Functional Significance of the UCSNP-43 Polymorphism in the CAPN10 Gene for Proinsulin Processing and Insulin Secretion in Nondiabetic Germans
Recently, an association of the G allele in UCSNP-43 of calpain 10 with type 2 diabetes and decreased glucose disposal was reported. Calpain 10 is also expressed in pancreatic islets. It is not known, however, whether and how this polymorphism contributes to the biological variation of β-cell function. We studied 73 nondiabetic subjects from the southwest region of Germany (G/G, n = 41; G/A, n = 29; and A/A, n = 3) using a modified hyperglycemic clamp (10 mmol/l glucose, added glucagon-like peptide 1, final arginine bolus). The genotype distribution was not different between subjects with normal glucose tolerance (n = 56) and those with impaired glucose tolerance (n = 17; P = 0.74, χ2 test). First-phase insulin secretion (adjusted for sex and insulin sensitivity from hyperglycemic clamp) was greater in G/G (2,747 ± 297 pmol/min) than in G/A + A/A (1,612 ± 156 pmol/min, P = 0.003). Insulin secretion in response to arginine (adjusted for insulin sensitivity) was also greater in G/G (9,648 ± 1,186 pmol/min) than in G/A + A/A (5,686 ± 720 pmol/min, P = 0.04). The acute poststimulus proinsulin-to-insulin ratio was lower in G/G (1.6 ± 0.4% first phase; 1.6 ± 0.2% arginine) than in G/A + A/A (4.0 ± 0.5% first phase, P < 0.001; 2.5 ± 0.4% arginine, P = 0.03). In conclusion, it appears unlikely that any association of the UCSNP-43 polymorphism alone with type 2 diabetes involves impairment of insulin secretion in our population of German Caucasians. This may be entirely different with specific haplotype combinations.
The etiology of type 2 diabetes involves both environmental and genetic factors. Genome scans in various populations identified a susceptibility locus (NIDDM1) on chromosome 2 (1). Following up on this region, Horikawa et al. (2) recently reported on a single-nucleotide polymorphism in intron 3 of the CAPN10 gene (UCSNP-43) in which the G allele appeared to be associated with type 2 diabetes. Although in diabetic Pima Indians, the G/G genotype was not associated with an increased prevalence of type 2 diabetes, in normal glucose tolerant Pima Indians, the G allele was associated with decreased glucose disposal (3). The CAPN10 gene encodes for calpain-10, a cysteine protease, and in muscle, reduced calpain-10 mRNA levels were observed in the G/G genotype (3). Calpain-10 is also expressed in pancreatic islets (2), but nothing is known about the functional significance of the UCSNP-43 polymorphism for β-cell function.
We therefore studied insulin secretion kinetics (modified hyperglycemic clamp) in 73 nondiabetic Caucasian subjects from the southwest region of Germany (Tübingen area) genotyped for UCSNP-43 (CAPN10-g.4852G/A). We found an allelic frequency of 0.76 for the G allele and a genotype distribution of G/G (n = 41), G/A (n = 29), and A/A (n = 3) that was in Hardy-Weinberg equilibrium (P = 0.7, χ2 test). The genotype distribution was not different between subjects with normal glucose tolerance (NGT) and those with impaired glucose tolerance (IGT) (P = 0.5, χ2 test) or between subjects with and without a family history of diabetes (P = 0.8, χ2 test). As it turned out, the demographic characteristics were reasonably well matched between the two genotypes (Table 1).
The insulin secretion kinetics are shown in Fig. 1. First- and second-phase insulin secretion was greater in G/G (2,747 ± 297 and 739 ± 60 pmol/min) than in G/A + A/A (1,612 ± 156 and 586 ± 33 pmol/min, P = 0.003 and P = 0.056, respectively) (Fig. 2A). First- and second-phase insulin secretion in response to glucagon-like peptide 1 (GLP-1) were also greater in G/G (4,224 ± 438 and 3,069 ± 194 pmol/min) than in G/A + A/A (2,515 ± 191 and 1984 ± 106 pmol/min, P = 0.05 for both). Maximal insulin secretion (in response to GLP-1 plus arginine) was also greater in G/G (9,648 ± 1,186 pmol/min) than in G/A + A/A (5,686 ± 720 pmol/min, P = 0.04). The proinsulin-to-insulin (PI/I) ratio was lower in G/G (1.6 ± 0.4% first phase, 1.6 ± 0.2 arginine) than in G/A + A/A (4.0 ± 0.5% first phase, P < 0.001; 2.5 ± 0.4 arginine, P = 0.03) (Fig. 2B). Insulin sensitivity was not different between G/G (0.17 ± 0.02 units) and G/A + A/A (0.15 ± 0.02 units, P = 0.5).
These results clearly show that in this nondiabetic German population, the G allele was associated with higher glucose-stimulated insulin secretion and more efficient proinsulin conversion to insulin compared with the A allele. Because both a reduced first-phase insulin response to glucose and a higher PI/I ratio are characteristic of type 2 diabetes and IGT (4,5), our findings are difficult to reconcile with the previously reported association of the G allele with type 2 diabetes (2). It is possible, however, that the increased insulin secretion is indicative of primary hypersecretion, which has been identified as an independent predictor of type 2 diabetes (6). In this scenario, the G allele might be involved in the pathomechanism of primary hyperinsulinemia.
It is also possible that for some reason our study population of German Caucasians contained a selection bias whose nature is not clear to us. Conceivably, in a larger and possibly more representative population, we may not see this difference in insulin secretion any more. Nevertheless, it can be excluded from our data with reasonable safety that the pathophysiological mechanism by which the UCSNP-43 polymorphism in calpain 10 infers an increased risk for type 2 diabetes involves impaired β-cell function. Assuming a 5% difference in β-cell function to be functionally relevant for increasing a population’s risk for type 2 diabetes, based on our results, we would have needed to study at least 3,000 subjects to demonstrate a significant difference in the opposite direction (α = 0.05, 1 − β = 0.80), i.e., insulin secretion in G/G being lower than in G/A + A/A. Partially consistent with our interpretation is the lack of a difference in the acute insulin response to intravenous glucose in Pima Indians with and without the A allele (3).
It is necessary to point out that we did not study specific haplotype combinations. In the original publication, individuals at higher risk for type 2 diabetes had a combination of two different haplotypes that both contained the G allele of UCSNP-43. However, because the low number of individuals carrying the high-risk haplotypes (∼1% in the German population ) makes the association approach impossible, we did not examine haplotype combinations. It cannot be excluded from our data that high-risk haplotypes (e.g., 112/121) have impaired β-cell function in a nondiabetic population.
Nevetheless, for our German Caucasian study population, it is possible to conclude that mechanisms by which the UCSNP-43 polymorphism specifically, irrespective of haplotype combinations, contributes to the development of type 2 diabetes is likely to involve insulin action or obesity-related factors, but not β-cell dysfunction.
RESEARCH DESIGN AND METHODS
All protocols are part of the Tübingen Family (TÜF) Study and were approved by the ethical committee of the University of Tübingen. Before the study, informed written consent was obtained from all participants. We studied 73 unrelated healthy volunteers (NGT n = 56, IGT n = 17; WHO criteria ), of whom 38 had a family history of type 2 diabetes using a modified hyperglycemic clamp. All subjects included in this analysis tested negative for GAD antibodies and did not take any medication known to affect glucose tolerance, insulin sensitivity, or insulin secretion. The study population was unselected and comprised every nondiabetic participant of the TÜF study undergoing a hyperglycemic clamp.
Genomic DNA extracted from peripheral blood lymphocytes was genotyped for UCSNP-43 (CAPN10-g.4852G/A) by sequencing a polymerase chain reaction (PCR)-amplified fragment. PCR primers were forward 5′-GCT GGC TGG TGA CAT CAG TG-3′ and reverse 5′-TCA GGT TCC ATC TTT CTG CCAG-3′, as previously described (3). PCR was performed in 25 μl volume with 100 ng of genomic DNA in a buffer containing 1.5 mmol/l MgCl 2, 1 μmol up-primer and 1 μmol down-primer, 0.2 mmol/l dNTP, and 0.2 μl (1U) of AmpliTaq Gold (Applied Biosystems, Foster City, CA) DNA polymerase. PCR conditions were as follows: 94°C for 10 min, followed by 33 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s, followed by a final extension at 72°C for 10 min. DNA was sequenced with the primer 5′-AGC AGG GTT GGA GCT TGA GAG-3′. DNA-cycle sequencing was carried out using the Big Dye Terminator on an automated DNA sequencer (model 310; Applied Biosystems).
Modified hyperglycemic clamp.
After an overnight fast, at around 8:00 a.m., a 200-min modified hyperglycemic clamp was performed as previously described (8). In brief, an intravenous glucose bolus was given to instantaneously raise blood glucose to 10 mmol/l. Subsequently, a glucose infusion was adjusted to maintain blood glucose at 10 mmol/l. After 120 min, GLP-1 (human GLP-1 (7-36) amide; Poly Peptide, Wolfenbhttel, Germany) was given as a primed-continuous infusion (0.6 pmol/kg; 1.5 pmol · kg−1 · min−1) during the next 80 min (9). At 180 min, a bolus of 5 g arginine hydrochloride (Pharmacia&Upjohn, Erlangen, Germany) was also injected. Sampling times are shown in Fig. 1.
Insulin secretion rates during the hyperglycemic clamp were calculated by deconvolution from C-peptide concentrations using standard kinetic parameters, as previously described (10,11). Phases of insulin secretion were calculated as previously described (12). Proinsulin conversion to insulin was assessed using the PI/I ratio (n = 71) immediately after the acute stimulus (2.5–5 min after the glucose and arginine bolus), when differences in elimination kinetics have negligible influence on concentrations (5,13). An insulin sensitivity index was calculated as the glucose infusion rate second phase divided by the insulin concentration between 60 and 120 min during the hyperglycemic clamp (8).
For statistical analyses, insulin secretion parameters were log transformed and adjusted for insulin sensitivity and sex. For statistical comparisons with the wild type (A/A), subjects heterozygous (G/A) and homozygous (A/A) for the mutation were combined. For statistical analysis, the secretion indexes were log transformed and linearly adjusted for the insulin sensitivity index. Comparisons between genotypes were made using unpaired Student’s t test or Wilcoxon’s rank test, where appropriate. A P value of <0.05 was considered to be statistically significant.
This study was supported in part by the European Community (QLRT-1999-00674) and the Deutsche Forschungsgemeinschaft DFG (FR 1561/1-1).
Address correspondence and reprint requests to Dr. Michael Stumvoll, Medizinische Universitätsklinik, Otfried-Müller-Str. 10, D-72076 Tübingen, Germany. E-mail:.
Received for publication 17 January 2001 and accepted in revised form 21 May 2001.
GLP-1, glucagon-like peptide 1; IGT, impaired glucose tolerance; NGT, normal glucose tolerance; PCR, polymerase chain reaction; PI/I, proinsulin-to-insulin; TÜF, Tübingen Family.