Reduction of Reactive Oxygen Species Ameliorates Metabolism-Secretion Coupling in Islets of Diabetic GK Rats by Suppressing Lactate Overproduction

Ebselen was purchased from Calbiochem (La Jolla, CA). HEPES, KCl, EGTA, glucose, NaCl, NaHCO₃, HClO₄, Na₂CO₃, H₂O₂, BSA, and the substrates used in ATP production, except glycerol phosphate, were purchased from Nacalai (Kyoto, Japan). [U-¹⁴C]-glucose was obtained from GE Healthcare (Uppsala, Sweden). Lactate dehydrogenase (EC 1.1.1.27) and Dowex 1 × 8 anion exchange resin (formate) (50–100 mesh) were obtained from Wako (Osaka, Japan). Hypoxia-inducible factor 1α (HIF1α) inhibitor (3-[2-(4-adamantan-1-yl-phenoxy)-acetylamino]-4-hydroxybenzoic acid methyl ester) was obtained from Merck Millipore (Darmstadt, Germany). All other reagents were obtained from Sigma-Aldrich (St. Louis, MO).

Male Wistar and GK rats were obtained from Shimizu (Kyoto, Japan). All experiments were carried out with rats 7–10 weeks of age. The body weight of GK rats used in the experiments was similar to that of Wistar rats (means ± SE; Wistar, 231 ± 3, vs. GK, 217 ± 3 g, not significant). The nonfasting plasma glucose of GK rats was higher compared with that of Wistar rats (means ± SE; Wistar, 5.72 ± 0.12, vs. GK, 11.57 ± 0.58 mmol/L, P < 0.01). The animals were maintained and used in accordance with the guidelines of the animal care committee of Kyoto University.

Pancreatic islets were isolated from Wistar and GK rats by collagenase digestion as described previously (16). Isolated islets were washed with Krebs-Ringer bicarbonate buffer (KRBB) containing 2.8 mmol/L glucose and cultured for 12 h in RPMI-1640 medium containing 10% FCS and 5.5 mmol/L glucose without or with 10 mmol/L tempol, 10 μmol/L ebselen, or combination of tempol and ebselen (TE treatment) at 37°C in humidified air containing 5% CO₂. Insulin content of GK islets was similar to that of Wistar islets (means ± SE; Wistar, 20.0 ± 2.2, vs. GK, 23.0 ± 2.6 ng/islet, not significant).

O₂⁻ and the H₂O₂ production were measured at the end of the 12-h culture in conditions described above. O₂⁻ generation was detected by nitroblue tetrazolium (NBT) assay (17) using KRBB supplemented with 0.2% BSA (fraction V) and 10 mmol/L HEPES adjusted to pH 7.4 (KRBB medium). Groups of 100 islets were batch incubated in tubes containing 0.5 mL KRBB medium supplemented with 5.5 mmol/L glucose with 0.2% (weight/volume) NBT at 37°C for 30 min. After the islets were centrifuged (1,000 rpm for 2 min at 4°C), the supernatant was removed and formazan-NBT (NBT reduced) was dissolved in 100 µL 50% (volume/volume) acetic acid by sonication (5-s pulse, five times). The sonicate was briefly centrifuged, and the absorbance of the supernatant was measured at 560 nm using a spectrofluorometer (Shimadzu RF-5000, Kyoto, Japan). H₂O₂ release from islets was measured according to the method previously described (18). In brief, islets were cultured for 12 h in various conditions as described above. Groups of 150 islets were batch incubated for 60 min in KRBB medium containing 5.5 mmol/L glucose with 0.1 mg/mL horseradish peroxidase (type II) and 0.44 mmol/L homovanillic acid. After the islets were centrifuged (1,000 rpm for 2 min at 4°C), the supernatant was removed. Fluorescence of the supernatant was measured at excitation and emission of 315 and 425 nm, respectively, using a spectrofluorometer (Shimadzu RF-5000). Standard curves were produced using samples containing known amounts of H₂O₂ without islets.

Insulin release from islets was monitored using batch incubation as described previously (16). ATP content in islets was determined as previously described (9). ATP contents were measured using ENLITEN luciferin-luciferase solution (Promega, Madison, WI) by luminometer (GloMax 20/20n; Promega). Glucose oxidation was carried out using the previously described method (19).

After GK rats were divided into two groups, one group received intraperitoneal injection of tempol (50 mg/kg) and ebselen (10 mg/kg) twice (12 and 6 h before intraperitoneal glucose tolerance tests [ip-GTTs]) during a 15-h fast (TE); the other group was treated similarly with vehicle alone. After these treatments, ip-GTTs (1 g/kg body weight) were performed. Plasma glucose and plasma insulin levels were measured in samples taken at the indicated times. Plasma glucose levels were determined by the glucose oxidase method. Plasma insulin levels were determined using an enzyme immunoassay (Shibayagi, Gunma, Japan).

The preparations for whole-islet lysate and for the mitochondrial fraction were described previously (20,21). Western blotting was performed as previously described (20) with the following antibodies: mouse monoclonal anti–complex I (39 kDa subunit), anti–complex III (core II), anti–complex IV (subunit I), and anti–complex V (subunit α) of the mitochondrial respiratory chain (1:1,000) from Invitrogen (Carlsbad, CA), mouse anti–β-actin (1:5,000) from Sigma-Aldrich, rabbit polyclonal anti–LDH-A (1:1,000) from Cell Signaling (Danvers, MA), rabbit anti–PDK-1 (1:1,000) and rabbit anti-HIF1α (1:250) from Abcam (Paris, France), rabbit anti-UCP2 (1:200) from Chemicom (Temecula, CA), mouse anti–hsp-60 (1:5,000) from BD Biosciences (Franklin Lakes, NJ), rabbit anti–PK-M2 (1:2,500) from Novus Biological, LLC (Littleton, CO), and horseradish peroxidase–conjugated anti-rabbit and -mouse antibody (1:5,000) from GE Healthcare. Proteins were detected using an enhanced chemiluminescence system (GE Healthcare). Protein density was quantified by densitometric analysis using Multi Gauge software (Fujifilm, Tokyo, Japan).

After islets were cultured with or without TE for 12 h, whole-islet homogenates and mitochondrial fractions from cultured islets were obtained. Glucokinase and hexokinase activities were determined using whole-islet homogenates by NADH formation in an enzyme reaction as previously described (22). The activity of mitochondrial glycerol phosphate dehydrogenase (mGPDH) in whole-islet homogenates of Wistar and GK rats was measured by the reduction of 2-p-iodo-3-p-nitro-5-phenyltetrezolium to indoformazan as previously described (23). The activity of mGPDH in the mitochondrial fraction from GK islets was measured by the method based on the generation of ³HOH from l-[2-³H]glycerol-3-phosphate (24). In brief, after mitochondrial fractions were incubated at 37°C for 30 min with 2.5 mCi/mmol l-[2-³H]glycerol-3-phosphate and the reaction was stopped by the addition of antimycin A, the mixture was immediately applied to a column of Dowex 1-X8 (formate), which was washed with 0.5 mL water. The radioactivity of the column effluent was measured by a liquid scintillation counter.

The mitochondrial fraction from islets and measurement of ATP production were performed as previously described (21).

Lactate production was measured as previously described (25). In brief, after preincubation, groups of 20 islets were batch incubated in KRBB medium containing 2.8 and 16.7 mmol/L glucose for 60 min at 37°C. The supernatant (200 µL) was mixed with 0.5 mL of 0.5 mol/L glycine/0.4 mol/L hydrazine buffer (pH 9.0) containing 2.4 mmol/L NAD and 7.2 units/mL lactate dehydrogenase, and was incubated for 30 min at 37°C. The fluorescence of NADH was then measured at an excitation of 340 nm and emission of 450 nm. Standard curves were produced from samples containing known amounts of lactate.

The data are expressed as means ± SE. Statistical significance was determined by unpaired Student t test. P < 0.05 was considered significant.

Tempol, an SOD mimic, decreased O₂⁻ production in Wistar (Fig. 1A) and GK islets (Fig. 1B) but had no effect on H₂O₂ production in either of the islets (Fig. 1C and D). Ebselen, a GPx mimic, had no reducing effect on O₂⁻ and H₂O₂ production in either of the islets (Fig. 1). Cotreatment with TE reduced O₂⁻ production by 27.0% but had no effect on H₂O₂ production in Wistar islets (Fig. 1A and C). On the other hand, in GK islets, TE treatment prominently reduced O₂⁻ and H₂O₂ production by 65.0 and 84.6%, respectively (Fig. 1B and D). Taken together, these findings indicate that TE treatment is effective in reducing oxidative stress in GK islets.

The effect of antioxidant mimics on β-cell function was investigated. In the presence of 16.7 mmol/L glucose, insulin secretion (IS) from GK islets was reduced compared with that from control Wistar rats (GK, 1.01 ± 0.06, vs. Wistar, 2.32 ± 0.11 ng/30 min/islet, P < 0.01). Although TE treatment had no effect on IS from Wistar islets (Fig. 2A), it restored high glucose–induced IS (Fig. 2B) and ATP content (Fig. 2C) in GK islets. We demonstrated previously that impaired glucose-induced IS in GK islets was significantly restored by acute exposure to PP2, a Src inhibitor (9). To examine the involvement of Src inhibition in the improvement of glucose-induced IS after TE treatment, TE-treated GK islets were incubated for 30 min with or without 10 μmol/L PP2 in the presence of 16.7 mmol/L glucose. In TE-treated GK islets, 30-min exposure to PP2 prominently enhanced IS at high glucose (control, 1.71 ± 0.07, vs. PP2, 4.06 ± 0.22 ng/30 min/islet, P < 0.01), which implies involvement of an independent mechanism of Src inhibition in the improvement of glucose-induced IS by chronic TE treatment.

To investigate the effects of TE treatment on β-cell function in vivo, a GTT was assessed after intraperitoneal TE administration (Table 1). In GK rats, TE treatment decreased the glucose level at 15 and 30 min and increased insulin levels at 30 min after glucose loading. Studies using isolated islets from GK rats after TE treatment in vivo revealed that high glucose–induced IS was increased by TE treatment (Table 2).

Immunoblot analysis showed that TE treatment had no effect on protein levels of UCP2 in protein extracts of the islet mitochondrial fraction (Supplementary Fig. 1A) and on those of complex I, III, IV, and V of the mitochondrial respiratory chain in lysates of whole islets (Supplementary Fig. 1B) in both GK and Wistar islets.

TE treatment had no effect on glucokinase and hexokinase activities, which determines the velocity of glycolysis in β-cells (26), in GK and Wistar islets (Table 3). By TE treatment, glucose oxidation at high glucose, an indicator of Krebs cycle velocity, was not affected in Wistar islets, whereas it was significantly increased in GK islets (control, 25.4 ± 4.1, vs. TE, 55.8 ± 12.2 pmol/islet/90 min, P < 0.05) (Fig. 3).

ATP production by mitochondria from control and TE-treated islets of Wistar and GK rats was measured in the presence of various substrates and inhibitors (Table 4), including succinate (electron transfer at complex I by NADH generation in the Krebs cycle and at complex II directly), succinate with rotenone (electron transfer at complex II and inhibition of electron transfer through NADH at complex I), glutamate plus malate (electron transfer mainly at complex I by NADH generation derived from reaction via glutamate dehydrogenase), pyruvate plus malate (electron transfer mainly at complex I by NADH generation derived from reaction via pyruvate dehydrogenase), ascorbate plus N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) (electron transfer at complex IV directly), and glycerol-3-phosphate (G3P) (electron transfer at complex II by FADH₂ generation derived from reaction via mGPDH). Control and TE-treated mitochondria showed similar rates of ATP production for all substrates tested, except for G3P. TE treatment promotes the rate of mitochondrial ATP production in the presence of G3P in Wistar and GK islets 3.3- and 2.2-fold, respectively. In the measurement of activity of mGPDH using whole islet homogenates, the value in GK islets was lower compared with that of Wistar islets, and enhancement of the activity by TE treatment was not significant in either Wistar or GK islets (Table 3). However, in the measurement of activity of mGPDH using intact mitochondria, TE treatment increased mGPDH activity in GK mitochondria (Table 3).

Lactate production in GK islets was significantly higher compared with that in Wistar islets at both basal and stimulated levels of glucose (at 2.8 mmol/L glucose: Wistar, 1.32 ± 0.15, vs. GK, 8.15 ± 1.36, P < 0.01; at 16.7 mmol/L glucose: Wistar, 3.09 ± 0.33, vs. GK, 14.08 ± 1.68, P < 0.01) (Fig. 4A). TE treatment suppressed lactate production at both basal and stimulated levels of glucose in both Wistar and GK islets; the most prominent reduction in these conditions was at 16.7 mmol/L glucose in GK islets (control, 14.08 ± 1.68, vs. TE, 5.71 ± 0.79, P < 0.01). Protein expression levels of HIF1α, a potential upstream regulator of lactate dehydrogenase (LDH), in INS-1 cells were gradually increased over 12 h in the presence of 200 μmol/L CoCl₂, a chemical inducer of HIF1α (27) (Supplementary Fig. 2). TE treatment time-dependently suppressed HIF1α levels in GK islets, and the reduction was significant after 9-h TE treatment (∼30% reduction at 9 h; ∼50% reduction at 12 h) (Fig. 4B). Protein expression levels of HIF1α and HIF1α downstream targets, including LDH-A and pyruvate dehydrogenase-1 (PDK-1), were examined 12 h after TE treatment. In GK islets, TE treatment significantly decreased HIF1α, LDH-A, and M2-PK expression (HIF1α, ∼50% reduction; LDH-A, ∼30% reduction; M2-PK, ∼30% reduction) but did not affect PDK-1 expression (Fig. 4C and Supplementary Fig. 3).

IS in the presence of 16.7 mmol/L glucose from GK islets was enhanced by HIF1α inhibitor treatment (control, 1.34 ± 0.07, vs. HIF1α inhibitor, 1.77 ± 0.09 ng/30 min/islet, P < 0.01). Lactate production in the presence of 2.8 and 16.7 mmol/L glucose was reduced by HIF1α inhibitor treatment (2.8 mmol/L glucose: control, 12.44 ± 1.12, vs. HIF1α inhibitor, 9.46 ± 0.89 pmol/islet, P < 0.01; 16.7 mmol/L glucose: control, 15.39 ± 1.02, vs. HIF1α inhibitor, 12.42 ± 1.40 pmol/islet, P < 0.01) (Fig. 5).

The current study demonstrates that combination treatment by tempol, an SOD mimic, and ebselen, a GPx mimic (TE treatment), efficiently suppressed both H₂O₂ and O₂⁻ production in GK islets. Moreover, impaired glucose-induced IS and ATP elevation in GK islets were significantly improved by TE treatment but were not improved by tempol or ebselen alone. These results are compatible with previous studies in which coexpression of the antioxidant enzymes in a β-cell line was more efficient than expression of either single enzyme in reducing ROS production and oxidative injury (7,28,29).

We then examined the precise mechanisms of the improvement in glucose-induced ATP elevation in GK islets by TE treatment. UCP2 decreases mitochondrial ATP production by reducing mitochondrial hyperpolarization derived from an increase in mitochondrial proton conductance. UCP2 negatively regulates metabolism-secretion coupling in β-cells (30,31). However, regulation of UCP2 expression may not play an important role in the mechanism, as the UCP2 expression level was not affected in GK islets by TE treatment. Western blot analysis revealed that the expression of respiratory chain proteins, including complex I, III, IV, and V, was not affected by TE treatment. In addition, the relative mRNA levels of nuclear factors, including NRF-1, NRF-2, and TFAM, which affect mitochondrial biogenesis, also were not altered by TE treatment (data not shown). These results indicate that the regulation of mitochondrial mass is not involved in the mechanism.

ATP production is driven by the H⁺ gradient across the mitochondrial membrane generated by the transport of high-energy electrons in the respiratory chain. These electrons are derived from NADH and FADH₂, derived from the Krebs cycle in the matrix and/or transferred from the cytosol by the shuttle system. To determine which step in mitochondrial carbohydrate metabolism is affected in TE-treated islets, ATP production in the presence of various substrates and inhibitors was measured using mitochondrial fractions from the cultured islets. In the presence of G3P, ATP production was increased by TE treatment in both the Wistar and GK mitochondrial fraction. Transfer of a reduced equivalent from cytosol to mitochondria by the glycerol phosphate shuttle may well play a more important role in mitochondrial ATP production in islets than in other tissues, since the activity of mGPDH, a key enzyme in the glycerol phosphate shuttle is ∼60-fold higher in islets than in other tissues (24). In addition, decreased activity is observed in diabetic islets (32). Although an increase in the activity of mGPDH by TE treatment was detected, the increase was similar in the mitochondrial fraction from both GK and Wistar islets. Therefore, an improvement of mGPDH activity alone may not account for the improvement in glucose-induced ATP elevation in GK islets by TE treatment.

TE treatment did not affect the activity of glucokinase, a rate-limiting enzyme in glycolysis (26). In contrast, glucose oxidation at high glucose, which reflects Krebs cycle velocity, was improved in GK islets by TE treatment. These results suggest that coupling between glycolysis and mitochondrial oxidation, which plays an important role in glucose-induced IS (33), is improved by TE treatment. Lactate production weakens coupling between glycolysis and mitochondrial oxidation by decreasing the pyruvate supply to mitochondria. In several studies, the overexpression of LDH-A in β-cells attenuated glucose-induced IS (34–36) with an increase in lactate production (34,37). In addition, lactate release from β-cells increases in parallel with a decrease in the glucose oxidation rate (36). These findings support the hypothesis that regulation of lactate production may play an important role in the improvement of metabolism-secretion coupling by TE treatment. In the current study, lactate production in GK islets was higher compared with that in control, as shown in a previous study (38), and it was reduced by TE treatment. The protein expression level of LDH-A, an enzyme that catalyzes the reaction from pyruvate and NADH to lactate and NAD⁺, was 8.6-fold higher in GK islets compared with control Wistar islets, and was reduced by ∼30% after TE treatment. In nondiabetic pancreatic β-cells, low LDH activity together with high mGPDH activity may act as a device to transfer NADH in the cytosol to the mitochondria and reduce lactate production by decreasing the NADH level in cytosol (39). In GK islets, enhanced LDH activity and reduced mGPDH activity might produce lactate overproduction and reduce mitochondrial oxidation by decreasing the supply of pyruvate and reducing equivalents to mitochondria and by an increase in the availability of cytosolic NADH to produce lactate from pyruvate as well. In this case, reduced LDH activity and increased mGPDH activity by TE treatment may well have decreased lactate overproduction and ameliorated metabolism-secretion coupling in GK islets.

These metabolic profiles in GK islets resemble the Warburg effect, characteristic aerobic metabolism in cancer cells by which lactate is overproduced from glucose even under nonhypoxic conditions (40). HIF1α is a transcription factor that functions as a master regulator of oxygen homeostasis and upregulates the expression of LDH (41). HIF1α activity is controlled at the level of protein stability. When oxygen is present, two proline residues in the oxygen-dependent degradation domain of HIF1α are hydroxylated by the prolylhydroxylase enzymes (PHDs) (42). These hydroxylated motifs allow capture by the von Hippel-Lindau protein that forms the recognition component of an E3 ubiquitin ligase complex, which targets the hydroxylated HIF1α for polyubiquitination and proteasomal degradation (43). In the current study, TE treatment showed a time-dependent decrease in HIF1α protein levels in GK islets and 50% reduction compared with that in nontreated GK islets at 12-h culture, which could reduce LDH-A protein in TE-treated GK islets. The Warburg effect is usually thought of as a high rate of glycolysis. Since TE treatment did not affect glucokinase activity, which determines glycolytic velocity in β-cells, alteration in the HIF1α level in the pathophysiological range may not affect glycolytic velocity in β-cells. In this respect, lactate overproduction in GK islets differs from the typical Warburg effect.

HIF1α downregulates the expression of GLUT2 and glucokinase and upregulates PDK-1, which inactivates pyruvate dehydrogenase and suppresses metabolism in the Krebs cycle (44) in islets of von Hippel-Lindau gene (vhl)-deficient mice (45). On the contrary, TE treatment did not affect the glucokinase activity and protein level of PDK-1 in GK islets. This discrepancy is not fully understood, but one possibility is that the expression level of HIF1α may be higher than the pathophysiological level in vhl-deficient islets. This is supported by the fact that glucose-induced IS is impaired severely and almost abolished in vhl-deficient islets. The transcript level of LDH-A in vhl-deficient islets is about 20-fold higher compared with control islets, which is much higher than that of PDK-1 (∼5-fold) (46). This suggests that the expression of LDH-A is more readily affected by HIF1α expression than that of PDK-1.

An increase in the basal HIF1α level in GK islets was not found compared with that in Wistar islets. In contrast, basal levels of LDH-A in GK islets was prominently increased compared with that in Wistar islets, suggesting that regulators other than HIF1α may play an important role in the enhancement of basal LDH-A in GK islets. However, our finding in the current study remains valid, as ROS-dependent regulation of LDH-A by HIF1α, which plays an important role in lactate overproduction and impaired glucose-induced IS in GK islets, was established. In addition, the involvement of HIF1α in lactate overproduction was shown clearly by the finding that HIF1α inhibitor treatment restored high glucose–induced IS and suppressed lactate production in GK islets. ROS inhibits PHD activity and stabilizes HIF1α (47,48). In GK islets, an excess amount of ROS may inhibit PHD activity and stabilize the HIF1α protein. Thus, TE treatment may recover PHD activity by reducing the ROS level and decreasing the HIF1α protein level to eventually ameliorate metabolism-secretion coupling in GK islets.

In conclusion, we show that 12-h suppression of endogenous ROS, including O₂⁻ and H₂O₂, improves impaired metabolism-secretion coupling in GK islets. Our results suggest that lactate overproduction plays an important role in impaired metabolism-secretion coupling in diabetic β-cells; ROS reduction improves mitochondrial metabolism by suppressing lactate overproduction through the inhibition of HIF1α stabilization.

This study was supported by a Research Grant on Nanotechnical Medicine from the Ministry of Health, Labour and Welfare of Japan; scientific research grants; a grant from the Innovation Cluster Kansai project of the Ministry of Education, Culture, Sports, Science and Technology of Japan; and a grant from CREST of the Japan Science and Technology Cooperation.

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

M.S. researched data, contributed to the discussion, wrote the manuscript, and reviewed and edited the manuscript. S.F. contributed to the discussion, wrote the manuscript, and reviewed and edited the manuscript. Y.S., Y.N., E.M., G.Y., H.S., Y.T., and K.O. researched data. K.N. and N.I. contributed to the discussion, and reviewed and edited the manuscript. S.F. 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 the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 72nd Scientific Sessions of the American Diabetes Association, Philadelphia, Pennsylvania, 8–12 June 2012.

The authors thank C. Kotake and M. Akazawa (Kyoto University) for technical assistance.