l-Cysteine Inhibits Insulin Release From the Pancreatic β-Cell : Possible Involvement of Metabolic Production of Hydrogen Sulfide, a Novel Gasotransmitter

Pancreatic islets were isolated from male ICR mice (SLC, Hamamatsu, Japan) aged 8–12 weeks by the collagenase digestion method. The solution used for the isolation was Krebs-Ringer bicarbonate HEPES buffer (KRBH) containing 119 mmol/l NaCl, 4.75 mmol/l KCl, 5 mmol/l NaHCO₃, 2.54 mmol/l CaCl₂, 1.2 mmol/l MgSO₄, 1.2 mmol/l KH₂PO₄, and 20 mmol/l HEPES (pH 7.4 adjusted with NaOH) supplemented with 3 mmol/l glucose. All experiments using laboratory mice in this study were approved by the University Committee on Animal Experiments in accordance with the Guideline for Animal Experimentation in Oita University.

Total RNA was extracted from various tissues of mice and the β-cell line MIN6 using an RNeasy Total RNA kit (Qiagen, Hilden, Germany). Reverse transcription and PCR were performed using a One-Step PCR kit (Qiagen) with the gene-specific primers (CSE forward, 5′-ATGGATGAAGTGTATGGA GG-3′; CSE reverse, 5′-ACGAAGCCGACTATTGAGGT-3′; CBS forward, 5′-ACTA CGATGACACCGCCGAG-3′; and CBS reverse, 5′-AGTCCTTCCTGTGCGATGAG-3′) and primers for glyceraldehyde-3-phosphate dehydrogenase as an internal control (forward, 5′-AACGACCCCTTCATTGAC-3′; and reverse, 5′-TCCACGACATACTCAGCAC-3′). mRNA samples obtained from mouse liver and kidney were used as positive expression controls for both CSE and CBS. The PCR products were separated by 1.5% agarose gel electrophoresis followed by staining with ethidium-bromide, and the intensity of each band was analyzed with the NIH image software.

The anti-CBS antibody used in this study was raised as described by Enokido et al. (13). Isolated islets were sonicated in an ice-cold homogenization buffer consisting of 20 mmol/l Tris-HCl (pH 7.4), 2 mmol/l EDTA, 2 mmol/l 2-mercaptoethanol, 50 μg/ml phenylmethylsulfonylfluoride, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 250 mmol/l sucrose. Protein samples were boiled in a sample buffer containing 62.5 mmol/l Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, and 0.002% bromophenol blue. They were then separated on a 10% polyacrylamide gel and transferred to polyvinylidine fluoride membrane. The membrane was blocked with 5% skim milk in a buffer containing 10 mmol/l Trizma base, 137 mmol/l NaCl, and 0.1% Tween 20 (Tris-buffered saline with Tween [TBST]) and incubated overnight with an anti-CBS antibody (diluted ×1,000 in TBST). After washings in TBST, the membrane was further incubated with a horseradish peroxidase–conjugated donkey anti-rabbit IgG (1:25,000; Amersham BioSciences, Piscataway, NJ), and immunopositive bands were visualized by a commercialized kit (ECL Plus Western Blotting Reagent Pack; Amersham Biosciences).

MIN6 cells (donated by Prof. J.-I. Miyazaki, Osaka University, Osaka, Japan) were cultured in Dulbecco’s modified Eagle’s medium (Sigma, St. Louis, MO) supplemented with 66 mg/l kanamycin sulfate (Sigma) and 15% fetal bovine serum (JRH Biosciences, Lenexa, KA) at 37°C in a humidified atmosphere of 95% air/5% CO₂. They were passaged and harvested using trypsin/EDTA. The culture medium was replaced every other day.

Groups of five size-matched islets were preincubated in 1 ml KRBH with 0.1% BSA (fraction V; Sigma) and 3 mmol/l glucose for 1 h at 37°C. These islets were then incubated for 1 h in various conditions. The solutions containing NaHS (Sigma) were prepared by diluting 1 mol/l NaHS stock solution just before its application. For MIN6 cells, the cells were seeded onto a 24-multiwell plate (Nalge-Nunc, Rochester, NY) at a density of 1 × 10⁵ cells per well. After 2 days of culturing in the culture medium described above, the secretion experiments were carried out as for isolated islets. At the end of the incubation, incubation media were collected. For MIN6 cells, the aliquots were briefly centrifuged to avoid any detached cells. Samples were kept at −20°C for later assay, and the amount of insulin in the media was measured by a radioimmunoassay kit (Shionogi, Osaka, Japan) using bovine insulin as a standard.

Experiments of insulin secretion from permeabilized islets were carried out as previously described (14). Groups of five size-matched islets were preincubated as described above. After preincubation, these islets were washed twice with 1 ml glutamate buffer containing 100 mmol/l potassium glutamate, 42 mmol/l sodium glutamate, 16 mmol/l HEPES (pH 7), 5 mmol/l BSA, 1 mmol/l EGTA, and 3 mmol/l MgATP (Sigma). The islets were then permeabilized by coincubation with 0.125 IU/ml streptolysin-O (SLO) (Nissui Pharmaceutical, Tokyo, Japan) in the same glutamate buffer, under various conditions. At the end of the 1-h incubation, incubation media were retained, and the amount of insulin in the media was measured by radioimmunoassay.

Experiments for the measurement of the intracellular Ca²⁺ concentration ([Ca²⁺]_i) using dispersed islet cells were carried out as described previously (15). Isolated islets were dispersed to single islet cells by dispase (Godo Shusei, Tokyo, Japan). Dispersed islet cells were plated onto coverslips and cultured overnight in RPMI-1640 (Sigma) supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin (Meiji Seika, Tokyo, Japan), and 100 units/ml penicillin (Meiji Seika). The cells were loaded with 4 μmol/l fura PE3-AM (Sigma) for 2 h at 37°C. They were then mounted in a chamber on the stage of an inverted microscope (Diaphot TMD 300; Nikon, Tokyo, Japan) and were continuously superfused with KRBH with 3 mmol/l glucose and 0.1% BSA for 30 min at 37°C using a peristaltic pump at a flow rate of 1 ml/min. [Ca²⁺]_i was measured with an Argus-50/CA system (Hamamatsu Photonics, Hamamatsu, Japan), with alternating excitation of cells at 340 and 380 nm and monitoring of the resultant emission at 510 nm. Pairs of 340- and 380-nm fluorescence images were captured every 10 s and converted to 340:380 ratio images after subtraction of the background fluorescence. At the end of these experiments, tolbutamide at 300 μmol/l was applied to verify that the recording was obtained from a β-cell.

MIN6 cells (2 × 10⁴ cells/well) were seeded and cultured for 2 days before experimentation. After preincubation in KRBH containing 3 mmol/l glucose and 0.1% BSA for 1 h, they were incubated for 60 min in the presence of 20 mmol/l glucose and various concentrations of l-cysteine or NaHS. Cellular ATP was extracted and measured using a commercialized kit (ATPlite; PerkinElmer Life and Analytical Sciences, Groningen, the Netherlands) according to the manufacturer’s instructions.

Mitochondrial membrane potential (Δψ_m) was monitored with rhodamine 123 (Sigma), a fluorescent indicator of Δψ_m (16,17). Cells were loaded by incubation with 10 μg/ml rhodamine 123 for 10 min in KRBH containing 0.1% BSA and 3 mmol/l glucose at 37°C. After washing, they were transferred to an incubation chamber mounted on the stage of a laser scanning confocal microscope (LSM510; Carl Zeiss; Oberkochen, Germany), and cells were perfused with the BSA-containing KRBH at a flow rate of 1 ml/min using a peristaltic pump. Rhodamine 123 fluorescence was excited with a 488-nm argon laser, and its resultant fluorescence was acquired through a 505- to 550-nm band-pass filter every 15 s. Calbonyl cyanide p-(trifluoromethoxy) phenyl hydrazone (FCCP; Sigma), a protonophore, was applied at the end of the experiment as an indicator of the maximal Δψ_m depolarization. When necessary, obtained data were expressed as a percentage of the maximal value obtained by the protonophore in each experiment.

The levels of H₂S were measured by the modified method previously described by Tang and Santschi (18). After 1 h of preincubation with 3 mmol/l glucose, MIN6 cells (1.85 × 10⁷ cells) were incubated for 1 h with 20 mmol/l glucose in the presence or absence of 10 mmol/l l-cysteine. They were then washed twice with PBS and centrifuged three times at 1,000 rpm for 10 min. The pellet of cells was resuspended in 0.8 ml 0.01 mol/l NaOH, followed by the addition of 80 μl 20 mmol/l N,N-dimethyl-p-phenylenediamine sulfate (Sigma) in 7.2 mol/l HCl and 80 μl 30 mmol/l FeCl₃ in 1.2 mol/l. After 20 min at room temperature, the mixed solution was centrifuged at 5,000 rpm for 10 min, and the supernatant was applied to a Waters Sep-Pak Light tC18 cartridge, which was conditioned by passing 1 ml methanol and 1 ml milli-Q water. After the cartridge was rinsed with 5 ml milli-Q water, 0.5 ml methanol was passed through the cartridge in the reverse direction to elute methylene blue. The eluent was dried with speedvac and dissolved in 50 μl milli-Q water. The levels of methylene blue were measured by a high-performance liquid chromatographer (Waters 2695; Waters, Milford, MA). An acetonitrile gradient from 30 to 50% with complementary 70 to 50% of 50 mmol/l ammonium acetate and 5 mmol/l pentanesulfonic acid sodium salt (Wako, Osaka, Japan) was performed on a C18 column (Waters Symmetry C18; 5 μm 4.6 × 250 mm) using a flow rate of 1.0 ml/min in 12 min. Absorbance was monitored at a wavelength of 668 nm with UV/Vis detector (Waters 2487).

Results were expressed as means ± SE. Statistical evaluations were made by one-way ANOVA followed by the Tukey’s method. For Fig. 8, statistical significance was assessed by the unpaired Student’s t test. A probability of P < 0.05 was accepted as the level of statistical significance.

We assessed mRNA expression levels of CSE and CBS in mouse pancreatic islets and MIN6 cells by RT-PCR. A PCR-amplified 384-bp fragment of CSE and a 337-bp fragment of CBS were detected in mouse islets, kidney, liver, and MIN6 cells (Fig. 1). Western blot analysis using the polyclonal anti-CBS antibody revealed an immunopositive band with an apparent molecular weight of 62 kDa in mouse islets, kidney, and liver (Fig. 1).

Expression of H₂S-generating enzymes in pancreatic islets and MIN6 cells led us to examine the effects of the enzyme substrate l-cysteine and the H₂S donor NaHS on insulin release from pancreatic islets and MIN6 cells. l-Cysteine within a millimolar range inhibited glucose (10 mmol/l)-induced insulin release from mouse pancreatic islets in a dose-dependent manner, and NaHS more potently inhibited glucose-induced insulin release (Fig. 2A). l-Cysteine and NaHS did not decrease insulin release in the copresence of a nonstimulatory concentration (3 mmol/l) of glucose (data not shown). Inhibition of glucose (20 mmol/l)-induced insulin release by these substances was also observed in MIN6 cells at similar ranges (Fig. 2B). These inhibitory effects of l-cysteine and NaHS disappeared immediately after the removal in the perifusion system (data not shown).

In smooth muscle cells, H₂S-induced relaxation has been reported to be due to its activating effect on the ATP-sensitive K⁺ (K_ATP) channel (19). In the β-cell, closure of this channel by intracellular ATP plays a crucial role in glucose-induced depolarization of the plasma membrane, which causes Ca²⁺ influx through the voltage-dependent Ca²⁺ channel (20). Exposure of isolated islets to tolbutamide (100 μmol/l), a K_ATP blocker, or α-ketoisocaproate (α-KIC; 10 mmol/l), a mitochondrial fuel, induces insulin release via the K_ATP closure. Both l-cysteine (3 mmol/l) and NaHS (100 μmol/l) strongly inhibited insulin release by these substances in the concomitant presence of 3 mmol/l glucose (Fig. 3). However, l-cysteine and NaHS also inhibited insulin release when the voltage-dependent Ca²⁺ channel was directly opened by high K⁺ (30 mmol/l) depolarization (Fig. 3). To ascertain that the inhibition involves K_ATP-independent mechanism, we further examined the effects of l-cysteine and NaHS on high K⁺ (30 mmol/l)-induced insulin release (Fig. 4). l-Cysteine (3 mmol/l) inhibited insulin release in the copresence of the K_ATP opener diazoxide (200 μmol/l). Such inhibition was also observed with NaHS. These results suggest that the inhibition of insulin release by l-cysteine and NaHS involves a mechanism independent of their effects on the K_ATP activity.

We examined the effect of l-cysteine and NaHS on glucose-induced [Ca²⁺]_i oscillation. In single β-cells, [Ca²⁺]_i was low and stable at 3 mmol/l glucose. When the extracellular glucose concentration was increased to 10 mmol/l, most β-cells showed [Ca²⁺]_i oscillation (Fig. 5). Exposure to l-cysteine (3 mmol/l) and NaHS (100 μmol/l) promptly suppressed glucose-induced [Ca²⁺]_i oscillation without obvious changes in the mean [Ca²⁺]_i value (Fig. 5). The oscillatory changes in [Ca²⁺]_i reappeared at 9.5 min after the removal of these substances. These results suggest that the inhibition of insulin secretion by l-cysteine and NaHS was accompanied by suppression of [Ca²⁺]_i oscillation in mouse pancreatic β-cells.

Treatment of living cells with SLO forms small pores in the plasma membrane and enables cell-impermeant molecules to access the intracellular space (21). Insulin release from SLO-treated islets was stimulated by Ca²⁺ at 10 μmol/l (Fig. 6). Thus Ca²⁺-stimulated insulin release from permeabilized islets was suppressed by l-cysteine (3 mmol/l; Fig. 6A) and NaHS (100 μmol/l; Fig. 6B). Insulin release from SLO-treated islets was also stimulated by 100 μmol/l guanosine 5′-0-3-thiotriphosphate (GTPγS) in the copresence of 100 nmol/l Ca²⁺ (0.230 ± 0.035 and 1.043 ± 0.060 ng · islet⁻¹ · h⁻¹ for control and GTPγS, respectively; n = 4). NaHS (100 μmol/l) inhibited GTPγS-induced insulin release from SLO-treated islets (0.615 ± 0.116 ng · islet⁻¹ · h⁻¹; n = 4, P < 0.01). Under this condition, [Ca²⁺]_i is not altered by either Ca²⁺ influx or Ca²⁺ mobilization because of its chelation by EGTA. Therefore, the inhibition of insulin release by l-cysteine and NaHS may, at least in part, be independent of the changes in [Ca²⁺]_i. This inhibition may not merely result from their suppression of ATP production, because metabolic inhibitors such as carbonylcyanide m-chlorophenylhydrazone and oligomycin failed to decrease Ca²⁺-induced insulin release from permeabilized islets (data not shown).

Figure 7A shows dose-dependent effects of l-cysteine and NaHS on ATP content in MIN6 cells. l-Cysteine progressively decreased ATP production from glucose (20 mmol/l). NaHS at lower concentrations also inhibited ATP production. Figure 7B and C depicts the effects of l-cysteine (3 mmol/l) and NaHS (100 μmol/l) on glucose metabolism in dispersed β-cells assessed by a fluorescent dye, rhodamine 123. Activation of the respiratory chain induces H⁺ transport through the inner mitochondrial membrane, resulting in Δψ_m hyperpolarization. Glucose (10 mmol/l) caused Δψ_m hyperpolarization, indicated by a fall in the intensity of the fluorescence. Application of l-cysteine caused a small but significant inhibition of glucose-induced hyperpolarization (Fig. 7B), and that of NaHS resulted in an immediate and more potent inhibition of glucose metabolism (Fig. 7C). When the peak value of the inhibition was expressed as a percentage of the FCCP-induced depolarization, 24.9 ± 4.2 and 82.0 ± 6.3% were obtained by l-cysteine and NaHS, respectively (n = 18 for l-cysteine and n = 7 for NaHS).

Figure 8 demonstrates the levels of H₂S in MIN6 cells. In the presence of 20 mmol/l glucose, the content of H₂S was 0.099 ± 0.006 nmol/10⁶ cells (n = 6), and when the cells were incubated with 20 mmol/l glucose and 10 mmol/l l-cysteine, the H₂S level were elevated to 0.20 ± 0.022 nmol/10⁶ cells (n = 6, P < 0.01 compared with the value without l-cysteine). This result and other findings described above indicate that inhibition of insulin release by l-cysteine involves its metabolic production of H₂S.

Historically, H₂S has been recognized as a toxic gas, and it produces serious intoxication in certain occasions such as an environmental pollution (22). However, recent studies show that H₂S may be endogenously produced as a metabolite of l-cysteine and that the gas may be involved in the regulation of various cellular phenomena, i.e., relaxation of vascular or visceral smooth muscle cells and cardiomyocytes (8,23–25), endotoxin-induced inflammation (26), and also some neural functions such as intracellular Ca²⁺ dynamics and long-term potentiation (27,28). H₂S has been suggested to be implicated in the pathogenesis of certain diseases. Changes in the tissue H₂S contents and activities of H₂S-producing enzymes have been reported in Down syndrome (29,30) and acute pancreatitis (31).

The present findings indicate that l-cysteine potently inhibited insulin release from both mouse islets and MIN6 cells. The inhibition is likely to result from l-cysteine metabolism and eventual H₂S production, because 1) the major H₂S-generating enzymes from l-cysteine, CBS and CSE, were distributed in the pancreatic islets and MIN6 cells; 2) l-cysteine inhibited insulin release in a manner similar to NaHS-induced inhibition; and 3) H₂S was produced from l-cysteine in MIN6 cells.

As has been suggested in literature, H₂S exhibits its toxicity via inhibiting cellular metabolism by suppressing the respiratory chain in the mitochondria (32). Suppression of glucose-induced insulin release by l-cysteine and NaHS may result from their inhibition of glucose metabolism as shown by the fluorometrical measurements in the present study. However, we consider that l-cysteine and NaHS may also act on the β-cell secretory machinery in addition to inhibiting glucose metabolism, because these substances also inhibited insulin release from SLO-permeabilized islets under the conditions where ATP was exogenously supplied. Moreover, it was reported that glucose metabolism is nullified in permeabilized islets (33), and metabolic inhibitors examined so far in this study failed to decrease insulin release from SLO-treated islets.

It must be noted that the effects of l-cysteine and NaHS on glucose-induced changes of the β-cell [Ca²⁺]_i were very distinctive. Exposure of dispersed β-cells to l-cysteine and NaHS resulted in the disappearance of the [Ca²⁺]_i oscillation, whereas the average value of the [Ca²⁺]_i did not decline significantly. These results may corroborate an idea that oscillatory changes of [Ca²⁺]_i are crucial in glucose-induced insulin release (34). In SLO-treated islets, [Ca²⁺]_i is clamped by the Ca²⁺-EGTA buffer . Therefore, the inhibition of insulin release from permeabilized islets also implies that these effects of l-cysteine and NaHS are at least partly independent of the modulation of the [Ca²⁺]_i levels.

Taking the results obtained in this study into our consideration, we suggest here that l-cysteine and NaHS may share multiple sites for their inhibitory actions on insulin release. However, inhibition of glucose metabolism by l-cysteine was much smaller than that by NaHS, whereas these substances at the tested concentrations inhibited insulin release to a similar extent. This may suggest to us that l-cysteine may inhibit insulin release via a mechanism independent of H₂S production and the resultant inhibition of glucose metabolism. For example, taurine, known to be produced from l-cysteine (35), was reported to inhibit glucose-induced insulin release from mouse islets (36).

Plasma l-cysteine levels in the normal subjects have been reported to be within a submillimolar range (1). However, we consider that insulin release in vivo can be regulated by endogenous H₂S. It is reported that 10–150 μmol/l H₂S occurs in various normal tissues (8), and the plasma l-cysteine levels are reported to become higher in diabetic patients with renal complications (1,37). The activities of the H₂S-producing enzymes CBS and CSE are elevated under diabetic conditions (10,11). Moreover, the tissue contents of H₂S were reported to be increased in the liver and pancreas in streptozotocin-treated diabetic rats (11). H₂S is known to penetrate cellular membranes freely as another gasotransmitter nitric oxide (NO) does (38). Therefore, it is likely that insulin release from the pancreatic β-cell may be modulated by H₂S, which is generated by pancreatic islets or by the surrounding tissues under diabetic conditions.

In the β-cell, NO, generated as an arginine metabolite, modulates insulin release in a bimodal fashion (15). Because it was previously reported that H₂S may regulate the production of NO in the vascular smooth muscle cells (24), we examined the effects of NaHS on the intracellular NO levels in the pancreatic β-cells using the NO-specific fluorescent dye, DAF-2. However, we could not obtain any evidence for the involvement of H₂S in the regulation of NO synthesis in the β-cells (data not shown).

Insulin release is sophisticatedly regulated by physiological substances, such as various hormones, autacoids, neurotransmitters, and nutrients. Most of these substances are considered to modulate the levels of their relevant intracellular signal molecules, which increase or enhance insulin release. However, an intracellular signal molecule that directly inhibits insulin release has never been reported. From this aspect, H₂S is the first found signal molecule that plays an inhibitory role in insulin release from the pancreatic β-cell. Inhibitory mechanisms of insulin release by physiological substances are poorly understood at present (39). Therefore, it is tempting to investigate the involvement of H₂S production in the hormones or neurotransmitters that decrease insulin release.

Here, we demonstrated that l-cysteine and NaHS, a donor of its metabolite H₂S, have a potent inhibitory effect on insulin release via multiple actions on the insulin secretory machinery. These findings may participate in the physiological and/or pathological regulation of insulin release in vivo. We speculate that it may explain why the β-cell suffers from progressive functional deterioration in type 2 diabetic patients.

We thank Prof. J.-I. Miyazaki (Osaka University) for the gift of MIN6 cells and Prof. K. Nakayama and Dr. T. Ishikawa (University of Shizuoka) for their valuable technical help and advice.