Blockade of Na⁺ Channels in Pancreatic α-Cells Has Antidiabetic Effects

Glucose homeostasis is regulated primarily by the opposing actions of insulin and glucagon (1). Glucagon secreted by pancreatic α-cells regulates glucose homeostasis by promoting hepatic glucose production. Plasma glucagon levels are increased under both fasting and postprandial states in diabetes (2,3). This hyperglucagonemia increases hepatic glucose production, thereby contributing importantly to diabetic hyperglycemia (4,5). Therefore, not surprisingly, the lowering of glucagon levels or antagonizing its actions via blockade of glucagon receptors can significantly reduce hyperglycemia (6,7).

Regulation of glucagon secretion from pancreatic α-cells is modulated by the autonomic nervous system, gut hormones, paracrine factors, and various ion channels. Among these various contributors, secretion of glucagon from α-cells depends on the generation of Na⁺-dependent action potentials (8). It is known that α-cells express tetrodotoxin (TTX)-sensitive voltage-gated Na⁺ channel (NaCh) isoforms, and that NaCh blockers like TTX inhibit glucagon secretion (8–10). Recently, it was reported that pancreatic α-cells of diabetic mice, with chronic hyperglucagonemia, had increased Na⁺ current (I_Na), action potential duration, amplitude, and firing frequency, and increased glucagon content, which primes the cells for increased glucagon release (11).

Ranolazine is an antianginal drug with cardioprotective properties (12). The proposed mechanism of its anti-ischemic effects is the inhibition of late I_Na due to blockade of the cardiac isoform of NaCh, Na_v1.5 (13). In addition, ranolazine has been shown to have antidiabetic effects in clinical and nonclinical studies (14–18). Data from the CARISA Study showed that ranolazine, in a dose-dependent manner, lowered glycated hemoglobin (HbA_1c) levels in patients with chronic angina and type 2 diabetes mellitus (T2DM) (14). Data from the MERLIN-TIMI-36 Study (Metabolic Efficiency with Ranolazine for Less Ischemia in Non-ST Elevation Acute Coronary Syndromes Study) demonstrated that ranolazine lowered HbA_1c levels in patients with diabetes and reduced the incidence of newly elevated HbA_1c levels in initially normoglycemic patients (15,16). In nonclinical studies, ranolazine was found to lower fasting and nonfasting glucose levels, and to preserve pancreatic β-cells in streptozotocin (STZ)-treated mice and Zucker diabetic fatty (ZDF) rats (17,18); however, the mechanisms underlying the improvement in glycemia were not elucidated in these studies. In the current study, we tested the hypothesis that the antidiabetic effects of ranolazine may be due to the inhibition of glucagon secretion from pancreatic islets via blockade of NaChs in α-cells.

The results of the current study show that NaCh blockers inhibit glucagon secretion by blocking the Na_v1.3 isoform of pancreatic α-cells, which leads to glucagon- and glucose-lowering effects in animal models of diabetes. Given the role of I_Na and Na⁺-dependent action potentials of pancreatic α-cells in the secretion of glucagon, Na_v1.3-selective NaCh blockers may provide a novel mechanism for the treatment of hyperglycemia.

All animal studies were performed according to the protocols approved by the Institutional Animal Care and Use Committee of Gilead Sciences. Sprague-Dawley (SD) rats, ZDF rats, and lean control rats were purchased from Charles River Laboratories (Hollister, CA/Wilmington, MA). Animals were housed on a 12-h light/dark cycle at 22–25°C with ad libitum access to food and water.

Unless noted, all reagents were purchased from Sigma-Aldrich (St. Louis, MO). Ranolazine and GS-458967 (triazolopyridine derivative, 6-(4-(trifluoromethoxy)phenyl)-3-(trifluoromethyl)-[1,2,4]triazolo[4,3-a]pyridine) (19) were synthesized by Gilead Sciences. TTX was purchased from Tocris Bioscience (Minneapolis, MN).

Pancreatic islets were isolated from male SD rats (300–400 g) using Liberase TL (Roche Diagnostics, Dallas, TX) and Ficoll gradients (Mediatech, Manassas, VA) (20). Adult human pancreatic islets were obtained from donors of either sex ranging in age between 32 and 65 years, with BMI ranging between 21 and 41 kg/m² (National Disease Research Interchange, Philadelphia, PA, or Prodo Laboratories, Irvine, CA). Islets were cultured in islet culture medium (RPMI 1640 medium containing 10% FBS, 11 mmol/L glucose, 100 units/mL penicillin, 100 μg/mL streptomycin, 2 mmol/L l-glutamine, 10 mmol/L HEPES, 1 mmol/L sodium pyruvate) for 2–7 days. Equal sized islets were transferred to a 96-well plate with 10 islets per well and then treated with various agents in 150 μL of Krebs-Ringer buffer (in mmol/L: 129 NaCl, 4.8 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, 5 NaHCO₃, and 10 HEPES, pH 7.4) with 0.1% BSA and 3 mmol/L glucose for 1 h. Supernatants were collected and stored at −80°C until analysis.

Cultured islets were dispersed by Accutase, and the cell suspension was filtered through a 40-μm cell strainer, and washed once with Dulbecco’s PBS and 4% BSA. Islet cells were resuspended in Opti-MEM I and transfected with 50 nmol/L small interfering RNAs (siRNAs; Santa Cruz Biotechnology, Dallas, TX) using Lipofectamine 2000 (Life Technologies, Carlsbad, CA). A portion of the transfected cells was collected 48 h later for quantitative PCR (qPCR). Another portion was seeded in a laminin/poly-d-lysine–coated, 96-well plate (BD Biosciences, San Jose, CA) at 1.5 × 10⁴ cells/well and cultured for 48 h for glucagon secretion experiments.

Total RNA was extracted using TRIzol reagent (Life Technologies). cDNA was synthesized using an iScript Reverse Transcription kit (Bio-Rad, Hercules, CA). qPCR was performed using SYBR Green PCR Master Mix reagent (Life Technologies) on a Stratagene Mx3000P qPCR system (Agilent Technologies, Santa Clara, CA). The expression level of a gene was normalized relative to β-actin. Primers for α-subunits of human sodium channel isoforms were designed with Beacon Designer 7.0 (Premier Biosoft, Palo Alto, CA), and sequences are shown in Supplementary Table 1.

Rat islets were dispersed as described above, plated on poly-d-lysine/laminin-coated coverslips (BD BioCoat; BD Biosciences), and transfected with a rat glucagon promoter-EGFP plasmid to identify α-cells (21). Where indicated, α-cells were also cotransfected with either Scn3a-targeted or control siRNA (Life Technologies). Membrane voltage (V_m) and peak I_Na were recorded 48–72 h after dispersion using the perforated patch configuration, the 200B Axopatch amplifier, and pClamp version 10.2. Cells without spontaneous action potentials were excluded from analysis, and all compounds/drugs were dissolved in the bath solution.

Recordings of V_m were made at 32°C using the current-clamp configuration. The bath solution contained the following (in mmol/L): 140 NaCl, 5 HEPES, 3.6 KCl, 2 NaHCO₃, 0.5 NaH₂PO₄, 0.5 MgSO₄, 2.6 CaCl₂, 10 dextrose, and 10 sucrose, pH 7.35. Pipettes (3.5–5.0 MΩ) were pulled from borosilicate glass and tip filled with an internal solution consisting of the following (in mmol/L): 76 K₂SO₄, 10 KCl, 10 NaCl, 5 HEPES, 1 MgCl₂, pH 7.35, and amphotericin B (0.3 mg/mL). Series resistance (R_S) was monitored using a voltage step from −70 to 0 mV (5 ms, 0.5 Hz) and was allowed to stabilize prior to the beginning of the experiment (R_S <30 MΩ). Recordings (30 s) were analyzed using the event detection (threshold 10 mV) to quantitate total charge movement (area), resting V_m, action potential amplitude, and action potential frequency. I_Na recordings were made at room temperature in the voltage-clamp configuration. The I_Na bath solution contained the following (in mmol/L): 130 NaCl, 5 HEPES, 3.6 KCl, 2 NaHCO₃, 0.5 NaH₂PO₄, 0.5 MgSO₄, 2.6 CaCl₂, 3 dextrose, 20 TEA-Cl, 10 4-AP, 2.5 CoCl₂, and 0.5 tolbutamide, pH 7.35. Leak currents were subtracted using an online P/4 procedure, and R_S was compensated. I_Na averaged over multiple sweeps was analyzed before and after the application of the compound or drug.

Male ZDF and lean rats (5 weeks old) were acclimatized for 1 week. Drugs were given to the animals in Purina 5008 diet for 10 weeks at doses of ∼170 mg/kg/d ranolazine, 0.6 mg/kg/d GS-458967, and 30 mg/kg/d sitagliptin (Discovery Fine Chemicals, Dorset, U.K.). This dose of ranolazine yielded plasma concentrations between 8 and 10 μmol/L, which are slightly higher than the therapeutic concentration (6–8 μmol/L). The dose of GS-458967 was selected to yield plasma concentrations of 1–2 μmol/L, which is close to the half-maximal inhibitory concentration value for the inhibition of Na_v1.3 (0.6 μmol/L). Blood samples were collected via retro-orbital bleed under nonfasting and fasting (4–6 h) conditions with addition of aprotinin (Thermo Fisher Scientific, Pittsburgh, PA).

Diabetes was induced in SD rats using a single intraperitoneal injection of STZ (65 mg/kg). Fasting plasma glucose (FPG) and glucagon levels were measured before and 2 weeks after STZ injection. Diabetic and control SD rats were randomized to different groups for the oral glucose tolerance test (OGTT). After an overnight fast, each group was treated with ranolazine (30 mg/kg p.o.), GS-458967 (0.5 mg/kg p.o.), or vehicle at −15 min followed by a glucose load (2 g/kg in water) at 0 min. Blood samples were obtained via retro-orbital bleed.

After an overnight fast, male SD rats (12 weeks old) were dosed with vehicle, ranolazine (30 mg/kg p.o., at −15 min), glibenclamide (5 mg/kg p.o., at −15 min), or GS-458967 (0.5 mg/kg p.o., at −60 min), followed by insulin infusion (5 mU/min i.v.) for 60 min.

Glucose and HbA_1c levels were measured by clinical chemistry analyzer (model AU400; Olympus). Glucagon levels were measured by an ELISA kit (R&D Systems, Minneapolis, MN). The sensitivity of the glucagon assay is 14.7 pg/mL, and the reading range is 31.2–2,000 pg/mL. The intra-assay variation (coefficient of variability) is 2.7–3.6%, and the interassay variation (coefficient of variability) is 5.8–8.9%.

Three pancreatic sections from six ZDF rats per group were analyzed. The sections were immunostained with primary antibodies (insulin and glucagon) followed by secondary antibodies (fluorescein isothiocyanate; Jackson ImmunoResearch, West Grove, PA). Stained sections were photographed under a fluorescent microscope and quantitated using ImageJ software (NIH, Bethesda, MD).

All results are presented as the mean ± SE. Data were analyzed using GraphPad Prism7 or OriginPro7. Differences among groups were analyzed using a Student t test, or one-way or two-way ANOVA followed by appropriate post hoc tests. A P value <0.05 was considered statistically significant.

The effect of NaCh blockers on glucagon secretion was determined in human and rat pancreatic islets. Ranolazine, GS-458967 (a novel and selective NaCh blocker [19]), and TTX (a broad-spectrum NaCh blocker) significantly reduced glucagon secretion in a concentration-dependent manner in both rat and human islets (Fig. 1A–C and G). Compared with the control, maximal reductions of glucagon secretion were observed with ranolazine at 10 μmol/L (25 ± 3%), GS-458967 at 3 μmol/L (51 ± 9%), and TTX at 0.1 μmol/L (64 ± 3%) in human islets. In rat islets, ranolazine decreased glucagon secretion by 36 ± 6% at 10 μmol/L.

In contrast, veratridine, a NaCh activator (opener), increased glucagon secretion in human pancreatic islets in a concentration-dependent manner (data not shown), with a 10-fold increase at 30 μmol/L (Fig. 1D–F). Ranolazine, GS-458967, and TTX significantly reduced the veratridine (30 μmol/L)-induced increase in glucagon secretion in a concentration-dependent manner (Fig. 1D–F). Glucagon secretion was reduced by 36 ± 11%, 71 ± 10%, and 79 ± 8%, respectively, with ranolazine (10 μmol/L), GS-458967 (3 μmol/L), and TTX (0.1 μmol/L). In rat islets, glucagon secretion was reduced by 69 ± 11%, 85 ± 7%, and 83 ± 12%, respectively, with ranolazine (10 μmol/L), GS-458967 (1 μmol/L), and TTX (30 nmol/L) (Fig. 1H and I). Insulin secretion was not affected by these NaCh blockers under these conditions (Supplementary Fig. 1).

Gene expression levels of various isoforms of NaChs in human and rat pancreatic islets were determined by qPCR. Data show that Na_v1.3 (SCN3A) and Na_v1.7 (SCN9A) are the two major isoforms expressed in human pancreatic islets, followed by the Na_v1.2 (SCN2A) and Na_v1.6 (SCN8A) isoforms, which are expressed at much lower levels (Fig. 2A), whereas in rat islets the only major isoform expressed Na_v1.3 (Scn3a) (Fig. 2B).

Functional roles of Na_v1.3 and Na_v1.7 in glucagon secretion were determined by silencing the expression of these two isoforms using the siRNA approach in cells dispersed from human islets. Specific knockdown of Na_v1.3 or Na_v1.7 resulted in an average decrease of 66 ± 6% and 56 ± 2%, respectively, in the gene expression of Na_v1.3 or Na_v1.7 compared with the control (Fig. 2C). Knockdown of Na_v1.3, but not Na_v1.7, caused a significant reduction of veratridine-induced increase of glucagon secretion. Glucagon secretion induced by 10 and 30 μmol/L concentrations of veratridine was attenuated by 98 ± 3% and 85 ± 5%, respectively, after knockdown of Na_v1.3 in human islets (Fig. 2D). These data suggest that Na_v1.3 represents the functionally important isoform of NaChs that mediates glucagon secretion from pancreatic α-cells.

Ranolazine and GS-458967 reduced the peak I_Na (no late component was observed) in α-cells by 25 ± 4% at 10 μmol/L and 55 ± 7% at 1 μmol/L, whereas 0.3 μmol/L TTX caused a complete block, which is consistent with the selective expression of TTX-sensitive NaChs in α-cells (Fig. 3A and B). The spontaneous electrical activity of α-cells was inhibited by 63 ± 9% with 10 μmol/L ranolazine, 64 ± 5% with 1 μmol/L GS-458967, and 89 ± 4% with 0.3 μmol/L TTX (Fig. 3D and E). Detailed analysis of the electrical activity revealed that NaCh inhibition predominately reduced the amplitude and frequency of the activity (Supplementary Fig. 2C and E).

Knockdown of Na_v1.3 with targeted siRNA in α-cells reduced the peak I_Na by 85% compared with a non-targeted siRNA control (−8 ± 5 vs. −57 ± 5 pA/pF, respectively; Fig. 3C). Accordingly, the reduction of I_Na by knockdown of Na_v1.3 caused a reduction in the number of cells showing spontaneous electrical activity (0 of 5 cells), but not in non-targeted siRNA (3 of 5 cells).

Because we found that veratridine increases the release of glucagon (Fig. 1), we investigated whether veratridine, which is known to increase I_Na, also increases the electrical activity of α-cells. Figure 3F illustrates that 15 μmol/L veratridine evoked an increase (9- ± 2-fold) in α-cell electrical activity. In the presence of veratridine, the inhibition of I_Na with either ranolazine or GS-458967 was sufficient to nearly normalize α-cell electrical activity. The average inhibition of electrical activity was 71 ± 11% for 10 μmol/L ranolazine and 88 ± 2% for 0.3 μmol/L GS-458967 (Fig. 3G). Veratridine evoked a small increase in the frequency of electrical activity, which was reversed by NaCh inhibition (Supplementary Fig. 2F).

The in vivo glucagon-lowering effects of ranolazine and GS-458967 were determined in diabetic (STZ-treated) rats. Two weeks after STZ injection, baseline glucagon and glucose levels were markedly elevated in STZ-treated rats compared with the untreated group (Fig. 4A and C). In response to an oral glucose load, there was no increase in insulin levels in STZ diabetic rats (Fig. 4E). Glucagon levels in the STZ-plus-vehicle group (Fig. 4A) were significantly higher, peaking at 15 min (307 ± 11 pg/mL), than in the control group (130 ± 10 pg/mL). Glucose levels were also significantly higher in the STZ-plus-vehicle group (Fig. 4C), peaking at 30 min at 751 ± 16 mg/dL compared with 256 ± 13 mg/dL in the control group. The short-term administration of both ranolazine and GS-458967 reduced the increase in glucagon levels in response to the oral glucose load, reaching significance at 15 min compared with the STZ-plus-vehicle group (Fig. 4A). Accordingly, the change in area under the curve (ΔAUC) for glucagon was significantly smaller in the ranolazine- and GS-458967-treated groups compared with the vehicle-treated group (Fig. 4B). This decrease in glucagon levels was associated with significantly lower glucose concentrations (Fig. 4C) at 15 and 30 min in compound-treated groups compared with the vehicle-treated group. Accordingly, the ΔAUC for glucose was also smaller in ranolazine- and GS-458967-treated rats (Fig. 4D). There was no effect of ranolazine or GS-458967 treatment on insulin levels (Fig. 4E and F).

Long-term treatment with ranolazine, GS-458967, or the dipeptidyl peptidase 4 inhibitor sitagliptin (used as a positive control) in diabetic ZDF rats reduced fasting glucose, glucagon, and HbA_1c levels. The vehicle-treated ZDF group had fasting hyperglycemia (from 105 ± 5 at baseline to 341 ± 12 mg/dL) and hyperglucagonemia (from 130 ± 14 at baseline to 312 ± 14 pg/mL) at 10 weeks. The ranolazine-, GS-458967-, and sitagliptin-treated groups had significantly lower FPG levels from week 4 to week 10 (Fig. 5A) compared with the vehicle-treated group. Long-term treatment with ranolazine or GS-458967 also delayed the onset of hyperglucagonemia during the development of diabetes starting between week 4 and week 10. These effects of ranolazine and GS-458967 were comparable to that of sitagliptin (Fig. 5B). The HbA_1c level was 3.7–3.8% at baseline in the five groups and increased to 10.3 ± 0.2% at 10 weeks in the vehicle-treated ZDF group. Consistent with the improvement in glucagon and glucose levels, HbA_1c levels were significantly lower in the groups treated with ranolazine (6.3 ± 0.4%), GS-458967 (6.2 ± 0.5%), and sitagliptin (7.4 ± 0.8%) than in the group treated with vehicle (10.2 ± 0.2%) at week 10 (Fig. 5C).

Pancreatic islets from rats treated with ranolazine or GS-458967 showed more insulin and less glucagon staining compared with the vehicle-treated rats (Fig. 5D). With the development of diabetes, the clear round boundary of the islets was destroyed in the vehicle-treated ZDF group compared with the lean rats (Fig. 5D). Relative to the lean rats, the insulin-positive area (91 ± 2%) was smaller, whereas the glucagon-positive area (156 ± 18%) was significantly increased in the vehicle-treated ZDF rats (Fig. 5E and F). Treatment with ranolazine or GS-458967 significantly increased the portion of insulin-positive cells (104 ± 2% and 103 ± 3%, respectively) and decreased the portion of glucagon-positive cells (70 ± 12% and 79 ± 23%, respectively), which led to a partial reversal of the abnormal ratio between the insulin- and glucagon-positive areas (vehicle 4.7 ± 1; ranolazine 12 ± 2; GS-458967 12 ± 3) (Fig. 5G). These effects of ranolazine and GS-458967 were comparable to those of sitagliptin (10 ± 2 ratio).

To evaluate the potential for NaCh blockers to interfere with hypoglycemia-induced glucagon release, the effects of ranolazine and GS-458967 were studied during insulin-induced hypoglycemia in normal rats (Fig. 6). As expected, glucagon levels in the vehicle-treated group increased significantly at 40 min (671 ± 80 pg/mL) in response to hypoglycemia (Fig. 6A and B). Treatment with ranolazine (Fig. 6B) or GS-458967 (Fig. 6C and D) did not suppress the increase in glucagon levels, whereas glibenclamide (Fig. 6B), an insulin secretagogue used in clinical practice, significantly decreased the hypoglycemia-induced glucagon response (331 ± 57 pg/mL at 40 min). In addition, ranolazine and GS-458967 treatment did not affect the recovery from hypoglycemia after stopping insulin infusion. The glibenclamide-treated group, on the other hand, showed a significant delay in recovery from hypoglycemia (Fig. 6A).

The current study elucidates the mechanism underlying the antidiabetic effect of a NaCh blocker, ranolazine. The main findings of the study are as follows: 1) ranolazine inhibits the secretion of glucagon from human pancreatic islets; 2) the inhibition of glucagon secretion is due to the blockade of α-cell TTX-sensitive peak I_Na and a concomitant decrease in electrical activity; 3) Na_V1.3 is the likely NaCh isoform in α-cells that mediates the glucagonostatic effect of NaCh blockers; and 4) treatment with ranolazine lowers glucagon and glucose levels in diabetic ZDF rats. Furthermore, another selective NaCh blocker, GS-458967, has similar glucagon-lowering effects in vitro and also has antidiabetic effects in rodent models of diabetes.

TTX-sensitive voltage-gated NaChs in pancreatic α-cells support the generation of electrical activity, which increases intracellular calcium levels and causes exocytosis of glucagon (10,22). All NaCh blockers inhibited the electrical activity of α-cells and reduced the release of glucagon in a concentration-dependent manner (Figs. 1 and 2), whereas the NaCh activator veratridine enhanced the electrical activity of α-cells and increased glucagon secretion from intact islets (Figs. 1 and 3). These findings are consistent with previous reports demonstrating that NaChs can directly modulate glucagon secretion in intact islets obtained from nondiabetic rodents and humans (8–11,22,23). A recent report demonstrated no effect of TTX on glucagon release (24); however, the reasons for this discrepancy are not well understood. Our data show that the inhibition of glucagon secretion by NaCh blockers correlates with the inhibition of I_Na. For example, ranolazine at 10 μmol/L caused a 25% inhibition of I_Na in α-cells and a 36% reduction in glucagon release from rat islets. Although the role of α-cell dysfunction supporting increased glucagon release in diabetes has not been well characterized, it was reported recently that α-cells from diabetic mice, with elevated circulating glucagon levels, have enhanced I_Na density, and increased action potential amplitude and firing frequency (11). These data suggest that an increase in I_Na may be an important mechanism underlying the hypersecretion of glucagon from α-cells and imply that the inhibition of I_Na could restore normal glucose homeostasis in diabetes.

Ranolazine and GS-458967 did not affect insulin secretion (Supplementary Fig. 1) under conditions where glucagon inhibition was observed. Although a previous study (18) with ranolazine showed an increase in insulin secretion under high glucose conditions, the mechanism of this effect is not well understood at this time, but seems to be independent of NaCh blockade because GS-458967 does not have any effect on insulin secretion even under high glucose conditions.

Among the NaCh isoforms present in the excitable cells/tissues, the NaCh isoforms expressed at the highest level in human pancreatic islets are Na_v1.3 and Na_v1.7. Of the two, the Na_v1.3 seems to be functionally responsible for glucagon secretion, as the knockdown of Na_v1.3 gene expression caused a significant reduction in I_Na, electrical activity, and glucagon release. Consistent with our findings, a recent report (25) identified Na_v1.3 as the isoform responsible for glucagon secretion in mouse islets. Ranolazine and GS-458967, which lack isoform selectivity, inhibit Na_v1.3 with half-maximal inhibitory concentration values of 12 μmol/L (26) and 0.6 μmol/L, respectively, which is consistent with the data for the inhibition of glucagon release.

Postprandial hyperglycemia and hyperglucagonemia are characteristics of patients with diabetes (1). Reducing the postprandial paradoxical increase in glucagon response is a potential target for lowering postprandial glucose levels. Both ranolazine and GS-458967, given in the short term, significantly reduced the glucagon response to the glucose load in diabetic rats. This is the first demonstration of a selective NaCh blocker reducing the increase in glucagon levels during an oral glucose load in diabetic animals. The effect was similar to that reported with GLP-1 receptor agonists, which also inhibit glucagon secretion, albeit via a different mechanism (27).

Long-term administration of ranolazine or GS-458967 significantly lowered glucagon, glucose, and HbA_1c levels in ZDF rats. Relative to the vehicle group, ZDF rats treated with ranolazine or GS-458967 for 10 weeks had more insulin-producing and fewer glucagon-producing cells. Even though NaCh blockers prevented the increase in glucagon levels, glucose levels continued to increase with time, suggesting that glucagon is only partially responsible for hyperglycemia in this model. At later time points, the improvement in β-cell mass observed may also contribute to the lowering of glucose levels in drug-treated groups. The magnitude of these effects of ranolazine was comparable to that of the effects of sitagliptin. In clinical studies, ranolazine has been reported to lower HbA_1c between 0.5 and 0.7% (14), which is similar in magnitude to what has been reported for sitagliptin.

Ranolazine and GS-458967 did not affect the hypoglycemia-induced secretion of glucagon or the recovery from hypoglycemia (Fig. 6). These data suggest that the effect of ranolazine and GS-458967 on glucagon is unlikely to be mediated via inhibition of the autonomic nervous system, because it has been shown that a substantial portion of the hypoglycemia-induced increase in glucagon levels is mediated via activation of the autonomic nervous system (28). This is consistent with the very low incidence of hypoglycemic events observed with ranolazine in clinical practice. In our study, 10 μmol/L ranolazine reduced normal or veratridine-evoked glucagon release from pancreatic islets by 25–40%, respectively, and 100 nmol/L TTX reduced glucagon release by no more than 80%. These results suggest that additional depolarizing currents, such as calcium-mediated depolarization could prevent the complete suppression of glucagon secretion after I_Na inhibition. Therefore, the glucagonostatic actions of NaCh blockers are not expected to compromise the physiologic defense against hypoglycemia.

The abnormal glucagon secretion in T2DM and its key role in the development of fasting and postprandial hyperglycemia has recently been a focus of attention (1,7,29). Therefore, the suppression of glucagon secretion or the inhibition of its action in the liver constitutes potential therapeutic targets for diabetes (30–32). The results of studies with glucagon receptor antagonists (peptide and small molecule), antisense oligonucleotides, and glucagon receptor knock-out mice show that elimination or inhibition of glucagon receptor signaling has strong anti-hyperglycemic effects in various animal models of diabetes (31,33–35). Significant dose-dependent reductions in HbA_1c levels have been reported (33,36) with glucagon receptor antagonists MK-0893 (0.6–1.5 ± 0.5%, compared with placebo) and LY-2409021 (0.7–1%) in patients with T2DM.

The glucose-lowering effects of glucagon receptor antagonists are also associated with increases in levels of glucagon, liver enzymes, and LDL cholesterol, and in α-cell hypertrophy and hyperplasia (37,38). Hyperglucagonemia, accompanied by hypertrophy and hyperplasia of the pancreatic α-cells, has been observed in glucagon receptor knock-out mice (34) and mice treated with glucagon receptor antisense oligonucleotides (35). In contrast, in our study ZDF rats treated with NaCh blockers for 8 weeks had no change in lipid levels (Supplementary Table 2), had lower circulating glucagon levels, and had fewer α-cells than vehicle-treated rats (Fig. 5), a finding that is consistent with the results of our previous study (18). Together, these data suggest that the normalization of circulating glucagon levels by correcting abnormal glucagon secretion by α-cells may be a better approach than blocking glucagon receptor signaling in the liver for the treatment of diabetes. It has been suggested that inappropriately high glucagon levels in diabetes are due to impairment in α-cell function (39). Furthermore, α-cells from diabetic mice have enhanced I_Na density, and increased action potential amplitude and firing frequency with no change in Ca²⁺ currents (11). The authors proposed that alterations in the electrical properties of α-cells may prime the cells for enhanced glucagon secretion, with no change in Ca²⁺ currents. These data suggest that an increase in I_Na may be an important underlying mechanism for glucagon hypersecretion in diabetes, and our results are consistent with this interpretation. This hypothesis is consistent with our data in that the effect of ranolazine on glucagon secretion in normal islets and in normal animals is modest to minimal (when NaCh function is normal), but the effect is enhanced when NaChs are activated by veratridine in islets or in diabetic animals. Although the findings in mouse α-cells need to be confirmed in human α-cells, the HbA_1c-lowering effect of ranolazine observed in clinical studies is also consistent with this hypothesis (14,15). Furthermore, clinical data about ranolazine show that the HbA_1c-lowering effect of ranolazine is greater in patients starting with higher levels of HbA_1c (16). Although glucagon levels in these studies were not measured, higher HbA_1c levels indirectly suggest a greater impairment of β-cell function (i.e., insulin loss), resulting in the hyperactivity of α-cells, which in turn leads to higher glucagon levels. In this context, ranolazine would be expected to be more effective in patients exhibiting abnormally elevated α-cell electrical activity and glucagon secretion.

The HbA_1c-lowering effect of ranolazine has been previously demonstrated in three clinical studies, but the mechanism of this effect remained unclear. The current study shows that NaCh blockers inhibit glucagon release from pancreatic islets and also have antidiabetic effects due to the direct inhibition of I_Na in α-cells. Overall, these data suggest that a major factor contributing to increased glucagon levels may lie at the α-cell level (i.e., hypersecretion of glucagon), which can be corrected by the blockade of Na_v1.3 channels. Although the role of NaChs in the pathophysiology of human diabetes requires further investigation, the findings from the current study suggest that the inhibition of α-cell I_Na could become an attractive drug target for combination therapy with other classes of antidiabetic agents.

Acknowledgments. The authors thank Dr. Filip K. Knop from University of Copenhagen for his critical review of the manuscript and Dr. Jeffery Chisholm from National Research Council of Canada for his early contribution to the project. The authors also thank Ruth Chu, Jenny Jiang, Michael Van Petten, and Cheng Xie for their technical assistance with collection of data.

Duality of Interest. All authors are employees of Gilead Sciences, Inc. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. A.K.D. provided input for the study design and the interpretation of the data, and wrote, reviewed, edited, and approved the manuscript. M.Y., Y.N., K.M.K., and M.K. designed and carried out the experiments; collected and analyzed the data; and reviewed, edited and approved the manuscript. S.R. and L.B. contributed to the study design and the interpretation of the data; and reviewed, edited, and approved the manuscript. A.K.D. and S.R. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.