Persistent Improvement of Type 2 Diabetes in the Goto-Kakizaki Rat Model by Expansion of the β-Cell Mass During the Prediabetic Period With Glucagon-Like Peptide-1 or Exendin-4

Diabetic GK rats were obtained from our local colony initiated in Paris with progenitors issued from the original colony established by Goto et al. (12). Nondiabetic Wistar rats were used as control animals. All animals were fed ad libitum with a commercial pelleted diet (Diet 113; Usine d’Alimentation Rationnelle, France). Female rats were caged with a male rat from the homologous strain for one night and pregnancy was detected by abdominal palpation 14 days later. Natural birth occurred 22 days after mating.

Four groups of rats were studied: two control groups, control Wistar and control GK, and two experimental groups GK/GLP-1 and GK/Ex-4. In the two experimental groups, GK rats received a daily subcutaneous injection of GLP-1 (400 μg/kg body wt; fragment 7–36 amide;Sigma, France) or Ex-4 (3 μg/kg body wt; Bachem, France) for 5 days (from days 2 to 6) after their birth.

Animals were killed by decapitation 7 days or 2 months after birth. Blood samples were collected after decapitation in 7-day-old rats or from caudal vessels in 2-month-old rats and were immediately centrifuged at 4°C and stored at −20°C until assayed. Basal plasma glucose was measured at 1400 in adult awake rats fasted from 0900 (postabsorptive state).

In each group, 4–15 animals were studied. Pancreases were removed and weighed. For measurement of insulin content, pancreases were homogenized and centrifuged in acid-alcohol solution (75% ethanol, 1.5% HCl 12N, 23.5% distilled water) and the supernatant was stored at −20°C. For immunohistochemical studies, pancreases were fixed in aqueous Bouin’s solution during 24 h and embedded in Paraplast (Labonord, France).

Each pancreatic block was serially sectioned (7 μm) throughout its length to avoid any bias from regional changes in islet distribution and islet cell composition, and the sections were mounted on slides. Then 10 sections were randomly chosen at a fixed interval throughout the block (every 35th section for young rats and every 70th section for adult rats) and were immunostained for insulin using a peroxidase indirect labeling technique, as previously described (14). Sections were incubated for 1 h with primary antibodies (guinea pig anti-porcine insulin; ICN Pharmaceuticals, France). Thereafter, peroxidase-conjugated secondary antibodies were applied for 45 min (rabbit anti-guinea pig IgG; Dako, France). Staining was visualized by incubation with 3,3′-diaminobenzidine-tetra-hydrochloride (DAB) (kit DAB; Valbiotech, France). After staining, sections were mounted in Eukitt (Labonord).

Quantitative evaluation of total β-cell areas was performed using a computer-assisted image analysis procedure based on an Olympus BX 40 microscope (Olympus-France, France) connected via video camera to a personal computer and using Visiolab 1000 software (Biocom, France). The areas of insulin-positive cells, as well as those of total pancreatic sections, were evaluated in each stained section. The relative volume of β-cells was determined by a stereological morphometric method, calculating the ratio between the area occupied by immunoreactive cells and that occupied by total pancreatic cells. Total β-cell mass per pancreas was derived by multiplying this ratio by the total pancreatic weight.

Individual β-cell area was determined on the insulin-stained sections by evaluating the mean cross-sectional area of individual β-cells. The β-cell nuclei on a random section were counted, and the area of β-cell tissue in that section was measured by planimetry, as described above. The individual β-cell size was determined by dividing the β-cell area by the number of nuclei. Using this technique, it must be recognized that the actual number of β-cells is probably higher than the number counted because not all β-cells are sectioned across their nuclei and, therefore, the size of β-cells may be overestimated (14).

For replication studies, four animals in each group were injected 1 h before decapitation with 5′-bromo-2′-deoxyuridine (BrdU) (Sigma) at a dosage of 50 mg/kg i.p. After their pancreases were removed and fixed, as previously described, sections were double-stained for BrdU using a cell proliferation kit (Amersham International, U.K.) and for insulin. Briefly, sections were incubated with a mixture of nuclease and mouse monoclonal anti-BrdU antibody for 1 h at room temperature, washed with Tris 0.05 mol/l (pH 7.6), incubated with a peroxidase anti-mouse IgG for 30 min, and stained with DAB using a peroxidase substrate kit. After BrdU labeling, the tissue sections were washed with Tris 0.05 mol/l (pH 7.6) and then stained for insulin. Insulin staining was done by a guinea pig polyclonal anti-insulin antibody (ICN) for 1 h, as described above. This antibody was labeled with a goat anti-guinea-pig alkaline phosphatase-conjugated antibody (Valbiotech) for 45 min. It was revealed with an alkaline phosphatase substrate kit (Valbiotech). Sections were counterstained with light green and mounted in Eukitt. In these double-stained sections, β-cells exhibited red cytosol and BrdU-positive β-cells exhibited a brown nuclei. To estimate the β-cell replication rate, β-cells and BrdU-positive β-cells were counted using an Olympus BX 40 microscope. Results were expressed as the percentage of BrdU-positive β-cells. For each pancreas, ∼1,000 β-cells were counted.

To obtain an estimate of β-cell neogenesis activation, we quantified both the number of single β-cells incorporated into the duct epithelium and the number of β-cell clusters (2–10 cells in the cluster) on close contact with ducts belonging to three different sections per pancreas (5–6 animals per group). Values for single β-cells and β-cell clusters are related to the total area of the pancreatic section and calculated per micrometer squared of the pancreatic area.

In each rat group, body weight, basal plasma glucose, and insulin levels were measured at 1400 once per week from day 7 (at the end of treatment) to adult age (2 months). Blood samples were taken from a tail incision and assayed for glucose and insulin.

Glucose was dissolved in 0.9% saline and given by the saphenous vein route (0.5 g/kg body wt) to rats under pentobarbital sodium anesthesia (4 mg/100 g body wt i.p.). The test was performed at 1400 in adult rats fasted from 0900 (postabsorptive state). Blood samples were collected sequentially by the tail vessels before and at 5, 10, 15, 20, and 30 min after the injection of glucose.

Blood samples (200 μl) were immediately centrifuged at 4°C and plasmas stored at −20°C until assayed. Insulin and glucose responses during the glucose tolerance tests were calculated as the incremental plasma insulin (ΔI in picomoles per liter) values integrated over the 30-min period after the injection of glucose and the corresponding incremental integrated plasma glucose values (millimoles per liter). The insulinogenic index (ΔI/ΔG) represents the ratio of these two parameters. The rate of glucose disappearance was calculated from the slope of the regression line obtained with the log-transformed plasma glucose values from 10 to 30 min after glucose administration.

Insulin secretion in vitro was investigated in 2-month-old using the isolated perfused pancreas technique. The pancreas was isolated with the proximal portion of the duodenum and separated from the spleen and stomach. The pancreas plus duodenum block was perfused through the celiac and mesenteric arteries via a cannula inserted into the aorta at a constant 37°C temperature without recycling the medium. The perfusate was a Krebs-Ringer bicarbonate buffer (KRBB) with the following components: 2.8 mmol/l d-glucose (Merck, Germany), 118 mmol/l NaCl, 4 mmol/l KCl, 2.5 mmol/l CaCl₂, 1.2 mmol/l MgSO₄, 1.2 mmol/l KH₂PO₄, 25 mmol/l NaHCO₃, 1.25 g/l fatty acid-free BSA (Sigma), and 40 g/l dextran T70 (Pharmacia, Sweden). Pancreases were first perfused during a 20-min equilibration period after the surgery. The effluent of the final 10 min was collected at 1-min intervals for determination of the basal rate of insulin release. The stimulatory effects of glucose and arginine were then investigated sequentially. After a 10-min basal perfusion, the pancreases were perfused for 20 min with KRBB containing 16.7 mmol/l of d-glucose (Merck), followed by KRBB supplemented with 19 mmol/l l-arginine. The arginine perfusion was preceded and followed by a 10-min period of perfusion with KRBB containing 2.8 mmol/l d-glucose. The complete effluent (3 ml/min) was collected from the cannula in the portal vein at 1-min intervals in chilled tubes and frozen for storage at −20°C until assay.

Plasma glucose was determined using a glucose analyzer (Beckman, Palo Alto, CA). Plasma insulin and pancreatic insulin content were determined by radioimmunoassay. Immunoreactive insulin was estimated using purified rat insulin as the standard (Clinisciences, France) and antibody to insulin and rat ¹²⁵I-monoiodinated insulin (Diasorin, Italy) as tracer. Charcoal (Sigma) was used to separate free from bound hormone. This method allows the determination of 0.1 ng/ml with a coefficient of variation within and among assay of 10% (21).

Data are expressed as means ± SE. The significance of differences between mean values was evaluated by one-way ANOVA followed by a Scheffe test.

Comparison of the characteristics between untreated Wistar and GK rats showed that GK newborns exhibited lower weights for body and pancreas and lower basal plasma insulin concentrations than the Wistar newborns at the same age (Table 1). Nevertheless, their basal plasma glucose concentrations were similar (Table 1). These parameters in GK neonates were not significantly modified by the 5 days of GLP-1 or Ex-4 treatment. The pancreatic insulin content was significantly lower in untreated GK neonates as compared to Wistar neonates (Table 1) (P < 0.001), and their total β-cell mass was sharply decreased, at only 26% of Wistar values. After 5 days of GLP-1 or Ex-4 treatment, GK rats exhibited a significantly increased pancreatic insulin content (P < 0.001 and P < 0.01, respectively) as compared with values in untreated GK rats. Moreover, β-cell masses in GK/GLP-1 and GK/Ex-4 rats represented 50 and 46%, respectively, of Wistar values (Table 1).

The individual cross-sectional area of β-cells in 7-day-old GK pancreases was not significantly different from the corresponding value in Wistar rats (103 ± 7 [W] vs. 102 ± 3 [GK] vs. 99 ± 5 [GK/GLP-1] vs. 100 ± 4 μm² [GK/Ex-4]). This indicated that the enhancement of the total β-cell mass found in the treated GK groups could be mostly ascribed to β-cell hyperplasia and not to β-cell hypertrophy. Expansion of β-cell mass can result from an increase in the replication of preexisting β-cells and/or of the generation of new β-cells by neogenesis. To determine which one of these mechanisms was stimulated by GLP-1 or Ex-4 treatment, we investigated the BrdU-labeling index rate and neogenesis rate at the end of GLP-1 or Ex-4 treatment. The BrdU-labeling index of the β-cells was determined by double immunocytochemical staining for insulin and BrdU. No statistically significant difference was observed when comparing BrdU β-cell labeling indexes in Wistar and untreated GK groups (Fig. 1). In both GK/GLP-1 and GK/Ex-4 groups, the index was higher than in the untreated GK group (P < 0.001 and P < 0.01, respectively) and Wistar group (P < 0.001 and P < 0.05, respectively) (Fig. 1 and Fig. 2A and B). To estimate activation of neogenesis, the number of single β-cells and β-cell clusters per micrometer squared of pancreatic tissue was quantified. The number of isolated β-cells or β-cells buds per micrometer squared of pancreatic tissue was the same in the Wistar and untreated GK groups. These two parameters were strongly increased in the GK/GLP-1 and GK/Ex-4 groups compared to the untreated GK group (P < 0.001 for isolated β-cells and P < 0.01 for β-cell buds) (Fig. 1 and Fig. 2C and D).

Given the above-reported β-cell growth data, we examined whether or not treatment by GLP-1 or Ex-4 during the prediabetic period exerts long-lasting beneficial effects on GK biological characteristics. Follow-up from day 7 to adult age (2 months) showed that GLP-1 or Ex-4 treatment did exert a long-term positive influence on glycemic control (Fig. 3). Basal hyperglycemia is usually detected in untreated GK pups at ages 3–4 weeks, corresponding to the weaning period (age 31 days). After the neonatal GLP-1 and Ex-4 treatment, the increase in basal plasma glucose at 3–4 weeks was clearly blunted in the two treated groups as compared to the untreated GK group (Fig. 3). Moreover, the reduction of basal hyperglycemia in the GK/GLP-1 and GK/Ex-4 groups persisted at least until age 2 months. From age 7 days to adulthood (2 months), no change was recorded when comparing longitudinal variations of body weight and basal insulinemia in untreated GK and GK/GLP-1 or GK/Ex-4 rats (Fig. 3).

The 2-month-old GK females, whether untreated or treated with GLP-1 or Ex-4, exhibited a significantly lower body weight compared to age-related Wistar females. When examined in the postabsorptive state, basal plasma glucose levels in GK/GLP-1 and GK/Ex-4 rats were found to be decreased compared to those in untreated GK rats, but they still remained slightly higher compared with Wistar rats (Table 2). There was no significant difference in basal plasma insulin levels between Wistar and untreated or treated GK rats, except that the GK/GLP-1 group exhibited a higher insulinemia compared with Wistar or untreated GK groups (P < 0.05) (Table 2). Pancreatic insulin content in GK/GLP-1 and GK/Ex-4 female rats was increased compared to that in untreated GK rats. However, the pancreatic insulin content in the two treated GK groups remained still significantly lower than that in the Wistar group (P < 0.001) (Table 2). In both GK/GLP-1 and GK/Ex-4 groups, β-cell mass was significantly increased (P < 0.001 and P < 0.05, respectively) compared with in the untreated GK group (Table 2), and represented 71 and 63%, respectively, of the β-cell mass in the Wistar group, whereas β-cell mass in the untreated GK group represented only 40% of the Wistar value. The increase of total β-cell mass in the two treated GK groups was related to β-cell hyperplasia and not to β-cell hypertrophy, because of the similar individual cross-sectional areas of the β-cells among groups (198.1 ± 2.1 [W] vs. 196.2 ± 8.4 [GK] vs. 201 ± 10.2 [GK/GLP-1] vs. 202 ± 9.7 μm² [GK/Ex-4]).

At age 2 months, in vivo glucose-induced insulin secretion in GK/GLP-1 rats (as attested by the ΔI and ΔI/ΔG) was significantly increased (P < 0.05) during the intravenous glucose tolerance test compared to that in untreated GK rats (Fig. 4). However, it remained lower than in the Wistar group (P < 0.001). Tolerance to glucose was improved in the GK/GLP-1 and GK/Ex-4 rats, as attested by the significantly decreased ΔG value (P < 0.05) compared to that in the untreated GK group (Fig. 4). However, it did not return to a normal value, as the difference remained significant (P < 0.001) compared with the ΔG value in the Wistar group.

The in vitro insulin release in response to glucose and arginine was studied with the isolated perfused pancreas technique. Basal insulin secretion with 2.8 mmol/l glucose in the basal perfusion medium was found to be similar in the untreated GK compared with Wistar rats (Fig. 5, top panel). Only a small acute phase (5 min) could be distinguished after exposure for 20 min to a 16.7-mmol/l concentration that elicited a typical biphasic pattern of insulin release in the Wistar rats (Fig. 5, top panel). Basal insulin release in the adult GK rats that previously received the neonatal treatment by GLP-1 or Ex-4 was similar to that in Wistar and untreated GK rats (Fig. 5, top panel). During exposure to 16.7 mmol/l glucose, a slight but significant early increase was detectable in the GK/GLP-1 group (P < 0.05 vs. untreated GK rats); however, the overall insulin response in the GK/GLP-1 group remained defective (P < 0.001) compared to the normal Wistar response (Fig. 5, Table 3). In GK/Ex-4 rats, the slight improvement of the acute phase of glucose-induced insulin did not reach significance. The insulin response to 19 mmol/l arginine was similar in treated GK rats to that in untreated GK and Wistar rats (Fig. 5, bottom panel; Table 3).

The possibility of using GLP-1 as a treatment for type 2 diabetes first arose from observations of its insulinotropic activity. GLP-1, or its long acting analog Ex-4, has been shown to acutely reduce plasma glucose levels by increasing insulin release and synthesis, inhibiting glucagon release, and decreasing gastric emptying and appetite (3–5). Long-term beneficial effects have also been shown in human and rodent models (3–5). More recent data have outlined additional biological effects associated with pancreatic cell proliferation and differentiation that are potentially beneficial in the context of long-term treatment (7–11,5). These last observations support the view that a pharmacological intervention using GLP-1 in diabetic rodents with a primary β-cell mass deficiency could represent a new approach to stimulating compensatory expansion of β-cells. The optimal design for such an intervention would be in the prediabetic period, to prevent or at least delay hyperglycemia onset.

The GK rat model of type 2 diabetes is especially convenient for studying potential therapeutic effect of molecules that block and/or delay the onset of overt type 2 diabetes, as all adult GK rats have a stable basal mild hyperglycemia and the installation of overt diabetes is preceded by a period of normoglycemia, between birth and weaning. In adult GK rats, total pancreatic β-cell mass is decreased in the range of the decrease in pancreatic insulin stores (14,16) and plasma insulin secretion in vivo in response to glucose is abolished (13,15). With regard to β-cell population, the earliest alteration so far detected in our colony of GK rats targets the size of the β-cell population. In GK neonates, total β-cell mass is clearly decreased compared to in Wistar rats (14,16). Knowing this, it was tempting to investigate in the GK model the impact of GLP-1 or Ex-4 therapy applied during the first postnatal week (i.e., during the prediabetic period) on the time course of type 2 diabetes development, with a special focus on β-cell mass enlargement, improvement of the β-cell function, and improvement of glucose homeostasis.

In our study, administration of GLP-1 or Ex-4 during the first 5 postnatal days did not modify body weight or basal plasma glucose or insulin levels in GK rats when tested at day 7. However, GLP-1 or Ex-4 treatment strongly improved pancreatic insulin stores and β-cell mass compared to untreated GK neonates. After transient treatment with these peptides, there was a doubling of β-cell mass in only 5 days. During the first postnatal week, β-cell growth was high because of sustained proliferation and neogenesis mechanisms (17,18). More specifically, in normoglycemic neonatal rats, total β-cell mass doubled during this period (19). Similar impressive enhancements of the β-cell have been previously reported by our group in another diabetic model of β-cell regeneration, the n0-STZ model, in response to neonatal treatment with insulin (20) or GLP-1 (21): both treatments almost doubled β-cell mass (mainly by increasing the β-cell number) in only 5 days.

Expansion of β-cell mass can result from an increase in cell size (hypertrophy) and/or an increase in number (hyperplasia), the latter of which is achieved through replication of preexisting β-cells and/or neogenesis. Neither GLP-1 or Ex-4 treatments resulted in hypertrophied β-cells, but both increased the β-cell number. Indeed, the increased β-cell mass after GLP-1 or Ex-4 treatment resulted from enhancement of both differentiation (neogenesis) and proliferation of β-cells. These results are in agreement with two other reports, one in the insulinoma cell line INS-1 (7) showing that GLP-1 induces proliferation and upregulates pancreatic duodenal homeobox 1 activity, a transcription factor known to be implicated in neogenesis, and the other in pancreatectomized adult rats (9) showing that Ex-4 stimulates β-cell replication and neogenesis, with a resulting increase in β-cell mass.

Because GLP-1 and Ex-4 were found to exert positive effects on β-cell expansion in the GK model, the second aim of this study was to investigate the effect of these peptides administered during the prediabetic period on the duration of the β-cell mass enlargement and the improvement of glucose homeostasis. Follow-up of the GK biological characteristics from day 7 to adult age (2 months) showed that both GLP-1 and Ex-4 treatments exerted a sustained favorable influence on basal glycemia during this period. In this spontaneous model of type 2 diabetes, the onset of the diabetic phenotype is around 3–4 weeks after birth, corresponding to the weaning period. Our GLP-1 and Ex-4 administration protocols obviously delayed the installation of basal hyperglycemia and limited its severity. In 2-month-old GK/GLP-1 or GK/Ex-4 rats, basal plasma glucose levels were significantly decreased, although not completely normalized as compared to values in Wistar rats. Their pancreatic insulin stores and β-cell masses were still significantly increased.

Persistence of increased β-cell mass in rats age 2 months can be related to the following points. 1) In growing GK rats, treated or not, β-cell mass continues to increase from day 7 to adulthood (16). Thus we can imagine that, because β-cell mass is more important in 7-day-old treated GK rats as compared with untreated GK rats, growth of that increased β-cell mass leads to an enhanced β-cell mass at adult age in GK/GLP-1 and GK/Ex-4 groups compared with untreated GK group. 2) Follow-up of treated GK rats showed that basal plasma glucose was improved compared with untreated GK rats (Fig. 3). The delay of hyperglycemia appearance in the GLP-1— or Ex-4—treated GK rats may be viewed as instrumental in maintaining enhanced β-cell mass, resulting in an increased β-cell mass 7 weeks after the end of GLP-1 or Ex-4 treatment, in accordance with the concept of glucose toxicity targeted to the β-cell mass (22). Such a concept is especially relevant to the situation in the GK model for two reasons: first, when adult GK rats are fed with a carbohydrate-rich diet for 6 weeks, hyperglycemia worsens and is accompanied by a further 50% reduction of β-cell mass compared with GK rats fed a normal diet (23), and second, treatment with an α-glucosidase inhibitor (voglibose) limits the reduction of β-cell mass in GK rats via a reduction of their basal hyperglycemia (24).

Glucose-stimulated insulin release, as evaluated in vivo during the intravenous glucose tolerance test, was significantly improved in both GK/GLP-1 and GK/Ex-4 rats. This improvement is correlated with a modest but significant early enhancement of the insulin release in vitro (isolated perfused pancreas) in response to glucose in the GK/GLP-1 group as compared to the untreated GK group. Therefore, our data suggest that the long-lasting glucose tolerance improvement after neonatal GLP-1 or Ex-4 treatment could be attributable, at least in part, to a slight enhancement of the functionally active pancreatic β-cell mass.

Whether GLP-1 or Ex-4 therapy improves the size of the functionally active β-cell population through a direct or indirect mechanism in GK rats is not clear. Several alternative but not exclusive explanations can be considered. First, because treated and untreated GK neonates had the same basal plasma glucose levels, we can eliminate the possibility that GLP-1 or Ex-4 exerted their effect via a reduction of glucotoxicity. Second, as both peptides strongly increase insulin secretion in rats (3–5), we can imagine that at least part of the effect of GLP-1 and Ex-4 on β-cell mass enlargement is mediated through an increased circulating insulin level acting as a growth factor. In support of this last contention, our group has reported an enhanced growth of β-cell mass in a diabetic model of β-cell regeneration after insulin treatment (20). In the present study, GK/GLP-1 rats exhibited an increased basal plasma insulin level by adult age. We cannot exclude the possibility that the transient increase of plasma insulin after GLP-1 injection acted as a trophic factor during the treatment period. Third, we cannot exclude a direct role of the two peptides on the enlargement of β-cell mass during the treatment. Indeed, GLP-1 and Ex-4 may exert their trophic action on β-cell mass by a direct effect via the β-cell GLP-1 receptor (25,26). The reported abnormal glucose tolerance and islet topography in the GLP-1 receptor null mouse are consistent with such a direct growth factor effect of the two peptides on the β-cell (27–29). Further support for β-cell mass enlargement being a direct effect of the two peptides comes from the finding that GLP-1−receptor mRNA and proteins are expressed in pancreatic ductal cells, which are a site for β-cell neogenesis (9), a pathway that is strongly increased in our study.

In the face of the improvements in basal plasma glucose and tolerance to glucose, and besides the glucose-stimulated insulin secretion, we cannot exclude a long-term effect of the peptide treatment on peripheral insulin action. However, because the peripheral effects of GLP-1 have been observed under only acute conditions (30), and because of the very short biological activity of the peptide, it is very unlikely that any improvement in insulin-dependent glucose utilization by GLP-1—treated GK rats could be attributable to a direct GLP-1 effect. Instead, we suggest the possibility that the long-term improvement in basal hyperglycemia exerts a favorable influence on peripheral and liver insulin action for the following reasons: 1) various experimental designs, with both in vivo and in vitro techniques, have demonstrated that chronic hyperglycemia per se is capable of inducing a state of insulin resistance (31); 2) several groups have reported the beneficial effect of a near-normalization of plasma glucose on peripheral insulin action in different animal models of type 2 diabetes (31,32); and 3) correction of hyperglycemia with phlorizin has been shown to be sufficient to normalize insulin action on glucose metabolism by the peripheral tissues in the GK model (33).

Therefore, these encouraging findings in the GK model of type 2 diabetes indicate for the first time that GLP-1 or Ex-4 treatment limited to the prediabetic period (first week after birth) delays the installation of spontaneous type 2 diabetes and limits its severity. Under these conditions, GLP-1 represents a unique therapy that can stimulate β-cell growth and/or differentiation in spontaneously diabetic rodents and may prove to be an invaluable agent for the prevention of human type 2 diabetes.

This work was partly supported by grants from MERCK-LIPHA laboratories.