Molecular Structure of K_ATP Channel and SUR1 Nucleotide Binding Domains

The molecular structure of K_ATP channels consists of hetero-octameric complexes formed by four pore-forming inward rectifier potassium channel (Kir6.x) subunits and four sulfonylurea receptor (SURx) subunits (Aguilar-Bryan et al., 1995). There are two isoforms of inward rectifier potassium channel Kir6.1 and Kir6.2 (Clement et al., 1997). The sulfonylurea receptor is an enzymatic member of the ABC (ATP-binding cassette) superfamily, which uses the energy of ATP binding and hydrolysis to transport substrates across the cell membrane. This receptor has two nucleotide binding sites (Wilkens, 2015).

There are three isoforms of the sulfonylurea receptor: SUR1, SUR2A, and SUR2B (Devaraneni et al., 2015). The sulfonylurea receptor is a regulatory subunit because it: 1) confers sensitivity to Mg-nucleotides, 2) is activated by K⁺ channel openers such as diazoxide and pinacidil, and 3) is inhibited by sulfonylureas such as glibenclamide and tolbutamide (Aguilar-Bryan and Bryan, 1999). In addition, SUR1 increases the open probability of Kir6.2 via phosphatidylinositol-4, 5-biphosphate (PIP2) (Song and Ashcroft, 2001) and may increase the sensitivity to ATP (Tucker et al., 1997).

The combination of Kir6.2 (encoded by KCNJ11) and SUR1 (encoded by ABCC8) subunits form the K_ATP channel in beta cells (Fig. 1), and this combination is also present in alpha and delta cells from the islet (Clark and Proks, 2010; Ashcroft and Rorsman, 2013). Kir6.2 is a member of the inward rectifier potassium channels superfamily formed by two transmembrane domains, M1 and M2. These domains are linked by a sequence of amino acids that partially enters the membrane to form the pore (loop P) and a signature sequence of glycine-phenylalanine-glycine that corresponds to the K⁺ selectivity filter (see Fig. 1) (Xie et al., 2007). Furthermore, the N and C termini are intracellular in Kir6.2 channels and are structurally important in the regulation by ATP and sulfonylureas.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Pancreatic K_ATP channel structure: (A) Membrane topology of Kir6.2 channel and sulfonylurea receptor (SUR1) subunits. Functional K_ATP (right) is a 4:4 octamer. Transmembrane domains (TMD0, TMD1, and TMD2): Residues marked in red are associated with neonatal diabetes mellitus (NDM) and those marked in purple are associated with congenital hyperinsulinism (CHI). Retention motif (RKR) is marked in brown. (B) Cooperativity between nucleotide binding domains (NBD1 and NBD2) to form ATP-binding sites, walker A and B motif, loop of glutamine (loop-Q), histidine (loop-H) and aspartate (loop-D). NBS1 binds ATP or MgATP and NBS2 binds MgADP or MgATP. ABC signature sequence (LSGGQ): (C) Pharmacologic modulators of pancreatic K_ATP channels.

The SUR1 receptor is formed by two transmembrane domains (TMD1 and TMD2), each one with six transmembrane helices, and two cytosolic H (NBD1 and NBD2) subunits between the TMDs (Aguilar-Bryan and Bryan, 1999). Moreover, the N-terminal domain (TMD0) has five transmembrane helices connected to the TMD1 through a long cytosolic loop known as the L0 or CL3 linker (Fig. 1) (Devaraneni et al., 2015).

Both NBDs have a highly conserved structure, and each one is formed by a catalytic and an α-helical subdomain. The catalytic subdomain contains Walker A and B conserved motifs that play important roles in ATP binding and hydrolysis. The A motif (P-loop) contains a highly conserved lysine residue that interacts with the β and γ phosphates of the nucleotides (Devaraneni et al., 2015).

The B motif contains an aspartate residue that coordinates Mg²⁺ binding (ter Beek et al., 2014) and a basic glutamate residue that binds the γ-phosphate of ATP through a water molecule (Beis, 2015). Downstream the Walker B motif, the catalytic base consists on a glutamate residue that both hydrolyzes ATP and forms a dyad with the H-loop histidine of the antiparallel NBD (Jones and George, 2012). In addition, the catalytic domain contains functionally relevant residues, such as aspartate (D-loop), that allow the formation of the hydrolytic-competent state and enable the ATP hydrolysis.

The α-helical domain contains both the ABC signature sequence (LSGGQ), also known as C-loop, and a Q-loop (Beis, 2015). Both NBDs are grouped in an antiparallel manner to form two ATP binding sites (NBS). Thus, Walker A and B of NBD1 and the signature sequence of NBD2 form the first ATP-binding site, where nucleotides are located at the central part of the sandwich (Fig. 1). While in the second binding site, A and B motifs of NBD2 are associated with the signature sequence of NBD1(Vedovato et al., 2015).

Pharmacology of K_ATP Channels in Pancreatic Beta Cells

Pancreatic K_ATP channelopathies

Functional and structural defects in K_ATP channels impair insulin secretion, leading to the onset of diseases. In 1981, it was first described that glucose failed to promote insulin secretion in a concentration-dependent manner in isolated tissue from a child with hyperinsulinemia-induced hypoglycemia (Aynsley-Green, 1981; Dunne, 2000). The persistent hyperinsulinemia hypoglycemia of infancy, also named congenital hyperinsulinism (CHI), is a complex disorder composed of clinical, morphologic, and genetic changes (Shah et al., 2014). Excessive insulin secretion characterizes CHI, despite low blood glucose levels in the early neonatal period. This disorder has an incidence of around 1 in 50,000 live births.

Clinically, CHI is classified in mild and severe forms of the disease. The first one appears during the newborn period, but also in infancy and childhood. In the severe form the symptoms develop after the first hours or days after birth. Failure in the diagnosis and treatment of hypoglycemia could lead to severe brain damage and epilepsy (Kapoor et al., 2009).

Histologically, CHI is divided into diffuse, focal, and atypical forms. The focal form is sporadic in inheritance and only affects discrete areas of the pancreas. The diffuse disease is inherited in an autosomal recessive or dominant way. The latter is the most common of CHI and affects the entire pancreas. Finally, the atypical forms may be diffuse along the gland with normal or abnormal islet anatomy (Shah et al., 2014).

The etiology of CHI may be due to rare variants in nine different genes, such as: 1) ABCC8 (SUR1 receptor), 2) KCNJ11 (Kir6.2 channel), 3) GCK (glucokinase), 4) GLUD1 (mitochondrial enzyme glutamate dehydrogenase), 5) SLC16A1 [monocarboxylate transporter 1 (MCT1)], 6) HADH (mitochondrial L-3-hydroxyacyl-coenzyme A dehydrogenase), 7) UCP2 (uncoupling protein 2), and 8 and 9) HNF4A and HNF1A (hepatocyte nuclear factor 4alpha and 1alpha). All of these genes encode for proteins involved in beta-cell insulin secretion (Roženková et al., 2015). The most frequent causes for CHI are variants in KCNJ11 and ABCC8, which account for ∼70% of all cases with more than 200 mutations reported to date, some of them shown in Fig. 1. Both are inherited either by autosomal recessive or dominant form (Remedi and Koster, 2010).

K_ATP channel variants can also be grouped into two functional classes: those that reduce: i) the cellular membrane expression by affecting either channel biosynthesis, trafficking, assembly; or ii) reduce the intrinsic channel activity (Remedi and Koster, 2010). Genetic variants in ABCC8 and KCNJ11 along with HNF1A and GCK are associated not only with CHI but also with maturity onset of diabetes in young and neonatal diabetes mellitus (NDM), each of them are monogenic forms of glucose-regulation disorders in children.

Neonatal diabetes mellitus (NDM), caused either by Kir6.2 (KCNJ11) or SUR1 (ABCC8) mutations, is a rare disorder that appears during the first 6–9 months of life, with an incidence of 1:300,000 live births. The patients typically have low weight at birth, and only one-third of them develop severe ketoacidosis. NDM can be permanent or transient (TNDM) and requires lifetime glycemic control or until a remission period. Moreover, some patients present a more severe form of NDM with severe developmental delay and epilepsy, whereas other patients show a milder phenotype without epilepsy.

NDM-associated K_ATP channel mutations cause the opposite phenotype of CHI. Therefore, they have an overactive channel that remains opened, either increasing the K_ATP activity mediated by Mg nucleotides or altering the intrinsic gating (Edghill et al., 2010; Remedi and Koster, 2010).

Gloyn et al. (2004) reported genetic variants in Kir6.2 as the potential cause of NDM for the first time in 2004. To date, over 30 rare variants in Kir6.2 and approximately 15 in SUR1 have been identified and associated with NDM. These are the most common cause of permanent NDM (approximately 30–50%) and involved in 15% of the TNDM cases (Shimomura, 2009; Shimomura et al., 2009). TNDM also correlates with mutations in hepatocyte nuclear factor-1beta (HNF1b), which can also cause maturity-onset diabetes of the young 5. Furthermore, the majority of cases of TNDM are due to abnormalities of an imprinted locus on chromosome 6q24 that results in the overexpression of a paternally expressed gene (Aguilar-Bryan and Bryan, 2008).

Electrophysiological studies have observed that Kir6.2 variants at Arg50, Ile192, Arg201, and Phe333 residues involved in ATP binding decrease the K_ATP sensitivity. In addition, variants of Val59, Cys166, Ile197, and Ile296 residues involved in the channel gating impair the open probability of the channel (Proks et al., 2005; Edghill et al., 2010).

SUR1 mutations affect either the Mg-nucleotide-mediated activation or the intrinsic gating. These mutations could be located along the protein length, including many instances of the following amino acid substitutions V1523A, V1524M, I1425V, and R1531A in NBD2 (Edghill et al., 2010).

Mutations and polymorphism in Kir6.2 (KCNJ11) have already been linked to type 2 diabetes mellitus (Gloyn et al., 2003). Several genetic studies consistently demonstrated an association of the E23K polymorphism with a reduced insulin secretion in glucose-tolerant and -intolerant cohorts as a risk allele in the development of type 2 diabetes (Nielsen et al., 2003; Chistiakov et al., 2009; Villareal et al., 2009).

The E23K polymorphism is due to a substitution of a glutamic acid (E) by a lysine (K) at codon 23 of the KCNJ11. This residue substitution is located in the N-terminal domain of the Kir6.2 subunit (Bonfanti et al., 2015).

Electrophysiological recording of reconstituted K_ATP channels in mammalian expression consistently shows hyperactivity of E23K channels, which has been suggested to increase the intrinsic channel open probability with a decrease in the sensitivity to ATP inhibition (Villareal et al., 2009; Remedi and Koster, 2010) or enhanced activation by long-chain acyl coenzyme A (CoA) esters with minimal effects on ATP sensitivity (Riedel and Light, 2005).

Impairment of the K_ATP channel activity can alter beta-cell electrical activity and insulin secretion sufficiently to cause diabetes. The Kir6.2-G324R mutation reduces the channel ATP sensitivity in beta cells; however, the difference in ATP inhibition between homozygous and heterozygous patients was remarkably small. Nevertheless, the homozygous patient developed neonatal diabetes, whereas the heterozygous parents were unaffected (Vedovato et al., 2016).

Endogenous Regulation of K_ATP Channels in the Beta Cell

The regulation of K_ATP channels by endogenous ligands could be required for the appropriate function under physiologic conditions. ATP/ADP ratio is the principal physiologic regulator of these channels; however, they can also be regulated by lipids, such as long-chain acyl-CoA esters (LC-CoAs), phosphatidylinositol-4, 5-biphosphate (PIP2) (Tarasov et al., 2004), syntaxin-1A (Syn-1A) (Pasyk et al., 2004), and leptin (Gavello et al., 2016).

Acyl CoAs are products of free fatty acid (FFA) esterification by acyl-CoA synthetase. Larsson et al. (1996) demonstrated the activation of K_ATP channels by LC-CoAs at a physiologic concentration in mice and clonal pancreatic beta cells. LC-CoAs esters are also potent stimulators of K_ATP channels activity in human beta cells and do not require the presence of Mg²⁺ or nucleotides (Bränström et al., 2004). The increase in the activity of K_ATP channels occurs by interaction of LC-CoAs in the carboxyl-terminal domain of Kir6.2 subunit (Gribble et al., 1998b), where the Lys332 residue is a key structural determinant (Bränström et al., 2007). Moreover, the effects of LC-CoAs on the activity of K_ATP channels depend on both the length and the saturation degree of the acyl chain. Saturated acyl CoAs are the most potent activators of K_ATP channels, followed by monounsaturated and polyunsaturated acyl CoAs, respectively (Riedel and Light, 2005).

In addition, the K_ATP channels with E23K (0.6471 allele frequency) and I337V (0.6448 allele frequency) Kir6.2 polymorphisms, which are frequent in Caucasian type 2 diabetes populations, showed significantly increased activity in response to palmitoyl-CoA, compared with wild-type K_ATP channels. These observations suggest that the diabetogenic effects of E23K and I337V polymorphisms occur through an increased K_ATP channel activity mediated by high levels of LC-CoAs (Riedel et al., 2003).

To date, there is clear evidence supporting the regulating role of phosphoinositides in the activity of many ion channels (Hille et al., 2015). Fan and Makielski (1997) showed for the first time that the K_ATP channels are regulated by the phosphatidylinositol 4-5 biphosphate (PIP2). PIP2 is a second messenger that increases the activity of K_ATP channels. Electrostatic interaction occurs between negative phosphate head groups of PIP2 and positively charged residues in Kir6.2 (Schulze et al., 2003), including Arg54 in N terminus and Arg176, Arg177, and Arg206 in C terminus. All of these residues are localized above the ATP binding site and near to the plasma membrane (Ribalet et al., 2005).

Several studies demonstrated that SUR1 plays an important role in stabilizing the interaction of PIP2 with the Kir6.2 subunit. Recently, a variant (E128K) in the TMD0 of SUR1 showed diminished the open probability and the responses to PIP2 (Pratt et al., 2011), raising the possibility that TMD0 of SUR1 modulates the Kir6.2 gating by reducing the response to PIP2. Thus, PIP2 may play an important role in regulating the activity of K+ channels and, therefore, insulin secretion. Furthermore, PIP2 can interact with other proteins that also regulate the K_ATP channel activity, and these proteins appear to be involved with the exocytotic machinery [e.g., synaptotagmins, calcium-dependent activator protein for secretion, and syntaxin-1A (Syn-1A)]. Syntaxin-1A is a target protein of SNARE complex that regulates the K_ATP channel trafficking (Pasyk et al., 2004) and is a potent inhibitor endogenous of K_ATP channels (Chang et al., 2011). PIP2 directly and indirectly regulates the interaction of syntaxin-1A on SUR1. PIP2 also affects the Syn-1A binding on NBDs of SUR1, albeit PIP2 acts indirectly on syntaxin-1A clusters on the plasma membrane to sequester Syn-1A and reduce its availability and interaction with SUR1 at the plasma membrane (Liang et al., 2014).

Leptin is an adipocyte-derived hormone that regulates food intake, body weight, glucose homeostasis (Friedman and Halaas, 1998), and was recently demonstrated as a regulator of K_ATP trafficking within beta cells (Holz et al., 2013). Moreover, leptin can activate different ion channels, including K_ATP channels, both in central neurons and beta cells (Gavello et al., 2016). The mechanism by which leptin activates K_ATP channels involves phosphatase and tensin homologe inhibition mediated by phosphoinositide 3 kinase. phosphatase and tensin homologe inhibition induces F-actin depolymerization and, consequently, K_ATP channel opening (Park et al., 2013b; Wu et al., 2015).

The Role of K_ATP Channels in the Pathophysiology of MS and T2DM

In obesity and type 2 diabetes, acyl CoAs, PIP2, and Syn-1A levels change. Chronic plasma levels of FFA in obese (Golay et al., 1987) and diabetic patients (Reaven et al., 1989) can increase the cytosolic acyl CoAs and K_ATP activity (Riedel et al., 2003). Long-chain polyunsaturated fatty acids maintain cell membrane fluidity and facilitate the cellular signaling mechanisms. Long periods of high-fat diet, particularly saturated fat and trans fatty acids, alter cell membrane lipid composition, compromising transmembrane insulin receptor signaling and consequently glucose uptake peripheral tissues (Wilcox, 2005). Furthermore, in diabetic human and rodent beta cells, the levels of Syn-1A and cognate SNARE proteins are reduced (Nagamatsu et al., 1999; Ostenson et al., 2006). The increase in adipose mass will elevate plasma leptin levels and this, in turn, will feed back to pancreatic beta cells to inhibit insulin secretion via stimulation of K_ATP channels (Campfield et al., 1995). This suggests that during pathophysiological conditions such as obesity, insulin resistance, or metabolic syndrome, the regulation of K_ATP channels could be affected and may contribute to the disease. Recent studies have evaluated the K_ATP activity during the metabolic syndrome (MS), which is a group of symptoms including central obesity, hyperinsulinemia, insulin resistance, high plasma levels of triglycerides, and high arterial blood pressure that predispose the subject to develop type 2 diabetes mellitus, cardiovascular diseases, and certain types of cancer. In contrast to type 2 diabetes mellitus, where there is an irreversibly reduced insulin secretion and pancreatic beta-cell exhaustion, the MS condition may be reversible, allowing the study of channel modifications before type 2 diabetes mellitus develops. A metabolic syndrome model in Wistar rats, developed by adding 20% sucrose to the drinking water for 6 months, was used to study the K_ATP channel activity and ATP sensitivity in inside-out patches (Velasco et al., 2012). During the metabolic syndrome, the ATP sensitivity of the K_ATP channels is higher than in the control. The dissociation constant (K_d) and Hill coefficients for ATP interaction with K_ATP channels were 18.3 ± 0.01 and 1 ± 0.01 μM for control and 10.1 ± 0.9 and 0.9 ± 0.01 μM for MS cells, respectively. It will be important in the future to investigate the mechanisms of this change, considering that the intracellular environment modulates K_ATP channels activity and eventually the excess of ATP cause the channel dysfunction in type 2 diabetes mellitus (Velasco et al., 2012).

Structural Determinants of VGSC

Voltage-gated Na⁺ channels (VGSC) are multimeric proteins. The core of the channel is formed by the alpha subunit, which encodes four homologous domains (I–IV), each of them comprising six membrane-spanning alpha helices (S1–S6) (Ahern et al., 2016). Nine members (Nav1.1–1.9 or Scn1–9a) constitute the gene family of VGSC alpha subunits (Catterall et al., 2005). Among them, Nav1.7 and Nav1.3 are widely expressed in pancreatic beta cells of rodents (Vignali et al., 2006; Zhang et al., 2014), whereas Nav1.6 and Nav1.7 are expressed in humans (Braun et al., 2008).

Alpha subunits share a prototypical architecture (Fig. 2A), where the segments S5–S6 of each domain and the extracellular loop that connects them (called P, denoting pore), constitute the hydrophilic pore that allows the movement of ions and determines the particular selectivity of the channel for Na⁺. Furthermore, the S4 segments serve as voltage sensors due to positively charged residues in these regions (Catterall, 2000a; Ahern et al., 2016). These channels activate in response to membrane depolarization, but their activity decreases during a sustained stimulus, a phenomenon known as inactivation (Fig. 2B) (Ahern et al., 2016). The biophysical properties of alpha subunits, as well as their trafficking, are regulated by the interacting VGSC beta subunits, which comprise 4 genes and 5 members, as one of them presents two splicing variants (O'Malley and Isom, 2015). Among the latter, the predominant beta subunit expressed in pancreatic beta cells is Scn1b (Ernst et al., 2009; Zhang et al., 2014).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Structure and biophysical properties of voltage-gated Na⁺ channels. (A) The cartoon represents the structure of the pore-forming alpha subunit. In each domain (D-I to D-IV), S4 segments show positive charges and the S5–S6 segments with the P-loop in between are highlighted in red. (B) Schematic representation of the closed, open, and inactivated states of VGSC. At hyperpolarized membrane potentials (i.e., the resting membrane potential of −70 mV in pancreatic beta cells) Nav channels remain in the closed state. A depolarizing stimulus causes the transition to an open state in Nav channels, where Na⁺ ions permeate through the channel pore. A sustained depolarization induces the inactivation of the channel, represented as a second gate that prevents the flow of ions even when the activation gate is still open. The recovery from inactivation occurs at hyperpolarized potentials. (C) The classic role for Nav channels in beta-cell physiology involves their activation after the high glucose-dependent K_ATP channel closure, which in turn enhances depolarization and spiking activity, promoting Ca²⁺ entry and insulin secretion. The scheme also summarizes some of the pharmacological manipulations that potentiate (right panel, sharp arrowhead) or inhibit (let panel, blunted arrowhead) VGSC activity and consequently modulate glucose-stimulated insulin release

Historical Overview of Nav Channels and Insulin Secretion

Direct reports on functional VGSC in insulin-secreting cells first appeared in the 1980s, when a transient inward Na⁺ current was described in RINm5F cells, which could be blocked by 2.5 µg/ml of tetrodotoxin (TTX) (Rorsman and Trube, 1986). This finding was in contrast to a previous report from the same group on native pancreatic beta cells from NMRI mice, where the authors found no effect of TTX (20 µM) on inward currents elicited by membrane depolarization using the whole cell patch clamp recordings (Rorsman and Trube, 1986). At that time, a plausible explanation was that voltage-activated Na⁺ currents could reflect a dedifferentiated state of RINm5F cells (Rorsman et al., 1986). However, because the RINm5F cell line was derived from a rat insulinoma, species-related differences could also account for this discrepancy.

The report of a TTX-sensitive Na⁺ current in isolated pancreatic beta cells (Plant, 1988), helped advance an understanding of previous negative results in beta cells from NMRI mice. This report determined that the voltage for half-maximal steady-state inactivation (V_0.5) was more negative than −100 mV at regular (2.6 mM) or low (0.2 mM) Ca²⁺ concentrations in the bath solution. The latter argued against a potential role of VGSC in insulin secretion in mice, because the resting membrane potential is around −70 mV at low glucose, a potential at which the channels of mice beta cells are expected to be fully inactivated. Consistently, the application of 1.5 µM TTX did not alter the electrical responses when glucose rose from 3 to 20 mM (Plant, 1988).

Almost at the same time, Hiriart and Matteson (1988) demonstrated that this was not the case for rat pancreatic beta cells. They consistently recorded voltage-activated Na⁺ currents with an almost complete steady-state inactivation around −40 mV, which was rightward shifted, compared with the study of Plant (1988) in mice. Accordingly, the average insulin release of individual beta cells was decreased by incubating with TTX (200 nM), as measured by a reverse hemolytic plaque assay at high glucose (15.6 mM) (Hiriart and Matteson, 1988).

Studies in other species (such as dogs and humans) confirmed the expression of VGSC (Philipson et al., 1993), as well as their contribution to the spiking activity of pancreatic beta cells. Similar results have been obtained in porcine beta cells (Pressel and Misler, 1990), where TTX (1 µM) blocks voltage-gated Na⁺ currents and action potentials elicited by 12 mM glucose (Silva et al., 2009). Even in mice, where VGSCs were supposed to marginally participate in beta-cell excitability because of inactivation at physiologic membrane potential (Plant, 1988), the genetic deletion of the ancillary beta 1 subunit (Scn1b) impairs the glucose-induced insulin secretion, causing glucose intolerance (Ernst et al., 2009).

Recently, Zhang et al. (2014) examined the role of VGSC in pancreatic beta cells of mice, combining several pharmacological and genetic tools and using state-of-the-art models for the assessment of beta-cell physiology (i.e., intact islet electrophysiology, insulin release measurements in isolated islets, and perfused pancreas). The authors found that although Scn9a/Nav1.7 is the isoform with the highest expression level in this model, the form Scn3a/Nav1.3 is also expressed and significantly contributes to VGSC-mediated currents in 30% of the beta-cell population (Zhang et al., 2014). Accordingly, glucose-induced insulin secretion was impaired in pancreatic islets from Scn3a−/− but not in Scn9a−/− mice (Zhang et al., 2014). These results support that a minor fraction of cells expressing higher relative amounts of Scn3a/Nav1.3 transcripts may exhibit greater excitability and mediate the recruitment through gap junction coupling of other beta cells with less sensitivity to rises in glucose concentration (Zhang et al., 2014).

Indeed, VGSC may be a clinically relevant target, because a decrease in their activity profoundly impacts beta-cell excitability and insulin release in humans. In an early paper, Pressel and Misler (1990) showed that replacing ∼94% of the extracellular Na⁺ with N-methylglucamine abrogated the action potentials elicited by 10 mM glucose in human pancreatic beta cells. Later, it was demonstrated that 1 µM TTX reduced the amplitude and broadened the width of the action potential of beta cells from human donors (Barnett et al., 1995). As expected, TTX application (as well as Na⁺ substitution) in the bath solution significantly reduced the voltage-activated Na⁺ currents in these cells, whereas the toxin impaired insulin secretion when islets were challenged from 3 to 6 mM glucose (Barnett et al., 1995). Braun and coworkers (2008) obtained similar results using TTX, which reduced insulin secretion in 6 and 20 mM glucose of isolated pancreatic islets from human donors. These authors also corroborated that TTX (0.1 µg/ml) decrease the amplitude of voltage spikes in pancreatic beta cells, as well as the voltage-gated Na⁺ whole cell currents (Braun et al., 2008).

The molecular identity of the channels mediating these currents has also been explored in human pancreatic islets, where the highest mRNA levels corresponded to Nav1.6 and Nav1.7 (Braun et al., 2008). However, only a minor component of the voltage-dependence inactivation of Na⁺ currents resembled that of beta cells of mice with an inactivation V_0.5 of −105 mV (Zhang et al., 2014). The hyperpolarizing shift in the voltage of half-maximal inactivation for Nav1.7 in pancreatic beta cells seems to occur through the interaction with a beta-cell-specific intracellular factor (Zhang et al., 2014), yet to be determined. Changes in the activity of Nav1.7 channels may be linked to beta-cell dysfunction and the susceptibility of developing painful neuropathy, because this channel is also expressed in dorsal root ganglion neurons (Hoeijmakers et al., 2014). Recently, it was shown that Nav1.7 channels are linked to insulin production, because pancreatic islets from Scn9a−/− mice exhibit increased insulin content compared with their wild-type counterparts (Szabat et al., 2015). The latter also offers new perspectives for drugs like carbamazepine, which inhibits the channel and dose dependently increases the mRNA expression of both insulin genes (Ins1 and Ins 2), as well as Pdx1 (a marker of pancreatic beta cells) in isolated islets from mice (Szabat et al., 2015).

Pharmacology of Nav Channels in Pancreatic Beta Cells

From the previous section, it is evident that tetrodotoxin, a potent blocker of voltage-gated Na⁺ channels, stands out as the first-choice pharmacological agent to use in the study of these proteins. Toxins have been extensively used to explore the structure function of ion channels (Morales-Lazaro et al., 2015) and also because their high potency and selectivity can be harnessed in a more complex pathophysiological context (Diaz-Garcia et al., 2015). The chemical structure and the mechanisms of action of these bioactive molecules are diverse. The toolkit of toxins with proven effects on insulin secretion via VGSC includes compounds isolated from plants, snakes, and scorpions, which modulate the biophysical properties of these channels instead of blocking the pore, as TTX does. The use of toxins like veratridine, which prevents channel inactivation (Ulbricht, 2005) through the neurotoxin interaction site 2 (Cestele and Catterall, 2000), established the presence of VGSC in pancreatic beta cells years before the first direct electrophysiological recordings. An early paper from Donatsch et al. (1977) showed that veratridine (100 µM) increases insulin secretion in isolated islets from albino rats perfused with 5.6 mM glucose to a similar extent to that observed with a rise in glucose to 16.7 mM. Moreover, these authors demonstrated that the secretagogue activity of veratridine could be reversibly inhibited by 3 µM TTX (Donatsch et al., 1977).

The effect of veratridine (200 µM) and TTX (3 µM) on insulin secretion was later corroborated by Pace (1979) in pancreatic islets from Sprague-Dawley rats. The author also observed a veratridine-induced depolarization of pancreatic beta cells in culture, as well as an increase in their spiking activity (Pace, 1979). In agreement with these findings, the ability of veratridine (30 µM) to induce Na⁺ oscillations in islets cells from ob/ob mice was prevented by TTX (Grapengiesser, 1996). A concentration of 10 µM veratridine was not as effective as a 20-fold higher concentration, but increased insulin secretion when 20 nM of Leiurus toxin was applied simultaneously (Pace and Blaustein, 1979). The scorpion toxin from Leiurus quinquestriatus also delays the inactivation of Na⁺ currents (Gonoi et al., 1984), which is a common property of alpha-scorpion neurotoxins and certain spiders' toxins interacting with the neurotoxin receptor site 3 in the alpha subunit of VGSC (Cestele and Catterall, 2000). Another alpha-scorpion neurotoxin, TsTx-V (5.6 µg/ml) from Tityus serrulatus has been shown to reversibly increase the electrical activity of pancreatic beta cells from Swiss mice perfused with 11 mM glucose, in a similar fashion as 110 µM veratridine (Goncalves et al., 2003). Moreover, TsTx-V also increases the insulin secretion of pancreatic islets from Wistar albino rats when incubated with 2.8 or 8.3 mM glucose. Crotamine, a toxin from the rattlesnake Crotalus durissus terrificus, also promotes insulin release by preventing VGSC inactivation. An active fraction (F2) obtained from reversed-phase high-performance liquid chromatograph, enhanced insulin secretion of rat islets incubated with 16.7 mM glucose, exhibiting a threefold augment in potency with respect to the native crotamine (Toyama et al., 2000).

The effect of certain drugs has been studied on the late Na⁺ currents from insulin secreting cells, which increase after long-term exposure to 33 mM glucose, a condition that resembles hyperglycemia (Rizzetto et al., 2015). Notably, the blockers ranolazine (10 µM) and TTX (0.5 µM) decrease the spiking activity of INS-1E cells when stimulated by high glucose or tolbutamide, as well as the tolbutamide-induced Ca²⁺ transients (Rizzetto et al., 2015). Also, both blockers attenuated the potentiating effect of veratridine (40 µM) on glucose-stimulated insulin secretion (Rizzetto et al., 2015).

Recently, VGSC were identified as regulators of cytokine-induced beta-cell death. This finding arose from the discovery of carbamazepine as a protecting agent against a cocktail of cytotoxic cytokines (tumor necrosis factor-alpha, interleukin-1beta, and interferon-gamma) (Yang et al., 2014b). In the study, the authors demonstrated that carbamazepine (0.01–100 µM) dose dependently reduced the cytokine-induced apoptosis in beta cells and expanded these findings to other VGSC blockers such as lidocaine and TTX (Yang et al., 2014b). Interestingly, carbamazepine (0.3–60 µM) also prevents the lipotoxicity of palmitate (0.5 mM) on beta cells (Cnop et al., 2014). These results shed new light on unexpected roles for VGSC in inflammatory and metabolic events that may occur in diabetes mellitus.

Nav Channels Modulation by Endogenous Molecules

The activity of VGSC can also be modulated by endogenous signals, usually acting as downstream effectors of signaling cascades. This phenomenon was observed since the initial studies on VGSC in insulin-secreting cells, because the chronic (48 hours) application of phorbol ester (TPA, a PKC activator) was shown to decrease the Na⁺ current density in RINm5F cells (Rorsman et al., 1986). Nav channel activity may predispose to beta-cell death after an inflammatory insult (Yang et al., 2014b). On the contrary, nerve growth factor (NGF) favors beta-cell survival and insulin secretion (Vidaltamayo et al., 2002; Navarro-Tableros et al., 2004). It has been demonstrated that rat beta cells exhibited higher Na⁺ current densities when incubated with 50 ng/ml of 2.5S NGF for 5 days, which could be prevented by either actinomycin D or cycloheximide (Vidaltamayo et al., 2002). The latter suggested that the NGF-mediated increment in Na⁺ currents may be due to an increased transcription and protein synthesis, which is supported by elevated Nav1.3 transcripts in the NGF-treated cells (Vidaltamayo et al., 2002).

The ability of NGF to enhance the activity of voltage-gated ion channels in pancreatic beta cells is not exclusive to the Nav family, because it has also been demonstrated for Ca²⁺ channels (Navarro-Tableros et al., 2007a). In both cases, the effects are mediated by TrkA, a high-affinity receptor with intrinsic tyrosine kinase activity (Rosenbaum et al., 2001). Remarkably, the inhibition of protein tyrosine kinases leads to impaired insulin secretion (Persaud et al., 1999), which agrees with the role of these signaling molecules in maintaining proper beta-cell excitability.

Nav channels may also serve as downstream effectors in G-coupled receptor cascades. The muscarinic agonist carbachol (100 nM), for instance, promotes insulin release from isolated rat beta cells at low glucose concentrations (0–5.6 mM), which can be prevented by the coapplication of 319 nM TTX (Hiriart and Ramirez-Medeles, 1993). However, the role of Nav in the muscarinic signaling, at high glucose (20.6 mM), is not straightforward. Strikingly, carbachol decreases insulin secretion in this condition, as well as its inhibitory effect, and is not additive with that of TTX (Hiriart and Ramirez-Medeles, 1993).

A recent work from Salunkhe and coworkers (2015) demonstrated that the microRNA-375 also modulates the activity of Nav channels in insulin-secreting cells. The authors observed that the overexpression of microRNA-375 in INS-1 832/13 cells led to a decrease in the protein levels of Scn3a and Scn3b. Nevertheless, the relevance of this regulation is not completely understood in a more physiologic context, because glucose-induced insulin secretion was not affected by knocking down microRNA-375 (Salunkhe et al., 2015).

Interestingly, Nav channels are also connected to ATP levels in beta cells. It was proposed that Na⁺ currents may be enhanced by an increase in glucose levels and further ATP production in beta cells (Zou et al., 2013). This idea is supported by whole cell recordings of beta cells in slices from rat pancreas, where concentrations of ATP (2–8 mM) in the recording pipette shift the inactivation curve to more positive potentials and accelerate the recovery from inactivation of VGSC (Zou et al., 2013). In the other direction, the TTX-sensitive depolarizing Na⁺ influx contributes to the cytosolic Na⁺ transients (Nita et al., 2014), which, together with Ca²⁺ transients, propagate to the mitochondria and modulate ATP synthesis. Thus, a glucose-induced uptake of Na⁺ through the mitochondrial Na⁺/Ca²⁺ exchanger promotes Ca²⁺ extrusion and reverses the divalent cation uptake by the mitochondrial Ca²⁺ uniporter (Nita et al., 2014).

After four decades of study on VGSC in pancreatic beta cells, we have advanced in the understanding of how Na⁺ currents contribute to cell depolarization, voltage and Ca²⁺ oscillations in beta cells, and, finally, in the promotion of glucose-induced insulin secretion (Fig. 2C). However, the regulatory pathways involving VGSC, from endocrine signals to epigenetic mechanisms, are still an incipient field of research. The evidence suggesting that Nav channels interact with metabolism in a bidirectional way is fascinating, because the role of these channels is not yet entirely understood in metabolic diseases such as diabetes or the metabolic syndrome. Indeed, VGSC regulates the physiology of pancreatic beta cells through mechanisms other than just a depolarizing effect, a reason to consider them as potential drug targets for treating related pathologies.

Molecular Structure of Cav Channels

VDCC consists of heteromeric complexes composed of Cavα1, Cavα2, Cavβ-, Cavδ-, and Cavγ-subunits (Fig. 3). The Cavα1 subunit forms the ion-conducting pore, and Cavα2, Cavβ-, Cavα2δ-, and Cavγ-subunits regulate its conductivity and trafficking (Yang et al., 2014a). The Cavα1 subunit is formed by four homologous domains that consist of six transmembrane segments (S1 to S6), a membrane-associated pore loop (P-loop) between each S5 and S6 segments, three intracellular linkers (LI-II, LII-III, LIII-IV), and N and C termini. As mentioned above, S5 and S6 segments and the P-loop form the Ca²⁺ conducting pore. The S4 segments work as the primary voltage sensor involved in the inactivation and the reactivation of the channel (Catterall, 2000b). The three-dimensional structure of Cav channels has been described, using electron microscopy and X-ray crystallography. The Cavβ subunit is intracellular; the Cavγ subunit is formed by four transmembrane segments, and the Cavα2δ is extracellular and attached to the membrane through a disulfide linkage to the transmembrane delta subunit (Catterall, 2000b).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Representation of the calcium channel subunit structure, signaling that regulates them, and hormone and drug effects on calcium influx through calcium channels. Greek letters depict each calcium channel subunit, and Roman numbers represent each domain in the alpha-1 subunit. PKA, protein kinase A; PKB, protein kinase B; PKC, protein kinase C; PKG, protein kinase G; SNARE, soluble N-ethylmaleimide sensitive factor attachment protein receptor; Gβγ, G protein subunits beta-gamma; CAMKII, calcium/calmodulin-dependent kinase II; GLP-1, glucagon-like peptide-1; NGF, nerve growth factor; IL-1, interleukin 1; TNF-a, tumor necrosis factor alpha; IFNg, interferon gamma. Green arrows/names represent an increasing effect and red lines/names depict an inhibitory effect on calcium influx through Cav channels. Question mark symbols represent an unknown effect on calcium channels. Magenta arrow describes physical interaction.

We can classify VDCC according to different criteria. A functional characterization considers their open probability, in low-voltage activated (T-type channels) and high voltage-activated (L-, N-, P/Q-, and R-type channels). A second structure-based classification considers the type of α1 subunits being expressed and subdivides Cav channels into three families, namely Cav1, Cav2, and Cav3. They are associated with L-, P/Q-, N-, R-, and T- type Cav currents (Hiriart and Aguilar-Bryan, 2008; Yang et al., 2014a; Zamponi et al., 2015). There are also several subtypes of L channels, depending on their alpha subunit, alpha1C (CACNA1C, Cav1.2) and alpha 1D (CACNA1D, Cav1.3). L-type Cav1.2 and Cav1.3 channels are expressed in rats and mice and in rats, mice, and humans, respectively. However, Cav1.3 contributes the most to glucose-stimulated insulin secretion in rats and humans and Cav1.2 in mouse beta cells (Hiriart and Aguilar-Bryan, 2008; Rorsman et al., 2012).

Beta cells from diverse species and cell lines express different combinations of VDCC. For example, human beta-cell expression of CACNA1D mRNA exceeds CACNA1C mRNA by 60 times (Reinbothe et al., 2013). Although the same predominance of Cav1.3 is observed in the rat, in the mouse, Cav1.2 channels are dominant (Hiriart and Aguilar-Bryan, 2008; Rorsman et al., 2012). Cav3.1 and Cav3.2 (T-type) currents are present in INS-1 cells, dog, human, and rat, but not in mouse. Moreover, human beta cells lack Cav2.3 (N-type) currents (Taylor et al., 2005; Hiriart and Aguilar-Bryan, 2008), and N-type Cav2.2 channels have been identified in adult rat beta cells (Navarro-Tableros et al., 2007b). Also, human beta cells express P/Q-type Cav2.1, and mouse beta cells show P/Q-type Cav2.1 and R-type Cav2.3 channels (Yang et al., 2014a).

To date, 10 mammalian α1 genes and four distinct Cavβ genes (Cavβ1, Cavβ2, Cavβ3, and Cavβ4) have been identified in beta cells. Rat and mouse native beta cells express β2 and β3 subunits, although β2 is more abundant than β3 in the rat (Yang et al., 2014a). To date, it has been suggested that Cavβ2 and Cavβ3 subunits modulate GSIS possibly by regulating the interactions between VGCCs and PKC isozymes (Rajagopal et al., 2014).

Pharmacology of Cav Channels in Pancreatic Beta Cells

Many pharmacological agents modulate the activity of Cav channels. For example, L-type Cav1 channels are sensitive to La³⁺, dihydropyridines, alkyl phenyl amines, and benzothiazepines. Recently, it was observed that the natural polyphenolic flavonoid quercetin (3,3′,4′,5,7-pentahydroxyflavone) enhances L-type Cav1 current, also stimulating insulin secretion in INS-1 cells. Quercetin interacts with L-type Cav1 in a different site from that of the agonist Bay K 8644 (1,4-Dihydro-2,6-dimethyl-5-nitro-4-[2-(trifluoromethyl)phenyl]-3-pyridinecarboxylic acid, methyl ester). The mechanism acts by shifting the voltage-dependent activation toward negative potentials (Bardy et al., 2013). L-type Cav1 channels are the major contributors to the whole cell Cav current and insulin granule exocytosis in human, rat, and mouse beta cells. Despite Cav1.2 being the principal L-type Ca²⁺ channel in mouse, Cav1.3 is the most important L-type channel in rat and human beta cells. Cav1.3 plays a role in cAMP homeostasis through adenylate cyclase 1 and may participate in insulin secretion in an indirect manner (Kitaguchi et al., 2013). Recent work suggests that Cav1.2 LII-III loop interactions may play a significant role in the glucose-stimulated excitability. They may regulate channel density on the plasma membrane, the Ca²⁺-induced-Ca²⁺ release from the endoplasmic reticulum (ER), as well as the activity of other channels such as Ca²⁺-activated potassium channels (Catterall, 2000b; Wang et al., 2014).

P/Q-type Cav2.1 channels are sensitive to ω-agatoxin IVA, N-type Cav2.2 channels are blocked by ω-conotoxin GVIA but are insensitive to dihydropyridines, alkyl phenyl amines, and benzodiazepines. R-type Cav2.3 channels are resistant to almost all the Cav channel blockers, but it is blocked by the toxin SNX-482. Although SNX-482 has shown a high affinity for Cav2.3 channels, it is capable of blocking other Cav channels Yang and Berggren, 2006; (Catterall, 2011; Yang et al., 2014a).

Calcium entry into the cell through P/Q-type Cav2.1 channels is the second most important entry route in rodent beta cells after glucose stimulation. Nevertheless, P/Q channels contribute in a similar fashion as L-type channels to the whole cell Cav current in human beta cells (Yang and Berggren, 2006; Yang et al., 2014a). N-type Cav2.2 channels are expressed in adult rat beta cells, although 50% of the protein is distributed in the membrane and 50% at the cytoplasm (Navarro-Tableros et al., 2007b).

T-type Cav3.1 and Cav3.2 channels play a significant role in the regulation of insulin secretion. In rat beta cells, they activate around −40 mV and generate a hump at the beginning of the whole cell I(Ca)-V relationship, with rapid inactivation and slow deactivation, and they may play the role of a pacemaker at glucose-stimulating levels (Hiriart and Matteson, 1988; Hiriart and Aguilar-Bryan, 2008). T-type Cav currents have been suggested to be involved in the beta-cell response to cytokines (Wang et al., 1999). Finally, R-type Cav2.3 channels are expressed rodent but not human beta cells (Hiriart and Aguilar-Bryan, 2008). They may play a role in the second phase of insulin secretion by supplying new secretory granules to the release sites (Jing et al., 2005; Yang and Berggren, 2006; Rorsman et al., 2012).

Endogenous Regulation of Cav Channels in the Beta Cell

Several mechanisms and pathways, including protein phosphorylation, Ca²⁺/calmodulin signaling, G protein-coupled receptors, and phosphorylated inositol may endogenously regulate Cav channels activity and density in the beta-cell membrane (Yang and Berggren, 2006). These regulatory mechanisms interact simultaneously to control insulin secretion in the body (Fig. 3).

The Role of Cav Channels in the Pathophysiology of MS and T2DM

VDCC alterations, because of mutation, the downregulation of their expression, activity, and/or density at the membrane, could lead to beta cell dysfunction (Yang and Berggren, 2006; Velasco et al., 2012). Changes in calcium influx through VDCC play a vital role in beta-cell dysfunction in metabolic syndrome (MS) and diabetes models. In an MS rat model induced by consumption of a 20% sucrose solution during 24 weeks, changes in whole cell calcium influx were observed, which may be related not only to beta-cell hypersecretion but dysfunction and exhaustion (Velasco et al., 2012). Additional studies have shown that a 48-hour islet exposure to 28 mM glucose increased insulin secretion due to an increased nifedipine-sensitive VDCC influx (Qureshi et al., 2015).

It is well accepted that beta-cell chronic exposure to high long-chain free fatty acids contributes to increasing basal insulin secretion and, at the same time, reduces GSIS during obesity (Zraika et al., 2002). In part, these effects involve a reduced calcium influx through VDCC, associated with an increased activity of the membrane metalloendopeptidase neprilysin (Zraika et al., 2013). Other groups have found that beta-cell long-term exposure to free fatty acids results in selective suppression of exocytosis elicited by brief depolarizations. They suggested that depolarization-induced Ca²⁺ influx is limited to discrete areas within beta cells due to calcium channel aggregation and that FFAs induce a diffused Ca²⁺ influx along the membrane surface. This effect could impair insulin exocytosis, because release-competent granules may not be exposed to exocytotic local levels of Ca²⁺ (Rorsman et al., 2012). Furthermore, two single-nucleotide polymorphisms in the CACNA1D gene (Cav1.3) reduce its mRNA expression, impair insulin secretion, and are associated with type 2 diabetes (Reinbothe et al., 2013). Finally, blocking L-type calcium channels with diazoxide could inhibit hyperglycemia-induced beta-cell apoptosis (Zhou et al., 2015).