000 and projections for 2030
.
Diabetes Care
2004
;
27
:
1047
1053
[PubMed]
2
O’Rahilly
S
,
Barroso
I
,
Wareham
NJ
.
Genetic factors in type 2 diabetes: the end of the beginning?
Science
2005
;
307
:
370
373
[PubMed]
3
Chauhan
G
,
Spurgeon
CJ
,
Tabassum
R
, et al
.
Impact of common variants of PPARG, KCNJ11, TCF7L2, SLC30A8, HHEX, CDKN2A, IGF2BP2, and CDKAL1 on the risk of type 2 diabetes in 5,164 Indians
.
Diabetes
2010
;
59
:
2068
2074
[PubMed]
4
Stancáková
A
,
Kuulasmaa
T
,
Paananen
J
, et al
.
Association of 18 confirmed susceptibility loci for type 2 diabetes with indices of insulin release, proinsulin conversion, and insulin sensitivity in 5,327 nondiabetic Finnish men
.
Diabetes
2009
;
58
:
2129
2136
[PubMed]
5
McCarthy
MI
.
Genomics, type 2 diabetes, and obesity
.
N Engl J Med
2010
;
363
:
2339
2350
[PubMed]
6
Sladek
R
,
Rocheleau
G
,
Rung
J
, et al
.
A genome-wide association study identifies novel risk loci for type 2 diabetes
.
Nature
2007
;
445
:
881
885
[PubMed]
7
Steinthorsdottir
V
,
Thorleifsson
G
,
Reynisdottir
I
, et al
.
A variant in CDKAL1 influences insulin response and risk of type 2 diabetes
.
Nat Genet
2007
;
39
:
770
775
[PubMed]
8
Saxena
R
,
Voight
BF
,
Lyssenko
V
, et al
Diabetes Genetics Initiative of Broad Institute of Harvard and MIT, Lund University, and Novartis Institutes of BioMedical Research
.
Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels
.
Science
2007
;
316
:
1331
1336
[PubMed]
9
Zeggini
E
,
Weedon
MN
,
Lindgren
CM
, et al
Wellcome Trust Case Control Consortium (WTCCC)
.
Replication of genome-wide association signals in UK samples reveals risk loci for type 2 diabetes
.
Science
2007
;
316
:
1336
1341
[PubMed]
10
Scott
LJ
,
Mohlke
KL
,
Bonnycastle
LL
, et al
.
A genome-wide association study of type 2 diabetes in Finns detects multiple susceptibility variants
.
Science
2007
;
316
:
1341
1345
[PubMed]
11
Frayling
TM
,
Timpson
NJ
,
Weedon
MN
, 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
:
889
894
[PubMed]
12
Zeggini
E
,
Scott
LJ
,
Saxena
R
, et al
Wellcome Trust Case Control Consortium
.
Meta-analysis of genome-wide association data and large-scale replication identifies additional susceptibility loci for type 2 diabetes
.
Nat Genet
2008
;
40
:
638
645
[PubMed]
13
Voight
BF
,
Scott
LJ
,
Steinthorsdottir
V
, et al
MAGIC investigators
GIANT Consortium
.
Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis
.
Nat Genet
2010
;
42
:
579
589
[PubMed]
14
Dupuis
J
,
Langenberg
C
,
Prokopenko
I
, et al
DIAGRAM Consortium
GIANT Consortium
Global BPgen Consortium
Anders Hamsten on behalf of Procardis Consortium
MAGIC investigators
.
New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk
.
Nat Genet
2010
;
42
:
105
116
[PubMed]
15
Unoki
H
,
Takahashi
A
,
Kawaguchi
T
, et al
.
SNPs in KCNQ1 are associated with susceptibility to type 2 diabetes in East Asian and European populations
.
Nat Genet
2008
;
40
:
1098
1102
[PubMed]
16
Yasuda
K
,
Miyake
K
,
Horikawa
Y
, et al
.
Variants in KCNQ1 are associated with susceptibility to type 2 diabetes mellitus
.
Nat Genet
2008
;
40
:
1092
1097
[PubMed]
17
Altshuler
D
,
Hirschhorn
JN
,
Klannemark
M
, et al
.
The common PPARgamma Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes
.
Nat Genet
2000
;
26
:
76
80
[PubMed]
18
Gloyn
AL
,
Weedon
MN
,
Owen
KR
, et al
.
Large-scale association studies of variants in genes encoding the pancreatic beta-cell KATP channel subunits Kir6.2 (KCNJ11) and SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is associated with type 2 diabetes
.
Diabetes
2003
;
52
:
568
572
[PubMed]
19
Sandhu
MS
,
Weedon
MN
,
Fawcett
KA
, et al
.
Common variants in WFS1 confer risk of type 2 diabetes
.
Nat Genet
2007
;
39
:
951
953
[PubMed]
20
Grant
SF
,
Thorleifsson
G
,
Reynisdottir
I
, et al
.
Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes
.
Nat Genet
2006
;
38
:
320
323
[PubMed]
21
Weedon
MN
,
McCarthy
MI
,
Hitman
G
, et al
.
Combining information from common type 2 diabetes risk polymorphisms improves disease prediction
.
PLoS Med
2006
;
3
:
e374
[PubMed]
22
Cauchi
S
,
Meyre
D
,
Durand
E
, et al
.
Post genome-wide association studies of novel genes associated with type 2 diabetes show gene-gene interaction and high predictive value
.
PLoS ONE
2008
;
3
:
e2031
[PubMed]
23
Lango
H
,
Palmer
CN
,
Morris
AD
, et al
UK Type 2 Diabetes Genetics Consortium
.
Assessing the combined impact of 18 common genetic variants of modest effect sizes on type 2 diabetes risk
.
Diabetes
2008
;
57
:
3129
3135
[PubMed]
24
Lyssenko
V
,
Jonsson
A
,
Almgren
P
, et al
.
Clinical risk factors, DNA variants, and the development of type 2 diabetes
.
N Engl J Med
2008
;
359
:
2220
2232
[PubMed]
25
Miyake
K
,
Yang
W
,
Hara
K
, et al
.
Construction of a prediction model for type 2 diabetes mellitus in the Japanese population based on 11 genes with strong evidence of the association
.
J Hum Genet
2009
;
54
:
236
241
[PubMed]
26
’t Hart
LM
,
Simonis-Bik
AM
,
Nijpels
G
, et al
.
Combined risk allele score of eight type 2 diabetes genes is associated with reduced first-phase glucose-stimulated insulin secretion during hyperglycemic clamps
.
Diabetes
2010
;
59
:
287
292
[PubMed]
27
Takeuchi
F
,
Serizawa
M
,
Yamamoto
K
, et al
.
Confirmation of multiple risk Loci and genetic impacts by a genome-wide association study of type 2 diabetes in the Japanese population
.
Diabetes
2009
;
58
:
1690
1699
[PubMed]
28
Yamauchi
T
,
Hara
K
,
Maeda
S
, et al
.
A genome-wide association study in the Japanese population identifies susceptibility loci for type 2 diabetes at UBE2E2 and C2CD4A-C2CD4B
.
Nat Genet
2010
;
42
:
864
868
[PubMed]
29
Horikawa
Y
,
Miyake
K
,
Yasuda
K
, et al
.
Replication of genome-wide association studies of type 2 diabetes susceptibility in Japan
.
J Clin Endocrinol Metab
2008
;
93
:
3136
3141
[PubMed]
30
Horikoshi
M
,
Hara
K
,
Ito
C
, et al
.
Variations in the HHEX gene are associated with increased risk of type 2 diabetes in the Japanese population
.
Diabetologia
2007
;
50
:
2461
2466
[PubMed]
31
Omori
S
,
Tanaka
Y
,
Takahashi
A
, et al
.
Association of CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8, and KCNJ11 with susceptibility to type 2 diabetes in a Japanese population
.
Diabetes
2008
;
57
:
791
795
[PubMed]
32
Tabara
Y
,
Osawa
H
,
Kawamoto
R
, et al
.
Replication study of candidate genes associated with type 2 diabetes based on genome-wide screening
.
Diabetes
2009
;
58
:
493
498
[PubMed]
33
Expert Committee on the Diagnosis and Classification of Diabetes Mellitus
.
Report of the expert committee on the diagnosis and classification of diabetes mellitus
.
Diabetes Care
2003
;
26
(
Suppl. 1
):
S5
S20
[PubMed]
34
Maeda
S
,
Tsukada
S
,
Kanazawa
A
, et al
.
Genetic variations in the gene encoding TFAP2B are associated with type 2 diabetes mellitus
.
J Hum Genet
2005
;
50
:
283
292
[PubMed]
35
Báez
S
,
Tsuchiya
Y
,
Calvo
A
, et al
.
Genetic variants involved in gallstone formation and capsaicin metabolism, and the risk of gallbladder cancer in Chilean women
.
World J Gastroenterol
2010
;
16
:
372
378
[PubMed]
36
Nielsen
DM
,
Ehm
MG
,
Weir
BS
.
Detecting marker-disease association by testing for Hardy-Weinberg disequilibrium at a marker locus
.
Am J Hum Genet
1998
;
63
:
1531
1540
[PubMed]
37
Pascoe
L
,
Frayling
TM
,
Weedon
MN
, et al
RISC Consortium
.
Beta cell glucose sensitivity is decreased by 39% in non-diabetic individuals carrying multiple diabetes-risk alleles compared with those with no risk alleles
.
Diabetologia
2008
;
51
:
1989
1992
[PubMed]
38
Haupt
A
,
Staiger
H
,
Schäfer
SA
, et al
.
The risk allele load accelerates the age-dependent decline in beta cell function
.
Diabetologia
2009
;
52
:
457
462
[PubMed]
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

Supplementary data