Ion Channels of the Islets in Type 2 Diabetes

Abstract

Ca²⁺ is an essential signal for pancreatic β-cell function. Ca²⁺ plays critical roles in numerous β-cell pathways such as insulin secretion, transcription, metabolism, endoplasmic reticulum function and the stress response. Therefore, β-cell Ca²⁺ handling is tightly controlled. At the plasma membrane Ca²⁺ entry primarily occurs through voltage-dependent Ca²⁺ channels (VDCCs). VDCC activity is dependent on orchestrated fluctuations in the plasma membrane potential or voltage, which are mediated via the activity of many ion channels. During the pathogenesis of type-2 diabetes (T2D) the β-cell is exposed to stressful conditions, which result in alterations of Ca²⁺ handling. Some of the changes in β-cell Ca²⁺ handling that occur under stress result from perturbations in ion channel activity, expression or localization. Defective Ca²⁺ signaling in the diabetic β-cell alters function, limits insulin secretion and exacerbates hyperglycemia. In this review, we focus on the β-cell ion channels that control Ca²⁺ handling and how they impact β-cell dysfunction in T2D.

Graphical Abstract

Introduction

The islets of Langerhans are micro organs embedded in the pancreas and contain endocrine cells which secrete hormones to regulate blood glucose. In rodents the most abundant cell type in the islet is the β-cell which secretes insulin to lower blood glucose and is located in the core of the islet. This is followed by the α-cell which is found in the mantle of the islet and secretes glucagon to elevate blood glucose levels. Other minor cell types, including δ-cells, ε-cells, and PP cells secrete somatostatin, ghrelin, and pancreatic polypeptide respectively. These hormones are released into the circulation to coordinate glucose storage and utilization in target tissues but also act in an autocrine or paracrine manner within the islets for feedback regulation to achieve glucose homeostasis in changing metabolic environment. Here, it is worth noting that compared to rodent islets, human islets have lower β- to α-cell ratios and β-cells form clusters that are interspersed with non-β cells rather than residing in the core; these distinct architectures likely impact signaling between cell types in the different species.

The secretory activity of each endocrine cell type is controlled mostly by ion channels. These ion channels control K⁺, Na⁺, Cl⁻, and Ca²⁺ fluxes to generate electrical and Ca²⁺ signals needed for secretion. The ionic mechanisms underlying secretion regulation is best understood in β-cells, as they are the sole source of insulin and indispensable for glucose regulation. β-cells are equipped with a large repertoire of channels that can be regulated at the level of transcription, translation, post-translational modifications and intracellular trafficking to respond to extracellular and intracellular signals and fine tune their secretory activities. Over sixty different ion channels are known to be expressed. Their roles in insulin secretion in physiological and pathological conditions are of primary interest in the islet biology field.

Type 2 diabetes mellitus is a complex chronic disease often associated with obesity. Initiation and progression of the disease likely involve many genetic and environmental factors. It is clear that both insulin resistance and failure of β-cells to secrete sufficient amount of insulin to curtail hyperglycemia are the two main drivers of T2D. In T2D, plasma glucose and insulin secretion become decoupled. Studies of extreme phenotypes such as monogenic forms of diabetes, genome-wide association studies, and profiling gene expression changes have implicated many ion channels and their regulatory proteins in the cause or acceleration of the disease. Understanding the ionic mechanisms that underlie this pathology is critically important for identifying therapeutic targets for the prevention and treatment of T2D.

This review will discuss ion channels that have been shown or proposed to contribute to β-cell dysfunction and T2D based on genetic and functional evidence from animal and human studies. In particular, we highlight ion channels that control Ca²⁺ entry necessary for secretion and how metabolic changes associated with insulin resistance alter ion channel regulation and Ca²⁺ handling to cause defective secretion. The primary focus will be on β-cells but other islet cell types and their involvement in T2D resulting from ion channel defects are also considered.

Ion channels in secretagogue-stimulated electrical excitability

β-cell plasma membrane electrical excitability is required for physiological glucose-stimulated Ca²⁺ entry and insulin secretion from β-cells. Electrical excitability results from changes in β-cell membrane potential (V_m) that is tightly controlled by the orchestrated opening and closing of many ion channels during glucose-stimulation (Figure 1). At rest the β-cell membrane sits near - 70 mV due to the activities of K_ATP channels and a slight sodium leak through other small conductance channels. During glucose-induced increases in the ATP/ADP ratio K_ATP channel closure leads to depolarization due to cessation of K⁺ efflux but continued Na⁺ leak[1]. Glucose metabolism also results in β-cell swelling that activates the volume regulatory Cl⁻ current through SWELL1, leading to Cl⁻ efflux and further V_m depolarization[2]. In human β-cells depolarization activates voltage-dependent Ca²⁺ channels (VDCCs) and voltage-gated Na⁺ channels (VGSCs) resulting in the upstroke of the action potential (AP); the AP is the primary electrical signal of the β-cell. The AP upstroke depolarizes the V_m and allows even more Ca²⁺ entry via VDCC activation activating big conductance K⁺ (BK) channels, which repolarizes the AP. During glucose-stimulation the V_m sits at a depolarized plateau potential where APs can rapidly fire leading to sustained Ca²⁺ entry and insulin secretion. A drop in blood glucose levels results in a reduction in the β-cell ATP/ADP ratio leading to activation of K_ATP channels, which causes V_m hyperpolarization and terminates electrical excitability.

Figure 1:

Ion channels that contribute to the human β-cell action potential. Glucose is taken up and metabolized in β-cells resulting in both an increase in the ATP/ADP ratio as well as cell swelling. The increase in the ATP/ADP ratio leads to K_ATP closure whereas swelling activates SWELL-1 Cl^- efflux. Reduced K_ATP activity together with the depolarizing influences of SWELL1 mediated Cl^- efflux and Na⁺ influx through Na⁺ leak channels results in V_m depolarization. V_m depolarization activates voltage-dependent Ca²⁺ channels (VDCCs) and voltage-gated Na⁺ channels (VGSCs) that are responsible for the upstroke of the action potential (AP); a representative human β-cell AP is shown on the right side of the β-cell. The rapid V_m depolarization as well as Ca²⁺ entry that occurs during the AP upstroke activates K⁺ channels including BK, K_V, SK and IK. BK channels are voltage and calcium activated, which provide the largest conductance and fastest activating K⁺ current that is responsible for the initiation of AP repolarization. K_V channels are activated by V_m depolarization and also contribute to AP repolarization. Finally, the Ca²⁺ activated K⁺ channels including SK and IK are smaller conductance K⁺ currents that stay active for longer following AP repolarization leading to afterhyperpolarization and a reduction in AP frequency.

β-cell excitotoxicity

Insulin resistance results in increased β-cell electrical excitability and insulin secretion during the pathogenesis of diabetes. Too much electrical excitability and excessive Ca²⁺ entry causes disruptions in β-cell Ca²⁺ handling resulting in a stressed condition which is referred to here as β-cell excitotoxicity. During β-cell excitotoxicity the resulting changes in Ca²⁺ homeostasis leads to changes in ion channel activity as well as significant changes in many signaling cascades (e.g. mitochondrial dysfunction). These stressful conditions result in elevated cytokine secretion (such as from immune cells) that further impact β-cell ion channel activity and Ca²⁺ handling, which negatively impacts insulin secretion. Moreover, the elevations in cytoplasmic Ca²⁺ during excitotoxicity result in transcriptional changes that also impact ion channel activity and thus Ca²⁺ handling. Single cell transcriptome analysis has found that the transcription of mouse β-cell ion channels changes dramatically in response to short term insulin resistance associated with a high fat diet [3]. For example, VDCCs such as Cacna1d that encodes Ca_v1.3 are significantly reduced [3]. Whereas K⁺ channel transcripts such as KCNK16 and Kcnb1 that encode TALK-1 and K_v2.1 respectively are increased[3]. β-cell transcriptome studies have also been performed on humans with or without T2D and ion channels transcript abundance are also different between these groups[4, 5]. As the human transcriptomes are from patients with long standing diabetes, the changes in β-cell ion channel transcript abundance are different from short term insulin resistance changes in rodents. For example the Transient Receptor Potential Cation Channel Subfamily M Member 3 (TRPM3) is significantly downregulated in human β-cells from T2D patients[4]. The stressful conditions of elevated β-cell electrical excitability in T2D lead to many changes that eventually impact the failure of β-cells to meet the increased demand for insulin[6]. Importantly, β-cell excitotoxicity is not only due to changes in ion channel function but also due to changes in β-cell nutrient utilization and ER stress that occur during the pathogenesis of T2D.

β-cell ion channels that control and disrupt calcium entry in T2D

K_ATP Channels

K_ATP channels are weakly inwardly rectifying K⁺ channels present in the β-cell plasma membrane. Activities of K_ATP channels are controlled by intracellular ATP and ADP, the ratio of which fluctuates according to blood glucose levels. This enables the channel to serve as a metabolic sensor and couples serum glucose to insulin secretion. Channel response to changing ATP/ADP ratios is coordinated by the two constituent protein subunits: an inward rectifier Kir6.2 and an ABC transporter protein sulfonylurea receptor 1 (SUR1). Recent high resolution cryo-EM structures show that the channel is a hetero-octamer with four Kir6.2 subunits in the center forming the K⁺ conducting pore and four SUR1 subunits in the periphery[7–9]. Each SUR1 protein interacts with a Kir6.2 via its N-terminal transmembrane domain (TMD0) which is connected to SUR1’s ABC transporter core structure consisting of two transmembrane domains and two nucleotide binding domains by a long intracellular loop (L0) (Figure 2).

Figure 2:

(A) CryoEM structure of the K_ATP channel viewed from the side (left) and the top (right). (B) Left: SUR1 topology showing the different domains. Right: Structure of SUR1 built from cryoEM density map showing the glibenclamide (GBC) binding site above NBD1 in the transmembrane bundle formed by transmembrane helices from both TMD1 and TMD2.

Both Kir6.2 and SUR1 interact with intracellular adenine nucleotides to regulate channel activity[10, 11]. ATP, and to a much lesser extent ADP, binds to a non-canonical nucleotide binding pocket formed by the N- and C-terminal cytoplasmic domains of two adjacent Kir6.2 subunits and closes the channel [12, 13]. The inhibitory effect of ATP occurs in isolated membrane patches even in the absence of Mg²⁺ thus does not involve ATP hydrolysis[10]. ATP and ADP also interact with the two nucleotide binding domains (NBDs) of SUR1 to stimulate channel activity in an Mg²⁺-dependent manner[14]. In the intracellular milieu where most ATP and ADP are in complex with Mg²⁺, binding of MgATP at NBD1 and MgADP at NBD2 causes dimerization of the two NBDs[9]. This conformational change is transduced to the Kir6.2 pore and antagonizes the ATP inhibition at Kir6.2, resulting in a net gain of channel activity. Thus, channel activities resulting from the tug-of-war between ATP inhibition at Kir6.2 and MgADP stimulation at SUR1 reflect the changes in ATP/ADP ratios in β-cells as the blood glucose concentration swings. It is worth noting that although NBD2 is capable of binding and hydrolyzing MgATP[15, 16], there is evidence that MgATP hydrolysis to MgADP at this site is not required for NBD dimerization and channel stimulation[17, 18]. Thus the mere increased occupancy of MgADP at NBD2 when MgADP concentrations increase following a fall of blood glucose likely underlies the MgADP stimulation effect.

The importance of K_ATP channels in insulin secretion and glucose regulation is evident from genetic variations in the channel genes (KCNJ11 for Kir6.2 and ABCC8 for SUR1) associated with a spectrum of insulin secretion disorders[19]. Loss-of-function K_ATP mutations were linked to congenital hyperinsulinism soon after the channel was cloned in 1995[20]. Although it was predicted that gain-of-function mutations would cause the opposite clinical phenotype, i.e. diabetes, direct demonstration did not come until almost ten years later[21]. A pioneer study showing that overexpressing an ATP-insensitive variant of Kir6.2 in mice resulted in a neonatal diabetes phenotype[22] paved the way for subsequent studies that identified many mutations in the Kir6.2 and SUR1 genes in human neonatal diabetes patients. Neonatal diabetes associated mutations in Kir6.2 typically reduce channel inhibition by ATP whereas those in SUR1 tend to increase channel stimulation by MgADP but some have also been shown to reduce ATP inhibition by allosteric mechanisms[23, 24]. In addition to neonatal diabetes, some mutations can also cause neurological defects including developmental delay and epilepsy, referred to as DEND syndrome[23]. This is most likely due to overactivity of these mutant channels in neurons[25] known to express K_ATP channels rather than secondary effect of hyperglycemia although detailed mechanisms are still not well understood. Aside from causing congenital diabetes, genetic variants in K_ATP channels have also been associated with MODY (maturity onset diabetes of the young) and one polymorphism in KCNJ11 (rs5219; resulting in E23K; Table 1) is associated with increased risk of T2D in multiple GWAS studies[26, 27]. The effects of the K allele on channel activity and insulin secretion assessed by glucose tolerance tests have been investigated in a number of studies[26, 28–30]. Although some reported a small effect, others failed to detect significant difference compared to the E allele. There is now increased appreciation that tiny effect sizes of many common risk alleles likely contribute to complex diseases such as T2D[31]. Despite strong genetic evidence, proof of genotype-phenotype association may be challenging.

Table 1.

Genetic Variations in Ion Channels Implicated in Diabetes ¹

Gene Name	Common Name	Proposed function in β-cells	Polymorphisms or mutations	Functional effects on the channel	Impact on blood glucose	References
KCNJ11	Kir6.2 (KATP pore subunit)	setting resting Vm	p.E23K	Gain of function	Increased risk of T2D	[26, 27]
ABCC8	SUR1 (KATP regulatory subunit)	setting resting Vm	p.R370S	Loss of function	Associated with hyperinsulinemia early in life but glucose intolerance or diabetes in adult carriers	[34]
ABCC8	SUR1 (KATP regulatory subunit)	setting resting Vm	p.E1507K	Loss of function	Associated with hyperinsulinemia early in life but glucose intolerance or diabetes in adult carriers	[32, 33]
KCNK16	TALK-1	Vm; ER K+ flux	p.A277E	Gain of function	Increased risk of T2D	[58, 59, 70]
KCNH6	hERG2 (Kv11.2)	AP repolarization?	P235L	Loss of function	Hyperinsulinemia hypoglycemia early in life and hyperinsulinemia and diabetes in adults	[103]
KCNJ9	Kir3.3 or GIRK3	GPCR activation induced Vm hyperpolarization	p.V366A	Not determined	Increased risk of T2D in PIMA Indians	[124]
SCN9A	Nav1.7	AP	p.I739V	Gain of function	Two thirds of these patients developed diabetes	[152]

¹Note numerous gain of function mutations causing neonatal diabetes have been identified in KCNJ11 and ABCC8; they are not listed in this table.

Worth noting, two dominant inactivating ABCC8 mutations, E1507K and R370S, have been associated with hyperinsulinism early in life but glucose intolerance or diabetes in adult carriers[32–34]. The glucose intolerance may be attributed to glucose-blindness (i.e. impaired insulin secretion responses to acute or ramp glucose stimulations), a phenotype well documented in mice lacking K_ATP channels[35, 36]. The impaired insulin secretion response to glucose may be a result of chronic elevations of cytosolic Ca²⁺. Curiously, later large scale studies of dominant congenital hyperinsulinsm-causing mutations did not find significantly increased risk for diabetes later in life[37]. An interesting possibility is that the different findings are due to differences in how the patients were treated for their hyperinsulinemia and hypoglycemia, for example with diazoxide (a K_ATP channel opener) versus glucose feeding, which would affect Ca²⁺ levels, hence β-cell function.

K_ATP channels have been a major therapeutic target for T2D for more than half a century[38]. Sulfonylureas such as glibenclamide (glyburide) and glinides such as repaglinide (prandin) inhibit K_ATP channel activity to stimulate insulin secretion[39] and remain a first-line T2D treatment, especially in low-income settings[40]. Although long-term use of sulfonylureas frequently results in accelerated β-cell failure likely caused by chronic hyperexcitability and ER stress[41–43], studies in mice suggest this failure is readily reversible without permanent damage to β-cells[44]. Whether periodic stops in sulfonylurea therapy can prolong drug effect in human patients remains to be seen[45]. Since the discovery that gain-of-function K_ATP mutations underlie neonatal diabetes, glibenclamide has also become a mainstay for treating K_ATP-neonatal diabetes and DEND syndrome[46]. In addition to sulfonylureas and glinides, carbamazepine, an anticonvulsant best known for its effect to block Na_v channels[47], was recently shown to also inhibit K_ATP channel activity[48]. Carbamazepine competes for glibenclamide binding to K_ATP channels with an estimated K_d of ~25 nM[49] and abolishes channel response to MgADP[50]. Structural studies using cryoEM have revealed that glibenclamide and carbamazepine as well as repaglinide share a common binding pocket in the SUR1 protein, consistent with these drugs acting via a similar mechanism to inhibit K_ATP channels[13, 51–53]. Given the similarity between carbamazepine and sulfonylureas in their effects on K_ATP channels, it would be predicted that carbamazepine would also stimulate insulin secretion. However, since carbamazepine also inhibits Na_v channels and may have other actions such as maintaining functional β-cell mass[54], its effect on insulin secretion will require further investigation.

Two-Pore-Domain Potassium (K2P) Channels

K2P channels are small conductance K⁺ channels that are active at all membrane potentials and thus termed “leak” K⁺ channels. Human β-cells show high expression of many K2P channels including: TASK-1, TALK-1, TALK-2, and TWIK-1. The K2P channels expressed in islets show outward rectification with small K⁺ currents at hyperpolarized potentials that increase during depolarization. During glucose-induced inhibition of K_ATP channels most K2P channels stabilize the plateau potential from where APs fire[35, 55, 56]. Therefore, K2P channels help to regulate the V_m and Ca²⁺ entry during GSIS (Figure 3)[57]. Although the roles of K2P channels in islet biology are still being elucidated, K2P channel polymorphisms and mutations have been identified that negatively impact β-cell function. For example, a GOF polymorphism in KCNK16 that encodes TALK-1(rs1535500 resulting in A277E) is associated with an increased risk of developing T2D[58, 59]. This is likely due to enhanced β-cell V_m polarization, which is predicted to reduce β-cell Ca²⁺ entry and insulin secretion. Furthermore, rs1535500 also results in increased expression of the downstream gene KCNK17 in human islets, which encodes TALK-2[60]. This suggests that higher levels of TALK-2 together with greater function of TALK-1 lead to a hyperpolarizing impact on β-cell V_m, which would limit VDCC activity and reduce insulin secretion. Interestingly a gain-of-function mutation in TALK-2 that results in severe cardiac conduction disorder does not cause overt diabetes[61]; however, patients with this mutation would be predicted to show decreased GSIS. Because KCNK16 (that encoded TALK-1) is the most abundant and most β-cell restricted K⁺ channel transcript, mutations that cause extensive TALK-1 channel gain-of-function would be predicted to cause neonatal diabetes or maturity onset diabetes of the young (MODY).

Figure 3:

Ion channels that control where the β-cell V_m sits. At rest the activity of K_ATP channels holds the β-cell V_m below the activation threshold for VDCCs. When blood glucose concentrations are elevated closure of K_ATP leads to V_m depolarization. During glucose-induced K_ATP closure the constitutive activity of small conductance K2P potassium channels such as TALK-1 and TASK-1 balance the depolarizing conductance from SWELL-1 and Na⁺ leak channels at a depolarized plateau potential from where VDCCs and VGSCs can efficiently activate, allowing AP firing. Ca²⁺ entry which occurs during VDCC activation leads to Ca²⁺ release from the endoplasmic reticulum (ER), which results in activation of Ca²⁺ activated K⁺ currents (K_slow) that hyperpolarize the V_m leading to VDCC closure. K_slow currents eventually inactivate in response to the resulting low Ca²⁺ levels caused by hyperpolarization-induced closure of VDCCs, which results in another wave of depolarization and Ca²⁺ influx. These β-cell oscillations in V_m and Ca²⁺ are responsible for pulsatile insulin secretion. K2P channels such as TALK-1 also form functional complexes on the ER membrane. ER TALK-1 channels provide a countercurrent for Ca²⁺ ions leaving the ER and potentiate ER Ca²⁺ release. This is important for increasing K_slow activation and therefore loss of TALK-1 channels impairs Ca²⁺ ER release limiting the K_slow current and increasing Ca²⁺ oscillation frequency. Other K2P channels expressed in the human β-cell (e.g. TWIK-1 and TALK-2) also presumably play roles in regulating V_m or ER Ca²⁺ handling.

Interestingly, TALK-1 and TASK-1 channels also form function complexes on the endoplasmic reticulum (ER) membrane (Figure 3). Maintenance of ER Ca²⁺ (Ca²⁺_ER) homeostasis requires that Ca²⁺ movement across the ER membrane is balanced with a simultaneous K⁺ flux in the opposite direction [62–64]. Similar to other ER K⁺ channels such as the trimeric intracellular cation channels, TRIC-A or TRIC-B [65, 66], TALK-1 and TASK-1 control Ca²⁺ movement across the ER membrane by balancing Ca²⁺_ER release with K⁺ flux in the opposite direction, [62–64]. Without this K⁺ countercurrent, Ca²⁺ release from the ER rapidly generates a negative charge on the inside of the ER membrane, inhibiting further Ca²⁺_ER release [67–69]. TALK-1 control of Ca²⁺_ER release also impacts the β-cell V_m response to glucose[70]. Glucose-induced Ca²⁺ entry results in Ca²⁺_ER release leading to activation of Ca²⁺-activated K⁺ (K_Ca²⁺) channels that repolarize the V_m during the silent phases between oscillations of membrane depolarization (these K_Ca²⁺ currents are collectively termed K_slow)[70]. Therefore, TALK-1 channel control of Ca²⁺_ER release slows β-cell Ca²⁺ oscillation frequency and thus is predicted to tune pulsatile insulin secretion. This suggests that polymorphisms or mutations that impact TALK-1 function would be predicted to impact Ca²⁺_ER handling as well as V_m. Indeed the T2D associated GOF polymorphism (rs1535500, TALK-1-A277E) in KCNK16 results in less Ca²⁺_ER[70]. TALK-1-A277E-induced reduction in Ca²⁺_ER storage also causes an increased ER stress response. Importantly, Ca²⁺_ER homeostasis becomes perturbed during the pathogenesis of T2D as well as T1D, which leads to β-cell dysfunction[71–74]. Furthermore, certain mutations cause monogenic neonatal diabetes by increasing ER stress (e.g. INS-1, and WFS-1)[75, 76]. Therefore, enhanced TALK channel activity at the ER membrane with rs1535500 would be predicted to contribute to ER stress during the pathogenesis of T2D. Other K2P channels may also help control the ER stress response by tuning β-cell Ca²⁺_ER handling.

Although K2P channels are K⁺ selective ion channels, one of the β-cell K2P channels, TWIK-1, is not only permeable to K⁺ but undergoes pore collapse under acidic conditions becoming permeable also to Na^{+ [1]}. In β-cells, TWIK-1 channels rapidly cycle between endosomes and the plasma membrane. When TWIK-1 channels reach the plasma membrane from acidic endosomes they are still in a conformation permeable to Na⁺. Interestingly, because conformational recovery from the pore collapse upon subsequent exposure to physiological extracellular pH (>7) as the channels arrive the plasma membrane is a slow process^[1], the TWIK-1 channels are recycled back to the endosomes before they ever lose Na⁺ permeability and become permeable exclusively to K⁺ at the plasma membrane ^[1]. Under low glucose during conditions when K_ATP channels are active the TWIK-1 Na⁺ permeability does not significantly influence the β-cell V_m. However, when K_ATP is inhibited during high glucose conditions the TWIK-1 Na⁺ permeability helps to depolarize the β-cell V_m. Thus, mouse β-cells deficient for TWIK-1 sit at a more hyperpolarized V_m under stimulatory glucose conditions and are predicted to have less VDCC activity and reduced insulin secretion[1]. This suggests that TWIK-1 could also provide an important Na⁺ leak current in human β-cells.

Calcium-activated K⁺ channels (K_Ca²⁺)

Calcium-activated K⁺ (K_Ca²⁺) channels play an important role in the repolarizing phase of the β-cell action potential. The K_Ca²⁺ channel with the highest expression in human β-cells is the BK channel, which is activated by both depolarization as well as Ca²⁺ binding. The BK channels are the largest and most rapidly activating K⁺ current that initiates during the AP upstroke and therefore these channels play an important role repolarizing the β-cell AP. Inhibition of BK channels results in increased human β-cell AP amplitude as well as increased depolarization-induced insulin secretion (during K_ATP inhibition)[77]. BK channels are also expressed in neurons and LOF mutations in these channels cause epilepsy with cerebellar atrophy due to increased neuronal electrical excitability[78], which may also be the case in β-cells of patients with BK channel LOF; this would be predicted to improve GSIS. Interestingly, GOF mutations in BK also cause epilepsy due to faster AP firing frequency[79], which has been suggested to be due to faster AP repolarization and increased ability to fire subsequent APs[79]. While the role of BK channel GOF mutations on glucose tolerance has not been determined, if BK GOF mutations cause greater AP firing frequency in β-cells as they do in neurons, then this would be predicted to lead to an improvement of glucose tolerance due to greater electrical excitability and Ca²⁺ entry following glucose stimulation.

Other K_Ca²⁺ channels are also expressed in human β-cells including the small-conductance Ca²⁺-activated potassium channel, SK3, and the intermediate conductance Ca²⁺-activated K⁺ channel, IK[80, 81]. These channels limit the frequency of β-cell APs similar to their action in neurons[82–85]. SK3 and IK activation occurs as a consequence of Ca²⁺ influx during the AP and remains open after termination of the AP leading to after-hyperpolarization (AHP)[86, 87]. AHP limits the firing of another AP and thus reduced AP firing frequency[87]. KCNN3 and KCNN4 the genes that code for SK3 and IK1 are expressed at fairly low levels in human β-cells[80]. Thus, human islet insulin secretion is only modestly increased by IK inhibition and does not change with SK3 inhibition[88]. Therefore, these channels may not contribute substantially to dysfunction of Ca²⁺ handling in T2D. Gain-of-function mutations in IK channels cause hereditary xerocytosis (HX)[89]. Although certain types of HX are associated with diabetes and overactivity of IK would be predicted to limit insulin secretion[88], HX-associated diabetes is primarily due to pancreatic hemochromatosis and iron deposition[90]. It has also been found that islet SK3 expression is increased in rats following intrauterine growth restriction; a condition which increases the risk of T2D development in adulthood[91]. Therefore, conditions that upregulate β-cell SK3 levels could potentially negatively impact β-cell function by reducing β-cell electrical excitability.

K_v channels

Many voltage-gated K⁺ (K_v) channels are expressed in human islets[4, 80]. Although K_v currents can readily be recorded from human β-cells, their role in regulating electrical excitability and insulin secretion is controversial[77, 92]. K_v2.1 and K_v2.2 currents are responsible for 65–80% of human β-cell K_v activity and can be inhibited with toxins[77, 92]. Interestingly, stromatoxin inhibition of both K_v2.2 and K_v2.1 has been shown to have little impact on human β-cell electrical activity and GSIS⁵⁵. However, another K_v2.1 and K_v2.2 toxin inhibitor (guangxitoxin-1E) causes substantial increases of human islet GSIS[92]. Furthermore, a small molecule (RY796) inhibitor of both K_v2.1 and K_v2.2 also enhances human islet GSIS[92]. While the exact roles of K_v2.1 and K_v2.2 channel activity on human β-cell GSIS remain to be determined, the KCNB2 gene that encodes K_v2.2 shows much higher expression (nearly 10 fold greater) levels in β-cells when compared to KCNB1 that encodes K_v2.1[4, 80]. This suggests that Kv2.2 channels may be more important in controlling human β-cell electrical activity; selective K_v2.2 knockdown in human β-cells will help uncover how this channel impacts electrical activity. Interestingly, the expression of human islet K_v2.1 is reduced in T2D[93]. However, as the KCNB1 is expressed at a low level compared to other K⁺ channel transcripts (e.g. encoding BK, Kir6.2 or TALK-1)[4, 80], the K⁺ conductance of K_v2.1 would be expected to only modestly impact human β-cell electrical activity[77]. Mutations in KCNB1 have been identified, which result in infantile epilepsy[94–97]. These mutations cause dominant negative inhibition of K_v2.1 or loss of K⁺ selectivity of the K_v2.1 channel complex[94–97]; both of which result in increased neuronal excitability that leads to generalized or focal seizures[94–97]. Importantly, a mutation in K_v2.1 that results in a constitutively activated[95], nonselective cation channel that is not voltage activated would be predicted to result in a large enhancement of insulin through increased Ca²⁺ influx. While this has yet to be assessed in serum from these patients, the outstanding resource for patients with KCNB1 mutations (KCNB1.org) could allow a clear assessment of the function or lack of function of K_v2.1 currents during human insulin secretion.

Other K_v channels are also expressed in the human β-cell, which presumably contribute to AP repolarization and could impact β-cell function in T2D patients. For example, the gene that encodes the K_v1.7 channel is in a region on chromosome 19 (19q13.3) that has SNPs linked to an increased susceptibility for developing T2D[98]. Furthermore, blocking Kv1.7 in rodent islets increases AP firing frequency and insulin secretion[99]. Other K_v channels that are expressed in and regulate human β-cell function are the human ether-á-go-go-(HERG) and related gene encoding K_v11.1 and K_v11.2 respectively[4, 80], which are also important cardiac or neuron repolarizing K⁺ channels[100, 101]. Importantly, loss-of-function K_v11.1 mutations that cause long-QT syndrome result in increased insulin secretion and hypoglycemia[102]. This suggests that K_v11.1 inhibits GSIS presumably through polarizing influences on the human β-cell V_m[102]. Interestingly, however, loss-of-function mutations in KCNH6 the gene encoding K_v11.2 causes maturity onset diabetes of the young[103]. The diabetic phenotype occurs in adults and individuals with these mutations also show hyperinsulinemia as children[103, 104]. Similarly, mice with a K_v11.2 LOF mutation show hyperinsulinemia after birth followed by hypoinsulinemia as adults[103]. Thus, K_v11.2 LOF induces enhanced Ca²⁺ entry into β-cells that results in increased insulin secretion as well as ER stress over time leading to β-cell failure and diabetes[103]. As K_v11.x GOF mutations would be predicted to limit VDCC activity and reduce insulin secretion, these mutations would be predicted to also result in a diabetic phenotype.

K_v proteins have also been shown to have roles independent of K⁺ flux. For example, K_v2.1 and K_v2.2 channels interact with vesicle-associated membrane protein-associated proteins isoforms A and B (VAPA and VAPB) to enable close association of the ER and plasma membrane[105, 106], which presumably could control Ca²⁺ handling and/or influence actin dynamics of the human β-cell. Furthermore, K_v2.1 has also been shown to interact with SNARE machinery (e.g. Syntaxin-1A and SNAP-25)[107, 108], recruit insulin granules to the plasma membrane, and enhance GSIS[109]. SNARE machinery such as Syntaxin-1A has also been shown to bind many other ion channels such as VDCCs and the K_ATP channel complex to modulate their function or surface expression, which may also impact insulin secretion[110, 111]. Although KCNB1 the gene encoding K_v2.1 is expressed at much lower levels than the other β-cell ion channels that bind syntaxin-1A, reduced T2D islet KCNB1 has been suggested to negatively impact human β-cell insulin secretion through reduced K_v2.1 syntaxin-1A interactions[93]. This suggests patients with a truncation of the C-terminus of K_v2.1 (such as chr20:47991009delG,p.S363fs*13; which does not contain a sytaxin-1A binding site) may show significantly impaired GSIS and a predisposition for developing diabetes[112](KCNB1.org).

G protein-gated inwardly rectifying K⁺ (GIRK) channels

GIRK channels are activated by Gi/o-coupled receptors when they release Gβγ[113, 114]. Two primary types of Gi-coupled receptors found in human β-cells are the somatostatin receptors (SSTRs) and the α2 adrenoreceptors (α2-ADRs)[115–118]. Therefore, somatostatin and epinephrine activate β-cell GIRK channels[119], which cause β-cell V_m hyperpolarization and reduced VDCC activity[113, 120]. The importance of GIRK control of β-cell Ca²⁺ handling is exemplified by the increase in Ca²⁺ entry and insulin secretion that occurs during Gi receptor signaling inhibition with pertussis toxin (PTX; also termed islet-activating protein)[121, 122]. Although GIRK has not been directly linked to T2D, a polymorphism in the α2A-ADR encoding gene ADRA2A (rs553668), is associated with an increased risk of developing T2D[123]. Human islets from carriers of the rs553668 polymorphism exhibit reduced insulin secretion and greater levels of α2A-ADR[123], which causes enhanced GIRK activation and thus hyperpolarizes the β-cell reducing VDCC activity. Mutations and polymorphisms in the genes that encode GIRK channels have also been identified, which may impact glucose homeostasis. For example, single nucleotide polymorphisms in KCNJ9 the gene encoding GIRK3 show significant correlation with increased risk of developing diabetes in a PIMA Indian population[124]. Furthermore, a gain-of-function mutation in KCNJ5 that encodes the GIRK4 channel causes Familial Sinus Node Disease[125] but has not yet been shown to disrupt glucose homeostasis in these patients. As GIRK4 is expressed in human β-cells[80, 120], over activity of this channel would be predicted to inhibit insulin secretion.

GIRK channels are also regulated by signals other than Gβγ such as phosphatidylinositides, cholesterol and intracellular Na^{+ [126–128]}. Thus their activity changes under conditions associated with T2D may also lead to perturbations in β-cell function. For example phosphatidylinositol 4,5-bisphosphate (PIP₂) is required for GIRK channel opening[129–131]. As insulin resistance that occurs during the pathogenesis of T2D leads to inhibition of insulin-induced activation of the PI3K/AKT pathway[132], this may reduce GIRK activity. Furthermore certain GIRK channels activity is enhanced during conditions of high intracellular Na⁺ [126, 133, 134]. Because intracellular Na⁺ concentration increases during excessive electrical excitation that occurs during the excitotoxic conditions associated with T2D, Na⁺-mediated activation of GIRK may eventually lead to β-cell polarization and limit VDCC activity[6]. On the other hand, GIRK activation under excitotoxic circumstances may also help protect β-cells from apoptosis by reducing the excessive Ca²⁺ entry[135]. Finally, cholesterol activates GIRK conductance and thus changes in cholesterol typically observed in patients with T2D could result in overactive β-cell GIRK, which would also reduce VDCC activity and insulin secretion[136].

Ca²⁺ channels

Voltage-dependent Ca²⁺ channels are the primary entry point for β-cell Ca²⁺ during GSIS. As Ca²⁺ entry is essential for β-cell GSIS, blocking VDCCs impairs insulin secretion[137]. The VDCCs that have been implicated in GSIS include: the L-type channels Ca_V1.2 and Ca_V1.3; the P/Q-type channel Ca_V2.1; and the T-type channel Ca_V3.2[138]. Indeed, gain-of-function mutations in VDCCs have been shown to result in hyperinsulinemia and hypoglycemia. For example a GOF mutation in the Ca_V1.2 channel associated with Timothy syndrome results in intermittent hypoglycemia, which may be due to enhanced insulin secretion[139]. Furthermore, a gain-of-function mutation in the Ca_V1.3 channel also results in hyperinsulinemia and hypoglycemia[140]. As enhanced VDCC activation causes increased insulin secretion, reduced VDCC activity would be predicted to reduce insulin secretion and result in diabetic phenotypes. Indeed, VDCC transcript abundance has been shown to be reduced in T2D. Expression of the CACNA1D gene that encodes Ca_V1.3 is reduced in T2D islets, which correlates with reduced insulin secretion[141]. This is due to polymorphisms in the promoter of CACNA1D that result in reduced expression of the CACNA1D gene[141]. The CACNA1D gene shows open chromatin structure in human islets and changes in transcription factor binding to this promotor are proposed to reduce Ca_V1.3 levels[141, 142]. As reduced Ca_V1.3 levels are associated with T2D, it would be predicted that a loss-of-function mutation in Ca_V1.3 would also disrupt insulin secretion and result in a diabetic phenotype[143]. Although Ca_V1.3 LOF mutations cause deafness, how Ca_V1.3 LOF mutations impact glucose homeostasis has not been assessed[143].

Aside from genetic contributions, during the pathogenesis of diabetes signals that regulate VDCC activity change and likely impact β-cell Ca²⁺ influx and insulin secretion. For example, PIP₂ β-cell increases VDCC activity[144]. As detailed above, PIP₂ signaling becomes disrupted under the insulin resistant conditions of T2D due to reduced insulin-induced activation of the β-cell PI3K/AKT pathway[132]. Therefore, reductions in β-cell PIP₂ levels would be predicted to impair VDCC activity in T2D. Furthermore, inflammatory cytokines are known to increase in T2D and cytokines impact VDCC activity. Short term exposure of β-cells to cytokines has been shown to increase low-voltage activated VDCC currents[145], which would be expected to actually enhance insulin secretion. However, elevated basal β-cell Ca²⁺ levels on a background of insulin hypersecretion may also have negative impacts such as enhancing β-cell ER-stress and mitochondrial dysfunction causing deterioration of β-cell health[146]. Indeed, VDCC inhibition with verapamil has been found to protect β-cells from the stressful conditions of T2D. While this is due in-part to verapamil-mediated repression of thioredoxin-interacting protein expression that results in reduced oxidative stress[135], other agents such as diazoxide that also limit VDCC activity by hyperpolarizing the β-cell V_m protect β-cells from failure under T2D conditions[147, 148]. It is known that during the early stage of T2D development, β-cells compensate for insulin resistance by secreting more insulin through elevated total Ca²⁺ influx. Thus, the relationship between VDCC activity and β-cell function in T2D may depend on the stage of the disease[6].

While changes in VDCC activity can result in T2D [6, 141], there are also changes in location of VDCCs to the insulin granule that are negatively impacted during the pathogenesis of T2D[149]. During insulin secretion VDCCs cluster near granule release sites[149]. β-cell L-type VDCC clusters of 15 to 20 channel complexes form near insulin granules during secretion[149]. VDCC clustering allows for greater depolarization-induced Ca²⁺ entry near docked insulin granules, which helps to promote insulin granule fusion with the plasma membrane[149]. In T2D β-cells Ca²⁺ influx near granules is disrupted and total depolarization-induced Ca²⁺ entry is reduced[149]. Ca²⁺ entry at the plasma membrane micro domains near insulin granules in response to membrane depolarization is also significantly reduced in T2D β-cells[149]. Furthermore, insulin granule release does not preferentially occur near Ca²⁺ entry sites in T2D[149]. Therefore, both Ca²⁺ handling and granule location to hot spots for Ca²⁺ entry following secretagogue stimulation are perturbed in T2D β-cells.

Na⁺ Channels

Voltage-gated Na⁺ channels play an important role in controlling action potential firing in human β-cells. Thus, inhibition of VGSCs results in reduced human β-cell AP amplitude, diminished Ca²⁺ entry, and reduced insulin secretion [77]. β-cell ablation of rodent VGSC encoding α-subunits (e.g. Scn3a) or β-subunits (e.g. Scn1b) also leads to reduced GSIS[150, 151]. While mutations or polymorphisms in VGSC encoding genes have not been shown to increase diabetes risk, a gain-of-function mutation in Na_v1.7 (I739V) has been identified that causes painful neuropathy[152]. Two thirds of the patients identified with this GOF Na_v1.7 (I739V) mutation also developed diabetes[152], which suggests that increased VGSC may result in a diabetic phenotype but further studies are required to support this. As blocking VGSCs causes reduces insulin secretion[77], it seems counterintuitive that more activity of VGSCs could cause β-cell dysfunction. However, blockade of VGSCs with carbamazepine has been shown to protect β-cells from inflammation associated β-cell destruction[153]. VGSC inhibition may reduce β-cell excitotoxicity in response to the increased electrical excitability of β-cells required to meet increasing insulin demand in T2D; indeed this has been shown in T2D after reducing β-cell electrical excitability with the K_ATP activator diazoxide[147]. However, there are many off target influences of carbamazepine such as VDCCs[154], which may also play a role in the β-cell protective influences of carbamazepine. Interestingly, there is heterogeneous expression of Scn3a in human islet β-cells, which potentially implicates that the different amounts of β-cell VGSCs could provide variability in β-cell electrical excitability and GSIS[155]. Heterogeneity of human β-cell Scn3a expression may also impact sensitivity to excitotoxicity that occurs during the pathogenesis of T2D.

Swell-Activated Cl⁻ Channels (SWELL)

During glucose stimulation β-cells swell leading to activation of volume-regulated anion current (VRAC), which is also termed the swell-activated Cl⁻ current (I_Cl-,SWELL)[156–158]. VRAC/SWELL activation allows for the regulatory volume decrease (RVD) that occurs following exposure of β-cells to hypoosmotic conditions[156]. Although VRAC/SWELL would be predicted to produce RVD following glucose-induced swelling, β-cells maintain an increased volume during the duration of glucose stimulation[156]. Thus, VRAC/SWELL currents activated during glucose-induced swelling remain active until euglycemia has been reached[156]. The leucine-rich repeat–containing (LRRC8) proteins form the channels termed SWELL1, which are responsible for the VRAC/SWELL currents in β-cells[159]. When SWELL1 is activated during glucose-induced β-cell swelling Cl⁻ efflux through these channels enhance depolarization leading to greater Ca²⁺ influx through VDCCs[156–158]. Although SWELL1 serves as a critical ion channel that controls glucose-stimulated β-cell depolarization, this channel complex has also been shown to transport neurotransmitters such as glutamate and γ-aminobutyric acid (GABA)[160, 161]. Because ionotropic glutamate and GABA receptors play an important role in regulating β-cell electrical excitability (detailed below), SWELL1 may contribute to controlling the activities of these receptors by influencing ligand release from β-cells. SWELL1 activity may change under diabetic conditions; for example, the GLUT1, GLUT-2 and glucokinase are reduced in T2D islets[162] and the resulting reduction in glucose transport and utilization in T2D β-cells could lead to reduced swelling as well as activation of SWELL1. Therefore, reduced SWELL1 in T2D could reduce glucose-stimulated depolarization, VDCC activation and insulin secretion.

Ionotropic Receptors

Neuronal innervation as well as paracrine mediated neurotransmitter release modulates islet cell function in part through ionotropic receptors, which becomes disrupted during the pathogenesis of diabetes. The main ionotropic receptors expressed in the islet are: GABA receptors, glycine receptors, purinergic receptors and glutamate receptors(Figure 4) [163–167], as we discuss below.

Figure 4:

Ionotropic receptors that control β-cell insulin secretion. Ionotropic receptors expressed in human β-cells are activated by ligands and selective to specific ions. Upon activation the ionotropic receptors modulate VDCC activity by controlling the V_m. Ionotropic receptors that are permeable to Cl⁻ such as the glycine receptors (GLYRs) and GABA receptors (GABAARs) lead to V_m depolarization upon activation due to Cl⁻ efflux, which results in enhanced VDCC activation and insulin secretion. Ionotropic receptors that are permeable to Ca²⁺ and Na⁺ such as the NMDARs and P2XRs lead to both Ca²⁺ entry and V_m depolarization upon activation. While P2XRs enhance Ca²⁺ entry and GSIS, NMDARs cause the opposite. Two possible mechanisms for NMDAR inhibition of GSIS are through excessive depolarization causing voltage-dependent inactivation of VDCCs or through enhanced trafficking of K⁺ channels (including K_ATP and K_V2.1) to the plasma membrane leading to V_m hyperpolarization. Ionotropic receptor ligands are released from the β-cell (autocrine) or are released from other nearby cells (paracrine) to control β-cell function.

GABA receptors

GABA receptors (GABARs) are also expressed on human β-cells. GABARs are chloride channels that when activated depolarize the β-cell and have been shown to stimulate VDCC activity leading to insulin secretion[163]. Depolarization by GABARs is due to Cl⁻ efflux from the β-cell (Figure 4); a high intracellular concentration of Cl⁻ in β-cells allows for Cl⁻ efflux at the V_m from where electrical excitability occurs[168]. However, there is also evidence that activation of GABARs can inhibit insulin secretion; this results either due to voltage-dependent inactivation of VGSCs and VDCCs or potentially via Cl⁻ influx through GABARs instead of efflux[169]. β-cells that are depolarized above the reversal potential of Cl⁻ (~−40mV) would be expected to show Cl⁻ influx through GABARs and this could be the case under conditions of excessive depolarization, which may occur under certain conditions such as T2D patients treated with sulfonylureas. Islets from T2D donors also show reduced expression of ionotropic GABA(A)Rs, which would be predicted to reduce GABA signaling[170]. GABA is produced and secreted from β-cells. GABA levels are reduced in T2D islets[169]. The enzyme GAD65 that produces GABA from glutamate shows reduced activity under the stressful conditions associated with T2D, which can be due to changes in transcription of GAD65 as well as perturbations in palmitoylation that change the intracellular localization of GAD65[171, 172]. Reduced GAD65 activity leads to reduced β-cell GABA in human islets from patients with T2D[169]. To compensate for the reduced GABA levels the islet GABARs show increased open probability and current in T2D β-cells when compared to nondiabetic (ND) β-cells[169]. This is not due to expression differences in the genes that encode GABARs as these receptors show equivalent expression in non-diabetic and T2D islets[169].

GABA has also been reported to have protective impacts on β-cell function, stimulate β-cell replication, as well as promoting α-cell neogenesis and conversion into β-cells[173–177]. GABA stimulated β-cell depolarization causes Ca²⁺ entry, which can activate PI3K and Akt signaling pathways that promote growth and survival[178]. While this may be one mechanism that promotes β-cell replication in response to GABA, it has also been shown that GABA promotes α-cell neogenesis and transdifferentiation to increase β-cell mass[174]. GABA induced β-cell formation through α-cells would be predicted to incorporate GABA-induced changes in V_m potentially through hyperpolarization induced inhibition of glucagon secretion, which could lead to α-cell hyperplasia that occurs during conditions of reduced liver glucagon signaling[179]. The positive impacts of GABA on β-cell number and function indicate that increasing GABA or GABA activity in islets during the pathogenesis of T2D would limit β-cell dysfunction and maintain β-cell mass[173–176]. For example, positive allosteric modulators that enhance GABA activity on GABARs result in enhanced human β-cell replication and GABAR activators (e.g. muscimol) stimulate human β-cell calcium entry and insulin secretion[180]. However, it is worth keeping in mind that recent studies call into question whether activation of GABARs enhance β-cell mass or function[181, 182]. Thus, more studies are needed to fully elucidate the functional role of GABAR in β-cells.

Glycine receptors

Human β-cells express ionotropic glycine receptors (GlyRs) including GlyRα1 and GlyRα3[164]. Glycine is also found in insulin secretory granules and can be released during GSIS[183]. GlyR activation by glycine results in β-cell Cl^- efflux leading to V_m depolarization and increased electrical excitability (Figure 4)[164]. Glycine-induced elevations of human β-cell Ca²⁺ enhance insulin-secretion and thus glycine provides autocrine feedback to enable greater GSIS. Furthermore, supplementation with glycine improves human glucose tolerance[184–188]. Patients with T2D show decreased expression of the GlyRα1 transcript as well as protein abundance[5, 164]. Reduced GlyRα1 in T2D β-cells would be predicted to reduce glycine-induced depolarization and the resulting amplification of insulin secretion. Indeed glycine induced Cl^- currents are reduced by ~50% in β-cells from T2D donors when compared to nondiabetic β-cells[164]. Moreover, T2D patients treated with conditions that increase plasma glycine levels show improved glucose handling and a delay in diabetes onset^{155; 156; 157; 158; 159}. Interestingly, insulin also stimulates GlyR activity[164, 189] and therefore the insulin resistance associated with T2D is predicted to reduce GlyR activation during glucose-stimulation. Taken together, these data suggest that glycine control of β-cell Ca²⁺ entry through activation of GlyRs is perturbed in T2D and contributes to reductions in β-cell Ca²⁺ handling and insulin secretion.

Purinergic receptors

The P2X purinoceptors (P2XRs) are activated by purinergic ligands such as ATP and are primarily permeable to Ca²⁺ (Figure 4). There are many P2X receptors expressed in human β-cells including P2X2 P2X3 P2X4 P2X6 and P2X7[165]. All P2X receptors would contribute to ATP-induced Ca²⁺ entry and enhancement of insulin secretion (Figure 4)[166]. As ATP is released from insulin granules during GSIS[190], P2XR activation augments glucose-stimulated Ca²⁺ entry and thus GSIS[166]. While the exact role for each P2X receptor of the human β-cell has yet to be conclusively established, changes in P2X receptor protein levels have been observed in T2D islets[165]. Specifically, P2X7 levels are reduced in β-cells of patients with T2D[165]. Interestingly, P2XR7 expression and P2X7 protein levels are also controlled by glucose and saturated fatty acid levels[165]. When glucose and palmitic acid are elevated human islets show increased P2X7 levels in the short term[165]. This suggests that increased P2X7 levels can help β-cells adapt to the toxic conditions that eventually cause T2D. As P2X7 levels eventually drop in T2D, reduced P2X7 during prolonged stress may lead to reductions in GSIS due to reduced Ca²⁺ entry.

Polymorphisms near or in the P2XR7 gene are also associated with changes in glucose homeostasis[191]. A GWAS study in a Chinese population identified many polymorphisms near P2XR7 that is associated with an increased risk of developing T2D[192]. Certain polymorphisms in or near P2XR7 were also found associated with changes in insulin sensitivity and variations in glucose homeostasis[191]. Moreover, a missense polymorphism (rs1718119) that results in A348T within P2RX7 was identified that is associated with increased insulin sensitivity and greater insulin secretion. Importantly, the P2RX7 A348T results in a gain-of-function that increases channel activity in response to ATP and is thus predicted to enhance β-cell Ca²⁺ entry and insulin secretion[191, 193]. On the other hand, rodents with a P2RX7 variant (P451L) that results in a loss-of channel function show impaired glucose tolerance and insulin resistance compared to controls (with a proline at position 451), which suggests a potential impairment in islet function[191]. Therefore, P2RX7 plays an important role in enhancing β-cell insulin secretion and reduced P2RX7 function leads to islet dysfunction and insulin resistance, which increases the risk of T2D.

Glutamate receptors

Many ionotropic glutamate receptors are expressed in the islet (reviewed in[194]). They are known to be responsible for controlling α-cell Ca²⁺ handling and glucagon secretion as well as δ-cell somatostatin secretion[195, 196]. In the β-cell the primary ionotropic glutamate receptor that controls membrane potential are N-methyl-d-aspartate receptors (NMDAR), which are composed of GluN subunits including the GluN1 obligatory subunit. Interestingly, NMDAR activation limits glucose-stimulated Ca²⁺ influx in human and mouse β-cells, which has been proposed to be due to activation of K⁺ conductance that leads to inhibition of VDCC activity[167]. Recently, NMDARs have been reported to regulate the trafficking of K_ATP channels and K_v2.1 channels in INS-1 and dispersed human β-cells[197]. Stimulation of NMDARs either directly by agonists or by leptin acting on β-cell leptin receptors increases the abundance of K_ATP and K_v2.1 channels at the cell surface and results in membrane hyperpolarization, which would explain how activation of NMDARs inhibits GSIS.

Patients with T2D show elevations in glutamate levels[198, 199]. This change has to do with differences in consumption but also changes in liver production of glutamate^{169; 170}. Thus, elevations in glutamate in obese individuals during the pathogenesis of T2D may lead to over activation of NMDARs. Moreover, hyperglycemic conditions have been shown to lead to increased β-cell glutamate release, which could also contribute to impaired β-cell function in T2D. Importantly, inhibition of NMDARs with dextromethorphan leads to both enhanced insulin secretion as well as protecting β-cells from failure in response to the stressful conditions associated with T2D[167]. Interestingly the enhanced insulin secretion effect by NMDAR inhibitors is dependent on the presence of K_ATP channels as mouse islets from Kir6.2 knockout animals failed to show such response, suggesting NMDAR inhibition stimulates insulin secretion by blocking NMDAR-dependent K_ATP channel trafficking to the β-cell membrane. Although NMDAR inhibition allows greater glucose-stimulated Ca²⁺ entry into β-cells stimulating GSIS, elevated NMDAR activation also promotes Ca²⁺ permeability through these ionotropic receptors. Thus, continual NMDAR Ca²⁺ permeability in T2D may perturb Ca²⁺ handling, which is predicted to reduce both insulin secretion as well as negatively impact β-cell health. This is supported by experiments on rodent β-cells that show reduced cell viability and decreased GSIS with chronic NMDAR activation. Treating diabetic rodents with an NMDAR antagonist (memantine) improved β-cell function and glucose homeostasis[198]. Together, these studies show that human β-cell NMDAR activity plays an important role in Ca²⁺ handling and suggest NMDARs as a therapeutic target for T2D.

Conclusions and Future Perspectives

Recent advances in human genetics, transcriptome and proteome analyses of normal and T2D islet cells reaching single cell resolution, and mechanistic analysis of ion channel regulations in cell lines, isolated islets and animal models have markedly improved our understanding of the role of ion channels in islet biology and how ion channel dysregulation contributes to dysfunction of the islets, in particular the ability of β-cells to secrete sufficient insulin in T2D. Despite the progress, many questions remain and translating the knowledge gained into clinical practice for the prevention and treatment T2D is still challenging.

From the genetic study perspective, while GWAS data offer potential T2D risk candidates they often need to be followed up by careful mechanistic studies to demonstrate causality. In this regard, studying human extreme phenotypes such as monogenic diabetes including neonatal diabetes and maturity onset diabetes of the young (MODY) have proven fruitful in identifying ion channels that are critical for insulin secretion. However, even when there is evidence for causality, genetic variations of a specific ion channel gene alone are unlikely to serve as a strong predictor of disease propensity since genetic background, epigenetic and environmental factors could impact how genetic variations manifest. As an example, genetic association studies in different ethnic populations often uncover different genetic risk factors. Most genetic studies thus far have focused on linkage with increased risk of T2D. It may be informative to also analyze genetic traits pertaining to ion channels of the islets that are protective against T2D. This approach has been successfully employed to identify genetic variants of proteins important for cardiovascular function, specifically in controlling cholesterol levels.

Mechanistic studies probing the functional role of ion channels have often been conducted in cell lines and transgenic mouse models. Results from these studies may or may not translate to humans as there is clear evidence that β-cell ion channel compositions are not identical in rodents and humans. Further, findings from mouse models can be clouded by the genetic background and strategies of genetic manipulations used. There is also the added caution of metabolic differences between rodents and humans. Efforts to study human islets are also met with potential variations from donors of different gender, age, ethnicity, and in the case of T2D, stage of the disease and other metabolic indicators such as obesity.

A central theme highlighted in this review is the importance of Ca²⁺ homeostasis in β-cell function and health. Intracellular Ca²⁺ is handled by not only ion channels present in the plasma membrane but also intracellular membranes including the ER, the mitochondria, and lysosomes. The coordinated actions of ion channels involved in Ca²⁺ handling and their regulation by signaling molecules afford β-cells the ability to adapt to increased insulin demands such as during pregnancy or mild insulin resistance. However, while enhanced Ca²⁺ signals are expected to augment insulin secretion and improve glucose tolerance, prolonged and unchecked Ca²⁺ signals can harm β-cells, resulting in β-cell excitotoxicity that reduces insulin secretion and causes β-cell demise. Therefore, strategies which enhance insulin secretion without causing permanent cell damage are needed. Better understanding of Ca²⁺ handling with increased temporal and spatial resolutions under physiological and pathological conditions will be important for such efforts.

Finally, pharmacological agents targeting β-cell ion channels are an essential part of the therapeutic development efforts for T2D. Currently available drugs or toxins have largely been discovered by serendipity or screening compound libraries. Rational drug design based on ion channel mechanisms has been limited by the difficulty in obtaining high resolution structures of these membrane proteins. Recent advances in single-particle cryo-electron microscopy techniques have made it possible to obtain high resolution structures of many ion channels at an astonishing pace. This presents an exciting opportunity for designing new and improved drugs that target β-cell ion channels for the treatment of T2D and other insulin secretion diseases.

• Ca²⁺ is an essential signal for pancreatic β-cell function including insulin secretion, transcription, metabolism, and the stress response.

• β-cell Ca²⁺ signals are governed by orchestrated activities of ion channels.

• In type 2 diabetes, disruption of ion channel function and regulation leads to β-cell excitotoxicity and Ca²⁺ mishandling to impair glucose-induced insulin secretion response.

• Understanding the ionic mechanisms that regulate insulin secretion will help identify therapeutic modalities for diabetes that target the excitation-secretion pathway.