Introduction

Type 2 diabetes mellitus (T2DM) is a complex metabolic disease in which multiple genetic effects and metabolic and environmental factors contribute to pathogenesis (McCarthy 2004). A genome-wide search for T2DM genes in Mexican-Americans identified a susceptibility locus on chromosome 2q37.3, NIDDM1, for T2DM (Hanis et al. 1996). Horikawa et al. (2000) identified the Calpain-10 Gene [CAPN10 (MIM 605286)] by fine mapping and positional cloning as a putative T2DM susceptibility gene. CAPN10 is comprised of 15 exons spanning 31 kb of genomic sequence and encodes a 672-amino-acid intracellular protease. The allele combination of CAPN10 (SNP-43, -19, and -63), is reported to be associated with increased risk of T2DM in many populations (Horikawa et al. 2000; Cassell et al. 2002; Elbein et al. 2002; Lynn et al. 2002; Malecki et al. 2002; Iwasaki et al. 2005), although this is somewhat controversial (Tsai et al. 2001; Daimon et al. 2002; Fingerlin et al. 2002; Chen et al. 2005; Wu et al. 2005). The 112/121 diplotype of CAPN10 is associated with a 3-fold increase in the risk of contracting T2DM in Mexican-Americans and in Northern European populations (Horikawa et al. 2000). However, the associations between these genetic variants of CAPN10 and diabetes have not been consistently observed in other populations, including Japanese (Evans et al. 2001; Tsai et al. 2001; Rasmussen et al. 2002; Horikawa et al. 2003; del Bosque-Plata et al. 2004). Further studies in other racial populations are still needed to confirm the role of CAPN10 polymorphisms in the pathogenesis of T2DM. The aim of this study was to determine whether there is any association between specific CAPN10 polymorphisms and increased T2DM risk in the Korean population.

Materials and methods

Subjects and study design

The study population consisted of 454 unrelated subjects with T2DM and 236 non-diabetic controls. The diagnosis of T2DM was made according to the criteria of the WHO (2003). The age at onset of T2DM in the diabetic subjects was <60 years. In order to rule out type 1 or maturity-onset diabetes of the young, the following were excluded from this study: subjects diagnosed before they were 25 years old or subjects receiving insulin therapy within 3 years of the onset of diabetes. The non-diabetic control population consisted only of individuals with a normal fasting glucose level (< 100 mg/dl), no family history of diabetes and who were older than 60 years old. Subjects with late onset (>60 years old) T2DM and younger non-diabetic controls (<60 years old) were excluded from the examination to increase genetic power. The study was approved by the Institutional Review Board of Inha University Hospital, and written informed consent was obtained from all participants.

Data collection

The patients’ medical and family histories were recorded. Their blood pressure, height, and weight were measured. Blood samples were collected after an overnight fast to determine fasting plasma glucose, HbA1c, total cholesterol, high-density lipoprotein (HDL)-cholesterol, and triglyceride levels. Concentrations of fasting plasma glucose, total cholesterol, and triglyceride levels were determined using an enzymatic colorimetric assay. The HDL-cholesterol concentration was measured using lipoprotein electrophoresis. The low-density lipoprotein (LDL)-cholesterol level was calculated using the Friedewald formula (Friedewald et al. 1972). The HbA1c value was determined using high-performance liquid chromatography (Greencross, Seoul, Korea). The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated by {[fasting insulin (μU/ml) × fasting glucose (mmol/l)]/22.5}. The HOMA-beta cell function was calculated by [20 × fasting insulin (μU/ml)/fasting glucose (mmol/l)−3.5] (Matthews et al. 1985; Kang et al. 2005a).

Genotyping analysis

Three polymorphisms in CAPN10 were genotyped for haplotype analyses as described previously (Evans et al. 2001; Horikawa et al. 2003): SNP-43 (g.4852, G>A, rs3792267), SNP-19 (g.7920 in/del32bp, rs3842570), and UCSNP-63 (g.1637C>T, rs5030952). The alleles were designated as described by Horikawa et al. (2000). A polymerase chain reaction was performed with genomic DNA using a sense primer, 5′-CACGCTTGCTGTGAAGTAATGC-3′, and an antisense downstream primer, 5′-CTCTGATTCCCATGGTCTGTAG-3 for SNP-43, 5′-GTTTG GTTCTCTTCAGCGTGGAG-3′ and 5′-CATGAACCCTGGCAGGGTCTAAG-3′ for SNP-19, and 5′-AGCACTCCCAGCTCCTGATC-3′ and 5′-AAGGGGGGCCAGGGC CTGACGGGGGTGGCG-3″ for SNP-63. NsiI (Promega, Madison, WI) and HhaI (Takara Bio, Shiga, Japan) were used to perform restriction fragment length polymorphism analysis.

Statistical analysis

Genotypic distributions were tested for deviation from the Hardy–Weinberg equilibrium using the chi-square goodness-of-fit test. A comparison of variables between the groups of genotypes was performed using a two-tailed Student’s t-test. Statistical differences in the genotype frequencies between T2DM subjects and the controls were assessed by a chi-square test. In diplotype analysis, odds ratio (OR) and 95% confidence interval (CI) were calculated by comparing each diplotype with all the other diplotypes as a group. The OR and 95% CI for the various genotypes were calculated using Fisher’s exact test. Statistical analysis of triglyceride, HDL-cholesterol, and LDL-cholesterol levels was performed using log-transformed values because the distribution was not normal. Pairwise linkage disequilibrium (LD) between three Calpain-10 loci was assessed using SAS Genetics (version 9.1, SAS Institute, Cary, NC) statistical package. Haplotype frequencies created by these three SNPs were inferred by a maximum likelihood method using the Haplotyper program (http://www.people.fas.harvard.edu/~junliu/Haplo/) (Niu et al. 2002; Kang et al. b). Haplotype distributions were analyzed by a likelihood-ratio test. Statistical power was computed using a power calculator program (http://calculators.stat.ucla.edu/powercalc/). P values <0.05 were considered significant.

Results

A total of 454 unrelated subjects with T2DM and 236 non-diabetic controls were genotyped. There were no gender differences between the two groups. HbA1c, fasting plasma insulin, HOMA-IR, and HOMA-beta cell values in the diabetic patients were 8.4±3.3%, 83.23±95.43 pmol/l, 4.2±5.4, and 62.7±79.9, respectively. The body mass index (BMI) and blood pressure were higher in the diabetes patients than that in the non-diabetic controls (Table 1). The diabetes group showed higher serum triglyceride levels and lower HDL-cholesterol levels than the control group. There was no significant difference in the frequency of SNP-43, -19, and -63 between the diabetic patients and the non-diabetic controls, which is in agreement with previous studies (Table 2) (Horikawa et al. 2000; Daimon et al. 2002). The distribution of each allele did not significantly deviate from Hardy–Weinberg equilibrium (Table 2). The Haplotyper program (Stephens et al. 2001) estimated the haplotype allelic distributions and identified eight alleles (Table 3). Haplotype frequencies were 47.7% for 121, 19.2% for 112, 14.7% for 111, 9.5% for 122, and 6.7% for 221. The most common haplotype in this study was 121, which is in agreement with other previous studies (Horikawa et al. 2000; Cassell et al. 2002; Fingerlin et al. 2002; Malecki et al. 2002; Wu et al. 2005). Calculations of the LD showed a weak positive LD between these SNPs (Table 4). LD among the three markers (D′) ranged from 0.2383 to 0.5464 (average pairwise D′ of 0.4423). Likelihood-ratio test revealed that 111 haplotype showed a high risk of T2DM (<0.0001; Table 3). The 121/121 diplotype was associated with significantly decreased risk of T2DM (OR=0.660, 95% CI =0.460–0.948, =0.029, power =63.5%; Table 5). The subjects with 111/121 diplotype showed a high risk of T2DM (OR =2.580, 95% CI =1.602–4.155, <0.001, power =78.7%; Table 5). The 112/121 diplotype, which is reported to be associated with T2DM in Mexican-American and Northern European population (Horikawa et al. 2000; Cox et al. 2004), was not significant in this study.

Table 1 Clinical characteristics of the study population. BMI Body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, FPG fasting plasma glucose, LDL low-density lipoprotein, HDL high-density lipoprotein
Table 2 Allele and genotype frequencies of the CAPN10 SNPs. Data are represented as mean ±SD. Alleles were designated as described by Horikawa et al. (2000): SNP-43 allele 1 = G, allele 2 = A; SNP-19 allele 1 = 2R (two repeats of 32-bp sequences), allele 2 = 3R (three repeats of 32-bp sequences); SNP-63 allele 1 = C, allele 2 = T. Haplotypes 211, 212, and 222 were not observed. P value assessed by Hardy–Weinberg equilibrium chi-square test
Table 3 Frequencies of haplotype and the risk of T2DM. The odds ratio and 95% confidence interval of each haplotype are relative to the other haplotypes as a group. Haplotypes with very rare frequencies (< 5%) are not shown
Table 4 Linkage disequilibrium (LD) between CAPN10 polymorphisms in the Korean populations
Table 5 Frequency of diplotype of CAPN10 and the risk of T2DM. The odds ratio (OR) and 95% confidence interval (CI) of each diplotype are relative to the other diplotype as a group. P values are calculated by Fisher’s exact test

Discussion

We previously reported that 111/121 diplotype is associated with metabolic syndrome in patients with T2DM (Kang et al. 2006). In this study we investigated whether the 111/121 diplotype is associated with diabetes or not. Our results show that a novel 111/121 diplotype, defined by the three polymorphic alleles designated SNP-43, -19, and -63 of CAPN10, is associated with increased risk of T2DM in the Korean population. This study also suggests that subjects homozygous for the 121 haplotype have a reduced risk of T2DM. Although Horikawa et al. demonstrated that the 112/121 diplotype confers the highest risk of diabetes in Mexican-Americans (OR =2.8) (Horikawa et al. 2000; Cox et al. 2004), we found no significant association between the 112/121 diplotype and increased susceptibility to T2DM (OR =0.693, =0.077). Wu et al. (2005) reported that 112/221 was associated with a reduced risk of T2DM in a Chinese population (OR =0.39). We did not find a reduced frequency of this diplotype 112/221 in the T2DM group compared to the control group (OR =0.518, =0.072). The 121 haplotype was reportedly associated with increased risk in European populations (Malecki et al. 2002; Orho-Melander et al. 2002) while being associated with a reduced risk in Japanese (Iwasaki et al. 2005). There was no significant association between diabetic risk and the 121 haplotype in this study; conversely, the 111 haplotype was shown to be associated with an increased risk of T2DM in this study (<0.0042). Our study also showed that the common 121/121 diplotype appeared to be associated with protection from T2DM in our population (OR =0.660, 95% CI =0.460–0.948, <0.001). We examined the association between the CAPN10 SNP-44 genotype and diabetes in a subset of our study population (96 diabetes and 96 controls). We found no significant association between CAPN10 SNP-44 and diabetes (=0.0.817).

Calpain is a member of a family of calcium-activated intracellular proteases. The CAPN10 gene is located on chromosome 2q37, encodes at least eight alternative splicing variants, and contains 15 exons spanning 31 kb. Calpain-10 is expressed in many tissues such as the heart, pancreas, brain, liver, skeletal muscle, and kidney (Horikawa et al. 2000), and is essential for multiple cellular functions. The exact molecular mechanisms of how some polymorphisms in the CAPN10 increase susceptibility to T2DM are unclear. However, the polymorphisms of this gene most likely influence glucose uptake in skeletal muscle and adipocytes (Paul et al. 2003; Ridderstrale et al. 2005; Turner et al. 2005) as well as glucose-induced insulin secretion in pancreatic β cells (Sreenan et al. 2001; Paul et al. 2003; Zhou et al. 2003; Ridderstrale et al. 2005; Turner et al. 2005).

Most of the haplotype frequencies in our study are similar to those reported by Horikawa et al. (2003) except for the frequency of the significant haplotype 111 (diabetes 16.48% vs 12%; control 7.95% vs 9%). It is possible that the younger age and increased obesity in our population (53.3±11.0 years old vs 62.0±11.0 years old; BMI, 24.4±3.5 kg/m2 vs 23.9±3.3 kg/m2) might have contributed to this discrepancy. Additionally, late onset (>60 years) diabetic subjects were excluded from our study, and our study included a larger number of diabetic subjects (454 vs 177) than the study of Horikawa et al. (2003). The susceptibility locus of the Mexican-Americans, NIDDM1, may also contribute to the development of T2DM in other populations, but each study localized susceptibility to different regions of the genome, which suggests that different combinations of the susceptibility gene contribute to the development of T2DM in each population.

In conclusion, we found that the novel 111/121 diplotype in the Calpain-10 gene is associated with an increased risk of T2DM and that the 121/121 diplotype was associated with a significantly decreased risk of T2DM in the Korean population.