The study population comprised of 3,757 Japanese individuals (1,444 subjects with type 2 DM and 2,313 controls) who either visited outpatient clinics or were admitted to the participating hospitals (Gifu Prefectural General Medical Center, Gifu; Gifu Prefectural Tajimi Hospital, Tajimi; Japanese Red Cross Nagoya First Hospital, Nagoya; Inabe General Hospital, Inabe; Hirosaki University Hospital, Reimeikyo Rehabilitation Hospital and Hirosaki Stroke Center, Hirosaki, Japan) between 2002 and 2012 due to various symptoms or for an annual health checkup. Written informed consent was obtained from all the participants and the Institutional Review Board of each participating hospital approved the study.

Type 2 DM is defined according to the criteria of the World Health Organization, as described previously (15,16). Subjects with type 2 DM had a fasting plasma glucose level of ≥6.93 mmol/l (126 mg/dl), a blood glycosylated hemoglobin (hemoglobin A1c) content of ≥6.5%, or were taking anti-diabetic medication. Individuals with type 1 DM, maturity onset diabetes of the young, DM associated with mitochondrial diseases or single-gene disorders, pancreatic diseases or other metabolic or endocrinological diseases were excluded from the study. Individuals on medication that may cause secondary DM were also excluded. The control individuals had a fasting plasma glucose level of <6.05 mmol/l (110 mg/dl), a blood hemoglobin A1c content of <6.2% and had no history of DM or of receiving anti-diabetic medication.

SNPs that were recently identified as susceptibility loci for CHD in Caucasian populations were searched for by meta-analyses of GWASs (13,14). These SNPs were examined with the dbSNP database (National Center for Biotechnology Information; http://www.ncbi.nlm.nih.gov/SNP/) to find SNPs with a minor allele frequency of >0.015 in a Japanese population. Finally, 29 SNPs (data not shown) were selected and the association with type 2 DM was examined. Wild-type and variant alleles of the SNPs were determined from the original sources.

Venous blood (7 ml) was collected into tubes containing 50 mmol/l ethylenediaminetetraacetic acid (disodium salt) and genomic DNA was isolated with a kit (Genomix; Talent, Trieste, Italy). Genotypes of the 29 SNPs were determined at G&G Science (Fukushima, Japan) by a method that combines polymerase chain reaction and sequence-specific oligonucleotide probes with suspension array technology (Luminex Corporation, Austin, TX, USA). The overall call rate of genotyping of 29 SNPs was 99%. The detailed genotyping methodology was performed as described previously (17).

The χ² test was used to compare the categorical variables, whereas the Mann-Whitney U test was used for analysis of the quantitative data. Allele frequencies of each SNP were compared between subjects with type 2 DM and controls by the χ² test. A false discovery rate (FDR) was calculated to compensate for multiple comparisons of genotypes, and FDR≤0.05 was considered to indicate a statistical significance for association. Multivariable logistic regression analysis was performed with type 2 DM as a dependent variable and age, gender (0, women; 1, men), body mass index (BMI) and the genotype of SNP as independent variables. The SNP was assessed according to dominant (the combined group of heterozygotes and variant homozygotes verses wild-type homozygotes), recessive (variant homozygotes verses the combined group of wild-type homozygotes and heterozygotes) and two additive [additive 1 (heterozygotes verses wild-type homozygotes) and additive 2 (variant homozygotes verses wild-type homozygotes)] genetic models. As fasting plasma glucose level and blood hemoglobin A1c content were not normally distributed (P<0.01 by the Kolmogorov-Smirnov Lilliefors test), these parameters were compared among genotypes by the non-parametric Kruskal-Wallis test. Statistical analysis was performed with JMP version 11 and JMP Genomics version 6.0 software (SAS Institute, Cary, NC, USA).

The clinical characteristics of the study subjects are shown in Table I. Age, the frequency of males, BMI, the prevalence of smoking, myocardial infarction, dyslipidemia and hypertension, as well as serum concentrations of triglycerides and creatinine, were higher, whereas the serum concentrations of high-density lipoprotein (HDL) cholesterol were lower in subjects with type 2 DM compared to controls.

Allele frequencies were compared between subjects with type 2 DM and controls by the χ² test and five SNPs with P<0.05 are shown in Table II. Among these SNPs, rs964184 (C→G) of the ZPR1 zinc finger gene (ZPR1) was significantly (FDR≤0.05) associated with the prevalence of type 2 DM. The genotype distributions of five SNPs were in Hardy-Weinberg equilibrium (P>0.05) among subjects with type 2 DM and controls.

Multivariable logistic regression analysis with adjustment for age, gender and BMI revealed that rs964184 of ZPR1 was significantly associated with type 2 DM in the dominant and additive 1 and 2 models, with the minor G allele representing a risk factor for this condition (Table III). As hypertriglyceridemia is an important risk factor for type 2 DM, additional multivariable logistic regression analysis was performed with adjustment for serum triglycerides concentrations or hypertriglyridemia (serum concentration of triglycerides ≥1.65 mmol/l or taking anti-dyslipidemic medication) in addition to age, gender and BMI (Table III). rs964184 was also significantly associated with type 2 DM in the dominant and additive 1 models in this analysis.

Finally, the associations of rs964184 genotypes to fasting plasma glucose level and blood hemoglobin A1c content were examined by the Kruskal-Wallis test (Table IV). rs964184 was significantly associated with the two parameters and the G allele was associated with the increases in fasting plasma glucose level and in blood hemoglobin A1c content.