Toxins that Modulate Ionic Channels as Tools for Exploring Insulin Secretion

Abstract

Glucose-induced insulin secretion is a cardinal process in glucose homeostasis and metabolic expenditure. Uncoupling of the insulin response to glucose variations may lead to type-2 diabetes mellitus. Thus the identification of more specific drugs to facilitate the study of insulin secretion mechanisms and to develop new pharmacological agents for therapeutics is fundamental. Venomous organisms possess a great diversity of toxic molecules and some of them are neurotoxins that affect membrane excitability. This article reviews properties of those toxins affecting ion channels pivotal for insulin secretion and the usefulness of such compounds in the study of pancreatic beta-cell physiology. Here we examine the major contributions of toxinology to the understanding of the ionic phase of insulin secretion, to the determination of ion channel composition in different insulin secreting cell-line models as well as from primary cultures of different mammal species. Finally, we present a summary of the many diverse toxins affecting insulin release and a brief discussion of the potential of novel toxins in therapeutics.

Introduction

Stimulus-secretion coupling of glucose-induced insulin release in pancreatic beta cells depends on a rise of intracellular Ca²⁺ after augmented glucose transport through the plasma membrane and further catabolism of this molecule, which in turn increases the ATP/ADP ratio and promotes ATP-sensitive K⁺ channel (K_ATP) closure. The latter causes a depolarization mainly dependent on the contribution of TRP channels, whose cationic nonselective currents causes a slow depolarization followed by a fast depolarization phase mediated by voltage-gated Na⁺ and T-type Ca²⁺ channels to a plateau stage determined by high voltage activation Ca²⁺ channels (mainly L-type), and superimposed bursts of action potential spikes. Finally, membrane potential decays by the activation of voltage- and calcium-gated K⁺ channels (Fig. 1). The complex balance among these ion channels determines an oscillating pattern in membrane potential during stimulation with high glucose (Hiriart and Aguilar-Bryan 2008).

Fig. 1.

Oscillatory activity of β-cell membrane potential at stimulating glucose levels. The figure shows electrical response of β cells to high glucose and those ion channels involved in each phase of this response are marked. ATP-sensitive K⁺ channel (K_ATP) closure and transient receptor potential (TRP) channel opening determine slow depolarization preceding fast depolarization induced by currents through voltage-gated Na⁺ (Na_v) and low voltage activated (LVA) Ca²⁺ channels. During the plateau and spike bursting, high voltage activated (HVA) Ca²⁺ channels (Ca_v), mainly L-type, contribute to the rise of intracellular Ca²⁺ and insulin release by exocytosis. Finally, repolarization is achieved by delayed rectifiers (K_v) and Ca²⁺-sensitive big conductance K⁺ channels (BK)

Insulin secretion is tightly regulated by a set of endogenous secretagogues, counteracting hormones and growth factors such as acetylcholine, glucagon, and the autocrine nerve growth factor (NGF; Rosenbaum et al. 1998, 2001). However, there are many exogenous substances that also affect beta-cell physiology, like toxins, a group of bioactive compounds from venomous living beings that usually modulate ion channel activity and thus excitability, including the mechanisms of insulin secretion.

The purpose of this article is to present a review about some of the reported toxins with actions upon insulin secretion and to discuss their potential use in basic and applied research directed towards understanding beta-cell physiology.

Toxins Affecting Insulin Secretion Through Voltage-Gated Sodium Channels

Tetrodotoxin (TTX) is a voltage-gated Na⁺ channel blocking alkaloid, firstly isolated from fish of the Tetraodontidae family and from which its name is derived. Although it is well known that rodent, canine, and human beta cells express sodium channel α₁-subunits (Philipson et al. 1993)

Hiriart and Matteson (1988) first demonstrated the presence of a TTX-sensitive Na⁺ current in rat beta cells. Using the reverse hemolytic plaque assay (RHPA), they showed a significant inhibition by TTX (200 nM) on average glucose-induced insulin secretion in individual cells, without an effect at 5 mM glucose or below. The role of Na⁺ channels was then related to plasma membrane depolarization and activation of voltage-gated Ca²⁺ channels and the question of whether this current was essential or not for insulin secretion was solved.

TTX has been used to establish the relevance of Na⁺ channels in pancreatic beta cells from other species aside from rat. In human tissue, TTX (0.1 μg/ml) abolished an inward transient current elicited at depolarized potentials, diminished glucose-induced insulin secretion in more than 50% (Hiriart and Matteson 1988; Braun et al. 2008) and at a concentration of 1 μM reversibly abolished action potentials in canine beta cells (Pressel and Misler 1990). Plant (1988) reported for the first time a Na⁺ current in mouse beta cells. He also described that these channels are fully inactivated at the resting potential which accounts for insulin secretion being resistant to TTX in mice. However, Ernst et al. (2009) found a significant decrease in insulin secretion in response to 11 mM glucose in mouse intact islets when incubated with 1 μM TTX. Moreover, in vivo glucose intolerance and reduced response to glucose was observed in homozygous Scn1b null mice (unpaired β1 subunit activity) validating the relevance of Na⁺ channels and their regulatory subunits (especially Nav1.7 and β1) in mouse glucose homeostasis.

Crotamine, one of the major toxins from the venom of the Brazilian rattlesnake Crotalus durissus terrificus, is also a peptic modulator of Na⁺ channels and its suggested mechanism is to prevent the transition of the channels from closed to inactivated state; favoring the transition to the open state (Matavel et al. 1998). Toyama et al. (2000) used it as a valuable tool for studying the physiological role of Na⁺ channels in insulin secretion. Using a two-step purification process of the venom, combined with gel filtration and RP-HPLC, two toxin isoforms (F2 and F3) were obtained in the fraction classically described as a unique, 4.9 kDa, 42 amino acid residue-long basic polypeptide (Laure 1975). Out of these toxin isoforms, only F2 increased insulin secretion at an external glucose concentration of 16.7 mM.

The scorpion α-neurotoxins also modulate Na⁺ channel inactivation. They have been described as enhancers of beta-cell depolarization and glucose-induced insulin secretion in isolated islets. Particularly, a 5.6 μg/ml concentration of TsTx-V from Tityus serrulatus, induces a twofold increase of insulin secretion measured by radioimmunoassay, in the presence of 8.3 mM glucose (Gonçalves et al. 2003). This effect is evident because a 25% increase was registered in the active phase, during beta-cell oscillatory electrical activity due to toxin application. This effect is similar to that of veratridine (110 μM), an steroid alkaloid isolated from lilaceous plants which causes a persistent activation of voltage-gated Na⁺ channels by slowing or even abolishing inactivation, and further depolarizing beta-cell plasma membrane (Pace 1979). Gonçalves et al. have also demonstrated the involvement of Na⁺ channels in insulin secretion, and pointed out to the possible use of these channels as useful targets for pharmacological interventions aimed to treating insulin secretion disorders.

Toxins Affecting Insulin Secretion Through Voltage-Gated Calcium Channels

Calcium channels are essential for the correct function of the insulin secretion machinery since they trigger the exocytic pathway. L-type channels constitute the most important component in almost every beta-cell model, including cell lines from different species. However, calcium channel subtypes vary in abundance and importance depending on the cell type. Electrophysiological identification and isolation of these minor fractions is a hard task to achieve, because of their low expression levels in the plasma membrane and due to the fact that their kinetics and voltage dependence overlap with those of other types of currents. Nevertheless, specific modulators are able to bypass this handicap.

In rat insulinoma cells RINm5F, after incubation with and specific binding of radiolabelled ω-conotoxin (a toxin isolated from cone snails), a minor fraction (15–25%) of high voltage activated calcium channels was identified by electrophysiology. Those ω-conotoxin-sensitive channels corresponded to N-type calcium channels, a common class of presynaptic proteins in the nervous system and which had not been described in insulin-secreting cells (Aicardi et al. 1991). Moreover, this toxin (0.3 μM) decreased insulin secretion by one half when stimulated with d-glyceraldehyde and around a third when stimulated with alanine or KCl (Sher et al. 1992) rendering this specific N-type channel blocker as useful in the study of the composition and physiology of voltage-gated calcium channels in insulin-secreting cells.

In human beta cells, N-type channels seem to exert a negligible contribution to calcium influx as ω-conotoxin GVIA (0.1 μM), causes an average inhibition of only a 5% of the peak current and a 3% inhibition of total carried charge. Moreover, the toxin does not affect insulin secretion neither at basal nor at stimulating glucose concentrations (Braun et al. 2008). In the same paper, authors demonstrated that although L-type channels account for almost 50% of the calcium peak current, and thus for beta-cell electrical activity. P/Q-type channels contribute most to Ca²⁺ entry and exocytosis, as evidenced form the 80% reduction in insulin secretion due to the use of 0.2 μM ω-agatoxin (P/Q-type channel blocking peptide derived from spider venom) compared to a 31% reduction in insulin secretion by 10 μM isradipine, an L-type calcium channel blocker.

Toxins Affecting Insulin Secretion Through Voltage-Gated Potassium Channels and Ion Channels Determining Beta-Cell Membrane Resting Potential

Ionic currents through K⁺ channels are responsible for spike repolarization during the active phase in response to an external glucose rise. In rodents and humans, the voltage-dependent Kv2.1/2.2 and BK channels exert the most important contribution (Hiriart and Aguilar-Bryan 2008) and thus their inhibition is expected to cause an increment in insulin release.

Kv channels are blocked by 1 μM of Hanatoxin (HaTx), a peptide toxin isolated from the tarantula Grammostola spatulata—former Phrixotrichus spatulata—(Swartz and MacKinnon 1995), which induces Ca²⁺ oscillations in human pancreatic islets at 20 mM glucose, as predicted by a dynamic model of beta-cell excitability (Tamarina et al. 2005). Toxins like this could help understand the involvement of ion channels in insulin secretion and also to explore the biophysical properties of those molecules, such as their subunit composition in beta cells (Herrington 2007). In fact, HaTx was useful to identify the major component of the delayed rectifying current of human β-cells, as its application at a concentration of 100 nM decreased in a 65% the Kv peak current. Moreover, the application of scorpion Kv1 blockers such as margotoxin (from Centruroides margaritatus), kaliotoxin (Androctonus mauretanicus mauretanicus), and agitoxin (Leiurus quinquestriatus herbraeus), and also ShK from the caribbean cnidarian Stychodactyla helianthus (Castañeda et al. 1995), a potent Kv1.3 and Kv3.2 blocker, showed no effects on beta-cell Kv-channel activity, confirming previous evidences of no contribution of those channel subtypes to the delayed rectifying currents (Herrington et al. 2005).

From other spider venom (Plesiophrictus guangxiensis), another gating-modifier peptide named guangxitoxin was isolated and then synthesized to yield a product with the same kinetics as those of the native toxin. This toxin causes a rightward shift in the voltage dependence of Kv2.1 channel activation and shows high specificity for Kv2 channels. Like other gating-modifiers from tarantulas, it acts upon Kv4 too, but with an eightfold decreased potency, which makes it an appropriate tool for studying Kv2 channels in beta-cell physiology (Herrington et al. 2006). The isoform GxTx-1E increases the spike duration and also the firing frequency of intracellular calcium oscillation enhancing insulin secretion in a glucose-dependent manner, when applied in the micromolar range.

Recently, another three toxins from the Chinese tarantula Chilobrachys jingzhao have been described. These toxins shift the activation of Kv2.1 channels to more depolarized voltages than the above mentioned gating-modifiers can achieve, and thus will help to elucidate the physiological role of such channels in beta cells (Yuan et al. 2007). Many tarantulas produce Kv gating-modifier toxins that are analogous to BDS and APTx from marine cnidarians Anemonia sulcata and Anthopleura elegantissima, respectively. One of their effects is to induce an apparent speeding of channel closing (reviewed by Swartz 2007).

When specific Kv2.1/2.2 and A-type K current blockers like the spider venom stromatoxin, are applied to human pancreatic islets, no significant effects upon electrical activity or secretion can be observed. However, iberiotoxin (a scorpion toxin which blocks large-conductance Ca²⁺-activated K⁺ channels), promotes insulin secretion due to an increased spike amplitude, an effect which is unequivocally reproduced but just in the presence of tolbutamide that closes K_ATP channels (Braun et al. 2008). The answer to this apparent contradiction could result from the activation of K_ATP channels by iberiotoxin that would hyperpolarize the membrane, effect that is blocked by tolbutamide.

Other channels contributing to membrane potential in beta cells are some members from the family of transient receptor potential (TRP) channels, which elicit a nonspecific cationic current that causes an initial depolarization after K_ATP closure. Some pain-related TRP channels like the vanilloid-receptor-like 1 (TRPV1) are sensitive to mechanical, thermal, and chemical stimuli and capsaicin (reviewed by Salazar et al. 2009). This bioactive compound is contained in the plants of the genus Capsicum, or chili peppers.

TRPV1 presence has been described in rat beta cells, and increase insulin secretion has been shown when capsaicin in low concentrations (10⁻¹¹ to 10⁻⁹ M) is applied to RINm5F cells and to elevate plasma insulin levels when administered subcutaneously at 10 mg/kg in fasted rats (Akiba et al. 2004). TRPV1 channel is also activated by vanillotoxins, peptidic toxins isolated from the West Indies tarantula Psalmopoeus cambridgei and by the crude extract of venom from the tarantula Ornithoctonus huwena, which increases calcium influx through TRPV1 expressing HEK-293 cells (Siemens et al. 2006).

It is worth noting that the main scope in the study of plant and animal toxins is to determine how they affect the activity of this ion channel and the role of this protein in painful processes (Cromer and McIntyre 2008). Recently, Cuypers et al. (2006) demonstrated a capsaicin anti-desensitizing activity in crude extract of various cnidarians venoms that are associated with the typical persistent burning-pain sensation after cnidarians stings, considering the role of such channels in nociception. Interestingly, Andreev et al. (2008) have isolated from the sea anemone Heteractis crispa venom a novel polypeptide, named APHC1, which acts as a partial antagonist of TRPV1.

It is interesting to note a communication by Alevizos et al. (2007) where authors comment on the contradictory evidence with respect to the metabolic effects of capsaicin, since most of the literature agrees with a potentiated insulin response except for a report of attenuated postprandial hyperinsulinemia due to regular consumption of chili. The authors note the heterogeneous conditions used in the analyzed studies and argue about the possible use of capsaicin analogues as drugs for the treatment of metabolic disorders such as diabetes mellitus.

Toxins Affecting Transducers and Second Messenger Pathways

Beta-cell physiology is subject to endogenous modulation through guanine nucleotide-binding protein coupled receptors (GPCR) acting on paracrine and endocrine regulation of insulin secretion by a large number of islet, adrenal and enteric hormones, and also by neurotransmitters from both branches of the nervous autonomous system. Nevertheless, there are various exogenous compounds that also interact with such receptors.

Many of these insulinotropic toxins are reviewed elsewhere (Holz et al. 2000), but some of them are worth mentioning here since they are representative of quite different pathways of GPCR modulation. Pertussis toxin (PT), from the bacteria Bordella pertussis, is one of the most studied toxins because its mechanisms mainly involve the inactivation of inhibitory G proteins (G_i) from α₂-adrenergic receptors and thus it promotes an indirect activation of cAMP production. However, experiments carried on isolated rat pancreas after animals were pretreated with the toxin showed a remaining effect of adrenergic stimulation, indicating that adrenergic inhibition of insulin secretion in vivo presents a PT-insensitive pathway (Hillaire-Buys et al. 1992). The effect of PT has been studied to determine the functional status of G_i protein in pancreatic islets from obese Zucker rats and although hyperinsulinemia does not seem to be related with a total loss of function of such proteins, there was a decreased ability to respond to PT with respect to lean rats, suggesting subtle alterations in obese-cell physiology (Cawthorn and Chan 1991).

Catecholaminergic stimulation, another glucose-induced endogenous inhibitor of insulin secretion, involves the activation of G protein-gated inwardly rectifying K⁺ potassium channels (GIRK), whose four subtypes are expressed in mice β-cells, and which hyperpolarizing adrenaline-induced currents are reversibly inhibited by 100 nM of the honey bee (Apis mellifera) toxin tertiapin-Q (Iwanir and Reuveny 2008). However, there is a possibility of certain degree of inhibition of the big conductance Ca²⁺-dependent K⁺ channels, since the toxin also blocks them in a use, concentration and voltage-dependent fashion (Kanjhan et al. 2005).

The effects of xanthine derivatives upon Ca²⁺ oscillations induced by glucose mediated mobilization of intracellular Ca²⁺ has also been studied. Roe et al. (1993) found a differential action of 5 mM caffeine and theophylline on intracellular Ca²⁺, determining a transient release at 2 mM glucose after rapid exposures (less than 2 s) and a slow increment during sustained incubation with the compounds. Nevertheless, they did not report an effect of 10 μM ryanodine, a contradictory result to that reported by Takasawa et al. (1993), where the authors employed a tenfold concentration of ryanodine (100 μM), a concentration at which it induced Ca²⁺ release from islet microsomes. Although caffeine does not affect membrane potential at low glucose, it induces an augmentation of electrical spike activity, and thus an elevation of intracellular calcium, when glucose rises up to 12 mM (Worley et al. 1994).

Another plant derived compound, thapsigargin, depletes intracellular calcium stores by inhibition of microsomal Ca²⁺ ATPases and, when applied at 200 nM it increases insulin secretion by 2.5-fold in isolated rat beta cells, as evidenced by a RHPA at 15.6 and 5.6 mM glucose concentrations; by an increase of 18 and 34% in the percentage of insulin-secreting beta cells, respectively. Even at zero glucose, individual cells increased their secretion by 126%, indicating a glucose-independent mechanism for thapsigargin effect (Cruz-Cruz et al. 2005).

The enhanced insulin release results from a thapsigargin-activated nonselective cationic current, mainly of Na⁺, that induces a slow depolarization to a plateau level with superimposed action potentials, favoring Ca²⁺ entrance and exocitosis. This phenomenon resembles the calcium release-activated nonselective cation current (I_CRAN) elicited by direct stimulation at the picomolar range with the dinoflagellate (Gambierdiscus toxicus) derived maitotoxin (Worley et al. 1994; Roe et al. 1998). This toxin has been proposed to share the same target of stimulation of glucagon-like peptide-1 (GLP-1) and pituitary adenylyl cyclase-activating polypeptide (PACAP) through coupling of GTP-binding proteins, adenylyl cyclase activation and further elevation in cytoplasmic cAMP (Leech and Habener 1997).

Amongst the compounds that lead to increased intracellular Ca²⁺ and finally insulin granule exocytosis, there is α-latrotoxin, a protein isolated from the black widow spider Latrodectus tredecimguttatus, which has two peculiarities: to form presynaptically Ca²⁺ permeable pores by oligomerization and also to posses two exendin-4-like domains, characteristic of the Glucagon-like peptide-1 (GLP-1) related family of hormones that increase insulin secretion (Holz and Habener 1998). The insertion of the oligomers in biological membranes requires of some receptors like neurexin, latrophilin, and protein-tyrosine-phosphatase σ (Hlubek et al. 2000; Ushkaryov et al. 2008; Krasnoperov et al. 2002). These proteins may also increase exocytosis by different mechanisms. In beta cells and other insulin-secreting cell lines, the GPCR latrophilin mediates a Ca²⁺-independent potentiation of insulin release by α-latrotoxin, which could be reversed by epinephrine (Lang et al. 1998).

The effects of muscarinic agonists of acetylcholine receptors on single beta cells are dependent depend on the extracellular glucose concentration. At low glucose, carbachol increases insulin secretion; while at stimulating glucose concentrations (15.6 and 20.6 mM) an inhibitory effect is observed. These effects are reverted by atropine, an alkaloid toxin from solanaceuos plants that acts as an antagonist of muscarinic receptors (Hiriart and Ramirez-Medeles 1993). Natural compounds have proved to be very useful for basic research and also for pharmacological treatments involving these receptors, since the agonist muscarine is an alkaloid too, obtained from Amanita muscaria.

Final Remarks and Perspectives

Kv channels are an attractive target for the treatment of type-2 diabetes (Herrington et al. 2006), because their blockers increase the duration of glucose-induced action potentials and augment insulin release.

The ability to synthesize toxins provides us with a readily available source of blockers and minimizes the ecological impact of collecting them from its natural sources. Nevertheless, there are some major issues to be considered, as the expression efficiency of and recombinant peptide folding. Moreover, the stability of compounds after administration should be examined during their development as therapeutic candidates.

The finding of more specific modulators of beta-cell ionic channels is still a promising field because they can be applied in basic research and for its potential role in therapeutics. There is a large diversity of venomous organisms already known to modulate insulin secretion but there is a great number of unexplored poisonous species to search for novel toxins, including almost an entire phylum where molecules acting on early or late stages of the ionic phase of insulin secretion have not been identified (Fig. 2).

Fig. 2.

Diversity of natural modulators of ion channels involved in insulin secretion. The scheme summarizes the mechanism by which the described toxins modulate insulin secretion through actions on ion channels. Solid and dashed lines (green and red in the online version) represent a positive or negative effect on insulin release, respectively. Arrowheads denote activation of ion channels while oval endings depict inhibition or blockade. To illustrate the diversity of sources from where bioactive compounds are derived, their names are accompanied by a bracket enclosed letter according to the taxa of its venom producing organisms. Letter A (light red in the online version) symbolizes kingdom Animalia, P (green in the online version) Plantae, F (yellow in the online version) Fungi, R (orange in the online version) Protista, and M (blue in the online version) Monera

The use of peptidic toxins, for instance, has been highlighted in studies focusing on ion channel structure and function relationships taking into account their potency and selectivity (Dutertre and Lewis 2010). It is important to mention that selectivity in the action of many toxins could avoid side effects and provide additional advantages, like enhancing glucose-induced insulin release, without inducing pain in type-2 diabetic patients, when treated with specific Nav1.6 agonists. It would also be desirable to achieve reduced insulin secretion and pain relief in patients with insulinoma being treated with Nav1.7 blockers, as the latter are considered analgesics (Braun et al. 2008).