Regulation of the T‐type Ca²⁺ channel Cav3.2 by hydrogen sulfide: emerging controversies concerning the role of H₂S in nociception

Abstract

Ion channels represent a large and growing family of target proteins regulated by gasotransmitters such as nitric oxide, carbon monoxide and, as described more recently, hydrogen sulfide. Indeed, many of the biological actions of these gases can be accounted for by their ability to modulate ion channel activity. Here, we report recent evidence that H₂S is a modulator of low voltage‐activated T‐type Ca²⁺ channels, and discriminates between the different subtypes of T‐type Ca²⁺ channel in that it selectively modulates Cav3.2, whilst Cav3.1 and Cav3.3 are unaffected. At high concentrations, H₂S augments Cav3.2 currents, an observation which has led to the suggestion that H₂S exerts its pro‐nociceptive effects via this channel, since Cav3.2 plays a central role in sensory nerve excitability. However, at more physiological concentrations, H₂S is seen to inhibit Cav3.2. This inhibitory action requires the presence of the redox‐sensitive, extracellular region of the channel which is responsible for tonic metal ion binding and which particularly distinguishes this channel isoform from Cav3.1 and 3.3. Further studies indicate that H₂S may act in a novel manner to alter channel activity by potentiating the zinc sensitivity/affinity of this binding site. This review discusses the different reports of H₂S modulation of T‐type Ca²⁺ channels, and how such varying effects may impact on nociception given the role of this channel in sensory activity. This subject remains controversial, and future studies are required before the impact of T‐type Ca²⁺ channel modulation by H₂S might be exploited as a novel approach to pain management.

Abbreviations

CBS

cystathionine β synthase

CSE

cystathionine γ lyase

DRG neuron

dorsal root ganglion neuron

HEK293 cells

human embryonic kidney cells

K_ATP

ATP‐sensitive K⁺ channel

3‐MST

3‐mercaptopyruvate sulfurtransferase

TPEN

N,N,N′,N′‐tetrakis(2‐pyridylmethyl)ethane‐1,2‐diamine

Introduction

Ion channel activity is central to a vast and diverse array of cellular functions in both excitable and non‐excitable cell types. For example, ion channel activity controls gene expression, apoptosis, proliferation, contractility, fluid transport, exocytosis, excitability and action potential propagation. It is therefore unsurprising that the number of genes encoding ion channel proteins (either the pore‐forming proteins or auxiliary subunits) runs into the hundreds. This diversity is further increased by splice variation and also the fact that subunits of different channel types can form functional channels by combining with other (albeit related) subunits to form heteromeric complexes. Of the many functions regulated by ion channels one of the most important is the control of [Ca²⁺]_i as this ion is a ubiquitous intracellular signalling molecule that mediates many of the ways in which channels influence the above‐mentioned fundamental cellular functions (Berridge et al. 2003; Clapham, 2007). Here, we discuss recently discovered new modes of regulation of one specific class of Ca²⁺ channel, the voltage‐gated T‐type Ca²⁺ channel.

Tailoring ion channel activity to serve specific and often co‐ordinated roles requires not only the availability of multiple channel types, but also that they can be dynamically regulated. This can occur by a plethora of means, for example by coupling to specific G proteins and via numerous post‐translational modifications, including phosphorylation, ubiquitination, sumoylation, nitrosylation, sulfhydration and S‐acylation (Rajan et al. 2005; Gonzalez et al. 2009; Mustafa et al. 2011; Lipscombe et al. 2013; Shipston, 2014). Such modifications can regulate ion channel trafficking and membrane insertion, as well as ongoing activity. In recent years, it has become apparent that ion channels are also modulated by endogenously generated, biologically active gases, often termed gasotransmitters, such as nitric oxide, carbon monoxide and hydrogen sulfide (NO, CO and H₂S, respectively). The roles of these gases, especially NO, in diverse physiological and pathological settings have in many instances been long‐established, but the concept that they represent a new class of ion channel regulators is currently emerging (Wilkinson & Kemp, 2011; Peers et al. 2012, 2014). In this article, we focus on H₂S specifically as a modulator of voltage‐gated T‐type Ca²⁺ channels.

T‐type Ca²⁺ channels

When first described at the single channel level, this class of voltage‐gated Ca²⁺ channel was named ‘T‐type’ since members had a relatively tiny conductance and gave rise to transient currents (Nowycky et al. 1985). These channels are also distinguished from other voltage‐gated Ca²⁺ channels by their rapid activation, slow deactivation and very negative activation thresholds, as low as −60 mV or even below (Carbone & Lux, 1984; Perez‐Reyes, 2003; Iftinca & Zamponi, 2009). Example currents are shown in Fig. 1 A, and a typical current–voltage relationship is presented in Fig. 1 B. Owing to this low threshold for activation, they also may display a significant window current (i.e. display tonic activity; Fig. 1 C, arrowed) at potentials close to the resting membrane potential (RMP), and so can provide a sustained route for Ca²⁺ entry into resting cells (exemplified in Fig. 1 D) and also contribute to setting the RMP (Perez‐Reyes, 2003).Three genes, CACNA1G, CACNA1H and CACNA1I give rise to pore‐forming α‐subunits of T‐type Ca²⁺ channels, which are nowadays referred to as Cav3.1–3.3, respectively (Catterall et al. 2005), with Cav3.3 showing relatively slower activation and inactivation kinetics than Cav3.1 or Cav3.2 (Fig. 1 A). Heterologous expression of individual Cav3 α‐subunits gives rise to currents which closely resemble native T‐type Ca²⁺ currents, suggesting that auxiliary subunits are not required for assembly of functional Ca²⁺ channels, although other channel properties and trafficking may indeed be affected by the auxiliary subunits (Dolphin et al. 1999).

Figure 1. Properties of T‐type Ca²⁺ channels .

A, example currents evoked by step depolarizations applied (according to the protocol shown above each trace) to HEK293 cells expressing each of the 3 classes of T‐type Ca²⁺ channel, as indicated. Note the relatively slow activation and inactivation of Cav3.3 as compared with Cav3.1 and Cav3.2. B, schematic current–voltage relationship typical of T‐type Ca²⁺ channels; note the low threshold for activation. C, superimposition of steady‐state inactivation plot (continuous line) and activation profile (dashed line) typical of T‐type Ca²⁺ channels. Note the pronounced window current (region of overlap, arrowed). D, fluorimetric recording of [Ca²⁺]_i in a Fura‐2‐loaded HEK293 cell expressing Cav3.2. For the period indicated by the shaded area, the T‐type channel blocker mibefradil (1 μm) was applied, causing a reduction in basal [Ca²⁺]_i (N.T. Hettiarachchi and C Peers, unpublished observations). E, schematic diagram of membrane depolarizations evoked by current injections in thalamic neurones. Note the evoked burst activity is almost fully suppressed in cells lacking the T‐type channel Cav3.1 (see Kim et al. 2001).

T‐type Ca²⁺ channels are widely distributed, but their physiological functions have sometimes proved difficult to resolve. This is certainly true in vascular smooth muscle cells (VSMCs), where Cav3.1 and Cav3.2 are expressed. Numerous attempts have failed to record ‘classic’ T‐type Ca²⁺ channel activity in VSMCs (reviewed in Kuo et al. 2011, 2014). Recent studies suggest this is because VSMCs primarily express T‐type Ca²⁺ channel splice variants which do not activate at the low voltages normally associated with T‐type Ca²⁺ channels. Instead, they activate over more depolarized voltage ranges as are observed for L‐type Ca²⁺ channels. Splice variation accounting for these differences occurs in Cav3.1 (and similarly in Cav3.2) primarily around exons 25 and 26, corresponding to the intracellular linker region between domains III and IV (Chemin et al. 2001; Latour et al. 2004; Emerick et al. 2006). Four variants have been identified termed 25a, 25b, 25bc and 25ac, with variant 25bc predominating in both the juvenile and mature systemic vasculature. These variants all contribute to vasomotor tone (Kuo et al. 2011, 2014), and may also have a pacemaker role, controlling slow Ca²⁺ waves and contractile oscillations (Cribbs, 2006).

Despite continued uncertainty around the physiological roles of T‐type Ca²⁺ channels in the vasculature, it appears that in several vascular diseases the influence of T‐type Ca²⁺ channels, particularly Cav3.1, becomes more prominent. Their increased expression and activity appears to be instrumental in pathological vascular remodelling in both the systemic and pulmonary circulation, providing a route for Ca²⁺ entry which is required for proliferation (Tzeng et al. 2012; Chevalier et al. 2014). Interestingly, numerous cancers also rely on the expression of T‐type Ca²⁺ channels for proliferation, and hence tumour growth (Dziegielewska et al. 2014). These findings suggest that they represent a promising therapeutic target in the treatment of various cancers and cardiovascular diseases.

In the nervous system, T‐type Ca²⁺ channels serve better defined roles. Thus, for example, in thalamic and corticothalamic neurones T‐type Ca²⁺ channels are responsible for pacemaker activity and low threshold spikes (Huguenard & Prince, 1994), as illustrated in Fig. 1 E. They contribute to ‘rebound’ bursts of spikes following a hyperpolarizing postsynaptic potential. Indeed, evidence suggests that all three isoforms of T‐type Ca²⁺ channels can contribute to this excitability (Kim et al. 2001; Joksovic et al. 2006; Lee et al. 2014). In the peripheral nervous system, T‐type Ca²⁺ channels are prominent in somatosensory fibres including small, capsaicin‐sensitive (presumed nociceptive) dorsal root ganglion (DRG) neurons (Jevtovic‐Todorovic & Todorovic, 2006; Nelson et al. 2006; Rose et al. 2013) as well as in two distinct populations of low‐threshold mechanoreceptors (LTMRs), Aδ‐ and C‐LTMRs, innervating skin hair follicles (Francois et al. 2015). Cav3.2 is the dominant form in DRG and may even be the exclusive form in some mechanosensitive neurons (Shin et al. 2003). Cav3.2 channels can control burst firing in DRG neurons (White et al. 1989) and so strongly influence excitability (Nelson et al. 2005, 2007 a), implying they are of central importance to nociception since stimulus intensity correlates with burst frequency. Indeed, the role of T‐type Ca²⁺ channels in pain is well recognized and has been covered in depth in several recent reviews (Bourinet et al. 2014; Todorovic & Jevtovic‐Todorovic, 2014; Francois et al. 2014; Waxman & Zamponi, 2014; Zamponi et al. 2015). Thus, conditional genetic deletion (Francois et al. 2015), or downregulation of Cav3.2 in DRG using intrathecal injection of antisense oligonucleotides (Bourinet et al. 2005; Messinger et al. 2009) produced strong anti‐nociceptive effects in rodent pain models of neuropathic and inflammatory pain. Conversely, T‐type Ca²⁺ currents are often increased in pathological conditions associated with chronic pain, such as diabetic neuropathy (Jagodic et al. 2007; Messinger et al. 2009), peripheral nerve injury or inflammation (Jagodic et al. 2008; Marger et al. 2011; Garcia‐Caballero et al. 2014). Both the enhancement of channel trafficking (Orestes et al. 2013; Weiss et al. 2013) and enhanced deubiquitination (Garcia‐Caballero et al. 2014) were reported as underlying mechanisms for the latter phenomenon.

It is not entirely clear how exactly T‐type Ca²⁺ channels participate in the nociceptive transmission. Cav3.2 channel expression has been detected in different compartments of afferent fibres, including peripheral nociceptive terminals and axons of skin afferents (Rose et al. 2013; Francois et al. 2015), nodes of Ranvier of Aδ fibres (Francois et al. 2015), as well as in the presynaptic terminals of nociceptive fibres in the spinal cord (Jacus et al. 2012; however, cf. Francois et al. 2015). Therefore, multiple mechanisms (or their combinations) are conceivable, including setting the threshold for action potential generation (at the nerve terminals) and propagation (at the nodes of thinly myelinated fibres), supporting burst firing, or indeed promoting synaptic activity at the central terminals of afferent fibres. Finally, since T‐type Ca²⁺ channels are expressed in skin terminals of low threshold mechanoreceptors, including D‐hair cells (Coste et al. 2007; Francois et al. 2015), a more direct role of T‐type Ca²⁺ channels in mechanotransduction also cannot be excluded.

In accordance with strong evidence for the physiological role of T‐type Ca²⁺ channels in pain, recent data clearly demonstrated that pharmacological inhibition of T‐type Ca²⁺ channel activity produces strong anti‐nociceptive effects in various rodent pain models (Todorovic et al. 2001, 2002, 2004; Latham et al. 2009). Together with N‐type Ca²⁺ channels, T‐type Ca²⁺ channels are clinically validated drug targets for pain (Bourinet et al. 2014). Accordingly, intense search is currently underway for novel pharmacological tools targeting T‐type Ca²⁺ channel activity. These need to be more selective than widely used blockers such as mibefradil which, despite its ability to inhibit T‐type Ca²⁺ channels, is not highly selective (Bezprozvanny & Tsien, 1995). This new generation of more selective T‐type inhibitors (e.g. TTA‐A2, TTA‐P2, KYS‐05090S) shows promising analgesic properties in animal models of pain (Choe et al. 2011; Francois et al. 2013). Another novel, selective and orally bioavailable T‐type Ca²⁺ channel blocker, Z944, which showed analgesic activity in animal models of inflammatory and neuropathic pain, is currently being tested in clinical trials as a first‐in‐class novel oral analgesic (Lee, 2014). Moreover, Zamponi's lab has recently developed small molecule modulators that prevent deubiquitination of the Cav3.2 channels by the deubiquitinase USP5, which in turn can prevent channel upregulation in chronic pain conditions (Gadotti et al. 2015).

A redox‐ and Zn²⁺‐sensitive ‘module’ within the Cav3.2 subunit

Studies in both native DRG neurons and recombinant expression systems have revealed a regulatory site within Cav3.2 which confers high sensitivity to redox agents, distinguishing it functionally from both Cav3.1 and Cav3.3. Thus, for example, T‐type Ca²⁺ currents recorded in native nociceptive neurones and recombinant Cav3.2 currents are enhanced by reducing agents (dithiothreitol (DTT) or l‐cysteine) and inhibited by the oxidizing agent 5,5′‐dithio‐bis(2‐nitrobenzoic acid) (DTNB) (Nelson et al. 2007 a). Sensitivity to these agents was exploited in order to demonstrate the importance of Cav3.2 in nociception: hindpaw injections of DTT or l‐cysteine induces thermal and mechanical hyperalgesia (Todorovic et al. 2001), and such effects are prevented with the Ca²⁺ channel inhibitor mibefradil; furthermore, analgesic effects are observed with DTNB (Todorovic et al. 2001). Such findings were confirmed in a subpopulation of nociceptive neurons expressing high levels of Cav3 channels (Nelson et al. 2005). An important breakthrough in the field was achieved by the joint work of Perez‐Reyes’ and Lee's groups which allowed the characterization of specific Cav3.2 residues involved in the metal‐induced inhibition of the Ca²⁺ channel. The redox‐sensitive module is located extracellularly, involving interaction of the extracellular IS1–IS2 linker region with the IS3–IS4 linker region (Kang et al. 2010), the latter region containing a histidine residue (H191) which is key to conferring high redox sensitivity to this channel. Intriguingly, the same H191 residue that is involved in modulation of T‐type Ca²⁺ channel activity by redox mechanisms also mediates T‐type Ca²⁺ channel inhibition by low (submicromolar) concentrations of extracellular Zn²⁺ (Kang et al. 2006, 2010). The full Zn²⁺ binding site also includes residues that precede H191, namely D189 and G190, as well as the negatively charged residues at the outer portion of the IS2 segment (Kang et al. 2010). It is, as yet, unclear how exactly the redox‐ and Zn²⁺‐mediated modulatory pathways converge at the same site but one hypothesis suggests that H191 may represent a general binding site for transition metals (e.g. iron, copper, zinc) and can be subject to oxidation via a metal‐catalysed oxidation (MCO) reaction (Stadtman, 2001; Nelson et al. 2007 a). It is conceivable therefore that either binding of metal (e.g. Zn²⁺) or Zn²⁺‐independent MCO of H191 may result in a similar inhibition of channel activity. An alternative hypothesis is that oxidative modification (possibly but not necessarily at H191) can increase the effect of metal binding either by increasing the affinity at the binding site or by enhancing coupling efficiency between the metal binding and channel inhibition. In such a scenario, oxidative modification could result in sensitization of the channel to Zn²⁺ (and potentially other transition metals) making the channel sensitive to ambient concentrations in biological fluids. Total Zn²⁺ concentration in human plasma is reported to be within the range of 5–20 μm (Moran et al. 2012). Although it is likely that the free [Zn²⁺] is considerably lower, it is still quite plausible that the extracellular free [Zn²⁺] can reach the nanomolar range at which changes in Zn²⁺ affinity/sensitivity of Cav3.2 subunits may result in noticeable changes in channel activity. While the exact mechanism of regulation of Cav3.2 activity via the H191‐containing regulatory site remains to be elucidated, it is clear that redox modulation is involved in the action on the T‐type Ca²⁺ channels of some pharmacological agents and physiological signalling cascades including nitrous oxide (Orestes et al. 2011), CO and thioredoxin (Boycott et al. 2013), as well as GABA_B receptors (Huang et al. 2015).

H₂S as a gasotransmitter

H₂S is regarded as the third ‘gasotransmitter’ since it is an endogenous, enzymatically generated, biologically active gas. Emerging evidence indicates that it has widespread physiological and pathophysiological importance throughout the body (Li et al. 2011). Unsurprisingly, therefore, it has received increasing interest as a potential therapeutic target over the past decade or so, as was the case previously for both NO and CO (Moore et al. 2003; Leffler et al. 2006; Szabo, 2007). H₂S is generated primarily by the action of two widely distributed enzymes, cystathionine γ lyase (CSE) and cystathionine β synthetase (CBS; Fig. 2). Both synthesize the gas from l‐cysteine. More recently, the mitochondrially located 3‐mercaptopyruvate sulfurtransferase (3‐MST) has also been found to generate H₂S in both the brain and vasculature (Fig. 2). 3‐MST generates H₂S from 3‐mercaptopyruvate which is itself generated by cysteine aminotransferase (CAT; Kimura, 2010). Red blood cells can generate H₂S non‐enzymatically from inorganic polysulfides, a finding which has prompted the idea that H₂S may mediate the beneficial vascular effects of dietary garlic (Benavides et al. 2007). H₂S is also known to be liberated in a redox‐ or pH‐sensitive manner from sulfur ‘stores’, i.e. sulfur bound to proteins in mitochondria or the cytosol (Kimura, 2010). Differential distribution of H₂S‐generating enzymes results in a predomination of CBS in the central nervous system, along with 3‐MST (Leffler et al. 2006), whereas CSE is dominant peripherally, generating the majority of H₂S elsewhere in the body, including the vasculature. More recently, H₂S has been shown to be generated from d‐cysteine (Fig. 2), particularly in the cerebellum and kidney, via a novel pathway involving 3‐MST and d‐amino acid oxidase (Shibuya et al. 2013).

Figure 2. Hydrogen sulfide synthesis and effector ion channels .

Schematic diagram illustrating the synthesis of H₂S from homocysteine, l‐ and d‐cysteine via enzymes shown in boxes (CBS, cystathionine β synthase; CSE, cystathionine γ lyase; 3‐MST, 3‐mercaptopyruvate sulfurtransferase; CAT, cysteine aminotransferase; DAO, d‐amino acid oxidase). Some of the target ion channels modulated by H₂S are also indicated (NMDAR, N‐methyl‐ d‐aspartate receptor; K_ATP, ATP‐sensitive K⁺ channel; BK_Ca, large conductance K⁺ channel; VGCC, voltage‐gated Ca²⁺ channel; TRP, transient receptor potential channel).

H₂S as a regulator of ion channels

Ion channels (specifically NMDA receptors; Abe & Kimura, 1996) were amongst the first family of cellular proteins recognized as being molecular targets of H₂S. Since then, reports of ion channel regulation by H₂S have grown rapidly (Peers et al. 2012; Kuksis & Ferguson, 2015; Zhang et al. 2015), as indicated by Fig. 2. It is clear that modulation by H₂S is not confined to any particular class of ion channels (such as voltage‐ or ligand‐gated channels) or to channels which are selective for specific ions (e.g. K⁺, Na⁺, Ca²⁺ or Cl⁻) (Zhang et al. 2015). H₂S is also known to interfere with a number of intracellular signalling pathways (Wang, 2012), and it is via these pathways that H₂S, in some cases, modifies ion channel activity. In other cases, direct post‐translational modification accounts for its effects. The best studied of these is the vascular smooth muscle ATP‐sensitive K⁺ channel (K_ATP channel), which becomes more active when sulfhydrated (i.e. when –SH groups within cysteine residues are altered to –SSH groups). This direct process (also known as persulfidation; see Paul & Snyder, 2015) contributes to vasorelaxation and thereby protects against hypertension (Mustafa et al. 2011; Paul & Snyder, 2012).

The list of ion channels regulated by H₂S is long and continues to grow. However, some areas of contention have arisen along the way. Thus, for example, inhibition by H₂S of the high‐conductance, Ca²⁺‐sensitive K⁺ channel has been described in detail (Telezhkin et al. 2010), yet others report activation of this channel by H₂S (Sitdikova et al. 2010; Jackson‐Weaver et al. 2013). Similarly, inhibition of L‐type Ca²⁺ channels in cardiac myocytes (Sun et al. 2008; Avanzato et al. 2014) contrasts with the observed augmentation of L‐type Ca²⁺ channel activity reported in astrocytes (Nagai et al. 2004). Such discrepancies may have simple explanations, but these need to be identified if H₂S modulation of ion channels is to be exploited therapeutically. The remainder of this review is focused on the known pro‐algesic effects of H₂S and considers the recent, conflicting data concerning the role of T‐type Ca²⁺ channels in this process.

H₂S as a regulator of Cav3.2

Our group has recently reported an inhibitory effect of micromolar concentrations of H₂S on recombinant and native T‐type Ca²⁺ channels in sensory neurons (Elies et al. 2014). Among Cav3 subunits, this effect was selective to Cav3.2 while Cav3.1 and Cav3.3 were insensitive to H₂S. In agreement with the predominant expression of Cav3.2 subunits in DRG neurons, H₂S also strongly inhibited endogenous low‐voltage activated (LVA) currents in these neurons. Interestingly, the same extracellular histidine residue, H191, which is necessary for Cav3.2 modulation by Zn²⁺ and redox agents, was also found critical for the effect of H₂S. Thus, the H191Q mutation (where histidine has been substituted by glutamine) in Cav3.2 abolished channel sensitivity to H₂S and the analogous reciprocal mutation in Cav3.1 (Q172H) conferred sensitivity to H₂S on this subunit (Elies et al. 2014). It is not entirely clear exactly how redox/Zn²⁺ binding of Cav3.2 is implicated in this H₂S‐mediated effect but one theory is that H₂S may increase channel sensitivity to extracellular Zn²⁺. Indeed, chelation of ambient Zn²⁺ with TPEN (N,N,N',N'‐tetrakis(2‐pyridylmethyl)ethane‐1,2‐diamine) fully reversed the H₂S‐mediated T‐type Ca²⁺ channel inhibition. Moreover, pretreatment with TPEN abolished H₂S‐mediated inhibition when H₂S was applied in the continued presence of TPEN. However, subsequent washing out of TPEN in the presence of H₂S allowed the inhibition to commence (Elies et al. 2014). Thus, the action of H₂S on Cav3.2 seems to depend on ambient Zn²⁺, assumed to be present in extracellular solutions in trace amounts. Indeed, according to our atomic absorption spectroscopy measurements, total Zn²⁺ levels can reach micromolar levels in nominally zinc‐free laboratory solutions (D. Huang & N. Gamper, unpublished observations); likewise, micromolar concentrations of Zn²⁺ in plasma are also reported (Moran et al. 2012). While levels of free Zn²⁺ are likely to be much lower as compared to the total zinc, it is still likely that this metal is present in the extracellular milieu at levels sufficient to affect channel activity. We hypothesize that the action of H₂S results in the potentiation of channel inhibition by such ambient Zn²⁺, via a mechanism which may share commonalities with the redox modulation of Cav3.2 (see Abstract figure).

Although our data strongly suggest that H₂S inhibits Cav3.2 and the endogenous T‐type Ca²⁺ currents in DRG neurons, there is an alternative theory suggesting that H₂S augments T‐type Ca²⁺ currents, possibly by chelating Zn²⁺, and that this augmentation underlies the pro‐algesic actions of this gasotransmitter (Kawabata et al. 2007; Maeda et al. 2009; Takahashi et al. 2010; Matsunami et al. 2011). Thus, injection of the H₂S donor NaHS into the rat hindpaw produced hyperalgesia which was abolished by the oxidizing agent DTNB (5,5′‐dithiobis‐(2‐nitrobenzoic acid)) and by the pharmacological inhibition of T‐type Ca²⁺ channels (Kawabata et al. 2007). Inhibition of endogenous H₂S production also produced an anti‐algesic effect. This and other studies by the same group led them to conclude that H₂S could activate or ‘sensitize’ Cav3.2 channels in order to account for their pro‐algesic effects (Kawabata et al. 2007; Takahashi et al. 2010). Furthermore, H₂S‐induced colonic pain could be mimicked by chelation of Zn²⁺ (Matsunami et al. 2011). Yet, most of the evidence suggesting that the pro‐algesic effects of H₂S are mediated by T‐type Ca²⁺ channel augmentation is somewhat circumstantial and direct electrophysiological evidence is sparse and insubstantial. For example, pre‐incubation of NG108‐15 (mouse neuroblastoma and rat glioma hybridoma) cells with 1.5 mm NaHS resulted in ∼20% augmentation of endogenous T‐type Ca²⁺ current while at 0.5 mm the effect was not significant (Kawabata et al. 2007). Another study demonstrated that pre‐incubation of human embryonic kidney (HEK) 293 cells exogenously expressing Cav3.2 channels with the CSE inhibitor propargylglycine resulted in a reduction of the current, which was interpreted as tonic augmentation of recombinant Cav3.2 by endogenous H₂S (Sekiguchi et al. 2014). Interestingly, only in the presence of this inhibitor were currents augmented by NaHS. These experiments are consistent with potential augmentation of T‐type Ca²⁺ currents by H₂S but the following has to be taken into account: (i) only very high, millimolar concentrations of NaHS were efficacious; (ii) due to the experimental protocol used (pre‐incubation) these experiments did not directly assess the acute effect of H₂S on T‐type Ca²⁺ channel activity.

In our hands, increasing the NaHS concentration to 3 mm did augment recombinant Cav3.2 currents, while even higher concentrations were needed to significantly augment native T‐type Ca²⁺ channel currents in DRG neurons (Elies et al. 2014). These experiments can, to some degree, reconcile conflicting experimental evidence from different laboratories and suggest a dual effect of H₂S on T‐type Ca²⁺ channels: H191‐mediated inhibition at low (micromolar) concentrations and potentiation (due to an unknown mechanism) at high (several millimolar) concentrations. It is important to point out, however, that the presence of H₂S at millimolar concentrations in mammalian tissues is highly unlikely. Initial reports estimated plasma H₂S levels within the range of 20–100 μm (Li & Moore, 2008), but even these values are now regarded as overestimations (Li et al. 2011). Therefore, physiologically relevant concentrations of H₂S are unlikely to be sufficiently high to produce T‐type Ca²⁺ channel augmentation. The expected effect of endogenous H₂S is, therefore, inhibition of T‐type Ca²⁺ channel currents, at least in cells and tissues that express significant amounts of Cav3.2 (such as sensory neurons). Therefore, the hypothesis that physiological levels of H₂S can trigger pro‐algesic effects via T‐type Ca²⁺ channel activation appears questionable at present. Furthermore, in our view, it is also unlikely that H₂S can act as a Zn²⁺ chelator (as was suggested by Matsunami et al. 2011) as (i) the chemical properties of H₂S do not suggest metal‐chelating properties (e.g. the ability to form polydentate complexes); (ii) acute application of H₂S leads to T‐type Ca²⁺ channel current inhibition, whereas Zn²⁺ chelation with TPEN causes current augmentation (Nelson et al. 2007 b; Elies et al. 2014); (iii) application of TPEN completely reverses H₂S‐induced T‐type Ca²⁺ channel inhibition; (iv) pre‐application of TPEN renders H₂S unable to inhibit T‐type Ca²⁺ channels.

At present, whilst it is hard to envisage how the painful/hyperalgesic effects of H₂S (especially those of endogenously produced H₂S) can be mediated by its effect on T‐type Ca²⁺ channels, other possible targets may account for this action. Thus, H₂S was suggested to produce some of its pro‐algesic effects via transient receptor potential channel TRPA1 activation (Andersson et al. 2012; Hsu et al. 2013). Additionally, H₂S was suggested to inhibit voltage‐gated K⁺ channels in trigeminal sensory neurons (most likely Kv1.1 and Kv1.4), thus producing depolarizing and excitatory effects (Feng et al. 2013).

Conclusions and perspectives

Tonic H₂S production clearly impacts a variety of both physiological and pathological processes, and probably does so in part via direct and indirect modulation of ion channel activity. The numerous and diverse pathways that can be modified by H₂S suggest that interventional control of this gasotransmitter could be therapeutically beneficial (e.g. in the cardiovascular system and peripheral sensory neurones). However, conflicts have arisen and must be resolved. We believe that at least some issues could be rectified if the experimental conditions through which we study the effects of H₂S are standardized (e.g. preparation of stock solutions of H₂S donors, flow rates of solutions containing donors, exposure time and obligatory use of agar bridges for reference electrodes).

Twenty years after the discovery of H₂S as an endogenously bioactive molecule (Abe & Kimura, 1996), one of the main challenges within the field remains the measurement of absolute concentrations of H₂S, in real time, both intracellularly and in extracellular compartments. For instance, development of appropriate intracellular fluorescence probes will expand not only the knowledge regarding intracellular concentrations of H₂S in space and time, but it will also help to correlate experimental data obtained from exogenously administered H₂S with experimental data obtained from endogenously generated H₂S. Additionally, future research regarding the physiological effects of H₂S must bear in mind the crosstalk signalling with other gasotransmitters (NO and CO). Only then can we obtain more comprehensive and uniformly acceptable data on which to build in order to design modulators of H₂S production and signalling for translational benefit.

Additional information

Competing interests

None declared.

Funding

This work was supported by grants from the British Heart Foundation (to C.P., J.P.B. and J.L.S.), the Medical Research Council (to C.P. and N.G.), and the Hebei Medical University (to N.G.).

Biographies

Jacobo Elies studied BSc Pharmacy and completed his PhD at the Department of Pharmacology of the University of Santiago de Compostela, Spain. In 2009 he joined Professor Chris Peers as a postdoctoral researcher. Since then he has been investigating the regulation of ion channels by gasotransmitters (CO and H₂S) in cardiovascular disease.

Nikita Gamper obtained his PhD in Physiology at the Sechenov Institute, St Petersburg, Russia. After postdoctoral work in Tübingen (Germany) and the University of Texas at San Antonio (USA) he joined the University of Leeds where he studies molecular and cellular mechanisms of nociception. His group investigates regulation of ion channels and G protein coupled receptors that control or influence excitability of peripheral ‘pain’ neurons.

Chris Peers obtained his BSc in Physiology and his PhD in Pharmacology at the University of London. He entered the world of oxygen sensing as a postdoctoral researcher in Piers Nye’s laboratory in Oxford. He then moved to the University of Leeds. His interests in the effects of hypoxia have extended to incorporate effects of gasotransmitters on ion channels and how this impacts on the cardiovascular and neurodegenerative diseases.