Insulin Receptor Substrate-2 (Irs2) in Endothelial Cells Plays a Crucial Role in Insulin Secretion

Type 2 diabetes is a heterogeneous disorder with varying degrees of insulin resistance and insulin secretion (1,2). The UK Prospective Diabetes Study clinical trial revealed a progressive impairment in pancreatic β-cell function during the course of the disease, implicating an important role of β-cell failure in the pathogenesis of type 2 diabetes (3). Actually, the progression of type 2 diabetes is associated with minimal changes in the degree of insulin resistance; however, insulin secretion is progressively blunted with the transition from prediabetes to overt diabetes (4). β-Cell failure or dysfunction is inherently associated with type 2 diabetes and may precede the onset of hyperglycemia. Thus, the amelioration of impaired insulin secretion might be a reasonable therapeutic target.

Several studies have reported that the islet blood flow is involved in the regulation of insulin secretion. Pancreatic islets are highly vascularized by a dense network of capillaries, and various mediators, such as insulin, regulate the islet blood flow (5,6). Iwase et al. (7) demonstrated that orally administered insulin secretagogues acutely increased the islet blood flow. Moreover, renin-angiotensin system (RAS) inhibitors that regulate the islet blood flow, such as ACE inhibitors and angiotensin II receptor blockers (ARBs), increased insulin secretion in response to glucose administration (8–10). In addition to RAS inhibitors, some vasoactive drugs enhance pancreatic islet blood flow, augment insulin secretion, and improve glucose tolerance (11,12). These findings suggest the involvement of the islet blood flow in insulin secretion.

Insulin receptor substrate-2 (Irs2) is the major insulin receptor substrate isoform expressed in endothelial cells (13). We previously reported that mice with the Irs2 deletion (knockout [KO]) in endothelial cells (ETIrs2KO mice) exhibited an attenuation of the insulin-induced capillary blood flow in skeletal muscle (14). In the current study, we used these mice to demonstrate the relationship between insulin secretion and the islet blood flow. Although insulin secretion from isolated islets was maintained, insulin secretion was significantly impaired in the ETIrs2KO mice. These mice showed significant decreases in the islet blood flow, similar to the results seen in skeletal muscle in a previous study (14). In addition, after the administration of enalapril maleate, which enhances the islet blood flow in these KO mice, insulin secretion was almost completely restored to levels equal to those in the control mice. These data suggest that the absence of Irs2 in endothelial cells impairs the islet blood flow, which may be one of the mechanisms responsible for the decrease in insulin secretion. We previously reported that hyperinsulinemia linked to obesity leads to the downregulation of Irs2 in endothelial cells (14). Thus, Irs2 in endothelial cells may serve as a novel therapeutic target for preventing and ameliorating type 2 diabetes and metabolic syndrome.

ETIrs2KO mice were generated as described previously (14). C57BL/6J mice were obtained from CLEA Japan (Tokyo, Japan). Mice were housed under a 12-h light-dark cycle and were given access ad libitum to normal chow consisting of 25% (w/w) protein, 53% carbohydrates, 6% fat, and 8% water (MF diet; Oriental Yeast Co., Ltd., Osaka, Japan). Male mice were used for all of the experiments in this study. The animal care and experimental procedures were approved by The University of Tokyo Animal Care Committee.

To evaluate the effect of an ACE inhibitor on the islet blood flow and glucose-induced insulin secretion, enalapril maleate (100 μg/kg body weight) or the corresponding volume of saline (0.1 mL) was injected intravenously.

Mice were denied access to food for 16 h starting at 1900 h on the previous evening and continuing until the end of the fasting period. Control and ETIrs2KO mice were intraperitoneally injected with glucose (3.0 g/kg body weight) to evaluate insulin secretion. Mice were also injected intraperitoneally with l-arginine monohydrochloride (2.1 g/kg body weight) and glucagon (10 mg/kg body weight; Glucagon G Novo, Novo Nordisk, Bagsvaerd, Denmark). Blood samples from tail snips were collected at the indicated times, and the blood glucose level was immediately measured using an automatic Glutest Pro blood glucose meter (Sanwa Kagaku Kenkyusho, Nagoya, Japan). Whole blood samples were collected and centrifuged in heparinized tubes, and the plasma samples were stored at −30°C. The insulin levels were determined using an insulin radioimmunoassay (RIA) kit (Institute of Isotopes, Budapest, Hungary) using rat insulin as the standard.

Pancreatic islets were isolated from 12-week-old mice using collagenase digestion, as described previously (15). Insulin secretion from the islets was measured under static incubation with Krebs-Ringer bicarbonate (KRB) buffer (129 mmol/L NaCl, 4.8 mmol/L KCl, 1.2 mmol/L MgSO₄, 1.2 mmol/L KH₂PO₄, 2.5 mmol/L CaCl₂, 5 mmol/L NaHCO₃, and 10 mmol/L HEPES; pH 7.4) containing 0.2% BSA. In the static incubation experiments, batches of 10 freshly isolated islets were preincubated at 37°C for 30 min in 500 μL KRB buffer containing 2.8 mmol/L glucose. The preincubation solutions were replaced with 500 μL KRB buffer containing the test agents, and the batches of islets were incubated at 37°C for 60 min. At the end of the incubation, aliquots of the buffer were immediately sampled and stored at −30°C until assay. For measurement of the islet insulin content, islets were solubilized in an acid-ethanol solution overnight at −30°C. The insulin concentration was measured using an insulin RIA kit, and the resulting concentration was corrected according to the DNA content, which was measured using the PicoGreen DNA assay kit (Invitrogen, Carlsbad, CA).

The kinetics of insulin secretion were studied in vitro using a perifusion system. Isolated pancreatic islets were used immediately after isolation. Size-matched islets (n = 50) were placed in each column. Then, the columns were gently closed with the top adaptors and immersed in a vertical position in a temperature-controlled water bath at 37°C. The perifusion medium was maintained at 37°C in a water bath. All of the columns were perifused in parallel at a flow rate of 0.6 mL/min. After 30 min of static incubation with KRB buffer (2.8 mmol/L glucose), the islets were stimulated by the continuous addition of 22.2 mmol/L glucose. Samples were collected every 2 min and stored at −30°C until further analysis.

Total RNA was extracted from the islets using an RNeasy kit (QIAGEN Sciences, Gaithersburg, MD), in accordance with the manufacturer’s instructions. After treatment with RQ1 RNase-Free DNase (Promega, Madison, WI) to remove genomic DNA, cDNA was synthesized using MultiScribe Reverse Transcriptase (Applied Biosystems, Foster City, CA), and TaqMan quantitative PCR (50°C for 2 min, 95°C for 10 min, followed by 40 cycles at 95°C for 15 s, and 60°C for 1 min) was then performed using the ABI Prism 7900 PCR system (Applied Biosystems) to amplify insulin1, insulin2, and cyclophilin cDNA. The primers that were used were purchased from Applied Biosystems. The relative abundance of the transcripts was normalized to the constitutive expression of cyclophilin mRNA.

Isolated pancreata were fixed with 4% paraformaldehyde at 4°C overnight. Tissues were routinely processed for paraffin embedding, and 4-μm sections were cut and mounted on silanized slides. Pancreatic sections were stained with anti-rabbit insulin antibodies (diluted 1:200; Santa Cruz Biotechnology, Dallas, TX). Images of the pancreatic tissue and islet β-cells were viewed on the monitor of a computer through a microscope connected to a camera with a charged-coupled device (Olympus, Tokyo, Japan). The areas of the pancreata and β-cells were traced manually and analyzed with WinROOF software (Mitani, Chiba, Japan), as previously described (15). At least 50 islets per mouse were analyzed. BrdU incorporation was analyzed as described previously (16). In brief, BrdU (100 mg/kg in saline; Sigma-Aldrich, St. Louis, MO) was injected intraperitoneally, and the pancreas was removed 6 h later. The sections were immunostained with anti-BrdU antibody (diluted 1:200; Dako, Glostrup, Denmark). BrdU-positive β-cells were quantitatively assessed as a percentage of the total number of β-cells by counting the cells in a minimum of 50 islets per mouse. Immunohistochemical staining for pimonidazole was done as described previously (17), with slight modifications. The oxygenation marker pimonidazole was injected intravenously into the tail vein (60 mg/kg). The animals were killed 2 h later, and their pancreata were removed and prepared for histological analysis. The sections were immunostained with anti-pimonidazole antibody (diluted 1:100; Hypoxyprobe, Burlington, MA). Pimonidazole-positive islets were quantitatively assessed as a percentage of the total number of islets by counting in a minimum of 50 islets per mouse.

Mice were used for the perfusion experiments after they had been denied access to food overnight for 16 h, as previously reported (18) with slight modifications. Briefly, the superior mesenteric and renal arteries were ligated, and the aorta was tied just below the diaphragm. The perfusate was infused via a catheter placed in the aorta and collected from the portal vein. The perfusate used was KRB HEPES buffer supplemented with 4.6% dextran and 0.25% BSA and gassed with 95% O₂ and 5% CO₂. The flow rate of the perfusate was 1 mL/min. Pancreata were perfused with KRB HEPES buffer containing 2.8 or 22.2 mmol/L glucose. The perfusion protocols began with a 10-min equilibration period with the same buffer used in the initial step (i.e., from 1 to 5 min), as shown in the figures. The insulin levels in the perfusate were measured using an RIA kit.

The experiments were performed according to a protocol described in detail in a previous report (19). Briefly, polyethylene catheters were inserted via the right carotid artery into the ascending aorta and into the femoral artery. After the blood pressure stabilized, nonradioactive Dye-Trak microspheres (Triton Technology, Los Angeles, CA), with a mean diameter of 10 μm, were injected for 10 s via the catheter placed with its tip in the ascending aorta. Starting 5 s before the microsphere injection and continuing for 60 s, an arterial blood sample was collected from the catheter in the femoral artery at a rate of ∼0.50 mL/min. The exact withdrawal rate was confirmed in each animal by weighing the sample. After the reference sample was obtained, another blood sample was drawn for the measurement of the blood glucose level. The whole pancreata were removed, blotted, weighed, and treated using a freeze-thawing technique to visualize the microspheres. The capillary blood flow values were calculated according to the formula: Qorg = Qref × Norg/Nref, where Qorg is the organ capillary blood flow (mL/min), Qref is the withdrawal rate of the reference sample (mL/min), Norg is the number of microspheres present in the organ, and Nref is the number of microspheres in the reference sample. The microsphere contents of the adrenal glands were used as a control to confirm that the microspheres had adequately mixed in the arterial circulation (19). A <10% difference in the numbers of microspheres between the right and left adrenal glands was taken to indicate sufficient mixing. The islet blood flow was expressed per islet weight estimated by multiplying the pancreatic weight with the islet volume fraction of the whole pancreas in each animal. To evaluate the effect of insulin on the islet blood flow, insulin (0.75 units/kg body weight) was injected intraperitoneally as described previously (20) with slight modifications. The anesthetized animals were left for 20 min. Before microsphere injection, blood glucose levels were determined using arterial blood samples.

Fluorescein-labeled lectin (Lycopersicon esculentum; Vector, Burlingame, CA), which binds specifically to endothelial cells and epithelial cells, was injected into the tail vein (1 mg/mL solution/0.1 mL/mouse) and allowed to circulate for 3 min. The pancreas was then excised and immersion-fixed in 4% paraformaldehyde for 16 h at 4°C. After fixation, the pancreata were immersed in 30% sucrose as a cryoprotectant, after which the tissues were embedded in optimal cutting temperature compound (Sakura Seiki, Tokyo, Japan). The block was sectioned into 15-μm-thick sections and collected on microscope slides. Images of the vasculature were viewed on the monitor of a computer through a charge-coupled device camera with Biozero (KEYENCE, Osaka, Japan). The areas of the vasculature and β-cells were traced manually and analyzed with ImageJ software. At least 50 islets were analyzed per mouse.

Values are expressed as the mean ± SE. The statistical significance of differences between groups was determined using a two-tailed indirect Student t test. Data involving more than two groups were assessed using an ANOVA.

The blood glucose levels at 5 and 15 min after glucose loading were significantly higher in the ETIrs2KO mice than in the control mice at 12 weeks (Fig. 1A). The plasma insulin levels at 2 and 5 min were significantly lower in the ETIrs2KO mice than in the control mice at 12 weeks (Fig. 1A). At 24 weeks, although the blood glucose level at 5 min was comparable to that in the control mice, the blood glucose level at 15 min was significantly higher in the ETIrs2KO mice (Fig. 1B). The plasma insulin levels at 2 min after glucose loading were significantly lower in the ETIrs2KO mice (Fig. 1B). In addition to glucose, glucagon-induced and arginine-induced insulin secretion was also significantly impaired in the ETIrs2KO mice (Fig. 1C and D). These findings suggest that insulin secretion induced by various secretagogues is impaired in the ETIrs2KO mice.

To clarify the molecular mechanism responsible for the impairment of insulin secretion in the ETIrs2KO mice, we next measured insulin secretion from isolated islets using static and perifusion analyses at 12 weeks. In contrast to the results of the in vivo study, glucose-stimulated insulin secretion was comparable between the control and the ETIrs2KO mice in a static incubation experiment (Fig. 2A). Moreover, the perifusion experiments also demonstrated that insulin secretion under perifusion with a stimulating glucose concentration was comparable between the control and the ETIrs2KO mice at 12 weeks (Fig. 2B). These data indicated that insulin secretion from isolated islets was not impaired in the ETIrs2KO mice. The gene expression levels of insulin1 and insulin2 (Fig. 2C) and the insulin contents (Fig. 2D) in the isolated islets were not significantly different between the control and the ETIrs2KO mice at 12 weeks.

Consistent with our previous study (14), the β-cell mass from the ETIrs2KO mice tended to be larger, but no statistical difference was seen between the two groups (Fig. 2E and F) at 24 weeks. In addition, the rate of BrdU incorporation into the pancreatic β-cell nuclei in the ETIrs2KO mice tended to be increased but was not significantly different compared with that in the control mice at 24 weeks (Fig. 2G). These data suggest that the impairment of β-cell function and/or mass often leads to impaired insulin secretion, which was not observed in the ETIrs2KO mice.

The impairment of insulin secretion observed in vivo, but not in vitro, prompted us to investigate the secretory responses of the pancreas in perfusion experiments that were capable of assessing insulin secretion via blood vessels. The insulin responses to glucose were significantly impaired in the ETIrs2KO mice pancreata at 12 weeks (Fig. 3A). In addition, the amount of secreted insulin measured by the area under the curve after glucose stimulation (from 5 to 30 min) was also significantly impaired in the ETIrs2KO mice (Fig. 3B). These data suggest that the impaired insulin secretion in the ETIrs2KO mice is caused by impairment in the blood circulation and not by dysfunction of the β-cells.

The mean blood pressure in anesthetized mice was similar between the control and the ETIrs2KO mice at 9–12 weeks (Fig. 4A). Although the difference in the pancreatic blood flow was not significant (Fig. 4B), the number of microspheres in islets and islet blood flow were significantly decreased in the ETIrs2KO mice compared with that in the control mice in the basal state (Fig. 4C and D). Moreover, in the insulin-treated state, islet blood flow was significantly decreased in the ETIrs2KO mice at 9 weeks (Fig. 4E and Table 1). These findings suggest that the absence of Irs2 in endothelial cells induces a reduction in the islet blood flow, similar to previous observations in skeletal muscle (14), leading to the decrease in insulin secretion in these mice.

The capillary area stained with lectin was significantly smaller in the islets from the ETIrs2KO mice compared with those from the control mice at 12 weeks (Fig. 5A and B). However, the number of capillaries in the islets was comparable between the control and the ETIrs2KO mice (Fig. 5C). Moreover, we performed immunohistochemical staining for pimonidazole to assess islet oxygenation. No difference in the frequency of pimonidazole-positive islets was observed between the control and the ETIrs2KO mice (Fig. 5D). These findings suggest that the capillary area, but not the number of capillaries in the islets or islet oxygenation, is reduced in the ETIrs2KO mice.

Recent studies have demonstrated that prevention of angiotensin II formation, by ACE inhibition with enalapril maleate, could preferentially increase the islet blood flow as a result of a vasodilating action (21). It had been demonstrated that islet microvessels may produce higher levels of angiotensin II than microvessels in the exocrine pancreas and, therefore, may be more sensitive to ACE inhibition in islets. In this study, the administration of enalapril maleate had no effect on the mean blood pressure (Fig. 6A), blood glucose concentrations (Table 2), or pancreatic blood flow (Fig. 6B), as reported previously (21). However, islet blood flow in the ETIrs2KO mice was restored significantly, almost to a level comparable with that in the control mice, at 9–12 weeks (Fig. 6C). Furthermore, although enalapril maleate had no effect on the glucose-induced insulin secretion from isolated islets (Fig. 6D), glucose-induced insulin secretion and glucose intolerance in the ETIrs2KO mice were significantly restored to levels equal to those in the control mice (Fig. 6E). These results suggest that the impairment of insulin secretion might have been caused by the impaired islet blood flow.

In this study, we demonstrated that the absence of Irs2 in endothelial cells impairs insulin secretion (Fig. 1A–D) by reducing the islet blood flow (Fig. 4D). In fact, after treatment with enalapril maleate, glucose-stimulated insulin secretion was restored to levels equal to those observed in the control mice (Fig. 6E), along with an improvement in the islet blood flow (Fig. 6C). These data suggest that Irs2 in endothelial cells regulates islet blood flow, mediating insulin secretion. Although several studies have suggested that impaired insulin secretion is predominantly caused by β-cell dysfunction (22,23), the absence of Irs2 in the capillaries also might cause insulin secretion in type 2 diabetes.

A recent study revealed that the correct integrity of the islet microvasculature is essential for normal islet function because it not only provides the means for the transport of nutrients and oxygen but also ensures adequate paracrine interactions within the individual islets. β-Cell–specific vascular endothelial growth factor-A gene ablation resulted in glucose intolerance and diabetes (24). These mice exhibited a decreased density of the microvasculature, and the capillaries exhibited an abnormal morphological appearance. Moreover, β-cell–specific Fyn-related kinase tyrosine kinase transgenic mice exhibited an impaired glucose-stimulated insulin secretion in vivo (25). Insulin secretion in isolated islets from these mice was similar to that of control mice; however, the islet blood flow and capillary lumen diameter in the islets were decreased. Although insulin secretion from isolated islets was maintained in these genetically modified animals, insulin secretion was significantly impaired in vivo. These data suggest that the disorganization of islet vascularization can impair glucose-stimulated insulin secretion, even if the function of β-cells is not impaired.

Moreover, recent investigations have demonstrated that pancreatic islet adaptation to insulin resistance is not limited to change within β-cells but also involves islet-specific neurovascular remodeling. To accommodate the increased demand for insulin delivery into the peripheral circulation, islet capillaries expand by dilation and not by angiogenesis (26). Given these findings, the correct integrity and function of the capillaries in the islets might both be essential for normal insulin secretion.

Why was the pancreatic blood flow similar but significantly decreased in the islets from ETIrs2KO mice? Although a previous study found the capillary blood flow after insulin treatment was impaired in the skeletal muscle of ETIrs2KO mice, this phenomenon was not observed during fasting (14). These data suggest that endothelial insulin signaling regulates capillary blood flow after insulin stimulation but not in the basal state. It is possible that the islet blood flow in the ETIrs2KO mice was decreased because the capillaries of the islets are constantly exposed to a high concentration of insulin, whereas the pancreatic blood flow is not because of the lack of insulin stimulation. In fact, the capillary blood flow was significantly higher in the islets than in the whole pancreas in the control mice (1.38 ± 0.16 mL/min/g islet vs. 0.73 ± 0.07 mL/min/g pancreas, P < 0.01; Fig. 4D and B).

In a present study, the islet blood flow was significantly decreased in the ETIrs2KO mice. Although precise mechanisms remain unclear, it seems more likely that this was due to microvascular dysfunction but not anatomical abnormalities. In fact, after treatment with enalapril maleate, the islet blood flow was restored to levels equal to those observed in control mice via vasodilating action (Fig. 6C). In addition, we did not detect anatomical abnormalities in islets of ETIrs2KO mice under the vascular staining (Fig. 5A and C) or the electron microscope (data not shown). Further study is needed to address this issue.

Recent trials have suggested that inhibitors of RAS, such as ACE inhibitors and ARBs, may reduce the incidence of new-onset type 2 diabetes in patients with or without hypertension who have a higher risk of developing diabetes (27–29). In addition to the amelioration of insulin secretion through the increased islet blood flow induced by an ACE inhibitor, as seen in this study, the blockade of the RAS during the development of diabetes has been attributed to improvements in peripheral insulin sensitivity and β-cell dysfunction (30,31). ACE inhibitors have been shown to improve whole-body and skeletal muscle insulin resistance in hypertensive subjects with or without type 2 diabetes (32–40). The reduction of angiotensin II–mediated vascular resistance by ACE inhibition may improve insulin-stimulated glucose transport activity in skeletal muscle (41). This improvement is associated with a favorable adaptive response in GLUT4 protein levels, glycogen storage, and the activities of relevant intracellular enzymes of glucose catabolism (42).

Moreover, locally generated and physiologically active RAS components have functions that are distinct from the classical vasoconstriction and fluid homeostasis actions of systemic RAS. Local RAS can affect islet cell function and structure in the adult pancreas as well as the proliferation and differentiation of pancreatic stem/progenitor cells during development (43,44). In fact, RAS blockade significantly attenuates islet damage and restores the β-cell mass by reducing oxidative stress, apoptosis, and attenuating profibrotic pathways (42,45). Thus, RAS inhibition may decrease the development and progression of diabetes through hemodynamic and nonhemodynamic effects or through the protection of β-cells and non–β-cells.

In conclusion, we demonstrated that the absence of Irs2 in endothelial cells impairs the islet blood flow, which may be one of the mechanisms responsible for the decrease in insulin secretion. Thus, Irs2 in endothelial cells may serve as a novel therapeutic target for preventing and ameliorating type 2 diabetes and metabolic syndrome.

Acknowledgments. The authors express sincere appreciation to Dr. Wakako Fujimoto (Division of Cellular and Molecular Medicine, Kobe University Graduate School of Medicine, Kobe, Japan) for his advice and assistance in the development of their perfusion experiments. In addition, the authors thank Emi Hashimoto-Koga, Katsuko Takasawa, Norie Kowatari-Ohtsuka, Eri Yoshida-Nagata, Ritsuko Hoshino, Eishin Hirata, Namiko Okajima-Kasuga, and Hiroshi Chiyonobu (Department of Diabetes and Metabolic Diseases, Graduate School of Medicine, The University of Tokyo) for their excellent technical assistance and animal care.

Funding. This work was supported by a grant for Translational Systems Biology and Medicine Initiative (TSBMI) from the Ministry of Education, Culture, Sports, Science and Technology of Japan and a Grant-in-Aid to T.Ka. for Scientific Research in Priority Areas (A) (18209033) and (S) (20229008) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.H., N.K., H.S., K.U., and T.Ka. designed the study and wrote the manuscript. S.H., N.K., H.S., M.S., I.T., T.Ku., and K.N. conducted the experimental research and analyzed the data. M.N. contributed to the data analysis and the preparation of the manuscript. T.Ka. is the guarantor of this work, and as such, had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis.