Specific and Combined Effects of Insulin and Glucose on Functional ...

0013-7227/03/$15.00/0 Printed in U.S.A.

Endocrinology 144(6):2717–2727 Copyright © 2003 by The Endocrine Society doi: 10.1210/en.2002-221112

Specific and Combined Effects of Insulin and Glucose on Functional Pancreatic ␤-Cell Mass in Vivo in Adult Rats MARYLINE PARIS, CATHERINE BERNARD-KARGAR, MARIE-FRANCE BERTHAULT, LUC BOUWENS, AND ALAIN KTORZA Laboratoire de Physiopathologie de la Nutrition (M.P., C.B.-K., M.-F.B., A.K.), Centre National de la Recherche Scientifique, Unite´ Mixte de Recherche, 7059, Universite´ Paris 7, 75251 Paris, France; and Cell Differentiation Group (L.B.), Diabetes Research Center, Free University of Brussels (VUB), Brussels 1050, Belgium We investigated the specific and associated effects of insulin and glucose on ␤-cell growth and function in adult rats. By combining simultaneous infusion either of glucose and/or insulin or glucose and diazoxide, three groups of rats were constituted: hyperglycemic-hyperinsulinemic rats (high glucose– high insulin), hyperglycemic-euinsulinemic rats (high glucose), and euglycemic-hyperinsulinemic rats (high insulin). All the infusions lasted 48 h. Control rats were infused with 0.9% NaCl (saline controls). In all groups, ␤-cell mass was significantly increased, compared with controls (by 70% in high glucose– high insulin rats, 65% in high glucose rats, and 50% in high insulin rats). The stimulation of neogenesis was suggested by the high number of islets budding from pancreatic ducts in high glucose– high insulin and high glucose rats and by the presence of numerous clusters of few ␤-cells within

F

OR A LONG TIME, it was believed that the endocrine pancreas belonged to a category of tissues that were finally differentiated and irreplaceable in the adult. This was mainly supported by the low replication rate of endocrine cells in adulthood (1). In the light of many recent data, this point of view has been drastically changed, and nobody disputes today that the endocrine pancreas is a plastic organ especially because of the high ability of the ␤-cell mass to change according to the insulin demand (review in Ref. 2). This property has been demonstrated in physiological as well pathophysiological conditions such as pregnancy (3) and obesity (4). Among the numerous substrates, hormones, and growth factors involved in endocrine pancreas plasticity and ␤-cell renewal, the role of glucose and insulin emerge and has been extensively studied (reviewed in Ref. 5). In several species including humans (6), glucose appears to play a key regulatory role in pancreatic plasticity because it is a potent stimulus of pancreatic ␤-cell growth both in vivo (7, 8) and in vitro (9, 10). The effect of insulin on ␤-cell growth in vivo is more controversial. ␤-Cell proliferation is stimulated by insulin treatment in fetal islets transplanted to diabetic rats (11, 12). Moreover, insulin therapy improves ␤-cell regeneration in newborn rats injected with streptozotocin on the day of birth (13) or adult mice with streptozotocin-induced experimental diabetes (14). On the contrary, other studies showed that high glucose rather than high insulin levels were crucial for islet growth in transplanted diabetic mice (15). Finally, a study of Abbreviation: BrdU, 5-Bromo-2⬘-deoxyuridine.

the exocrine pancreas in high insulin rats. ␤-Cell hypertrophy was observed only in high glucose– high insulin rats. The rate of ␤-cell proliferation was similar to that of controls in high glucose– high insulin rats after a 48-h glucose infusion, dropped dramatically in high insulin rats, and dropped to a lesser extent in high glucose rats. In high glucose– high insulin and high glucose rats, ␤-cell mass increase was related to a higher ␤-cell responsiveness to glucose in vitro as measured by islet perifusion studies, whereas in high insulin rats, no significant enhancement of glucose induced insulin secretion could be noticed. The data show that glucose and insulin may have specific stimulating effects on ␤-cell growth and function in vivo in adult rats independently of the influence they exert each other on their respective plasma concentration. (Endocrinology 144: 2717–2727, 2003)

Koiter et al. (16) stressed the interplay between glucose and insulin for the control of islet cell proliferation in vivo. Elucidating the precise role of insulin and interplay between insulin and glucose is of particular importance because a series of recent studies showed the tight dependence of ␤-cell mass homeostasis and function on insulin receptor and insulin receptor substrates insulin receptor substrate-1 and -2 (17–19). The purpose of our study was to appreciate the respective role in vivo of glucose and insulin on short-term ␤-cell mass changes in rats and some of the mechanisms underlying these changes. We also addressed the question of the functional consequences of these changes by measuring islet responsiveness to glucose. In this way, by combining under various experimental conditions simultaneous infusion during 48 h either of glucose and/or insulin or glucose and diazoxide, a potent inhibitor of insulin secretion, three groups of rats were constituted: hyperglycemic-hyperinsulinemic rats (high glucose– high insulin), euglycemic-hyperinsulinemic rats (high insulin), and hyperglycemic-euinsulinemic rats (high glucose). ␤-Cell mass, the main parameters of ␤-cell mass homeostasis, and insulin release in vitro (perifusion procedure) were further investigated. Materials and Methods Animals Three-month-old male Wistar rats weighing 280 –300 g were used. They had free access to water and standard laboratory diet pellets (no. 113, UAR, Villemoisson-sur-Orge, France).

2717

2718

Endocrinology, June 2003, 144(6):2717–2727

Paris et al. • Insulin, Glucose, and ␤-Cell Growth in Vivo

Infusions

␤-Cell replication

Rats were randomly divided into five groups as follows: 1) 0.9% NaCl-infused rats (saline controls rats), 2) glucose-infused rats (hyperglycemic-hyperinsulinemic: high glucose– high insulin rats), 3) insulin ⫹ glucose-infused rats (euglycemic-hyperinsulinemic; high insulin rats), 4) diazoxide-infused rats (diazoxide controls rats), and 5) glucose ⫹ diazoxide-infused rats (hyperglycemic-euinsulinemic: high glucose rats). All infusions lasted 48 h. The long-term infusion technique in unrestrained rats was used, as previously described (20). Briefly, 2 d before infusion, rats were fitted with an indwelling jugular vein catheter, and during the infusion period, rats were permanently connected to a pump via a device fitted with a water-tight swivel. In high glucose– high insulin rats, hypertonic (30% wt/vol) glucose (Chaix & Du Marais, Paris, France) was infused at an initial rate of 50 ␮l/min to produce hyperglycemia around 22 mm throughout the infusion period. In the high insulin group, euglycemia and hyperinsulinemia were induced using insulin infusion at a rate of 30 ␮U/min (Novo Nordisk, Bagsvaerd, Denmark) to produce a hyperinsulinemia around 15-fold higher than that of high glucose– high insulin rats. Simultaneous glucose infusion allowed maintaining euglycemia (around 5 mmol/ liter). In high glucose rats, hyperglycemia and euinsulinemia were induced by a simultaneous infusion of diazoxide (Sigma, St. Louis, MO) and hypertonic (30% wt/vol) glucose, which was infused at the same flow rate as in high glucose– high insulin group. Diazoxide solution (added to a bicarbonate-phosphate buffer, pH 9.5) was infused in the diazoxide controls group at a flow rate of 5 mg/kg⫺1䡠h⫺1. During infusion periods, plasma glucose and insulin concentrations were measured on arteriovenous blood collected from tail vessels by tail snipping five times daily in high glucose– high insulin, high glucose, and high insulin rats. This daily control allowed glycemia and insulinemia to be maintained in the required ranges by adjusting infusion flow rates. In saline controls and diazoxide controls groups, because glycemia and insulinemia remained rather stable, blood was collected only twice daily. Rats in which glycemia and insulinemia did not stay within these ranges were discarded. At the end of the 48-h infusion, pancreases were removed and fixed for morphometric investigations.

Sections that had not already been used for morphometric studies were used to measure ␤-cell replication rates. ␤-Cell replication was evaluated using the measurement of the 5-bromo-2⬘-deoxyuridine (BrdU) incorporation. BrdU was injected at a dose of 100 mg/kg ip 6 h before the rats were killed. A 6-h BrdU incorporation interval was chosen to avoid the possibility of including daughter cells (22). Sections were double stained for BrdU using a cell proliferation kit (Amersham International, Amersham, UK) and for insulin. For BrdU detection, sections were incubated with a monoclonal antibody anti-BrdU diluted in a nuclease solution (according to the kit protocol) for 1 h at room temperature and washed with Tris (0.05 m), pH 7.6. Thereafter they were incubated with a peroxidase antimouse IgG and stained with diaminobenzidine using a peroxidase substrate kit. For insulin detection, sections were then incubated with guinea pig antiinsulin antibody for 1 h as described above and then with an alkaline phosphatase-conjugated goat antiguinea pig IgG for 45 min (final dilution 1:150, Sera Lab, Carpinteria, CA). The alkaline phosphatase activity was revealed with an alkaline phosphatase substrate kit (Biosys-Vector, Compie`gne, France). Sections were then counterstained with hematoxylin and mounted in Eukitt. On these sections, ␤-cells showed red cytosol and BrdU⫹ cells appeared with brown nuclei. A minimum of 1100 ␤-cells nuclei was counted per section at a final magnification of ⫻1000. The proportion of BrdU⫹ ␤-cell nuclei was calculated. Results were expressed as the percentage of replication of ␤-cells in a 6-h interval.

Immunochemistry and morphometry

Evaluation of ␤-cell apoptosis

Pancreases were excised and weighed after their fat and lymph nodes had been removed. For each rat the splenic part of the pancreas was fixed in aqueous Bouin’s solution and embedded in paraffin. Each pancreatic block was then serially sectioned (7 ␮m) throughout its length to avoid any bias because of changes in islet distribution or cell composition and then mounted on slides. For each pancreas, seven sections were randomly chosen at a fixed interval through the block (every 35th section). This procedure has been shown to ensure that the selected sections are representative of the whole pancreas (8, 13). Sections were immunostained for insulin, cytokeratin-20 (CK-20), and Glut-2 as previously described (8, 24, 21). Quantitative evaluation of the ␤-cell mass was performed using a BH2 microscope (Olympus Corp., Melville, NY) connected via a color video camera to a Compaq PC computer and using the imagenia 2000 software (Biocom, Les Ulis, France). Area of insulin-positive cells, as well as that of total pancreatic sections, was evaluated in each stained section. ␤-Cell area was then determined by calculating the ratio between the area occupied by immunoreactive ␤-cells and that occupied by total pancreatic cells according to stereological methods. Finally, total ␤-cell mass per pancreas was derived by multiplying this ratio by the total pancreatic weight.

Sections from the same blocks used for ␤-cell mass measurement and replication were used to study ␤-cell apoptosis rates. Apoptotic cells were detected with the ApopTag in situ apoptosis detection kit (Applige`ne-Oncor, Illkirch, France) as previously described (24). Because apoptosis is a rapid process with less than 1 h of morphological evidence (25), measurement of apoptosis ␤-cell rate with a direct insulin staining could lead to an underestimation of this rate. Therefore, we used the staining of non–␤-cells to surround the core of the islet and identify the ␤-cells.

Individual ␤-cell area ␤-Cell size was measured on insulin-stained sections by evaluating the mean cross-sectional area of individual ␤-cells. The ␤-cell nuclei on a random section were counted, and the ␤-cell area in that section was measured by planimetry as described above. The ␤-cell area was divided by the number of nuclei to calculate the area of individual ␤-cells. Using this technique, it must be recognized that the actual number of ␤-cells is probably higher than the number counted because not all ␤-cells are across their nuclei; therefore, the size of ␤-cells is overestimated.

Evaluation of ␤-cell neogenesis To evaluate ␤-cell neogenesis, two parameters were used: 1) the number of ␤-cell clusters budding from the ducts (from 2–15 ␤-cells in the cluster) and the number of single ␤-cells incorporated into the duct epithelium and 2) the number of small islets (from 2 to 15 ␤-cells) (23) in each section. Quantification was performed on sections used for ␤-cell mass measurements (seven different sections per pancreas, five to six animals per group). Results were expressed as the number of single ␤-cells and ␤-cell clusters inside or budding from the ducts and isolated islets.

Insulin release under perifusion experiments Kinetics of insulin release in vitro was studied using the perifusion procedure. Each perifusion apparatus consisted of a small glass column with volume-reducing adaptors at both ends and contained a Bio-Gel (P-2, 200 – 400 wet mesh, Bio-Rad Laboratories, Inc., Richmond, CA). Bio-Gel beads were preswollen with the perifusion medium at 4 C overnight. On the day of the experiment, Bio-Gel beads were equilibrated with fresh perifusion medium by connecting the columns to the perifusion circuit for 30 min. Four columns were run at the same time. One hundred isolated freshly isolated islets were then carefully placed on the top of the Bio-Gel of each column and covered with an approximate volume of 100 ␮l Bio-Gel beads. Then the columns were gently closed with the top adaptors, immersed in vertical position in the water bath at 37 C. The perifusion medium was maintained at 37 C in a water bath and constantly gassed throughout the period of perifusion. The medium containing the basal buffer (2.8 mm glucose) was supplemented with glucose (16.7 mm glucose). The perifusion was then started and the medium was pumped to the columns at a final rate of 1ml/min using a peristaltic pump (Gilson). The perifusion fluid was collected in graduated centrifuge tubes at 1-min intervals using a fraction collector (Retriever II, Isco, Lincoln, NE). They were kept at low temperature until

5.80 ⫾ 0.39 5.63 ⫾ 0.35 18.4 ⫾ 0.57a

Data are means ⫾ SE, n ⫽ 8 in each group. Saline control, Rats infused with 0.9% NaCl; High glucose-high insulin, rats infused with 30% glucose; high insulin, rats infused with glucose and insulin; diazoxide control, rats infused with diazoxide; high glucose, rats infused with 30% glucose and diazoxide. a P ⬍ 0.01% compared to saline control.

16.10 ⫾ 2.60a 0.36 ⫾ 0.05 0.30 ⫾ 0.05 19.3 ⫾ 2.90a 0.31 ⫾ 0.03 0.42 ⫾ 0.04 22.50 ⫾ 1.90a 0.32 ⫾ 0.03 0.38 ⫾ 0.05 18.17 ⫾ 1.81a 0.29 ⫾ 0.02 0.48 ⫾ 0.04 0.52 ⫾ 0.01 0.34 ⫾ 0.06 0.30 ⫾ 0.03 3.39 ⫾ 0.20 6.20 ⫾ 0.50 17.20 ⫾ 1.60a 5.84 ⫾ 0.49 5.90 ⫾ 0.09 22.5 ⫾ 1.10a

48 h

5.12 ⫾ 0.19 5.80 ⫾ 0.20 5.20 ⫾ 0.26

5.69 ⫾ 0.44 5.70 ⫾ 0.60 22.60 ⫾ 1.60a

0.31 ⫾ 0.02 1.15 ⫾ 0.01

30 h 24 h

0.29 ⫾ 0.03 1.25 ⫾ 0.10 0.30 ⫾ 0.07 1.36 ⫾ 0.12 0.29 ⫾ 0.05 0.32 ⫾ 0.03

Plasma insulin (nmol/liter)

6h 0h 48 h 30 h 24 h

Plasma glucose (mmol/liter)

5.50 ⫾ 0.13 19.3 ⫾ 1.05a

Islets from high glucose– high insulin rats were more numerous and larger than those of saline controls rats. Moreover, these islets were often very large and many of them showed a multilobular aspect (Fig. 1B). In high glucose rats, although the general islet morphology was similar to that of high glucose– high insulin rats. Some important differences must be noticed: in high glucose the sections were characterized in this group by the presence of a larger number of clusters of ␤-cells, forming small islets (see Fig. 5B), and islet staining pattern for insulin was much higher in high glucose than in high glucose-high insulin rats, thus suggesting higher insulin contents (Fig. 1C).

5.70 ⫾ 0.25 5.90 ⫾ 0.10

Morphological changes

6h

All the data are summarized in Table 1. In the saline controls group, plasma glucose and plasma insulin remained stable during the 48-h infusion. In high glucose– high insulin rats, glucose infusion led to a rapid increase in glycemia, which was maximal at 6 h and stabilized at 20 –22 mmol/liter until the end of infusion. As a result, insulinemia increased rapidly reaching a mean value around 1 nmol/liter. In high glucose rats, plasma glucose levels were as expected very similar to that of high glucose– high insulin group, whereas plasma insulin concentrations remained close to basal values throughout infusions. Diazoxide infusion did not influence plasma parameters because no differences in both glycemia and insulinemia were observed in the diazoxide controls group, compared with saline control rats. In the high insulin group, plasma glucose was close to the values of saline controls rats. It dropped to around 3.5 mm between 30 h and 48 h of infusion, but the difference with controls did not reach statistical significance. Plasma insulin level was 15 times higher than high glucose– high insulin rats.

0h

Results Plasma glucose and insulin concentrations during infusions

Groups

Data are presented as means ⫾ se. Statistical significance was determined with the ANOVA test. P ⬍ 0.05 was considered significant.

TABLE 1. Time course of plasma glucose and insulin concentrations during infusions of 48 h in the different groups of rats

Data presentation and statistical methods

5.60 ⫾ 0.20 19.80 ⫾ 1.00a

Blood glucose was determined by a glucose analyzer (Glucotrend, Boehringer, Manheim, Germany). Plasma insulin was measured by RIA kits (DiaSorin, Inc., Rome, Italy). Pancreatic insulin content was determined using the same kit as above. For measurement of insulin content, one hundred freshly isolated islet were homogenized in 2 ml distilled water, centrifuged 15 min at 4 C, and the supernatant was stored at ⫺20 C. Islet insulin content was determined using an insulin-CT kit (CIS-Bio International), and experiments were expressed as nanograms per nanograms DNA⫺1䡠min⫺1. The islet insulin secretion rate during the perifusion experiments was expressed as picograms per nanograms DNA⫺1䡠min⫺1. DNA content per islet was determined according to a fluorometric method with bisbenzimidazole as fluorochrome (26).

5.70 ⫾ 0.09 22.8 ⫾ 1.10a

Analytical methods

5.70 ⫾ 0.13 21.90 ⫾ 1.30a

the end of the experiment, and they were stored at ⫺20 C until assayed for insulin. Insulin in perifusion fluid was determined by RIA, using an insulin-CT kit (Cis Bio Internationnal, Gif-sur-Yvette, France).

0.17 ⫾ 0.07 1.12 ⫾ 0.09

Endocrinology, June 2003, 144(6):2717–2727 2719

Saline control High glucose-high insulin High insulin Diazoxide control High glucose


2720



FIG. 1. Islet morphology in control (A), high glucose– high insulin (B), and high glucose (C) rats. D, A small cluster of ␤-cells, a typical feature of islet morphology in high insulin rats. ␤-Cells were immunostained for insulin and revealed with peroxidase method. Final magnification, ⫻200. n ⫽ 4 rats for saline controls and high glucose– high insulin; n ⫽ 5 for high insulin and diazoxide controls; n ⫽ 8 for high glucose.

In high insulin rats, except for a few islets displaying normal aspect and size, a high number of very dark small islets containing ␤-cells distributed in the whole pancreatic section was observed (Fig. 1D). ␤-Cell mass and ␤-cell size

The ␤-cell mass of high glucose– high insulin rats was increased by 70%, compared with that of saline controls groups (Fig. 2A). There was a 30% increase of individual ␤-cell size indicating ␤-cell hypertrophy (Fig. 2B). In the high glucose group, the ␤-cell mass increased by 65%, compared with saline controls group, therefore reaching values similar to that of high glucose– high insulin rats (Fig. 2A). No modification of the individual ␤-cell size was observed in this group (Fig. 2B). In high insulin rats, the increase of the ␤-cell mass was about 50%, compared with the saline controls group (Fig. 2A). The individual ␤-cell size was similar to that of the saline controls group (Fig. 2B).

␤-Cell neogenesis and replication

In high glucose– high insulin rats, neogenesis was evidenced mainly by the presence of many duct-associated ␤-cell buds whose number was increased by 400%, compared with that of saline controls rats (Figs. 3A and 4E), whereas there was no significant increase in the number of isolated ␤-cells, compared with saline controls rats (Fig. 3B and 4, A and D). Moreover, the majority of duct-associated ␤-cell buds were stained for cytokeratin 20 and insulin (Fig. 5A). In many ducts Glut-2 ⫹ insulin-positive cells were observed (Fig. 5B). After the 48-h infusion period, compared with the saline controls group (Fig. 6A), ␤-cell replication (Fig. 7C) was unchanged. Shorter periods of infusion (8, 16, and 24 h) resulted in a sharp decrease in ␤-cell replication (Fig. 6B). The same feature could be observed in high glucose rats, i.e. a large amount of duct-associated ␤-cell buds with no significant increase in the number of isolated small islets, compared with saline control rats (Figs. 3A and 4, B, D, and F). ␤-Cell proliferation was much lower than in controls (Fig. 6). In high insulin rats, the presence of many isolated small


FIG. 2. Total ␤-cell mass and ␤-cell area after 48 h of NaCl (saline controls), glucose (high glucose– high insulin), glucose plus insulin (high insulin), diazoxide (diazoxide controls), and glucose plus diazoxide (high glucose) infusions. A, Total ␤-cell mass (milligrams per pancreas). n ⫽ 4 rats for saline controls and high glucose– high insulin; n ⫽ 5 for high insulin and diazoxide controls; n ⫽ 8 for high glucose. B, Individual ␤-cell area (micrometers squared per ␤-cell). n ⫽ 4 rats for saline controls; n ⫽ 3 for high glucose– high insulin; n ⫽ 5 for high insulin and diazoxide controls; n ⫽ 5 for high glucose. Data are means ⫾ SE. *, P ⬍ 0.05.

islets (60% increase, compared with saline control rats) argued for the activation of the neogenic process (Figs. 3B and 4C). The number of duct-associated ␤-cell buds were not significantly modified, compared with control rats (Fig. 3A). ␤-Cell replication was dramatically decreased because it dropped to 90% of the values in saline control rats (Fig. 6). In the diazoxide control group, ␤-cell mass, individual ␤-cell size, and the parameters of neogenesis were similar to that of controls (Figs. 2 and 3). However, ␤-cell replication dropped drastically, compared with the saline control group (Fig. 6). ␤-Cell apoptosis

Apoptotic cells were detectable with the DNA breakage labeling method in the pancreases of each group infused rats. Figure 7 (A and B) represents an illustrative fragment of histological staining for apoptotic cells located within the islets. Whatever the experimental group, the number of apoptotic ␤-cells was comprised between 0 and 2.


FIG. 3. ␤-Cell neogenesis activation in rats after infusion of 48 h of NaCl (saline controls), glucose (high glucose– high insulin), glucose plus insulin (high insulin), diazoxide (diazoxide controls), and glucose plus diazoxide (high glucose) infusions. A, Duct-associated single ␤-cells and ␤-cell buds (number per millimeter squared pancreas). B, Isolated small islets (⬍15 ␤-cells; number per millimeter squared pancreas). n ⫽ 4 rats for saline and diazoxide controls; n ⫽ 3 for high glucose– high insulin; n ⫽ 5 for high insulin and diazoxide controls; n ⫽ 7 for high glucose. Data are means ⫾ SE. *, P ⬍ 0.05.

previous blockade of insulin secretion by diazoxide. It was unchanged in high insulin rats. There was no effect of diazoxide on insulin content (Fig. 8C). Kinetics of glucose-induced insulin release in vitro

Basal rate of insulin release (in response to 2.8 mmol/liter glucose) was very similar in high glucose– high insulin, high insulin, diazoxide controls, and saline control rats. It was largely increased in high glucose rats (Fig. 8A). Insulin response to 16.7 mmol/liter glucose was slightly lower in diazoxide controls than in saline controls, but the difference was not significant (Fig. 8, A and B). Glucose-induced insulin release was lower in high insulin rats than in control groups, whereas it was increased to a much larger extent than in saline controls in high glucose– high insulin and high glucose rats (Fig. 8B).

Islet insulin content

Islet insulin content was decreased in high glucose– high insulin rats, whereas it was largely increased in high glucose rats, compared with saline controls rats, likely because of the

Discussion

The first step of this work was to constitute groups of rats with well-defined glycemic and insulinemic profiles. A 48-h

2722



FIG. 4. Presence of a high number of small islets in high insulin rats (C) compared with high glucose– high insulin (A), high glucose (B), and saline controls (D). Evidence of islet budding from ducts in the pancreas of high glucose– high insulin (E) and high glucose (F) rats. ␤-Cells were immunostained for insulin and revealed with peroxidase method. Final magnification, ⫻20 (A–D) and ⫻200 (E and F).

FIG. 5. Immunohistochemical visualization of an insulin-cytokeratin 20-positive duct bud (A) and insulin-Glut2-positive cells within a pancreatic duct (B). Final magnification, ⫻400. A, Insulin in red (white arrow); CK20 in brown (black arrow). B, Insulin in red (white arrow); Glut2 in brown (black arrow).

glucose infusion induced marked hyperglycemia and hyperinsulinemia in agreement with our previous studies using the same experimental approach (8, 24). Simultaneous infusion of glucose and insulin allowed us to maintain a very high plasma insulin level concomitantly with euglycemia. The infusion of glucose and diazoxide resulted in an increase in plasma glucose concentration similar to that of high glucose– high insulin rats, whereas plasma insulin levels remained close to basal throughout the infusion period. This is in agreement with other studies in which diazoxide was also successfully used to reduce insulin secretion (27, 28).

In high glucose– high insulin rats, there was a clear increase in ␤-cell mass in agreement with our previous studies using a similar experimental model (8, 24) and studies from others (7, 29, 30). This confirms the remarkable plasticity of the endocrine pancreas in adult rats. Interestingly, in high glucose rats, the ␤-cell mass was increased to a similar extent as in high glucose– high insulin rats in the absence of any increase in plasma insulin concentration. Although an increase in intraislet insulin concentration cannot be completely excluded, this stresses the potent tropic effect of glucose per se. To our knowledge in all previous studies showing


FIG. 6. ␤-Cell replication rates expressed as percentage of BrdUpositive ␤-cells per 6 h in rats after 48-h infusion of saline (saline controls), glucose (high glucose– high insulin), glucose plus insulin (high insulin), diazoxide (diazoxide controls), and glucose plus diazoxide (high glucose) infusions. n ⫽ 5 rats for saline and diazoxide controls; n ⫽ 3 for high glucose– high insulin; n ⫽ 5 for high glucose. B, ␤-Cell replication rates expressed as percentage of BrdU-positive ␤-cells per 6 h in rats after 8-, 16-, 24-, and 48-h infusion of glucose (high glucose– high insulin) and after 48 h of saline (saline controls). n ⫽ 5 for saline controls; n ⫽ 4 for high glucose– high insulin 8 h, 16 h, and 24 h ; n ⫽ 3 for high glucose– high insulin 48 h. Data are means ⫾ SE. *, P ⬍ 0.05.

a stimulating effect of glucose on ␤-cell growth in vivo, hyperglycemia was associated with hyperinsulinemia (7, 8, 23, 29, 30). We demonstrate here that prolonged exposure of the endocrine pancreas to high glucose levels may provoke a rapid compensatory ␤-cell growth independently of changes in plasma insulin levels. It is noteworthy that the blockade of insulin secretion by diazoxide did not prevent the stimulating effect of glucose on ␤-cell mass expansion. It suggests that the increase in ␤-cell growth did not result from signals originating from the activation of the secretory process, at least downstream of K⫹ channel closure. The growing evidence of the important role of insulin and insulin receptor signaling in ␤-cell growth and homeostasis (17–19, 31, 32) stimulated us to address the question of the precise role of insulin in ␤-cell mass expansion. Several stud-


ies showed that insulin treatment (11–14) accelerated ␤-cell regeneration in diabetic rats and mice. Whether insulin exerts a direct effect on ␤-cell growth or indirectly through an alteration in plasma glucose concentrations remains questionable. Our finding that high plasma insulin levels lead to a substantial increase in the ␤-cell mass even in the presence of euglycemia provides clear evidence of the ability of insulin to directly promote ␤-cell expansion in vivo independently of its modulating effect on plasma glucose concentration. In a previous study, we failed to demonstrate any effect on ␤-cell growth of hyperinsulinemia maintained at a level similar to that of high glucose– high insulin rats (8). However, exogenous infusion of insulin cannot completely mimic the effect of endogenous hyperinsulinemia, especially the increase in intraislet insulin concentration in response to glucose stimulation with possible autocrine and/or paracrine effects are missed. Previous studies showed that the dramatic increase in intraislet insulin induced by high glucose was crucial for autocrine and paracrine interactions (33, 34). Therefore, we decided here to infuse insulin at a high flow rate in an attempt to increase insulin concentration enough to come near the autocrine/paracrine and endocrine situation induced by glucose infusion in high glucose– high insulin rats. Under these conditions, we cannot rule out the possibility that insulin effects could be ascribed, at least in part, to binding of the hormone to IGF-I receptor. However, this is rather unlikely because the affinity of IGF-I receptor for insulin is very low, and it was recently shown that ␤-cell mass was not affected by deletion of IGF-I receptor (35). ␤-Cell size was only slightly increased in high glucose– high insulin rats in agreement with our previous study (8) and another study using a similar protocol (7). It remained unchanged in the other groups of infused rats, thus stressing that the association of hyperglycemia and hyperinsulinemia is required for ␤-cell hypertrophy. The apoptosis rate was very low, and no difference could be observed among the experimental groups. Therefore, changes in the apoptosis rate could not contribute to ␤-cell expansion. In all groups of rats, there was no increase in ␤-cell proliferation, which was even decreased in high insulin and high glucose groups. We cannot exclude that in the latter group, the low proliferation rate could partly be ascribed to diazoxide in infusion because ␤-cell proliferation was decreased in diazoxide control rats. However, it is unlikely that the use of diazoxide makes us miss a possible stimulating effect of glucose on ␤-cell proliferation because proliferation rate was no higher than in controls in the high glucose– high insulin group. Although these data agree with those of Lipsett et al. (36), who showed a drop in ␤-cell proliferation during the first 24 h of glucose infusion in rats, this situation is rather paradoxical in regard to the well-documented stimulating effect of glucose and insulin on ␤-cell replication (6, 7, 9, 10). Although the precise role of insulin in ␤-cell proliferation is difficult to appreciate in vivo for the reasons mentioned above, a stimulating effect of insulin on ␤-cell replication has been reported in newborn rats islets in culture (37). For the present data, we can propose the following very speculative explanation: In high glucose– high insulin and high glucose rats, the abrupt and important rise in plasma glucose con-

2724



FIG. 7. A and B, Immunohistochemical staining of apoptotic ␤-cells (arrows) in pancreatic islets of high glucose– high insulin rats after 48-h glucose infusion. Magnification, ⫻200 (A), ⫻400 (B). C, BrdU immunostaining in ␤-cells in high glucose– high insulin rats after 48-h infusion. Magnification, ⫻400.

centrations maintained ␤-cells in a phenotype turned toward insulin biosynthesis and secretion, which is not compatible with cell replication. At 48 h, the ␤-cell replication rate increased, compared with time 24 h, probably because the new ␤-cell that appeared started to proliferate. In high insulin rats, plasma glucose was rather low during the last part of infusion. The requirement of a threshold in glucose concentration for the stimulation of ␤-cell proliferation by insulin was demonstrated by Koiter et al. (16), who showed that the stimulating effect of hyperinsulinemia on ␤-cell replication was hampered in the presence of even mild hypoglycemia. In all groups, some evidence indicates that neogenesis could contribute to the increase in ␤-cell mass. In a previous study, we showed that neogenesis was the main process leading to ␤-cell growth in glucose-infused rats (24). In an attempt to quantify neogenesis, we used the current definition that includes endocrine cells budding from pancreatic ducts and/or clusters of a few ␤-cells (review in Refs. 1 and 38). According to these criteria we found that both hyperglycemia and hyperinsulinemia stimulated the neogenic process. The importance of neogenesis in ␤-cell mass increase in adult rodents emerges from many studies especially during

pancreas regeneration (2) and recent studies stress that the neogenenic process may account for ␤-cell mass expansion in the absence of increased ␤-cell replication (14, 31, 39). Neogenesis was mainly evidenced by islets budding from pancreatic ducts in hyperglycemic rats, whereas it was characterized by the presence of numerous small clusters of ␤-cells within the exocrine pancreas in high insulin rats. Although we cannot ascertain that these different features correspond to different mechanisms of ␤-cell expansion, it can be assumed that glucose and insulin may stimulate neogenesis via specific ways. It must be recognized that neogenesis is a process very difficult to identify and quantify. Moreover, the quantitative data related to neogenesis at 48 h cannot account for the dramatic increase in ␤-cell mass that we observe at this time. However, neogenesis is a dynamic process that probably started a long time before 48 h of infusion, and we have only a static view of it at 48 h. In a previous study, we showed that neogenesis was much higher after a 24-h infusion than a 48-h infusion with glucose (24). Alternatively, but not contradictorily, the contribution of duct-unrelated neogenesis may be important as suggested by the presence of small ␤-cell clus-



FIG. 8. A, Dynamics of insulin release in response to 16.7 mmol/liter glucose from freshly isolated islets of saline controls, high glucose– high insulin, high insulin, diazoxide controls, and high glucose rats. Data are means ⫾ SE. *, P ⬍ 0.05 as compared to saline control groups; §, P ⬍ 0.05 as compared to high glucose– high insulin groups. n ⫽ 5 in each group. B, Glucose-induced insulin release above basal in saline controls, high glucose– high insulin, high insulin, diazoxide controls, and high glucose rats. Data are means ⫾ SE. *, P ⬍ 0.05 of five rats in each group. C, Islet insulin content in saline controls, high glucose– high insulin, high insulin, diazoxide controls, and high glucose rats. Data are means ⫾ SE. *, P ⬍ 0.05 of five rats in each group.

ters (21, 40, 41). Neogenesis from nonductal progenitors has been demonstrated in models of pancreas regeneration (42– 44). Interestingly, Lipsett and Finegood (39) showed that the increase in ␤-cell mass induced by continuous glucose infusion in rats was due mainly to acinar cell transdifferentiation into ␤-cells. This is at variance with our data showing only duct-related neogenesis in high glucose groups. Differences in experimental conditions (higher glucose levels and Sprague Dawley rats vs. Wistar rats in our study) may be of importance to explain the discrepancies between the data from both studies. The differential effects of glucose and insulin were also observed when considering ␤-cell function. In high glucose– high insulin and high glucose rats, ␤-cell growth correlated with a marked increase in islet responsiveness to glucose in agreement with other studies, showing sustained ␤-cell ac-

tivation in vitro after prolonged exposure to glucose in rats (8, 45, 46) as well as humans (47). Therefore, whatever the insulin status, ␤-cell mass changes induced by hyperglycemia result in a enhancement of islet secretory capacity. In contrast to high glucose, high insulin levels did not induce any improvement of ␤-cell function. The quantity of insulin released during the stimulation period was even lower than in controls. This suggests that in high insulin rats, the pool of newly formed ␤-cells did not achieve functional maturation. In addition, the poor insulin response to glucose in high insulin rats may be related to the rather low prevailing glucose concentration that is crucial for the further insulin response to glucose stimulation. In this group, as mentioned above, plasma glucose concentrations were close to basal values and even lower than 5 mmol/liter at the end of infusion.

2726



In conclusion, the data show first that glucose and insulin may have specific stimulating effects on ␤-cell growth in vivo in adults. Moreover, the occurrence of ␤-cell hypertrophy and the specific morphological changes in hyperglycemichyperinsulinemic rats suggest that the combined effects of glucose and insulin on ␤-cell mass homeostasis are not the simple addition of their respective isolated effects. In addition, the correlation between ␤-cell growth and responsiveness to glucose in hyperglycemic rats independently of their insulin status contrasts with the lack of effect of ␤-cell mass enlargement induced by isolated hyperinsulinemia on ␤-cell function. Although the involvement of low prevailing levels of plasma glucose cannot be excluded, this suggests that the new ␤-cells recruited by insulin infusion were not fully functional or poorly functional. These cells may lack one or several factors required for the full maturation of the process of responsiveness to glucose. The search and characterization of these factors may provide insights to the mechanisms underlying the maturation of ␤-cells. Acknowledgments The authors acknowledge the expert advice provided by Ce´ cile Tourrel and Kyung-Ah Sohn concerning the perifusion experiments. We thank Ms. Emmy De Blay for expert technical assistance concerning histological studies and Nadim Kassis for his assistance concerning DNA islet quantification. Received October 25, 2002. Accepted February 3, 2003. Address all correspondence and requests for reprints to: Maryline Paris, Laboratoire de Physiopathologie de la Nutrition, Centre National de la Recherche Scientifique, Unite´ Mixte de Recherche 7059, Universite´ Paris 7, case 7126, 2 place Jussieu, 75251 Paris, France. E-mail: [email protected].

14. 15. 16. 17. 18.

19.

20.

21. 22. 23. 24. 25. 26.

References 1. Swenne I 1992 Pancreatic ␤-cell growth and diabetes mellitus. Diabetologia 35:193–201 2. Bonner-Weir S 2000 Life and death of the pancreatic B cell. Trends Endocrinol Metab 11:375–378 3. Scaglia L, Smith FE, Bonner-Weir S 1995 Apoptosis contributes to the involution of ␤ cell mass in the post partum rat pancreas. Endocrinology 136: 5461–5468 4. Klo¨ppel G, Lohr M, Habich K, Oberholzer M, Heitz PU 1985 Islet pathology and the pathogenesis of type 1 and type 2 diabetes mellitus revisited. Surv Synth Pathol Res 4:110 –125 5. Bernard-Kargar C, Ktorza A 2001 Endocrine pancreas plasticity under physiological and pathological conditions. Diabetes 50(Suppl 1):S30 –S35 6. Tyrberg B, Eizirik DL, Hellerstrom C, Pipeleers DG, Andersson A 1996 Human pancreatic ␤-cell deoxyribonucleic acid-synthesis in islet grafts decreases with increasing organ donor age but increases in response to glucose stimulation in vitro. Endocrinology 137:5694 –5699 7. Bonner-Weir S, Deery D, Leahy JL, Weir G 1989 Compensatory growth of pancreatic ␤-cells in adult rats after short-term glucose infusion. Diabetes 38:49 –53 8. Bernard C, Thibault C, Berthault MF, Magnan C, Saulnier C, Portha B, Pralong WF, Penicaud L, Ktorza A 1998 Pancreatic ␤-cell regeneration after 48-h glucose infusion in mildly diabetic rats is not correlated with functional improvement. Diabetes 47:1058 –1065 9. Chick WL 1973 Beta cell replication in rat pancreatic monolayer cultures. Effects of glucose, tolbutamide, glucocorticoid, growth hormone and glucagon. Diabetes 22:687– 693 10. Schuppin GT, Bonner-Weir S, Montana E, Kaiser N, Weir GC 1993 Replication of adult pancreatic-␤ cells cultured on bovine corneal endothelial cell extracellular matrix. In Vitro Cell Dev Biol Anim 29A:339 –344 11. McEvoy RC, Hegre OD 1979 Syngeneic transplantation of fetal rat pancreas. Effect of insulin treatment on the growth and differentiation of the pancreatic implants after reversal of diabetes. Diabetes 28:141–146 12. Brown J, Heininger D, Kuret J, Mullen Y 1981 Islet cells grow after transplantation of fetal pancreas and control of diabetes. Diabetes 30:9 –13 13. Movassat J, Saulnier C, Portha B 1997 Insulin administration enhances growth

27. 28. 29. 30. 31. 32. 33. 34. 35.

36. 37.

38.

of the ␤-cell mass in streptozotocin-treated newborn rats. Diabetes 46:1445– 1452 Guz Y, Nasir I, Teitelman G 2001 Regeneration of pancreatic B cells from intra-islet precursor cells in an experimental model of diabetes. Endocrinology 142:4956 – 4968 Andersson A 1983 The influence of hyperglycaemia, hyperinsulinaemia and genetic background on the fate of intrasplenically implanted mouse islets. Diabetologia 25:269 –272 Koiter TR, Wijkstra S, Van Der Schaaf-Verdonk CJ, Moes H, Schuiling GA 1995 Pancreatic ␤-cell function and islet-cell proliferation: effect of hyperinsulinaemia. Physiol Behav 57:717–721 Withers DJ, Gutierrez JS, Towery H, Burks DJ, Ren JM, Previs S, Zhang Y, Bernal D, Pons S, Shulman GI, Bonner-Weir S, White MF 1998 Disruption of IRS-2 causes type 2 diabetes in mice. Nature 26:900 –904 Terauchi Y, Iwamoto K, Tamemoto H, Komeda K, Ishii C, Kanazawa Y, Asanuma N, Aizawa T, Akanuma Y, Yasuda K, Kodama T, Tobe K, Yazaki Y, Kadowaki T 1997 Development of non-insulin-dependent diabetes mellitus in the double knockout mice with disruption of insulin receptor substrate-1 and ␤ cell glucokinase genes. Genetic reconstitution of diabetes as a polygenic disease. J Clin Invest 99:861– 866 Kulkarni RN, Bruning JC, Winnay JN, Postic C, Magnuson MA, Kahn CR 1999 Tissue-specific knockout of the insulin receptor in pancreatic ␤ cells creates an insulin secretory defect similar to that in type 2 diabetes. Cell 96:329 –339 Ktorza A, Girard JR, Kinebanyan MF, Picon L 1981 Hyperglycaemia induced by glucose infusion in the unrestrained pregnant rat during the last three days of gestation: metabolic and hormonal changes in the mother and the fetuses. Diabetologia 569 –574 Wang RN, Kloppel G, Bouwens L 1995 Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats. Diabetologia 38:1405– 1411 Swenne I 1982 The role of glucose in the in vitro regulation of cell cycle kinetics and proliferation of fetal pancreatic B-cells. Diabetes 31:754 –760 Garofano A, Czernichow P, Breant B 1998 ␤-cell mass and proliferation following late fetal and early postnatal malnutrition in the rat. Diabetologia 41:1114 –1120 Bernard C, Berthault MF, Saulnier C, Ktorza A 1999 Neogenesis vs. apoptosis as main components of pancreatic ␤ cell ass changes in glucose-infused normal and mildly diabetic adult rats. FASEB J 13:1195–1205 Coles HS, Burne JF, Raff MC 1993 Large-scale normal cell death in the developing rat kidney and its reduction by epidermal growth factor. Development 118:777–784 Clark A, Wells CA, Buley ID, Cruickshank JK, Vanhegan RI, Matthews DR, Cooper GJ, Holman RR, Turner RC 1988 Islet amyloid, increased A-cells, reduced B-cells and exocrine fibrosis: quantitative changes in the pancreas in type 2 diabetes. Diabetes Res 9:151–159 Sako Y, Grill VE 1990 Coupling of ␤-cell desensitization by hyperglycemia to excessive stimulation and circulating insulin in glucose-infused rats. Diabetes 39:1580 –1583 Leahy JL, Bumbalo LM, Chen C 1994 Diazoxide causes recovery of B-cell glucose responsiveness in 90% pancreatectomized diabetic rats. Diabetes 43: 173–179 Finegood DT, Scaglia L, Bonner-Weir S 1995 Dynamics of ␤-cell mass in the growing rat pancreas. Estimation with a simple mathematical model. Diabetes 44:249 –256 Steil GM, Trivedi N, Jonas JC, Hasenkamp WM, Sharma A, Bonner-Weir S, Weir GC 2001 Adaptation of B-cell mass to substrate oversupply: enhanced function with gene expression. Am J Physiol Endocrinol Metab 280:E788 –E796 Bernal-Mizrachi E, Wen W, Stahlhut S, Welling CM, Permutt AM 2001 Islet B-cell expression of constitutively active Akt1/PKB a induces striking hypertrophy, hyperplasia, and hyperrinsulinemia. J Clin Invest 108:1631–1637 Tuttle R, Gill NS, Pugh W, Lee JP, Koeberlein B, Furth EE, Polonsky KS, Naji A, Birnbaum MJ 2001 Regulation of pancreatic ␤-cell growth and survival by the serine/threonine kinase Akt1/PKB. Nat Med 7:1133–1137 Weir GC, Bonner-Weir S 1990 Islets of Langerhans: the puzzle of intraislet interactions and their relevance to diabetes. J Clin Invest 85:983–987 Weir GC, Knowlton SD, Atkins RF, McKennan KX, Martin DB 1976 Glucagon secretion from the perfused pancreas of streptozotocin-treated rats. Diabetes 25:275–282 Kulkarni R, Holzenberger M, Shih D, Ozcan U, Stoffel M, Magnuson M, Kahn RC 2002 ␤-Cell-specific deletion of the IGF1 receptor leads to hyperinsulinemia and glucose intolerance but does not alter ␤-cell mass. Nat Genet 31:111–115 Lipsett M, Dhatt N, Finegood DT 1998 Pancreatic ductule formation induced by glucose infusion in adult rats. Diabetes 47(Suppl 1):A257 (Abstract) Rabinovitch A, Quigley C, Russell T, Patel Y, Mintz DH 1982 Insulin and multiplication stimulating activity (an insulin-like growth factor) stimulate islet (␤-cell) replication in neonatal rat pancreatic monolayer cultures. Diabetes 31:160 –164 Bonner-Weir S 2000 Islet growth and development in the adult. J Mol Endocrinol 24:297–302

Paris et al. • Insulin, Glucose, and ␤-Cell Growth in Vivo 39. Lipsett M, Finegood DT 2002 ␤-cell neogenesis during prolonged hyperglycemia in rats. Diabetes 51:1834 –1841 40. Bouwens L, Pipeleers DG 1998 Extra-insular ␤ cells associated with ductules are frequent in adult human pancreas. Diabetologia 41:629 – 633 41. Wang RN, Bouwens L, Kloppel G 1996 ␤-cell growth in adolescent and adult rats treated with streptozotocin during the neonatal period. Diabetologia 39: 548 –557 42. Rosenberg L 1995 In vivo cell transformation: neogenesis of ␤ cells from pancreatic ductal cells. Cell Transplant 4:371–383 43. Fernandes A, King LC, Guz Y, Stein R, Wright CV, Teitelman G 1997 Differentiation of new insulin-producing cells is induced by injury in adult pancreatic islets. Endocrinology 138:1750 –1762


44. Bouwens L 1998 Transdifferentiation versus stem cell hypothesis for the regeneration of islet ␤-cells in the pancreas. Microsc Res Tech 43:332–336 45. Thibault C, Guettet C, Laury M, NⴕGuyen JM, Tormo MA, Bailbe´ D, Portha B, Pe´ nicaud L, Ktorza A 1993 In vivo and in vitro increased pancreatic ␤-cell sensitivity to glucose in normal rats submitted to 48-h hyperglycemic period. Diabetologia 36:589 –595 46. Ling Z, Kiekens R, Mahler T, Schuit F, Pipeleers-Marichal M, Abdullah S, Gunther K, Malaisse W, Pipeleers D 1996 Effects of chronically elevated glucose levels on the functional properties of rat pancreatic ␤-cells. Diabetes 45:1774 –1782 47. Ling Z, Pipeleers D 1996 Prolonged exposure of human B cells to elevated glucose levels results in sustained cellular activation leading to a loss of glucose regulation. J Clin Invest 98:2805–2812