Phenotyping of Individual Pancreatic Islets Locates Genetic Defects in Stimulus Secretion Coupling to Niddm1i Within the Major Diabetes Locus in GK Rats

Reagents of analytical grade and deionized water were used. Collagenase, HEPES, and bovine serum albumin (BSA; fraction V) were obtained from Boehringer Mannheim (Mannheim, Germany). Tetramethylbenzidine, insulin-peroxidase, and KIC were bought from Sigma (St. Louis, MO). The rat insulin standard was from Novo Nordisk (Bagsvaerd, Denmark). IgG-certified microtiter plates were purchased from Nunc (Roskilde, Denmark). Tolbutamide was a gift from Hoechst (Frankfurt, Germany). The mouse insulin antibodies were raised in our laboratory from guinea pigs.

GK and F344 rats were obtained and bred as described previously (17,18). GK-derived genetic intervals were transferred onto the F344 background by 10 successive backcrosses followed by intercrosses between heterozygous animals to establish homozygous congenic strains. At each generation, genetic markers from the Niddm1 region were used to verify the integrity of the GK-susceptibility haplotype. The GK-derived chromosome regions are 7.6 ± 1 cM in Niddm1f (defined by the markers D1Mit7 and D1Mgh25) and 22 ± 1 cM in Niddm1i (defined by the markers D1Arb42a and D1Rat90).

Pancreatic islets were isolated from male Niddm1f, Niddm1i, and GK rats aged 16 weeks by shaking excised and minced pancreata in a buffer containing collagenase (6.7 mg/ml) at 37°C for ∼15 min. Age-matched male F344 rats were used as controls. Islets were cultured for 24 h in RPMI 1640 medium supplemented with 10% fetal calf serum, 100 IU/ml penicillin, and 5.5 mmol/l glucose. Perifusion of the islets for insulin and oxygen tension measurements was performed in a medium containing 1 mg/ml BSA and (in mmol/l): NaCl 125, KCl 5.9, MgCl₂ 1.2, CaCl₂ 1.28, and HEPES 25, titrated to pH 7.4 with NaOH. Glucose, tolbutamide, and KIC were added to the perifusion medium in concentrations indicated in the tables and figures.

Individual islets were placed in a 10-μl chamber and perifused at 37°C in the presence or absence of 3 mmol/l glucose. The flow rate was kept constant throughout each experiment with the aid of a peristaltic pump placed before the islet. Variation in the flow rate of 150–200 μl/min was allowed between experiments. These variations did not affect the insulin release pattern. After 60 min, collection of perifusate in 20-s fractions began. The samples were immediately cooled on ice. For islets perifused in the presence of 3 mmol/l glucose, the sugar concentration was raised to 11 mmol/l, and subsequently 1 mmol/l tolbutamide was added to the perifusion medium. For islets perifused in the absence of glucose, 11 mmol/l KIC was added to the perifusion medium. The perifusates were analyzed for insulin by a competitive enzyme-linked immunosorbent assay with the insulin-capturing antibody immobilized directly in the solid phase (19). The rate of insulin release was normalized to islet dry weight after freeze-drying and weighing using a quartz fiber balance. The islets weighed 1–2 μg.

Individual islets were attached to a poly-l-lysine–coated coverslip, used as a bottom in an open Sykes-Moore perifusion chamber (20). The chamber was thermostated to 37°C. The islets were perifused at a rate of 150–160 μl/min. Oxygen tension in the isolated islets was measured by a modified Clark microelectrode (21,22). The tip of the microelectrode was positioned into the islets with the aid of a micromanipulator. A stereomicroscope was used to control the penetration depth of the electrode, which was 25–50 μm. The electrode was polarized at −0.8 V, which gave a linear response between the oxygen tension and the electrode current. The electrical current was measured by a picoamperemeter (University of Aarhus, Aarhus, Denmark). After insertion of the electrode, the islets were allowed to equilibrate for 50 min in the presence or absence of 3 mmol/l glucose before sampling began. The sampling rate of the A/D converter was four data points per second. Averages of data points acquired during 20 s were calculated and used for further data presentation. For islets perifused in the presence of 3 mmol/l glucose, the sugar concentration was raised to 11 mmol/l. For islets perifused in the absence of glucose, 11 mmol/l KIC was added to the perifusion medium. Oxygen tensions at 11 mmol/l glucose and 11 mmol/l KIC were obtained 15 min after raising the glucose concentration or adding KIC, when oxygen tension had stabilized. The electrodes were calibrated in water saturated with Na₂S₂O₅ or air at 37°C before and after the experiments. The drift of the microelectrodes was <0.5% / h.

The data points for insulin secretion in the figures represent three-point moving averages. To analyze the oscillations of insulin release, we collected perifusates from islets exposed to 11 mmol/l glucose and tolbutamide to ensure sufficient amounts of insulin. Frequency determination of oscillations in insulin release was performed by Fourier transformation on both original and autocorrelated data using Igor software (Wave Metrics, Lake Oswego, OR). Results are expressed as means ± SE. Differences in mean insulin secretory rates and oxygen tension between groups were evaluated with one-way analysis of variance with Fisher’s post hoc test. Differences within groups were evaluated with Student’s t test for paired data.

Insulin release from individual islets was measured at 3 and 11 mmol/l glucose (Table 1, Fig. 1). At basal glucose, insulin release was 5.3 ± 1.5 and 4.2 ± 1.1 pmol · g⁻¹ · s⁻¹ from islets isolated from congenic strains Niddm1f and Niddm1i, respectively. Similar results were obtained from islets isolated from control F344 and diabetic GK rats. When the glucose concentration was increased to 11 mmol/l, the rate of insulin release doubled in islets from Niddm1f and normoglycemic F344 rats. In contrast, no change in the insulin secretory rate was observed from Niddm1i and GK islets when the glucose concentration was raised to 11 mmol/l.

The effect of the nonmetabolizable secretagogue tolbutamide on insulin release from islets was evaluated in the presence of 11 mmol/l glucose (Table 1, Fig. 1). The addition of 1 mmol/l of the sulfonylurea to the perifusion medium increased insulin release threefold in islets from all four strains. However, the secretory rates in the presence of 11 mmol/l glucose and tolbutamide differed substantially, with a significantly (P < 0.01) lower absolute rate in Niddm1i islets as compared with F344 islets. Islets from all four strains released insulin in pulses, with frequencies ranging from 0.37 to 0.39 min⁻¹ (Table 1).

The effect of KIC on insulin release from individual islets was assessed in the absence of glucose (Table 2, Fig. 2). In the absence of glucose, insulin release was similar to that observed at 3 mmol/l glucose. Introduction of 11 mmol/l KIC to the perifusion medium caused a prompt rise in insulin release in Niddm1f, Niddm1i, and control F344 islets. In these islets, insulin release increased approximately eightfold during the first 5 min upon exposure to KIC. In contrast, insulin release from GK islets was not affected by KIC.

Oxygen tension was measured from isolated islets in the presence of 3 and 11 mmol/l glucose (Table 3, Fig. 3). At low glucose, pO₂ in islets from congenic Niddm1f and Niddm1i rats was 121.8 ± 3.9 and 125.1 ± 3.0 mmHg, respectively. Similar results were obtained for control and GK islets. Raising the glucose concentration to 11 mmol/l caused a significant decrease of pO₂ by ∼6% in islets from Niddm1f, Niddm1i, F344, and GK rats.

The effect of KIC on pO₂ in islets from the four rat strains was determined in the absence of glucose (Table 4, Fig. 4). In the absence of glucose, pO₂ was similar to that recorded at 3 mmol/l glucose. In islets from Niddm1f and Niddm1i rats, 11 mmol/l KIC decreased pO₂ by ∼15%. Whereas a similar reduction in pO₂ was observed in islets from F344 rats, pO₂ was not significantly decreased in islets from GK rats.

The GK rat was established as a diabetes model by selection for offspring with hyperglycemia in successive generations, followed by sister–brother inbreeding to obtain a homozygous strain (23). During the selection process, alleles that contribute to the diabetes phenotype were fixed along several different pathways, e.g., deranged insulin secretion, decreased insulin action in target tissues, dysregulated suppression of hepatic glucose production, or feeding behavior. Exploration of the major diabetes QTL (Niddm1) (6) has led to the identification of one series of congenic rat strains that characterizes a sublocus with an insulin-resistance type of diabetes, Niddm1b/Niddm1f/Niddm1e. The corresponding congenic strains displayed postprandial hyperglycemia, fasting hyperinsulinemia, increased body weight and fat, reduced lipogenesis in isolated epididymal fat cells, and blunted insulin-stimulated glucose uptake in skeletal muscle (17,18). The disease-causing gene(s) was mapped to a 1-cM region that co-localized with the gene for insulin degrading enzyme. This region is common to all three congenic rat strains. The GK-derived enzyme encoded two amino acid substitutions that reduced enzyme activity by 30%.

In the present report, we characterized the insulin secretory phenotype of Niddm1f and Niddm1i rats. Congenic rats were therefore studied at the level of individual pancreatic islets by measuring insulin release and oxygen tension under various conditions. The results show that changes in insulin release and oxygen tension in response to glucose, tolbutamide, and KIC in islets from Niddm1f rats are comparable to those in islets isolated from normoglycemic F344 rats. The findings confirm that the diabetes phenotype of the Niddm1f sublocus is not associated with any major defects in insulin secretion.

The other sublocus of Niddm1, Niddm1i, encodes reduced lipogenesis in isolated fat cells and a recessive postprandial hyperglycemia associated with whole-body postprandial serum insulin deficiency (17). Congenic Niddm1i rats contain a 22-cM GK-derived chromosome region that is not contained in Niddm1b (17). Our present results, which show that islets from Niddm1i congenic rats are unable to respond with increased insulin secretion rate to a rise in glucose concentration from 3 to 11 mmol/l, strongly support the notion that Niddm1i encodes a defect in insulin secretion. Indeed, in ∼50% of the Niddm1i islets, the increase in the glucose concentration caused a reduction of insulin release. This phenomenon has been observed in isolated islets as well as in the perfused pancreas from the GK rat (24). This paradoxical insulin response to glucose has been explained by an imbalance between the action of glucose to increase the cytoplasmic Ca²⁺ concentration ([Ca²⁺]_i) by depolarization and influx of the ion and to decrease [Ca²⁺]_i by promoting intracellular sequestration and outward transport of the ion (25). If depolarization is impaired and influx of Ca²⁺ is reduced, then the lowering effect of glucose on [Ca²⁺]_i will manifest itself in a reduction of insulin release. A transient lowering in plasma insulin is a common observation in patients with type 2 diabetes (26) and may be related to this phenomenon.

The reason for the inability of glucose to stimulate insulin release adequately in the GK rat has been associated with defects in glucose metabolism in the pancreatic β-cell (27,28). Deficient expression of pyruvate dehydrogenase (29), increased secretion of islet amyloid polypeptide (30), alterations in lysosomal enzymes (31), lowered cAMP levels (32), and low levels of mitochondrial glycerol phosphate dehydrogenase (33) all have been suggested in this context. Glycerol phosphate dehydrogenase was mapped to rat chromosome 3 in the vicinity of a region linked to diabetes in GK rats (34). However, overexpression of the enzyme in islets from GK rats was not able to alleviate the impaired glucose-induced insulin secretion (35). Although glucose metabolism may be altered in the GK rat, normal Krebs cycle function has been observed in GK islets (36,37), which is compatible with our observation of a glucose-induced lowering of pO₂ in GK and Niddm1i islets. Such an increase in respiration, which is normally coupled to increased insulin release from the islet (22), in combination with the observed secretion impairment, could be explained by decreased coupling between oxidative metabolism and phosphorylation in the β-cells in Niddm1i islets. In support of this view, glucose-induced ATP production was lower in GK islets as compared with control islets (37). A potential candidate for this action is uncoupling protein 2, the expression of which has been demonstrated to affect glucose-stimulated insulin release (38,39,40).

The secretory defect of the Niddm1i congene was further characterized by exposing the islets to tolbutamide. This hypoglycemic sulfonylurea acts via closure of the ATP-sensitive K⁺-channels (41). The stimulatory effect of tolbutamide on Niddm1i islets, although less pronounced than in control F344 islets, indicates that the channels are functional in Niddm1i islets as well as in GK islets, as shown previously (37,42). When the mitochondrial substrate KIC (43) was added to the perifusion medium, insulin release was fully restored in Niddm1i islets. KIC is the deamination product of l-leucine, a reaction catalyzed by branched-chain 2-ketoacid dehydrogenase (BCKDH). The ketoacid is then directly metabolized in the β-cell mitochondria. This finding suggests that the decreased amount of insulin secreted in response to tolbutamide is probably due to insufficient ATP generation (37) for the energy-requiring process of recruiting granules for exocytosis (44,45), rather than decreased insulin content. In GK islets, the leucine deamination product produces a small (46) or, as observed in the present study, almost absent secretory response. This may be related to reduced generation of acetyl-CoA from KIC as a result of defective decarboxylation of this ketoacid by the mitochondrial BCKDH in GK rats (46). However, because Niddm1i islets respond normally to KIC, the genetic defect for the impaired insulin response to KIC in GK islets is unlikely to be related to Niddm1i and, according to syntenic conservation, the BCKDH genes are not located in the Niddm1i region.

In type 2 diabetes, not only are the plasma insulin levels decreased, but also the release kinetics of the hormone are affected. Regular plasma insulin oscillations, which are present in normal individuals (47), are replaced by an irregular plasma insulin pattern (48). The glucose-lowering effect of insulin depends on the release kinetics of the hormone (49,50). The disturbed plasma hormone pattern observed in patients with type 2 diabetes (48) has been connected with the downregulation of insulin receptors (51). It has been speculated that the disturbed plasma insulin pattern may be caused by loss of pulsatile release from the β-cell (26). The present results, that pulsatile insulin release from control F344 islets is comparable with regard to frequency to that observed from GK and the congenic islets, make this speculation less plausible. Instead, less regular plasma insulin oscillations may be caused by the decrease in glucose-induced insulin release manifested as a reduction of the amplitude of the insulin pulses (52).

In summary, genes that influence stimulus secretion coupling have been located to the chromosome region encompassed by the Niddm1i locus. High-resolution genetic mapping of the locus in combination with whole-animal and islet functional studies will provide unique opportunities to locate and study susceptibility genes involved in the cause of impaired insulin secretion in diabetes.

The study was supported by grants from the Swedish Medical Research Council (72X-14019, 9109, and 12202), the Swedish Diabetes Association, the Novo Nordisk Foundation, the Family Ernfors Foundation, the Marcus and Amalia Wallenberg Foundation, the Swedish Foundation for Strategic Research, the Magnus Bergvalls Foundation, the Swedish Society of Medicine, and the Göran Gustafsson Foundation.