Genetic Factors and Insulin Secretion: Gene Variants in the IGF Genes

Participants were selected from three independent studies in the Netherlands and Germany.

We have studied 76 subjects with NGT from Germany according the protocols of the Tübingen Family Study for Type 2 Diabetes (20). Newspaper ads and word-of-mouth proposing diabetes screens were used to recruit healthy volunteers for this study. Glucose tolerance status was assessed using OGTTs, and all subjects were GAD antibody negative and unrelated to each other. The subjects participating in the study had a mean age of 36 ± 12 years.

The Dutch NGT subjects (n = 67) were partly recruited as first-degree relatives from type 2 diabetic subjects (n = 44) (18,21). The other part of this cohort consisted of matched normoglycemic subjects. Subjects were unrelated to each other, and their mean age was 46 ± 6 years. A standard OGTT was used to define glucose tolerance status in all subjects.

Subjects with IGT (n = 94) were all of Dutch origin (19). The IGT subjects (aged 57 ± 7 years) were detected by population screening in the city of Hoorn. Subjects with a fasting glucose >5.5 mmol/l were invited for two OGTTs on separate days. Those with a mean postload glucose level between 8.6 and 11.0 mmol/l were included in the study. Also in this cohort, all subjects were unrelated to each other. The Dutch NGT and IGT groups were recruited independently in different regions of the country. Details of the study groups are described previously (18–20). Informed consent was obtained from all participants, and the appropriate local medical ethics committees approved the protocol.

Hyperglycemic clamps were performed at 10 mmol/l in all subjects. The Dutch NGT and IGT subjects underwent a 3-h clamp. In the German NGT subjects, the clamp lasted for 2 h. After an overnight fast, the subjects received an intravenous glucose bolus to acutely raise glucose levels to 10 mmol/l. Blood glucose levels were measured at the appropriate intervals to maintain a constant blood glucose during the clamp. Blood samples for insulin were drawn at 2.5-min intervals during the first 10 min of the clamp and at 10- to 20-min intervals during the remainder. Exact details of the clamping procedures in the different study groups were described previously (18–20,22). First-phase insulin secretion was defined as the sum of the insulin levels during the first 10 min of the clamp. Second-phase insulin secretion was defined as the mean of the insulin values during the last 40 min (80–120 min, NGT group, Germany) or the last 30 min (150–180 min, NGT and IGT groups, the Netherlands) of the clamp. The insulin sensitivity index (ISI) was determined by relating the glucose infusion rate to the plasma insulin concentration during the last 40 min (NGT, Germany) or 30 min of the clamp. The disposition index (DI) was calculated as the product of first-phase insulin secretion and ISI according to Bergman et al. (23).

Classification of the repeat length in the IGF-I gene was as described previously (10); carriers of the normal 192-bp allele had a repeat length of 19 CA repeat units. All other repeat lengths were classified as noncarrier (10). Genotypes of the ApaI polymorphism in the IGF-II gene were determined by a PCR-restriction fragment length polymorphism (RFLP)-based method or direct sequencing as described previously (13). The most prevalent GG genotype was used as a reference.

All data are presented as means ± SE or median with interquartile range. ANOVA or Mann-Whitney U test was used for general comparisons between the different genotypes. Variables were log-transformed before analysis if necessary. Adjustments for age, sex, BMI, and study center in the pooled analyses were done in separate general linear regression analyses for all parameters. A priori power calculations showed that the design used in this study would allow the detection of a difference in first or second phase of insulin secretion between 15 and 30% with 90% power (P ≤ 0.05). Results were regarded significant at P ≤ 0.05. Statistical analysis was performed with the SPSS version 10.0 software (SPSS, Chicago).

We used three different cohorts in our studies. One of the cohorts consisted of randomly recruited NGT subjects from Germany. The second cohort of NGT subjects was in part recruited as first-degree relatives of type 2 diabetic individuals. The third cohort consisted of subjects with IGT (18–20). The last two cohorts were independently selected from two different regions in the Netherlands and had both an increased risk for type 2 diabetes and evidence for β-cell dysfunction (22). As can be expected, the IGT group had the lowest insulin secretion levels, ISIs, and DIs compared with the other two NGT groups (21). Classification of the repeat length in the IGF-I gene was as described previously (10); carriers of the normal 192-bp allele had a repeat length of 19 CA repeat units. All other repeat lengths were classified as noncarrier (10). Genotypes of the ApaI polymorphism in the IGF-II gene were determined by a PCR-RFLP-based method as described previously (13). Genotype frequencies for the IGF-I and IGF-II variants were comparable between the different cohorts and are comparable to those reported in other Caucasian populations (Tables 1 and 2) (10–12,15). All genotype distributions were in the Hardy-Weinberg equilibrium (data not shown).

We have investigated whether the IGF-I gene variant is associated with insulin secretion in two independent populations from the Netherlands. Glucose-stimulated insulin secretion during the hyperglycemic clamp was not significantly different between the different genotypes, both under a recessive or dominant model (Table 1). However, we do see a reduced ISI and DI in IGT noncarriers of the 192-bp allele (P for trend <0.05, Table 1). Combined analysis of the Dutch NGT and IGT groups resulted in a significant association of the polymorphism with clamp-derived insulin sensitivity (ISI) (P = 0.01 for 192 bp/192 bp + non-192 bp/192 bp vs. non-192 bp/non-192 bp, with adjustment for age, sex, BMI, and study center). Furthermore, we noted a decreased DI in noncarriers of the 192-bp allele (P = 0.03 for trend, with adjustment for age, sex, BMI, and study center). Other diabetes-related parameters such as glucose and insulin levels were not significantly different between the different IGF-I genotypes (Table 1). There was no evidence for an interaction with BMI (data not shown). Data about the IGF-I polymorphism in the German NGT cohort are not available.

Because of the low number of AA carriers, we pooled the AA and AG genotypes of the IGF-II gene variant in most of our analyses (Table 3). Separate testing of the AA carriers did not essentially change our findings. Glucose and insulin levels during OGTTs were not significantly different between the two genotype groups in either of the study populations (data not shown). Also, the first phase of the glucose-stimulated insulin secretion, as measured by hyperglycemic clamps, was not different between the genotypes (adjusted for age, sex, and BMI) (Table 3). We did see, however, a decreased second-phase insulin secretion in the NGT group from the Netherlands (P = 0.04 for GG vs. AA + AG, adjusted for age, sex, and BMI, Table 3). This was, however, not replicated in the two other cohorts (Table 3). If BMI is a mediating factor in the causal chain of the gene variant leading to alterations in glucose metabolism, adjustment for BMI would be an over-adjustment. However, when we repeated the tests without BMI in the model, the results did not change.

Insulin sensitivity, as assessed with the hyperglycemic clamps, was not significantly different between the genotypes in all our cohorts (Table 3). Furthermore, there was no evidence for a gene-environment interaction with factors such as BMI and triglycerides (data not shown). However, trends toward lower body weight and BMIs were observed in carriers of the AA genotype without reaching statistical significance in any of the populations (Table 2). A pooled analysis of all three study cohorts did not change our findings (P > 0.1, data not shown). Furthermore, we did not observe significant changes in fasting glucose and (pro)insulin levels, lipid profiles, and other diabetes- and/or obesity-related parameters (Table 2, data not shown).

It has been shown previously that IGF-I and IGF-II levels are determinants of fetal growth and birth weight in both humans and animals. Reduced birth weight is a known risk factor for the development of diabetes and other features of the metabolic syndrome (7). Genetically determined low levels of IGFs may in part explain these observations. Previously it was shown that both the IGF-I and IGF-II gene variants examined in this study are associated with reduced levels of the respective growth factors (10,11). IGFs are also important factors for the delicate balance between cell survival and cell death in the pancreatic β-cell (4). A disturbed balance between β-cell replication and apoptosis might result in an altered setting of the β-cell, eventually resulting in β-cell failure. IGF-I and IGF-II bind to specific receptors on the pancreatic β-cell. One of these receptors is the IGF-I receptor (8). Studies using β-cell-specific IGF-I receptor knockouts have shown that these animals have an almost absent first and second phase of glucose-stimulated insulin secretion (8,9). Because β-cell mass is unaffected, this might be caused by altered glucose-sensing capacities of the β-cells lacking the IGF-I receptor (8). Together, this has led us to speculate that genetic defects in both IGF-I and IGF-II alter glucose-stimulated insulin secretion in humans. Preliminary data suggested a secretion defect in both carriers of the IGF-I and IGF-II gene variants (13,14).

In this study, we have examined glucose-stimulated insulin secretion in relation to these gene variants. From our data, we conclude that the two gene variants in the IGF genes are not associated with the magnitude of glucose-stimulated insulin secretion during hyperglycemic clamps in independent Dutch and German subjects. This suggests that the mild variations in IGF levels associated with the gene variants in IGF-I and IGF-II do not affect glucose-stimulated insulin secretion. This is further corroborated by the fact that IGF-I levels are not associated with glucose-stimulated insulin secretion (24). We have examined three cohorts with either NGT or IGT originating from three independent populations in the Netherlands and Germany. A limited part of our populations consisted of first-degree relatives of type 2 diabetic individuals and subjects with IGT (18,19). Both groups have an increased risk of developing type 2 diabetes (22). One of the early defects in these at-risk individuals is a reduced first-phase insulin secretion (22). If mutations in the IGF genes are associated with altered glucose-stimulated insulin secretion, we expect this should be first manifest in these at-risk individuals. With respect to changes in (first-phase) insulin secretion, we did not observe any association. We cannot, however, exclude the possibility that changes in insulin secretion only become manifest after stressing glucose homeostasis, for instance during overfeeding. Previously, it has been shown that carriers of the IGF-II gene variant have an altered insulin response during an OGTT after a 100-day period of overfeeding (13).

We observed a reduced ISI and DI associated with the CA repeat polymorphism in the IGF-I gene, suggesting an altered balance between insulin sensitivity and insulin secretion. Previously, it has been shown that the ISI as measured with the hyperglycemic clamp technique correlates very well with the values obtained by the “gold standard” hyperinsulinemic-euglycemic clamp (25). It has also been shown that a hyperbolic relationship exists between insulin secretion and insulin sensitivity. This hyperbolic relationship can be mathematically described by the multiplication of indexes for (first-phase) insulin secretion and insulin sensitivity, which is referred to as the disposition index (23). Subjects with type 2 diabetes have a lower DI than normoglycemic subjects. A low DI indicates an inability of the β-cell to adapt adequately to changes in insulin sensitivity. It is conceivable that, under circumstances of deterioration of insulin action, as occurs in obesity, this situation will lead to (further) deterioration of glucose homeostasis and may consequently lead to type 2 diabetes. Whether this is the case in noncarriers of the normal CA repeat allele of the IGF-I promoter gene is so far not known.

The exact mechanism by which IGF-I affects insulin sensitivity is unclear but might involve shared components of the IGF-I/insulin-signaling pathway (6). This phenomenon has been observed in mice lacking IGF-I expression in the liver. These mice show muscle insulin resistance most likely due to defects at the level of the insulin receptor (26). Furthermore, it is known that IGF-I levels affect insulin sensitivity in humans as well (3). Further studies are necessary to examine the relationship between the CA repeat in the IGF-I promoter and insulin resistance and reduced DI in detail.

We conclude that gene variants in the IGF-I and IGF-II genes are not associated with detectable alterations in glucose-stimulated insulin secretion in three independent study populations from the Netherlands and Germany. However, our findings of a reduced ISI and a reduced DI in noncarriers of the normal CA repeat allele in the IGF-I promoter may point to a disturbed adaptation of β-cells to insulin action and warrants further study, especially in obese (insulin-resistant) subjects.

The Dutch work described in this article was supported in part by the Diabetes Fonds Nederland (Amersfoort, the Netherlands) and the European Union Cost B17 program. The German studies were supported in part by a grant from the Deutsche Forschungsgemeinschaft to A.F. (Fr 1562/1-1). M.S. is currently supported by a Heisenberg-Grant from the Deutsche Forschungsgemeinschaft.

The authors thank all participants for their cooperation.