Angiotensin II Type 1 Receptor Blockade Improves β-Cell Function and Glucose Tolerance in a Mouse Model of Type 2 Diabetes

Genetically diabetic C57BL/KSJ +db/+db mice and their age-matched nondiabetic littermates C57BL/KSJ m+/+db were used for the experiments. The animals were obtained from the Laboratory Animal Services Centre of the Chinese University of Hong Kong. The experimental procedures were approved by the Animal Experimentation Ethics Committee of the Chinese University of Hong Kong.

Pancreatic islets were isolated as described previously (9). Briefly, male diabetic and control mice aged 10 weeks were killed and the pancreata were dissected. The pancreata were then placed in cold Hanks’ solution (Sigma Aldrich, St. Louis, MO) and were cut into small pieces of similar size about 1 mm³. These pieces were transferred to vials containing collagenase solution (Roche Molecular Biochemicals, Mannheim, Germany). The digest was then washed three times by filling the vial with Hanks’ solution. The islets were then selected under a light microscope and cultured free floating for 4–7 days in nonadherent culture dishes in RPMI 1640 medium (Sigma Aldrich) supplemented with 10% (vol/vol) fetal bovine serum (Sigma Aldrich). Culture medium was changed every other day.

Islet insulin release was measured as described previously (9). Groups of ten islets were transferred in duplicate to Falcon 24-well culture plates containing 0.25 ml Krebs-Ringer bicarbonate buffer (KRBB) supplemented with 10 mmol/l HEPES and 2 mg/ml of BSA. Islets were incubated for 1 h at 37°C (O₂/CO₂ 95:5) in medium containing 1.7 mmol/l glucose and then incubated for an additional hour in 0.25 ml of KRBB containing 16.7 mmol/l glucose. Glucose-induced insulin release from isolated islets was determined in the presence of 0.1, 1, 10, and 100 nmol/l angiotensin II (Sigma Aldrich). Angiotensin II was added to the KRBB medium with 16.7 mmol/l glucose during the 2nd h of incubation. The specific AT1R antagonist losartan (Merck, Whitehouse Station, NJ; 1 μmol/l) was administered 10 min before angiotensin II treatment or alone. After the incubation, the medium was collected for the measurement of insulin release using a Mouse Insulin ELISA Kit (Mercodia, Uppsala, Sweden).

Duplicate groups of ten islets were incubated, with or without 100 nmol/l angiotensin II or 1 μmol/l losartan, at 37°C in 100 μl of KRBB containing 50 μCi/ml of l-[4,5-³H]leucine (Amersham-Pharmacia, Buckinghamshire, U.K.) and 1.7 or 16.7 mmol/l glucose in an atmosphere of humidified air plus 5% CO₂. After a 2-h experimental treatment, the islets were washed in Hanks’ solution containing nonradioactive leucine (10 mmol/l) and sonicated in 200 μl of double-distilled water. The amount of labeled (pro)insulin was determined by an immunoabsorption assay (14), and total protein biosynthesis was measured in trichloroacetic acid precipitates of the islet homogenate.

Immunofluorescent double labeling (15) was employed to determine the specific localization of AT1R in islets. Pancreata were fixed with 4% (vol/vol) chilled paraformaldehyde and embedded in paraffin. Sections (5 μm thick) were mounted on glass slides, deparaffinized, and processed for indirect immunofluorescent double labeling. After several washes with PBS (pH 7.4), each slide was incubated with 4% (wt/vol) normal donkey serum (Jackson ImmunoResearch, West Grove, PA) for 1 h at 37°C to block nonspecific antibody binding. Excess blocking solution was poured off and the slide was incubated overnight at 4°C with both rabbit anti-AT1 serum (1:25) (Santa Cruz Biotech, Santa Cruz, CA) and goat anti-insulin serum (1:50) (Santa Cruz Biotech.). After three washes with PBS, the primary antibodies were detected using an anti-rabbit antibody labeled with rhodamine and an anti-goat antibody labeled with aminomethylcoumarin acetate (insulin) (Jackson ImmunoResearch) at room temperature for 1 h. Preadsorption and omission of primary antibodies were used as negative controls. AT1R (red) and insulin (blue) immunolabeling was detected with a fluorescent microscope equipped with a DC 200 digital camera (Leica Microsystems). At least 10 islets per mouse pancreas and five mice per group were randomly chosen, and therefore, at least 50 islets per group were analyzed. Islet β-cell mass was assessed by determining the proportion of area occupied by blue florescence within each islet (40× objective) using Leica Qwin image analysis software (Leica Microsystem) (16).

Real-time quantitative RT-PCR was performed using an ABI PRISM 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA) as described previously (9). Total RNA was extracted from pooled islets from control or type 2 diabetic mice (groups of 8–10 mice), using the Absolutely RNA Microprep Kit (Stratagene, Kirkland, WA) according to the manufacturer’s instructions. Total RNA was used as the template in one-step TaqMan amplification reactions. TaqMan primer and probe for AT1R and other RAS components were designed from mouse cDNA sequence using Primer Express Software purchased from Applied Biosystems Perkin-Elmer (9). For an internal control, 18S rRNA was used. TaqMan reactions were set up in a reaction volume of 25 μl with TaqMan PCR reagents. Each reaction consisted of 12.5 μl PCR master mix, 0.9 μmol/l of each amplification primer, 0.2 μmol/l corresponding TaqMan probe, and 30 ng RNA template. Each sample was run in duplicate with an initial 30-min period at 48°C and a 10-min period at 95°C to enable reverse transcription, followed by 40 cycles at 95°C for 15 s and a final 60°C step at 1 min. Amplification data were collected by the 7700 Sequence Detector and analyzed with Sequence Detection System software. The RNA concentration was determined from the threshold cycle at which fluorescence was first detected, the cycle number being inversely related to RNA concentration. The fold changes in AT1R in db/db islets were calculated using the 2^−ΔΔCT method, as described below in the statistical data analysis section.

Four-week-old obese db/db mice were randomly assigned to losartan dose groups. Each group consisted of 6–10 mice. The mice received 0, 1, 5, 10, 20, or 30 mg · kg⁻¹ · day⁻¹ of losartan dissolved in their drinking water for 8 weeks. The optimal concentration of losartan (10 mg · kg⁻¹ · day⁻¹) for suppressing blood glucose concentration was then selected for further analysis. Two groups of db/db mice consisting of 10–11 mice were randomly treated with losartan (10 mg · kg⁻¹ · day⁻¹) or water only. An additional two groups of age-matched control mice received the same treatment as the experimental groups.

Blood from the mouse tail vein was withdrawn twice a week to measure plasma glucose levels using a glucometer (Bayer Corporation, Tarrytown, NY). After 8 weeks of losartan treatment, oral glucose tolerance test (OGTT) was performed on mice after a 16-h overnight fast. The mice were gavaged with glucose (1 g/kg, dissolved in water). Blood glucose levels were assessed by collecting tail blood, and glucose level data were recorded immediately before and 15, 30, 60, 90, and 120 min after glucose administration. For the insulin tolerance test, the mice were injected (intraperitoneally) with 2 units/kg human biosynthetic insulin (Novo Nordisk, a generous gift from Dr. J.C. Chan at Prince of Wales Hospital, The Chinese University of Hong Kong), and blood glucose levels were detected immediately before and 15, 30, 60, 90, 120, 180, and 210 min after injection as described above.

Results are expressed as means ± SE for all groups. Multiple comparisons between groups were performed using an ANOVA followed by Tukey’s post hoc test or, when comparisons were made only relative to controls, by Dunnett’s test. When only two groups were compared, probabilities of chance differences between the experimental groups were calculated with Student’s unpaired two-tailed t test. For all comparisons, P < 0.05 was considered statistically significant. For real-time RT-PCR, the relative expression was normalized as percentage of 18S rRNA and calculated using the comparative threshold cycle method of 2^−ΔΔCT, as described previously (9).

Double immunolabeling was employed to localize precisely AT1R expression in the pancreatic islets of control and db/db mice (Fig. 1). Both AT1R and the β-cell marker insulin were present in control (Fig. 1A and B) and db/db islets (Fig. 1C and D), respectively. There was, however, an overall reduction by a factor of 2.2 in the percentage of proportional area positively labeled for insulin in db/db islets (Fig. 1B and D; control 0.38 ± 0.012; db/db 0.17 ± 0.014; n = 5 for both; P < 0.001). AT1R-labeling was more intense in db/db islets relative to control tissue (Fig. 1A and C). Immunolabeling specificity was validated by control experiments either with preadsorption of excess antigens for AT1R (Fig. 1E) and insulin (Fig. 1F) or with omission of primary antibodies (data not shown).

AT1R expression in β-cell was examined by real-time quantitative RT-PCR (Table 1). In order to normalize the β-cell number for comparison, the expression of AT1R in db/db islets was multiplied by a factor of 2.2 according to the intensity of β-cell insulin immunolabeling. As a result, the relative expression levels of AT1R mRNA to 18S rRNA in db/db islets was increased by about threefold relative to that in control tissue (Table 1 and Fig. 2). Other RAS components including AT2R and ACE were also found to be upregulated in db/db islets, even without multiplying by the correction factor of 2.2 (data not shown).

Insulin release from the isolated control islets was, as expected, markedly enhanced when glucose concentration in the incubation medium was increased from 1.7 to 16.7 mmol/l (Fig. 3A). Addition of angiotensin II inhibited insulin release in the high glucose concentration condition (16.7 mmol/l). At the highest concentration of angiotensin II used (100 nmol/l), the glucose-induced insulin release was completely prevented. Pretreatment of the isolated control islets with 1 μmol/l losartan, a specific AT1R antagonist, before the addition of angiotensin II (100 nmol/l) completely rescued glucose-induced insulin secretion. On the other hand, losartan alone did not affect glucose-stimulated insulin secretion (Fig. 3A). In islets isolated from obese db/db mice, insulin release was also augmented when glucose concentration was increased from 1.7 to 16.7 mmol/l (Fig. 3B). Similar to our observations with control tissue, exposure to the highest concentration of angiotensin II (100 nmol/l) prevented glucose-induced insulin release. Pretreatment of isolated db/db islets with 1 μmol/l of losartan before the addition of angiotensin II (100 nmol/l) not only completely rescued glucose-induced insulin secretion, but also tended to increase insulin release to an even higher level. Indeed, unlike in the control tissue, losartan alone increased glucose-induced insulin release in the db/db tissue (Fig. 3B). Note that the level of glucose-induced insulin release from the db/db islets (Fig. 3A and B) was about one-fifth of that from the control islets (Fig. 3C).

Islet (pro)insulin biosynthesis was, as was the insulin release, markedly higher at 16.7 than 1.7 mmol/l glucose in both control and db/db islets (Fig. 4). Angiotensin II inhibited islet (pro)insulin biosynthesis at 16.7 mmol/l glucose but not at 1.7 mmol/l glucose. Administration of the highest dose of angiotensin II (100 nmol/l) blocked glucose-induced (pro)insulin biosynthesis in both control and db/db islets. Pretreatment of isolated islets with 1 μmol/l of losartan before the addition of angiotensin II (100 nmol/l) restored the glucose-induced (pro)insulin biosynthesis. In db/db islets, but not in control islets, losartan alone increased the (pro)insulin biosynthesis to a level higher than that induced by glucose alone (Fig. 4B). Glucose-stimulated (pro)insulin biosynthesis in db/db islets was lower than that in control islets (Fig. 4A and B). However, losartan treatment restored db/db islet glucose-stimulated (pro)insulin biosynthesis to a level similar to that of control islets (Fig. 4C). Islet total protein synthesis was not affected by losartan treatment (data not shown).

Obesity-induced hyperglycemia in db/db mice was ameliorated by the administration of losartan in a dose-dependent manner up to 10 mg · kg⁻¹ · day⁻¹ for 8 weeks (Fig. 5A). The highest concentration (30 mg · kg⁻¹ · day⁻¹) however aggravated the hyperglycemia in the db/db mice (data not shown). In order to further examine the in vivo effects of losartan on glucose homeostasis, losartan (10 mg · kg⁻¹ · day⁻¹) was administrated to four different groups of mice: water-treated db/db mice (n = 10), losartan-treated db/db mice (n = 10), water-treated control mice (n = 11), and losartan-treated control mice (n = 11), as shown in Fig. 5B. Initially, all four groups had similar blood glucose levels. After 1 week, losartan-treated db/db mice still had normal blood glucose levels (7.7 ± 0.3 mmol/l), while water-treated db/db mice had slightly elevated blood glucose levels (9.9 ± 0.9 mmol/l). After 3 weeks of treatment, the beneficial effect of losartan treatment on glucose homeostasis was more pronounced (12.6 ± 1.2 and 17.8 ± 1.3 mmol/l for losartan- and water-treated db/db animals, respectively). By the end of the 8-week experiment, the plasma glucose levels of the losartan-treated db/db mice (17.6 ± 0.4 mmol/l) were lower than those of the water-treated db/db mice by nearly half (28.5 ± 0.8 mmol/l). There was no difference in the blood glucose levels of water-treated and losartan-treated control mice (7.9 ± 0.1 and 7.9 ± 0.2 mmol/l at 8 weeks treatment, respectively).

Oral losartan treatment also reduced glucose intolerance in db/db mice compared with that in water-treated db/db mice, as evidenced by the OGTT data (Fig. 6A). Losartan-mediated glucose tolerance improvement was readily observable when glucose response was expressed as the area under the curve (Fig. 6B). The 120-min area under the curve value of the losartan-treated db/db mice (1,991 ± 178 mmol/l) was between the values for the water-treated db/db control mice (3,231 ± 168 mmol/l) and the water-treated control mice (738 ± 46 mmol/l). On the other hand, insulin sensitivity of the peripheral tissue in db/db mice appeared not to be affected by chronic losartan treatment (10 mg · kg⁻¹ · day⁻¹) treatment. This was evidenced by the insulin tolerance test, as the net change of blood glucose level did not differ between the losartan-treated or water-treated db/db mice after an intraperitoneal injection of insulin (Fig. 7).

We recently identified a local RAS in the pancreatic islet and found that exogenously administered angiotensin II inhibits glucose-stimulated insulin secretion through binding to AT1R located on β-cells (9). This inhibitory action of AT1R was shown to be due, at least in part, to changes in (pro)insulin biosynthesis. However, endogenous levels of angiotensin II did not seem to play a role in regulating these functions during normal conditions. Importantly, the AT1R was found to be upregulated in islet β-cells from obesity-induced type 2 diabetes mice in the present study to such a degree that disturbances of insulin secretion and (pro)insulin biosynthesis were apparent with endogenous levels of angiotensin II. Indeed, inhibition of endogenous AT1R activity in islets of the type 2 diabetes db/db mouse by a specific antagonist, losartan, improved glucose-induced insulin secretion and (pro)insulin biosynthesis by about 45 and 42%, respectively.

The mechanism(s) by which local tissue RAS component expression is regulated is not well understood. Moreover, it is likely that such regulatory mechanisms differ among tissues and organs. However, at least in renal tissue, hyperglycemia and hyperlipidemia activate the local RAS (17). Moreover, in humans, weight gain and obesity per se seem to activate the RAS (18). Note that the upregulation of AT1R expression in db/db islets from type 2 diabetes observed in the present study is similar to that observed following transplantation of normal islets (19). In that case, not only a disturbance in (pro)insulin biosynthesis regulation but also a defect in islet blood flow were attributed to impaired glucose-induced insulin release. The heavy degranulation of the β-cells in db/db islets has precluded our investigation of islet blood flow, at least by standard techniques (20,21).

Several clinical trials (Captopril Prevention Project, Heart Outcomes Prevention Evaluation, Losartan Intervention For Endpoint reduction, The Study on Cognition and Prognosis in the Elderly, Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial, STOP Hypertension-2, Antihypertensive Treatment and Lipid Profile in a North of Sweden Efficacy Evaluation, The Valsartan Antihypertensive Long-term Use Evaluation, and Studies of Left Ventricular Dysfunction and Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity) have shown that RAS blockers can prevent type 2 diabetes development in high-risk individuals, e.g., in patients with arterial hypertension or chronic heart failure (22–31). All of these randomized clinical trials had cardiovascular prognosis as a primary end point but included analysis of the incidence of type 2 diabetes as a secondary end point, or as a post hoc analysis, after a mean follow-up of 1–6 years. From these studies, a recent meta-analysis calculated that the mean weighted relative risk for developing type 2 diabetes was reduced by 25% in those patients treated with a RAS blocker (32). The beneficial effect was similar with ACE inhibitors and with AT1R antagonists, as well as in patients with hypertension and those with heart failure, and was present regardless of the comparator (placebo, β-blockers/diuretics, or amlodipine). In an ongoing clinical trial called Nateglimide And Valsartan in Impaired Glucose Tolerance Outcomes Research, results thus far with type 2 diabetes incidence as the primary end point assessment show that an AT1R blocker, valsartan, seem to protect patients with impaired glucose tolerance from developing overt type 2 diabetes (33). In the present study, in vivo treatment of young diabetes-prone db/db mice with losartan did not ultimately prevent the development of diabetes but delayed the onset of hyperglycemia and ameliorated the hyperglycemia and glucose intolerance observed in these mice. This difference may, at least in part, be explained by the extremely prominent disposition for diabetes in these genetically obese leptin receptor-deficient mice. That is, their obese phenotype, which is already apparent by 3–4 weeks of age, and progressive increase in plasma glucose levels from 4 to 8 weeks of age may lead inevitably to diabetes (13,34).

The clinical benefits of RAS inhibition have been suggested to be the result of an improved delivery of insulin and glucose to peripheral muscles and direct effects on peripheral glucose transport and insulin signaling pathways. However, at least initially, a reduced glucose sensitivity in β-cells appears to be predominant over insulin resistance in the genesis of impaired tolerance to glucose (35). Our observation in db/db mice that islet AT1R may become upregulated and thus gain an increased importance in regulating islet functions that is detrimental for glucose-induced insulin release and (pro)insulin biosynthesis, suggests an alternative explanation for the clinical findings. Although AT1R blockers could improve insulin sensitivity (36,37), their effects in this aspect remain controversial and varied (12,38,39). Meanwhile, losartan has been reported to abolish the insulin sensitivity–enhancing effects of angiotensin II in vitro and in vivo (39). However, very high doses of losartan seem to be needed to activate PPAR-γ, an important mediator in insulin sensitivity regulation (38). According to our insulin tolerance test results, losartan (10 mg · kg⁻¹ · day⁻¹) is unlikely to exert an effect on insulin sensitivity in our animal model.

RAS inhibition also seems to be beneficial for the islets in other animal models of type 2 diabetes, such as the Zucker diabetic fatty rat (40) and the Otsuka Long-Evans Tokushima fatty rat (41). These benefits seemed to be due to preservation of the islet architecture and β-cell mass. Indeed, RAS activation is closely associated with oxidative stress–induced cell loss/failure in type 2 diabetes (42,43), with particular emphasis on pancreatic islet RAS-induced oxidative stress (44). Nevertheless, the mechanistic pathways involved in AT1R-mediated islet cell dysfunction in type 2 diabetes have yet to be fully resolved.

In conclusion, the present study in db/db mice provides proof of principle that pancreatic islet AT1R, though not having any obvious effects on normal islet function, may become upregulated during certain conditions, and this upregulation appears to have deleterious effects on insulin release and (pro)insulin biosynthesis. These findings provide a novel and at least partial explanation for the reduced incidence of type 2 diabetes that has been observed in a number of clinical trials applying AT1R inhibition to individuals at high risk for this disease.

We gratefully acknowledge the support by the Competitive Earmarked Research Grant from the Research Grants Council of Hong Kong (Project Nos. CUHK4537/05M and CUHK4364/04M), by the Direct Grant of Research from the Chinese University of Hong Kong (Project No. 2041158), by the MERK-COZAAR Medical School Grant Program #25868 (Project No. TM042204) awarded to P.S.L., and by the Swedish Research Council (72XD-15043) and European Foundation for the Study of Diabetes/Novo. Losartan was generously provided by Merck, Whitehouse Station, New Jersey.