Introduction

Cystathionine γ-lyase (CSE) uses l-cysteine as a substrate to generate H2S in many types of mammalian cells [1, 2]. CSE is mostly responsible for the endogenous production of H2S in pancreatic tissues and cloned pancreatic beta cell lines [28]. The inhibition of CSE activity by dl-propargylglycine (PPG) significantly decreased production of H2S and increased plasma insulin levels in Zucker diabetic fatty (ZDF) rats [6], while increased Cse expression and H2S formation were observed in streptozotocin (STZ)-induced diabetic rats [4, 6, 8]. Genetic deletion of Cse (also known as Cth) in mice markedly reduces H2S levels in different tissues, leading to increased blood pressure, accumulation of homocysteine and impairment of vascular endothelial functions [3]. When stimulated with high glucose (20 mmol/l), insulin secretion from pancreases of Cse-knockout (KO) mice is significantly increased [4]. Compared with wild-type (WT) mice, Cse-KO mice that receive STZ injections have a delayed onset of diabetic status [4]. STZ significantly increases pancreatic H2S production in WT mice, but not in Cse-KO mice, and it causes more apoptotic beta cell death in WT than in Cse-KO mice.

Voltage-dependent Ca2+ channels (VDCCs) in pancreatic beta cells play a critical role in the regulation of insulin secretion via controlling Ca2+ influx. It has been reported that H2S stimulates ATP-sensitive K+ (KATP) channels, which hyperpolarises cell membrane and indirectly inactivates L-type VDCCs, leading to reduced insulin secretion [5, 911]. Interaction between H2S and KATP channels in insulin-secreting cells constitutes an important regulatory mechanism for insulin secretion [10, 11]. On the other hand, the role of L-type VDCCs in H2S-regulated insulin secretion is unknown. One report described an inhibitory effect of H2S on L-type VDCCs in isolated rat cardiomyocytes [12]. Another study in cerebellar granule neurons reported the contradictory result that H2S increased intracellular Ca2+ concentration ([Ca2+]i) by stimulating L-type VDCCs [13]. Whether and how H2S interacts with L-type VDCCs in beta cells is intriguing and warrants in-depth investigation.

In the present study, Cse-KO and WT mice were used to isolate pancreatic beta cells. The effects of NaHS, a donor of H2S, were first determined on the electrophysiological properties of L-type VDCCs. Changes in L-type VDCCs were also tested in the presence of a CSE inhibitor (PPG) or in the absence of endogenous H2S. The insulin secretion from and H2S production in islets or pancreas were measured with or without blocking L-type VDCCs.

Methods

Animal preparation

The Cse-KO mice were generated and house bred as previously described [3]. The fourth generation of 10- to 16-week-old male Cse-KO offspring and age-matched male WT littermates on a C57BL/6 J × 129SvEv background were used. All animal experiments were approved by the Animal Use Committee of Lakehead University, ON, Canada.

Isolation of intact islets and beta cells from mouse pancreas

Briefly, the mice were anaesthetised with a single intraperitoneal injection of ketamine (2.2 mg/10 g body weight) and xylazine (0.44 mg/10 g body weight) before the opening of the abdominal wall. The pancreas was distended by injection of 1 ml Krebs–Ringer bicarbonate (KRB) containing 0.5–1 mg/ml collagenase type V (Sigma, USA), 0.5–1 μl/ml DNase I (Sigma, USA) and bovine serum albumin (0.5 mg/ml) through a cannula inserted into the common bile duct. KRB comprised: NaCl 129 mmol/l; KCl 4.8 mmol/l; KH2PO4 1.2 mmol/l; NaHCO3 5 mmol/l; MgCl2 1.2 mmol/l; HEPES 10 mmol/l; and CaCl2 2.5 mmol/l. The pancreas was then removed carefully and transferred to a 50 ml Falcon tube with the above enzyme solution, and incubated for 5–10 min in a water bath at 37°C. Cleaned islets of Langerhans were hand-picked under a dissecting microscope after ×3 centrifuge washout (at 1,500 g for 3 min at 4°C) with Ca2+/Mg2+-free KRB. The dispersed islets were dissociated into single cells by vigorous shaking in Ca2+/Mg2+-free KRB supplemented with ×1 enzyme-free cell-dissociation reagent (HiMedia Lab Pvt. Ltd, VWR, Mississauga, ON, Canada). The dispersed islet cells were cultured on small glass cover slips pre-coated with poly l-lysine, in RPMI 1640 medium supplemented with 10% (vol./vol.) fetal bovine serum, 100 U/ml penicillin, and 10 μg/ml streptomycin in humidified air at 37°C containing 5% CO2. Primary-cultured (1–3 days) islet cells were used in the patch-clamp experiments.

Measurement of insulin secretion from isolated pancreatic islets

Freshly isolated islets (ten size-matched for each batch) were washed and pre-incubated with glucose-free RPMI 1640 medium in 24 well plates [6]. After pre-incubation for 30 min, islets were treated for 30 min at 37°C with different chemicals in the presence of either 2.8 or 20 mmol/l glucose. In other experiments, insulin release from isolated islets was detected in KRB. At the end of each incubation period, the medium was collected and centrifuged for 10 min at 239 g to remove islet debris. The insulin level in the supernatant fraction was determined using the mouse insulin ELISA kit (Mercodia AB, Sylveniusgatan, Uppsala, Sweden). After insulin measurement, the islets were scraped from 24 well plates on ice into the centrifuge tube; protein concentration was then measured using Bradford Reagent (Sigma, St Louis, MO, USA).

Electrophysiological recordings

The patch-clamp technique was used to record whole-cell VDCC currents in the voltage-clamp mode [14, 15]. The glass cover slip with primary-cultured pancreatic beta cells attached was inserted into the perfusion chamber, which was mounted on the stage of an inverted phase-contrast microscope (Olympus IX71, Olympus, Tokyo, Japan). Cells were washed with bath solution in the chamber for 3–5 min before starting an experiment. Patch pipettes were pulled from borosilicate glass capillaries with filaments (Sutter Instrument, Novato, CA, USA) with tip resistance of 2–5 MΩ when filled with different pipette solutions. Currents were recorded with an Axopatch 200-B amplifier (Axon Instruments, Molecular Devices, LLC, Sunnyvale, CA, USA), controlled by a Digidata 1200 interface and a pCLAMP software (Version 6.02, Axon Instruments). Membrane currents were filtered at 1 kHz with a four-pole Bessel filter. At the beginning of each experiment, the patch-clamp amplifier was adjusted to set I = 0. No leakage subtraction was performed for the original recordings, and all cells with visible changes in leakage currents during the course of study were excluded from further analysis. The VDCC currents were recorded by a 300 ms step pulse ranging from −80 to +60 mV with 10 mV increments. The holding potentials (HPs) were set at −80 or −40 mV, at which the outward voltage-dependent K+ currents or inward T-type VDCC currents were largely inactivated, respectively. The pipette solution for recording VDCC currents contained: Cs2-aspartate 70 mmol/l, EGTA 10 mmol/l, MgATP 4 mmol/l, MgCl2 5 mmol/l, CaCl2 1 mmol/l, Na-pyruvate 5 mmol/l, K-succinate 5 mmol/l, glucose 25 mmol/l, HEPES 10 mmol/l, Na-creatine phosphate 5 mmol/l, creatine kinase 50 U/ml, Na2GTP 0.3 mmol/l (pH = 7.2). The bath solution for recording VDCC currents included: Trizma-HCl 110 mmol/l, TEA-Cl 5 mmol/l, BaCl2 20 mmol/l, CsCl 5 mmol/l, HEPES 20 mmol/l, glucose 5 mmol/l (pH = 7.4). Giga-ohm seal was made in normal extracellular high-Na+ solution: NaCl 135 mmol/l, KCl 4.8 mmol/l, MgCl2 1.2 mmol/l, CaCl2 2.5 mmol/l, HEPES 10 mmol/l. In some experiments, tetrodotoxin (TTX) at 1 μmol/l was included in the bath solution to block any Na+ channel currents. The osmolality of recording solutions was adjusted to 290 mOsmol/l, and pH to 7.2–7.4. Cells were continuously superfused with the bath solution containing the test chemicals at the desired final concentrations. All experiments were performed at room temperature. In electrophysiological recordings, beta cells were distinguished from other islet cells by large size (usually capacitance ≥7 pF), no detectable Na+ current when holding at −70 mV, and characteristic oscillatory or bursting electrical activity when exposed to 10 mmol/l glucose [16, 17].

The current density of L-type VDCCs was calculated by normalising the peak amplitude of current with cell capacitance (pA/pF). The channel recovery from depolarisation-induced inactivation was also produced by a double-pulse protocol stepped to 0 mV from −70 mV with an HP of −80 mV. An increase in the interval between the conditioning and test pulses resulted in a recovery of VDCC current (I Ca), which can be fitted by a mono-exponential equation: \( I/{{I}_{{\max }}} = 1 - \exp \left( { - t/\tau } \right) \), where I max is the current measured during the first pulse, I is the current measured during the second pulse, t represents the value of the interval, and τ represents the time constant of I Ca recovery from inactivation.

Chemicals and statistical analysis

Nifedipine, Bay K-8644, PPG, and NaHS were purchased from Sigma Chemicals, glibenclamide from Research Biochemicals International (Natick, MA, USA), and ACS 67 from Cayman Chemical Company (Ann Arbor, Michigan, USA). Stock solutions of nifedipine, glibenclamide and ACS 67 were made in DMSO and diluted to the desired concentrations immediately prior to use. DMSO alone was without effect at the concentration used (up to 0.3%, vol./vol.).

Data are expressed as means ± SEM and were analysed using paired Student’s t test and analysis of variance in conjunction with the Newman–Keuls test where applicable. Differences were considered statistically significant at the level of p < 0.05.

Results

Electrophysiological features and pharmacological sensitivity of VDCC currents

To separate the whole-cell I Ca components—peak I Ca (I Ca,peak) with fast inactivation and a steady-state I Ca current (I Ca,SS) without inactivation—different HPs were employed. At the HP of –80 mV, I Ca current traces were recorded by a 300 ms pulse stepped from −80 to +60 mV with 10 mV increments, exhibiting both I Ca,peak and I Ca,SS components (Fig. 1a). Changing the HP from −80 to −40 mV would inactivate T-type VDCCs [18, 19]. In our study, this HP change reduced the I Ca,peak density from −52.2 ± 5.5 to −36.7 ± 3.7 pA/pF (at +10 mV, n = 8, p < 0.05) (Fig. 1a,b). With the HP at −40 mV, the whole-cell inward I Ca showed only one I Ca,SS component (Fig. 1a–c). This I Ca,SS had the same current density with −80 or −40 mV as the HPs (n = 8, p > 0.05) (Fig. 1c). To confirm whether I Ca,SS represents mostly the activity of L-type VDCCs, a specific agonist and an antagonist for L-type VDCC were applied. Bay K-8644, an agonist of L-type VDCC, enhanced the ICa,SS by 53.3 ± 4.2% and 53.8 ± 6.4% (at +20 mV, n = 6, p < 0.05) at HPs of −80 and −40 mV, respectively (Fig. 1d). The stimulatory effect of Bay K-8644 on I Ca,SS was blocked by nifedipine by 81.1 ± 7.2% and 72.3 ± 6.5% at −80 and −40 mV HPs, respectively (Fig. 1d). However, both Bay K-8644 and nifedipine failed to affect the I Ca,peak (Fig. 1d). Thus, I Ca,SS was referred to as L-type VDCCs in this study.

Fig. 1
figure 1

VDCC currents in mouse pancreatic beta cells. (a) Original VDCC traces recorded from −80 mV to +60 mV with HPs of −80 mV (a) and −40 mV (b), respectively, and the differential currents (ab) generated after subtracting the current recorded in (b) from those in (a). (b) The IV relationship of I Ca,peak recorded with different HPs (−80 and −40 mV) is summarised (n = 8, p < 0.05). The I Ca,peak is measured at the maximum amplitude of I Ca at each testing potential. (c) The IV relationship of I Ca,SS recorded with different HPs of −40 mV (black squares) and −80 mV (white squares) is summarised (n = 8). The I Ca,SS is measured at the end of each testing pulse. (d) Original VDCC traces recorded at the testing potential of +10 mV with different HPs in the absence and presence of 0.2 μmol/l Bay K-8644 and 1 μmol/l nifedipine. The dotted line indicates zero current. The summary shows I Ca,SS before and after the application of Bay K-8644 and nifedipine (test potential, +10 mV, n = 6, *p < 0.05)

Inhibition of L-type VDCCs by exogenous H2S in pancreatic islet beta cells

With the HP at −80 mV, NaHS significantly inhibited the I Ca,SS by 31.3 ± 2.4% (n = 12, p < 0.05), but not the I Ca,peak (at +10 mV, n = 12, p > 0.05) (Figs 2a,c and 3a). The inhibitory effect of NaHS on L-type VDCCs was reversible after washing out NaHS (Fig. 2a). NaHS did not alter the reversal potential and the voltage-dependence of L-type VDCCs (Fig. 2a). The IC50 of the inhibitory effect of NaHS on L-type VDCCs was 65.4 ± 5.6 μmol/l (Fig. 2b). The I Ca,SS was inhibited by NaHS by 43.5 ± 4.4% and further inhibited by nifedipine by 32.3 ± 3.5% (n = 5, p < 0.01) (Fig. 2c). After changing the HP to −40 mV from −80 mV, only the I Ca,SS existed and NaHS significantly inhibited I Ca,SS by 45.6 ± 4.4% (Fig. 3b,c), confirming that NaHS targeted only L-type VDCCs. In general, NaHS did not shift the current–voltage (IV) curve of L-type VDCCs with the HP at −80 mV (Figs 2 and 3). It appears that with a HP at −40 mV NaHS caused a marginal, but inapparent, shift of the IV curve of L-type VDCCs to the left (Fig. 3b).

Fig. 2
figure 2

The inhibitory effects of NaHS on VDCC currents in mouse pancreatic beta cells. (a) Original VDCC traces sequentially recorded in the same cell from –80 mV to +60 mV with an HP of –80 mV in the absence (white squares) and presence of different concentrations of NaHS (at 10 [black circles], 30 [black squares], and 100 [black diamonds] μmol/l) and chemical washout (white circles). The corresponding IV relationship of I Ca,SS is shown in the right panel. (b) The concentration-dependent inhibition of I Ca,SS by NaHS (at +10 mV, HP = –80 mV, n = 12) with IC50 of 65.4 ± 5.6 μmol/l. (c) Original VDCC traces in the left panel showing that I Ca,SS was inhibited by NaHS (100 μmol/l), NaHS and nifedipine (1 μmol/l) as well as by nifedipine alone (TP = −80 to +60 mV, HP = −80 mV). The summary of these data is shown in the right panel, with current intensity measured at +10 mV (HP = −80 mV, n = 5, **p < 0.01)

Fig. 3
figure 3

The inhibition of VDCC currents by NaHS at different HPs in mouse pancreatic beta cells. (a) Original VDCC traces recorded from −80 to +60 mV with an HP of −80 mV before (a) and after (b) the application of NaHS (at 100 μmol/l). The corresponding IV relationship of I Ca, peak before (white squares) and after (black squares) the application of 100 μmol/l NaHS, and that of I Ca,SS before (white circles) and after (black circles) NaHS application. (b) Original VDCC traces recorded with an HP of −40 mV before (c) and after (d) NaHS application. The corresponding IV relationships of I Ca, peak and I Ca,SS are shown in the right panel. (c) Summary of the inhibitory effect of NaHS on the peak and steady-state I Ca currents at +10 mV at different HPs (−80 and −40 mV) (n = 10, **p < 0.01). (d) The effect of NaHS on the recovery kinetics from VDCC inactivation. A double-pulse protocol is shown with the testing potentials stepped from −70 to 0 mV at an HP of −80 mV. The interval between two pulses ranges from 0 ms to 200 ms at 10 ms increments. A downward shift of the recovery curve was observed in the presence of NaHS (black squares) in comparison with the curve in the absence of NaHS (white squares) (n = 6, *p < 0.05)

To evaluate the recovery from inactivation, currents were elicited by the double-pulse protocol, with the interval separating two pulses increased progressively from 0 to 200 ms (Fig. 3d). When the interval was short, the I Ca was small and decayed more slowly. An increase in the interval between the conditioning pulse and test pulse resulted in a recovery of ICa. The time constants (τ) of I Ca recovery from inactivation in the HP of −70 mV were 89.4 ± 6.3 and 115.4 ± 8.5 ms in the control and NaHS group, respectively (Fig. 3d). As the time interval (t) increased to 80–200 ms, the I/I max values in the NaHS-treated group were significantly decreased in comparison with those of the control group (n = 6, p < 0.05) (Fig. 3d). It appears that L-type VDCCs in INS-1E cells take more than 70 ms for recovery from inactivation. The mechanism underlying this delayed recovery is not clear yet.

ACS 67 is an analogue of latanoprost (an F-series prostaglandin analogue) that contains an H2S-releasing component conjugated to the latanoprost carboxyl group, and is therefore a lipid-soluble, slow-releasing H2S donor [20]. ACS 67 (100 μmol/l) also inhibits L-type Ca2+ channels. It decreased L-type VDCC currents by 18.3 ± 2.8% at a membrane potential of +10 mV (n = 5, p < 0.05). However, the inhibitory effect of ACS 67 was weaker than that of NaHS at the same concentration (45.4 ± 4.6% inhibition).

Effects of endogenous H2S on L-type VDCCs in pancreatic islet beta cells

Without PPG dialysis, NaHS at 100 μmol/l significantly inhibited I Ca,SS by 27.5 ± 3.2% (at +20 mV, n = 10, p < 0.05) in WT beta cells (Fig. 4a,c). By including 2 mmol/l PPG in the pipette solution, NaHS also inhibited I Ca,SS by 43.8 ± 5.6% (at +20 mV, n = 10) (Fig. 4a,c). On the other hand, NaHS inhibited I Ca,SS in Cse-KO mouse beta cells with and without PPG dialysis by 25.2 ± 3.0% and 32.2 ± 4.3%, respectively (at +20 mV, n = 10, p < 0.05), but PPG dialysis did not alter basal I Ca,SS level in Cse-KO mouse beta cells (Fig. 4b,c). To test the effect of endogenous H2S on I Ca,SS, Cse-KO mice were used to isolate pancreatic beta cells. NaHS significantly suppressed I Ca,SS by 29.7 ± 3.5% and 55.6 ± 6.4% in WT and Cse-KO mouse beta cells, respectively (n = 10, p < 0.01) (Fig. 5a–c). The basal current density of I Ca,SS in beta cells from Cse-KO mice was greater than that from WT mice (p < 0.05) (Fig. 5c).

Fig. 4
figure 4

The effect of NaHS on VDCC currents in the absence and presence of PPG dialysis of pancreatic beta cells from WT and Cse-KO mice. (a) The representative VDCC traces for inward I Ca,SS currents recorded in WT mouse pancreatic beta cells with an HP at −80 mV before and after the application of NaHS in the absence and presence of PPG dialysis. The corresponding IV relationship curves of I Ca,SS are shown in the right panel, either without PPG in the absence (white squares) and presence (black squares) of 100 μmol/l NaHS or with 2 mmol/l PPG dialysis in the absence (white circles) and presence (black circles) of NaHS. (b) The representative I Ca,SS current traces in Cse-KO mouse pancreatic beta cells with the same recording conditions as in (a). The corresponding IV relationship curves of I Ca,SS are shown in the right panel. (c) Summary of the effects of NaHS on I Ca,SS, measured at +10 mV without or with 2 mmol/l PPG dialysis of WT and Cse-KO pancreatic beta cells (HP = –80 mV, n = 10, *p < 0.05 for 100 μmol/l NaHS vs control, p < 0.05 for 2 vs 0 mmol/l PPG dialysis)

Fig. 5
figure 5

The effect of NaHS on VDCC currents in pancreatic beta cells from Cse-KO mice. (a) The representative inward I Ca,SS current traces in WT mouse pancreatic beta cells (HP = –80 mV) before and after the application of NaHS. The corresponding IV relationship curves before (white squares) and after (black squares) the application of 100 μmol/l NaHS are shown in the right panel. (b) The representative I Ca,SS current traces in Cse-KO mouse pancreatic beta cells with the same recording condition as in (a). The corresponding IV relationship curves are shown in the right panel. (c) Summary of the effects of NaHS on I Ca,SS currents, measured at +10 mV, in pancreatic beta cells from WT and Cse-KO mice (HP = –80 mV, n = 10; **p < 0.01 for 100 μmol/l NaHS vs control; p < 0.05 for Cse-KO vs WT)

Effects of the CSE/H2S system on insulin release from pancreatic islets

Following incubation with culture medium and normal glucose (at 10 mmol/l) for 30 min, insulin release from freshly isolated islets from Cse-KO mice was much higher than that from WT islets (n = 5, p < 0.05) (Fig. 6a). Insulin secretion from islets was inhibited by nifedipine by 45.7 ± 3.6% and 41.2 ± 4.5% in WT and Cse-KO islets, respectively. NaHS at 100 μmol/l and 300 μmol/l inhibited glucose-stimulated insulin secretion by 26.7 ± 4.5% and 49.5 ± 6.5% (n = 5, p < 0.05 vs control), respectively (Fig. 6b). NaHS or nifedipine alone inhibited insulin release, and their co-application had an increased effect (n = 5, p > 0.05) (Fig. 6c). Bay K-8644 alone enhanced insulin secretion in mouse islets, but co-application of NaHS or nifedipine counteracted the stimulatory effect of Bay K-8644 (Fig. 6c). These results suggest that nifedipine and H2S may inhibit the same L-type VDCCs. Glibenclamide alone increased insulin secretion, which was also inhibited by the addition of NaHS (n = 5, p < 0.05) (Fig. 6c). This result indicates that KATP channel activation is involved in the regulation of insulin release. Figure 6d further shows the inhibitory effect of endogenous H2S on insulin release from islets of WT or Cse-KO mice, assayed in KRB.

Fig. 6
figure 6

Insulin secretion and H2S production of pancreatic islets from WT and Cse-KO mice. (a) Glucose (10 mmol/l)-stimulated insulin secretion from WT and Cse-KO mouse islets and its inhibition by nifedipine (1 μmol/l) (n = 5, *p < 0.05 for KO vs WT; p < 0.05 nifedipine vs control). (b) The effects of NaHS on insulin secretion from WT mouse islets (n = 5, *p < 0.05 for 20 mmol/l vs 2.8 mmol/l glucose; p < 0.05 for 100/300 vs 0 μmol/l NaHS). (c) Effects of NaHS (100 μmol/l), nifedipine (1 μmol/l), Bay K-8644 (0.2 μmol/l), and glibenclamide (30 μmol/l) on 20 mmol/l glucose-stimulated insulin secretion from WT mouse islets (n = 5, *p < 0.05 for different treatments with NaHS, nifedipine, Bay K-8644, and glibenclamide vs control without treatment; p < 0.05 for NaHS + Bay K-8644 vs Bay K-8644, and NaHS + glibenclamide vs glibenclamide). Mouse islets were incubated with RPMI-1640 medium in experiments shown in (a–c). (d) Insulin secretion from mouse islets incubated in KRB; n = 4; *p < 0.05 for 16.8 mmol/l or 6.4 mmol/l vs 2.8 mmol/l; p < 0.05 for KO vs WT at 6.4 or 16.8 mmol/l glucose

Discussion

We and others have previously shown that H2S functions as an endogenous opener of KATP channels in different types of cell, including smooth muscle cells [2, 21], HEK-293 cells [22] and pancreatic beta cells [5, 11], independent of the activation of cytosolic second messengers. Interaction between H2S and KATP channels in insulin-secreting cells may underlie an important mechanism for the regulation of insulin secretion from pancreatic beta cells. Our present study examined the effects of exogenous and endogenous H2S on L-type VDCCs in mouse pancreatic beta cells. We have found that: (1) the H2S donors NaHS and ACS 67 reversibly inhibited L-type VDCCs; (2) PPG dialysis to inhibit CSE activity and lower endogenous H2S level increased the basal amplitude of L-type VDCCs in WT mouse beta cells; and (3) the density of L-type VDCCs in beta cells from Cse-KO mice, in which the endogenous H2S level is minimised, is greater than that from WT mice. These data represent the first evidence that H2S has an inhibitory role in the regulation of L-type VDCCs in mouse pancreatic beta cells.

Glucose stimulation of pancreatic beta cells involves membrane depolarisation and Ca2+ influx. An upsurge in [Ca2+]i triggers the exocytotic machinery, with subsequent insulin release. Ca2+ influx in beta cells occurs primarily through VDCCs, especially L-type VDCCs [2325]. L-type VDCCs, a large hetero-oligomeric complex consisting of α1-, α2/δ-, β- and γ-subunits, is characterised by high-voltage-activated persistent Ca2+ currents that are without inactivation or inactivate slowly. L-type VDCCs are blocked by nifedipine or verapamil and activated by Bay K-8644 or CGP28391 [26]. N- and R-type VDCCs have been reported in mouse beta cells [27]. These VDCCs belong to the same group of high-voltage-activated VDCCs as L-type channels. However, N and R channels can be blocked by ω-conotoxin and SNX482, respectively, but L-channels are sensitive to dihydropyridines. In our study, the blockade of L-type VDCCs by nifedipine and its stimulation by Bay K-8644 indicate that the L-type VDCCs currents recorded are unlikely to have been contaminated by N- or R-type VDCCs. On the other hand, the effects of H2S on N- and R-type VDCCs in beta cells cannot yet be excluded.

T-type VDCCs have been suggested to involve modulation of general membrane electrical activity to enhance insulin secretion [28]. They are activated at more negative membrane potentials and are inactivated quickly and completely [29]. T-type VDCCs have been recorded in mouse beta cells by others [18, 30, 31] and are likely to have been recorded in our present study. As we focused on the effects of H2S on L-type VDCCs, no further effort was taken to characterise T-type VDCCs in our study.

Both CSE and cystathionine β-synthase (CBS) are produced in rodent pancreatic tissues or cloned islet beta cell lines, while CSE plays a more important role in endogenous H2S production in beta cells. The expression level of Cse mRNA was significantly higher than that of CBS in pancreatic islets and PPG abolished most of the H2S production in pancreatic islets and INS-1E cells [5, 6]. Furthermore, H2S production in pancreatic beta cells is decreased by increased blood glucose level [7]. Overexpression of Cse inhibited insulin release from INS-1E cells, and lowering endogenous H2S production by PPG or Cse-targeted small interfering RNA had the opposite effect [5]. We previously showed that PPG ameliorated high blood glucose of ZDF rats to near normal levels, while amino-oxyacetate, a specific blocker of CBS, did not significantly change the hyperglycaemic status of ZDF rats [6]. Furthermore, H2S production rate was significantly decreased in pancreatic islets from Cse-KO mice and PPG treatment drastically reduced H2S production rate in pancreatic islets from WT mice [4]. In addition, l-cysteine and H2S inhibit glucose-stimulated insulin secretion from insulin-secreting beta cell lines (INS-1E, MIN6 and HIT-T15) and isolated rat islets [5, 11]. All these data suggest a critical role of CSE in producing H2S in beta cells, thus regulating insulin secretion under physiological or pathological conditions.

L-type VDCC play a crucial role in the regulation of insulin secretion in pancreatic beta cells. The decay of I Ca,SS during depolarisation is called inactivation, and is regulated by elevation of [Ca2+]i as well as by membrane depolarisation [32, 33]. Therefore, [Ca2+]i in the vicinity of the channels critically controls their activity during stimulus–secretion coupling. The Ca2+-dependent process of inactivation may serve as a negative feedback mechanism for regulating Ca2+ entry into pancreatic beta cells and insulin granule exocytosis [32]. The recovery from inactivation of I Ca,SS is affected by different factors, such as the HP, the voltage and duration of the conditioning pulse, the frequency of the coupled pulses and the extracellular ionic milieu. H2S inhibits I Ca,SS and delays the channels’ recovery from inactivation in beta cells, producing a negative feedback regulation on [Ca2+]i.

Glucose-stimulated insulin secretion was examined in our study with different manipulations of H2S level and function status of L-type VDCCs. NaHS at different concentrations had no statistically significant effect on basal insulin secretion with 2.8 mmol/l glucose (Fig. 6b). On the other hand, a significant portion of insulin release from pancreatic islets of Cse-KO mice remained in the presence of nifedipine (Fig. 6a). This result is not in conflict with the notion that H2S inhibits L-type VDCCs to decrease insulin release. H2S has multiple effects on insulin metabolism in the pancreas. We have previously demonstrated that H2S induces pancreatic beta cell apoptosis and decreases beta cell mass, thus reducing insulin production [4]. The diminished endogenous H2S in Cse-KO islets would increase pancreatic beta cell mass and more insulin would be produced, so that with the same concentration of nifedipine treatment more insulin would be available for release. Glibenclamide alone increased glucose-stimulated insulin release, which was partially suppressed by NaHS. This suggests that H2S-inhibited insulin release relies on KATP-channel-dependent and -independent mechanisms. The former is mediated by the indirect inactivation of L-type VDCCs by H2S via activation of KATP channels and membrane hyperpolarisation, whereas the latter involves direct inhibition of L-type VDCCs by H2S. No matter whether KATP channels or L-type VDCCs or both are affected by H2S, the inhibition of L-type VDCCs, leading to lowered [Ca2+]i, is the final coupling node between the cellular electrophysiological changes and the reduced insulin release induced by H2S.

In summary, we report an original observation that H2S, applied exogenously or generated endogenously, inhibits L-type VDCCs and regulates insulin secretion. This inhibitory effect of H2S on L-type VDCC in pancreatic beta cells is independent of the effect of H2S on KATP channels. As such, our report adds to current understanding of the regulation of pancreatic function by H2S through its effects on different ion channels in beta cells. This will pave the way for targeting different ion channels when devising selective therapeutic strategies to correct abnormal H2S metabolism in the pancreas under different pathological conditions.