Inhibition of Mitochondrial Na⁺-Ca²⁺ Exchanger Increases Mitochondrial Metabolism and Potentiates Glucose-Stimulated Insulin Secretion in Rat Pancreatic Islets

Pancreatic islets from adult male Sprague Dawley (SD) rats (250–300 g; Harlan, Indianapolis, IN) were isolated using collagenase (type P; Roche Molecular Biochemicals, Indianapolis, IN) as described previously (36). All animal procedures were approved by the institutional animal care and use committee. Isolated islets were cultured for 2–7 days in CMRL-1066 medium (Sigma, St. Louis, MO) containing 5.5 mmol/l glucose, 10% fetal bovine serum (FBS), penicillin (100 units/ml), and streptomycin (100 μg/ml) at 5% CO₂-95% air at 37°C as described previously (36). INS-1 cells were maintained in RPMI-1640 medium (Mediatech, Herndon, VA) supplemented with 10% FBS, 2-mercaptoethanol (0.0004%), HEPES (10 mmol/l), pyruvate (1 mmol/l), penicillin (100 units/ml), and streptomycin (100 μg/ml) at 5% CO₂-95% air at 37°C.

mNCE activity was measured in INS-1 cells that were treated with digitonin to permeabilize the plasma membrane without affecting the mitochondrial membranes. INS-1 cells in suspension were incubated with digitonin (0.007%) in sucrose buffer (250 mmol/l sucrose, 2.5 mmol/l KH₂PO₄, 10 mmol/l HEPES, pH 7.4) containing 1 μmol/l rotenone, 1 μmol/l cyclosporin A, 0.05 μmol/l calcium green, and 5 mmol/l succinate. The mitochondria were then loaded with 60 μmol/l calcium and 1 μmol/l ruthenium red. The permeabilized cell suspension was transferred to a 96-well plate in the absence or presence of CGP37157. After addition of NaCl (20 mmol/l), the rate of change in fluorescence was measured in an Fmax fluorescence plate reader, using wavelengths of 485 nm for excitation and 538 nm for emission (Molecular Devices, Sunnyvale, CA).

Measurements were performed as described previously by Maechler et al. (37). Briefly, INS-1 cells that express mitochondrially targeted aequorin (INS-1/EK-3 cells) were seeded onto glass coverslips and maintained for 2–3 days in RPMI-1640 containing 10% FBS. The cells were incubated with 2.5 μmol/l coelenterazine (Calbiochem, San Diego, CA) in Krebs-Ringer bicarbonate (KRB) buffer (135 mmol/l NaCl, 3.6 mmol/l KCl, 10 mmol/l HEPES, 5 mmol/l sodium bicarbonate, 0.5 mmol/l NaH₂PO₄, 0.5 mmol/l MgCl₂, 1.5 mmol/l CaCl₂, and 0.1% BSA, pH 7.4) containing 2.8 mmol/l glucose for 2 h at 37°C before placement into a thermostatically controlled chamber for perifusion. Luminescence was measured using a photomultiplier tube/photon-counting apparatus (model P10232; Electron Tubes, Rockaway, NJ), and [Ca²⁺]_m concentration was calculated from the luminescent readings as described previously (37).

Islets were dispersed using EDTA and trypsin as described previously (36). The imaging procedure has been described (38). Briefly, dispersed islet cells were plated onto glass coverslips and incubated at 37°C in CMRL-1066 culture medium. Cell viability as assessed by trypan blue exclusion was 95–98% for each of the islet cell groups studied. INS-1 cells, single islet cells, or islets on glass coverslips were loaded with fura-2/AM (1 μmol/l; Molecular Probes, Eugene, OR) for 20 min in KRB buffer (pH 7.4), then transferred to a closed perifusion chamber (Bioptechs, Butler, PA) on an inverted microscope. Fluorescence was measured by dual-excitation fluorimetry, with excitation and emission wavelengths set at 340/380 and 510 nm, respectively. Images were taken every 2 s for INS-1 cells and single islet cells and 20 s for islets.

For measurement of plasma membrane sodium-calcium exchanger (NCX) activity, NaCl was iso-osmotically replaced by sucrose, and NaHCO₃ was replaced by choline bicarbonate. To avoid cholinergic effects, the Na⁺-free buffer contained atropine (10 μmol/l). Cytosolic calcium ([Ca²⁺]_i) was calculated as described previously (39). KB-R7943 (Tocris, Ballwin, MO), an inhibitor of the NCX, was used as a positive control.

NAD(P)H autofluorescence was measured as described by Staddon and Hansford (40). INS-1 cells were starved in glucose-free KRB buffer for 1 h at 37°C, then harvested by trypsinization. The cells were resuspended in KRB buffer, and fluorescence was measured using an excitation wavelength of 340 nm and an emission wavelength of 470 nm in a Shimadzu RF-5301PC Spectrofluorophotometer (Shimadzu, Columbia, MD). Maximal fluorescence changes were monitored after the addition of 5 μmol/l rotenone and 20 mmol/l glucose.

INS-1 cells were starved for 30 min in glucose-free KRB buffer containing 1 μmol/l CGP37157 or vehicle and subsequently stimulated with various concentrations of glucose for 15 min at 37°C. For measurement of ATP in islets, islets were incubated with 15 mmol/l glucose for 30 min as described previously (41). The islets were then incubated for 30 min in glucose-free KRB buffer before addition of 1 μmol/l CGP37157 or vehicle and glucose for 30 min. ATP content was measured using Cell Titer Glo reagent (Promega, Madison, WI) according to the manufacturer’s instructions.

For measurement of reactive oxygen species (ROS) by dichlorofluorescein-diacetate (DCFC-DA) oxidation (Molecular Probes), INS-1 cells were suspended in glucose-free KRB buffer and incubated 1 h at 37°C. DCFC-DA (100 μmol/l) was added to all cells; compound and glucose were added as appropriate. After 2 h at 37°C, the fluorescence was read at an excitation wavelength of 485 nm and an emission wavelength of 530 nm. For measurement by Amplex-Red oxidation, INS-1 cells were incubated for 2 h with compounds and glucose before addition of the Amplex reagent according to the manufacturer’s instructions (Molecular Probes). Fluorescence was measured at an excitation wavelength of 530 nm and emission wavelength of 590 nm.

For static measurements, isolated islets were preincubated in oxygenated KRB buffer (pH 7.4) containing 16 mmol/l HEPES, 0.01% BSA, and 5.5 mmol/l glucose for 60 min in the absence or presence of CGP37157, followed by a 20-min treatment in fresh buffer with glucose and CGP37157. Insulin release was expressed as the difference between buffer insulin content at 20 min and at time 0. INS-1 cells were starved for 2 h in glucose-free RPMI-1640 medium containing 1% FBS before measurement of insulin secretion in the presence of glucose and CGP37157 over a 30-min period at 37°C. For islet perifusion, groups of 100 islets were preperifused for 60 min with oxygenated KRB buffer (pH 7.4) containing 2.8 mmol/l glucose, 0.1 mmol/l 3-isobutyl-1-methylxanthine, and 0.5% BSA at a rate of 0.5 ml/min. Fractions were collected every 2 min for insulin measurements by insulin enzyme-linked immunosorbent assay kit (Crystal Chem, Chicago; ALPCO, Windham, NH) using rat insulin as standards.

CGP37157 was administered intravenously at 1 mg/kg or orally at 1, 10, and 30 mg/kg to adult male SD rats (250 g; Harlan). Blood samples were collected at various times after administration. Plasma CGP37157 concentrations were measured by liquid chromatography/mass spectrometry.

Male Wistar rats (Charles River Laboratories, Worcester, MA), weighing 250–300 g, were housed individually under controlled light (12/12 h) and temperature conditions and given access to diet and water ad libitum. All procedures were in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Animal Subjects Committee of the University of California, San Diego. After a 2- to 3-day acclimation period, two catheters were placed in the right jugular vein (Micro-Renathane MRE-033, 0.033 in outer diameter and 0.014 in inner diameter; Braintree Scientific, Braintree, MA), and another was placed in the left carotid artery under general anesthesia. The anesthetic cocktail consisted of ketamine HCl (50 mg/kg) (Fort Dodge Animal Health, Fort Dodge, IA), acepromazine maleate (1 mg/kg) (Butler, Columbus, OH), and xylazine (4.8 mg/kg) (Butler) given intramuscularly. Four days after catheter placement, animals were randomly assigned to two groups. CGP37157 or vehicle (10% ethanol, 30% polyethylene glycol-400, 15% propylene glycol, and 45% water) was infused at 3.5 mg · kg⁻¹ · h⁻¹ for 5 h after a loading dose of 4.3 mg/kg. A glucose infusion (25 mg · kg⁻¹ · min⁻¹) was then begun. Blood samples (200 μl) were collected into tubes containing aprotinin (250 KIU/ml; Calbiochem) at the beginning of the experiment, before the glucose infusion, and at intervals during the glucose infusion. Blood glucose was measured using a B-Glucose Analyzer (HemoCue, Mission Viejo, CA). Plasma insulin concentrations were measured using a rat insulin enzyme-linked immunosorbent assay. Plasma C-peptide concentrations were measured using a radioimmunoassay (Linco, St. Charles, MO).

Data are given as means ± SE. The differences between treatment groups were analyzed by the Student’s t test or one-way ANOVA, with post hoc analysis using the Student-Newman-Keuls multiple comparison test. For animal studies, two-way ANOVA was used. P < 0.05 was accepted as significant.

To ensure that mNCE activity was present in INS-1 cells and to characterize the sensitivity of the activity to CGP37157, a fluorescent assay was used to measure Na⁺-dependent Ca²⁺ efflux from INS-1 mitochondria. CGP37157 (0.01–100 μmol/l) inhibited mNCE activity in digitonin-permeabilized INS-1 cells in a dose-dependent fashion (Fig. 1). Half-maximal inhibition was observed at 1.5 μmol/l, and complete inhibition of the mNCE activity was achieved at 10 μmol/l.

The resting [Ca²⁺]_m in INS-1/EK-3 cells ranged from 100 to 200 nmol/l (Fig. 2A). Glucose stimulation of the cells elicited an immediate and transient increase in [Ca²⁺]_m (Fig. 2A), as previously described by others (17). When the cells were treated with CGP37157, the peak increase was higher and the duration longer (Fig. 2A) than in control cells, such that the area under the curve (AUC) was increased 2.1 times by the mNCE inhibitor. Glucose increased [Ca²⁺]_m in INS-1/EK-3 cells in a dose-dependent manner (Fig. 2B). The glucose dose-response curve was shifted to the left, and the maximal [Ca²⁺]_m increased slightly by CGP37157 (1 μmol/l).

Acute changes in the NADH content of cells can be used as an indicator of the flux through the Krebs cycle. To determine whether inhibition of the mNCE enhanced Krebs cycle flux, NADH autofluorescence was measured in INS-1 cells under a variety of conditions. Basal autofluorescence was unchanged in INS-1 cells after the addition of glucose-free KRB buffer to the cells (Table 1). Glucose increased NADH autofluorescence in a dose-dependent manner. Preincubation of the cells with 0.1 or 1 μmol/l CGP37157 had no effect on basal autofluorescence but enhanced the glucose-induced rise in NADH in the presence of 7 or 15 mmol/l glucose (Table 1). These observations show that CGP37157 increases Krebs cycle flux when glucose is supplied but not in the absence of oxidizable substrate.

CGP37157 increased the glucose-induced rise in ATP content in INS-1 cells and rat islets (Table 2). In INS-1 cells, glucose alone caused a concentration-dependent increase in cellular ATP content. Addition of CGP37157 augmented the rise in ATP by 13% at 3.5 and 5.5 mmol/l glucose but had no effect in the absence of glucose. Glucose also increased ATP production in rat islets in a dose-dependent manner. Addition of CGP37157 (1 μmol/l) further increased ATP production by 45% in the presence of 5.5 mmol/l glucose and 19% at 8 mmol/l glucose.

To determine whether inhibition of the mNCE by CGP37157 increased ROS production, INS-1 cells were incubated in the presence of CGP37157 and glucose in various concentrations. In some culture dishes, antimycin was added to enhance ROS production. ROS production was unchanged by CGP37157, either when measured by Amplex-Red oxidation, which measures primarily H₂O₂ production, or by DCFC-DA oxidation, which is a more general indicator of ROS (Table 3). Antimycin-mediated ROS production was also unaffected by CGP37157.

Inhibition of the mNCE with CGP37157 increased insulin secretion from rat islets in both static and perifusion experiments (Fig. 3). In the presence of 5.5 mmol/l glucose, CGP37157 had no significant effect on insulin secretion at concentrations <10 μmol/l. In contrast, when incubations were performed with 8 mmol/l glucose, CGP37157 at 10 μmol/l increased GSIS to 166% of the basal level. The stimulatory effect of CGP37157 on GSIS exhibited a half-maximally effective concentration (EC₅₀) of 0.06 ± 0.02 μmol/l, and the increase was maximal at 0.1 μmol/l CGP37157 (Fig. 3A). When the CGP37157 concentration was held constant (0.1 μmol/l) and the glucose concentration varied, CGP37157 enhanced GSIS and shifted the glucose dose-response curve to the left (Fig. 3B). CGP37157 did not increase GSIS beyond the maximal effect observed with 30 mmol/l glucose. CGP37157 also enhanced insulin secretion in a dose-dependent manner in INS-1 cells (Fig. 3C). In the absence of glucose, CGP37157 at concentrations <10 μmol/l did not change insulin secretion. However, CGP37157 increased GSIS (8 mmol/l) to 156% of the control level at 0.1 μmol/l and 145% at 1 μmol/l.

In perifused rat islets, CGP37157 (0.1 μmol/l) did not change insulin secretion in basal glucose (2.8 mmol/l) but enhanced GSIS (AUC = 146% of glucose alone; n = 5) (Fig. 4). When islets were perifused with glucose-free buffer, CGP37157 (0.1 μmol/l) did not change insulin secretion (data not shown).

To determine whether CGP37157 alters [Ca²⁺]_i in pancreatic β-cells, [Ca²⁺]_i was measured in INS-1 cells and in single islet cells. CGP37157 induced an increase in [Ca²⁺]_i in the form of discrete transients and/or spikes in single islet cells (Fig. 4A; representative tracings from four experiments) and INS-1 cells (Fig. 4B). The effects of CGP37157 were reversible upon its removal from the perfusate and upon restimulation with tolbutamide (0.5 mmol/l), which increased [Ca²⁺]_i in single islet cells and in INS-1 cells (Fig. 4A and B). Pretreatment of islet cells with potassium cyanide (KCN) (100 μmol/l), an inhibitor of mitochondrial electron transport chain complex IV, abolished CGP37157-induced increases in [Ca²⁺]_i in single islet cells (Fig. 4C; representative tracing from four experiments).

NCX activity was measured in intact islets by removing sodium from the external medium, forcing the exchanger to function in reverse mode to expel sodium from the cells in exchange for calcium. NCX activity, as measured by AUC during sodium removal, remained constant in untreated cells during three successive measurements (Fig. 5A). Nimodipine decreased the AUC to 76 ± 5.6% of the control value, indicating that calcium uptake in this system was mediated in part by l-type calcium channels (Fig. 5B). KB-R7943, a specific inhibitor of the NCX, eliminated the residual cytosolic calcium increase (Fig. 5B). In contrast, CGP37157 had no effect on cytosolic calcium uptake compared with that in the presence of nimodipine alone (Fig. 5C; 76 ± 1.0 vs. 73 ± 1.2% of initial AUC, respectively). Hence, CGP37157 had no effect on plasma membrane NCX activity.

GSIS was measured in vivo by performing hyperglycemic clamp studies in normal Wistar rats. CGP37157 has low solubility, has a plasma half-life of 0.9 h, and is not orally bioavailable in rats (data not shown). Therefore, the compound was infused intravenously for 5 h to achieve sustained plasma concentrations of 1.6 ± 0.2 μmol/l. After the 5-h compound infusion, glucose was infused through a second intravenous catheter at a rate (25 mg · kg⁻¹ · min⁻¹) previously determined to yield blood glucose levels of 200 mg/dl. During a 50-min glucose infusion, blood glucose levels did not differ between vehicle- and CGP37157-infused animals (Fig. 6A). Plasma insulin (Fig. 6B) and plasma C-peptide (Fig. 6C) concentrations were significantly increased by CGP37157 during the hyperglycemic portion of the experiment but were unchanged at basal glycemia. The maximal increase in plasma insulin was observed at 25 min after the glucose infusion was begun.

We used CGP37157, an inhibitor of the mNCE, to show that increasing [Ca²⁺]_m in response to a glucose load enhances GSIS both in vitro and in vivo. In particular, CGP37157 inhibited mNCE activity in INS-1 cells, increased the glucose-induced [Ca²⁺]_m transient in INS-1 cells, and increased the glucose-induced rise in cellular ATP concentration in both INS-1 cells and rat islets. The compound also stimulated GSIS in a dose- and glucose-dependent manner in islets and in INS-1 cells. Finally, infusion of CGP37157 into normal rats increased plasma insulin and C-peptide levels during hyperglycemic clamp experiments. These observations illustrate that increasing [Ca²⁺]_m through mNCE inhibition activates flux through the Krebs cycle, thereby increasing ATP production, which in turn augments GSIS.

The coordinate rise of [Ca²⁺]_m and cellular ATP in response to glucose in cell culture models of the pancreatic β-cell has been well documented (17,19,28,37), as has the stimulatory effect of [Ca²⁺]_m on ATP production. Activation of pyruvate dehydrogenase, isocitrate dehydrogenase, and α-ketoglutarate dehydrogenase by calcium has been described in several tissues, including heart, liver, and some endocrine tissues (42–45), and occurs at [Ca²⁺]_m in the approximate range of 0.1–2.0 μmol/l (46–48). The peak [Ca²⁺]_m achieved in the present study in glucose- and glucose plus CGP37157-treated cells was within this range, consistent with our conclusion that the elevated [Ca²⁺]_m was responsible for the enhanced Krebs cycle flux and oxidative ATP production that we observed.

The increase in [Ca²⁺]_m in the present study occurred mainly as an increase in the peak concentration, with a smaller effect on the duration of the calcium transient. Maechler et al. (37) reported that addition of CGP37157 to permeabilized INS-1 cells in the presence of succinate prolonged the [Ca²⁺]_m transient but did not increase the peak [Ca²⁺]_m. The slight difference in results compared with the present study is likely due to the difference in substrate and to the use of permeabilized cells in the previous study. Using cardiac myocytes, Cox and Matlib (23) demonstrated that inhibition of the mNCE with CGP37157 increased [Ca²⁺]_m to a magnitude similar to that observed in the INS-1 cells in the present study. The increase in [Ca²⁺]_m in those studies was also associated with increased Krebs cycle flux and increased oxidative phosphorylation. We have extended the observation that mNCE inhibition increases oxidative metabolism to show that a downstream physiological consequence, namely GSIS, was also stimulated in islets and INS-1 cells. Although we did not measure [Ca²⁺]_m in islets in the present study, the similar increases in insulin secretion and cellular ATP in INS-1 cells and islets incubated with CGP37157 suggest that mNCE inhibition acts by increasing [Ca²⁺]_m in islets as well as in INS-1 cells.

Most of the effects of mNCE inhibition by CGP37157 in β-cells occurred only in the presence of supraphysiological glucose. Neither NADH production nor cellular ATP changed when glucose concentrations were <5.5 mmol/l. Importantly, insulin secretion was also unaffected at low glucose concentrations in INS-1 cells and in islets. During perifusion experiments, CGP37157 had no effect on insulin release at 2.8 mmol/l glucose but stimulated insulin secretion significantly at 11 mmol/l glucose. Similarly, in static incubations, CGP37157 did not affect insulin secretion from islets in the presence of 5.5 mmol/l glucose until the compound concentration was ≥10 μmol/l. Finally, during the 5-h intravenous infusions of CGP37157 in rats, plasma insulin levels were unaffected until hyperglycemia was induced by glucose infusion. We predicted that the secretagogue effect of CGP37157 would be glucose dependent because enhanced oxidative metabolism increases ATP production only when oxidizable substrate is available to the Krebs cycle. At glucose concentrations <5 mmol/l, little flux occurs through glucokinase, which determines the glucose concentration at which insulin secretion is initiated (49). As the glucose concentration rises above the glucokinase threshold, glycolytic metabolism of glucose produces pyruvate, providing substrate for the Krebs cycle and initiating the feed-forward regulatory cycle of [Ca²⁺]_m. Hence, the small increase in [Ca²⁺]_m that was induced by CGP37157 in the absence of glucose (Fig. 2B) would not be expected to increase the Krebs cycle flux, ATP, or GSIS because insufficient oxidizable substrate was present.

The CGP37157-induced increases in cytosolic calcium ([Ca²⁺]_i) in the present study are consistent with the proposed feed-forward mechanism of [Ca²⁺]_m in insulin secretion. That is, increased oxidative production of ATP in the presence of the mNCE inhibitor causes further depolarization of the plasma membrane and increases voltage-dependent calcium channel-mediated calcium influx into the cytoplasm. We demonstrated that, because the mitochondrial poison KCN abolished the CGP37157-mediated increase in [Ca²⁺]_i, the change in [Ca²⁺]_i depended on mitochondrial oxidative metabolism (Fig. 5C). To address the specificity of CGP37157 for the mNCE, we measured its effects on the NCX (50), an activity that is pharmacologically distinct from the mNCE. GP37157 had no measurable effect on the islet NCX, as predicted by earlier measurements that showed no effect of CGP37157 on the activity of the NCX in cardiac myocytes (35). CGP37157 had no effect on the activity of the calcium uniporter or in binding displacement assays of l-type calcium channels or K_ATP channels (data not shown). It is therefore unlikely that CGP37157 increased cytosolic calcium by blocking the NCX, inhibiting the calcium uniporter, blocking K_ATP channels, or activating l-type calcium channels.

A potential concern associated with enhancing flux through the respiratory chain is increased production of ROS, which are produced as byproducts of the electron transfer chain, primarily from complex I and ubiquinone (51). CGP37157 treatment did not increase in ROS production in INS-1 cells under conditions that enhanced oxidative metabolism, probably because compensatory metabolism of the radicals by the cellular antioxidant systems occurred. Because the β-cells, which have relatively poor antioxidant capacity compared with other tissues (52), showed no ROS accumulation, it is unlikely that other tissues would accumulate ROS during systemic administration of an mNCE inhibitor.

The tissue distribution of the mNCE and its relative contribution to [Ca²⁺]_m flux in various tissues are incompletely defined. A single report claimed to have purified the mNCE protein from cardiac tissue (53). However, neither the protein sequence nor the gene that encodes the protein has been published. Based on activity measurements, the mNCE is highly expressed in heart and muscle tissues, is in low abundance in liver (54), and is intermediate in neuronal tissue (31,55,56). However, net calcium efflux from mitochondria is determined not only by the mNCE, but by a second non-sodium-dependent exchanger as well (30,54,55). Hence, although systemic administration of an mNCE inhibitor would almost certainly have effects in tissues other than the pancreatic β-cell, the significance of those effects would depend on the relative activity of the mNCE, the activity of the non-sodium-dependent calcium egress pathway, and the possible existence of tissue-specific isoforms of the mNCE that possess different sensitivities to inhibitor compounds. Further studies are required to address nonpancreatic effects of mNCE inhibition. In addition, future studies will be required to define whether inhibition of the mNCE affects specific intracellular ATP pools (27) or alters the periodicity of calcium fluxes.

In conclusion, a novel glucose-dependent insulin secretagogue mechanism has been described. Using a prototype compound to inhibit the mNCE, we have shown that modulation of mitochondrial oxidative metabolism can favorably alter insulin secretion. The mNCE represents a novel target for type 2 diabetes drug discovery efforts.

The authors thank Dr. Rosario Rizutto for helpful discussions about [Ca²⁺]_m measurement and Dr. Neil Howell for critically reading the manuscript.