Voltage-Gated Proton Channels as Novel Drug Targets: From NADPH Oxidase Regulation to Sperm Biology

Unique features of H_V1

H_V1 belongs to a superfamily of voltage-gated ion channels, but it has a number of unique features that distinguishes it from other voltage-gated ion channels (Fig. 1). From a structural standpoint, the H_V1 channel lacks the prototypical pore domain of ion channels (114, 122). Voltage-gated ion channels usually consist of six transmembrane (TM) segments with the segments S1-S4 constituting the VSD that detects the changes in membrane potential and S5-S6 forming the ion conductance pore domain responsible for selective ion permeation. H_V1 channels only have the first four TM segments (S1-S4), which function both as VSD and as pore, the channel assembling as a dimer (72, 77, 139) with each subunit containing its own permeation pathway (72, 139). From a functional point of view, H_V1 channels are uniquely selective for protons with essentially no permeability to other ions, a feat that no other ion channel can achieve (30). This perfect selectivity is conferred by a single acidic residue Asp¹¹² whose neutralization remarkably converts the hH_V1 channel into an anion channel (102). This structure enables H_V1 channels to efficiently transport protons despite their usually very low concentration (100 nM at pH 7.0) in physiological solutions. H_V1 channel gating is also peculiar, with the channel lacking inactivation and sensing both the TM voltage and pH gradient. Activation, therefore, is favored by depolarizing voltages, by acidic intracellular pH, and by alkaline extracellular pH. Three charged Arg residues in the S4 TM segment confer the voltage dependency similar to the VSD of potassium channels, while the structural basis of the pH sensing is not known. H_V1 proton currents are also highly sensitive to temperature (33, 74, 75) and have a small (15 fF) unitary conductance (21), which can be accounted for by the low concentration of protons in physiological saline solutions. Finally, from a pharmacological perspective, H_V1 channels of most species are potently blocked by Zn²⁺ at concentrations ranging from 100 nM to 1 mM depending on the extracellular pH and on the presence of other polyvalent cations. In mammals, Zn²⁺ shifts the current–voltage relationship positively and slows the kinetics of H_V1 channel activation, two distinct effects that can be accounted for by the binding of polyvalent cation to two histidine residues (His¹⁴⁰ and His¹⁹³) located at the interface between the channel monomers (77, 101). A number of studies revealing the gating mechanism and structural features of H_V1 have been published recently (1, 11, 63, 73, 77, 97, 102, 112, 114, 138), and the known structural elements are shown in Figure1. A detailed discussion of H_V1 structural and functional aspects is beyond the scope of this review, and all the pertinent information can be found in a recent exhaustive review (31). Therefore, only structural aspects relevant for the development of H_V1 inhibitors will be described in the HV1 Pharmacology section.

OPEN IN VIEWER

FIG. 1.

Scheme of human H_V1. The human H_V1 protein consists of four transmembrane (TM) segments with the N-terminus and C-terminus being on the cytoplasmic side. Known structural elements and putative inhibitor binding sites are shown (11, 50, 73).

Mechanisms and regulation of H_V1 activation

The primary mechanism of H_V1 channel activation is voltage, the channel being activated by membrane depolarization, but the actual voltage required for activation is not fixed but set by the TM pH gradient (ΔpH, the pH difference between the two compartments separated by the conductive membrane). Thus, both the intracellular and extracellular pH (which is functionally equivalent to the intraluminal pH if the channel is located on an intracellular, vesicular membrane) are primary determinants of H_V1 channel activation. When the cytosol becomes more acidic relative to the extracellular or intraluminal side (as a result of either cytosolic acidification or extracellular/vesicular alkalinization), the entire conductance-voltage relationship of the channel shifts by 40 mV to more negative voltages for each unit increase in ΔpH. Conversely, when the extracellular or intraluminal side becomes more acidic than the cytosol (as a result of either cytosolic alkalinization or extracellular/vesicular acidification), the conductance-voltage relationship shifts by 40 mV to more positive voltages for each unitary change in ΔpH (20). In practical terms, this implies that three parameters should be known to determine whether an H_V1 will be activated in a given membrane: (i) the membrane voltage (measured on the cytosolic leaflet, relative to the extracellular or luminal side); (ii) the cytosolic pH, whose acidification favors activation at any given voltage; and (iii) the extracellular or intra-luminal pH, whose acidification opposes activation at any given voltage. In most organisms and tissues, the channel opens at voltages higher than +20 mV when the pH on both sides of the transporting membrane is identical (e.g., pHi/pHo 7.0/7.0). This implies that H_V1 essentially drives the efflux of protons from cells or the influx of protons inside membrane-enclosed compartments, provided that the voltage and pH are permissive for channel activation. These conditions are, in fact, rarely met in the plasma and organellar membranes of mammalian cells. Under physiological conditions, the cytosolic pH is only slightly more acidic than the extracellular pH, and ΔpH typically averages 0.3 pH units (acidic inside). A channel present at the plasma membrane will, therefore, only be activated at voltages higher than 8 mV (+20mV−(40mV/pH*0.3 pH)=8 mV), a value that is exceeded transiently in excitable cells during action potentials, in phagocytes during the respiratory burst (5, 47, 65, 113), and in eosinophils (6, 41). Since most intracellular organelles are acidic, the conditions are even less permissive for channel activation in endomembranes. For instance, a channel expressed in the membrane of early endosomes will face a luminal pH of 6.0 and a cytosolic pH of 7.0, that is, a ΔpH of −1 pH unit, which will shift the voltage threshold for activation to +60 mV (+20mV−(40mV/pH*(−1 pH))=+60 mV). Whether voltages higher than +60 mV can be reached in the membrane of organelles is not known, as direct recordings of the potential of these structures have not been performed, but at first glance, the pH conditions do not appear permissive for H_V1 channel activation. So far, the only evidence for H_V1 activity has been gathered either at the plasma membrane or at the membrane of phagosomes, phagocytic vacuoles forming around particles ingested by cells of the innate immune system such as neutrophils and macrophages. In these membranes, channel activation is facilitated by phosphorylation of the channel and by the functional coupling of the channel to nicotinamide adenine dinucleotide phosphate (NADPH)-dependent oxidase (NOX) enzymes, as will be discussed next.

NOX enzymes

NOX enzymes are a family of seven members (NOX1–NOX5 and DUOX1-2) that produce superoxide (O²⁻) or hydrogen peroxide (H₂O₂) using NADPH and molecular oxygen as substrates [for a review, see Bedard and Krause (9)]. These enzymes are widely expressed through multiple species and tissues and have varying functions depending on the localization, expression levels, and developmental stage. Reactive oxygen species (ROS) production by NOX enzymes has been implicated in host defense, redox-dependent signaling and also contributes to the progression of multiple diseases that are characterized by oxidative stress. NOX enzymes share a common structure: NOX1–5 have 6 and DUOX1-2 have seven TM domains. NOX1, NOX2, and NOX3 require cytosolic subunits for catalytic activity; NOX4 is constitutively active; and NOX5 and DUOX1-2 are Ca²⁺ dependent. NOX2 is probably the best characterized member of the family. It is expressed at high levels in neutrophils, and electron transfer by NOX2 mediates respiratory burst in neutrophils that are necessary for bacterial killing.

Relationship between H_V1 and NOX enzymes

During the reaction catalyzed by NOX enzymes, NADPH provides two electrons that are translocated across the plasma membrane to reduce molecular oxygen to O²⁻, generating two protons in the cytoplasm (Fig. 2). Based on the amount of O²⁻ generated by phagocytes, it has been calculated that NOX activity could lead to a drop in cytosolic pH by more than five units (i.e., to lethal levels around pH 2) (40). A variety of proton extrusion mechanisms prevent this lethal acidification, with H_V1 channels probably being particularly implicated in proton extrusion during NOX enzyme activation. The extrusion of protons through an electrogenic channel, rather than through an electroneutral exchanger, has a second advantage that charges are compensated and massive depolarization through isolated electron transport is avoided. The connection with O²⁻ producing NOX was made 5 years after the discovery of H⁺ channels in snail neurons (137), when Henderson et al. suggested that a voltage-gated proton channel contributes to the pH homeostasis in neutrophils, by counterbalancing acid production and depolarization through the phagocyte NOX (58). In 1993, voltage-gated H⁺ currents were reported in human neutrophils (32) and in the myeloid cell line HL-60, which expresses an NOX (40), and numerous subsequent studies reported that high levels of NOX expression coincide with high levels of H⁺ channel activity in phagocytes. A causal genetic link was soon refuted, however, as voltage-gated H⁺ currents could be recorded in monocytes (103) and granulocytes (37) from NOX2-deficient patients suffering from the chronic granulomatous disease (CGD). However, proton currents in NOX2-containing, but not in CGD, eosinophils were enhanced by phorbol-12-myristate-13-acetate (PMA) (5). A truncated form of NOX2 conferred increased H⁺ ion fluxes (as detected by fluorescent probes and by patch-clamp) to transfected cells, and mutations of critical histidines within this truncated NOX2 prevented the H⁺ fluxes (3, 4, 55, 86). These findings revived the interest in the theory that the phagocyte NOX2 might either function as a proton channel or be closely associated with it. Cloning of the HVCN1 gene coding for H_V1 proton channel in 2006 clarified the situation by showing that the H_V1 protein supports H⁺ fluxes, and that H_V1-deficient phagocytes are completely devoid of H⁺ currents, even in conditions favoring NOX activation (44, 90). Taken together, it is now clear that NOX enzymes are not H⁺ channels, but they play a modulatory role in H⁺ channel function. Along these lines, it has been shown that up-regulation of H_V1 and accumulation of H_V1 channels in phagosomes do not require functional NOX2 (110).

OPEN IN VIEWER

FIG. 2.

H_V1 function. Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase enzymes use NADPH as an electron donor and molecular oxygen as an electron acceptor to produce O²⁻. This activity leaves two H⁺ in the cytoplasm and depolarizes the membrane. H_V1 extrudes these protons to maintain physiological pH in the cytoplasm and to repolarize the membrane. H⁺ ions extruded by H_V1 are used in the reactions of generation of other reactive oxygen species (ROS) from O²⁻. NADPH oxidases, in particular NADPH-dependent oxidase 2 (NOX2), are a well-recognized source of cytosolic acidification. However, other sources, such as nonoxidative glucose break-down in cancer cells, might also be relevant.

Of note, the functional link between NOX expression and H⁺ channel activity is well established for NOX2, but from a theoretical point of view it could apply to all NOX isoforms. Indeed, such an association has been described for the Ca²⁺-activated NOX5 isoform in sperm cells [see section Spermatozoa, (99)]. However, many cell types that exhibit robust H⁺ channel activity do not show any evidence for NOX expression (e.g., snail neurons, basophil granulocytes). Thus, there are most likely NOX-independent H⁺ channel functions.

The HVCN1 gene: gene structure and splice variants

Human HVCN1 gene is located in chromosome 12 (12q24.11) and spans 41,127 bp (Fig. 3). It is neighboring with PPP1CC (protein phosphatase 1, catalytic subunit, and gamma isozyme) and with TCTN1 (tectonic family member 1). There is no indication of a thematic gene cluster in the area surrounding H_V1.

OPEN IN VIEWER

FIG. 3.

Genomic structure and regulation of human hydrogen voltage-gated channel 1 (HVCN1). (A) Putative HVCN1 promoter region. High probability transcription factor binding sites are shown (chromatin immunoprecipitation–sequencing [ChIP-Seq] data from ENCODE). (B) HVCN1 transcript variants. Exons are shown as boxes, and introns are shown as thin lines. White boxes represent UTR exons. Gray boxes represent translated exons. Three splice variants of HVCN1 are shown. The maps are based on National Center for Biotechnology Information (NCBI) Gene databases entries. (C) Predicted microribonucleic acid (miRNA) binding sites (TargetScan). (D) H_V1 protein. Two methionines corresponding to alternative ATG start codons are underlined. single-nucleotide polymorphism (SNPs) are shown in bold (see also Table 1).

Single-nucleotide polymorphisms

A database search (www.ncbi.nlm.nih.gov/snp) resulted in 608 total single-nucleotide polymorphisms (SNPs) for human HVCN1. An analysis of coding nonsynonymous validated SNPs restricted these findings to seven SNPs, of which only two have a frequency above 1% (Table 1 SNPs and Fig. 3D). One of these polymorphisms has been described in a study investigating the role of H_V1 in acid secretion by lung epithelia (63). The authors tested 95 human genomic DNA samples from Coriell genomic DNA SNP500V panel and found one HVCN1 allele that was heterozygous for M91T mutation. Airway epithelial cells naturally expressing this mutation had a higher threshold for pH activation than the wild-type cells (0.5 pH units more were required for Zn²⁺ sensitive acid secretion than for wild type). These data suggest that the M91T mutation would produce an H_V1 with reduced activity. However, no clinical data are available so far to prove this hypothesis. No mutation leading to H_V1 deficiency in human patients has been described so far.

Methods to measure H_V1 proton channel activity

There are essentially two ways to measure the function of voltage-gated proton channels: direct electrophysiological recordings of voltage-gated proton currents or indirect recordings of proton channel activity, obtained by measuring the changes in either the intracellular pH, the extracellular pH, or the pH of intracellular organelles. The two approaches have been extensively used, and each of them has their own advantages and disadvantages. Electrophysiological recordings offer a high degree of control and provide high-quality data but are technically demanding, while simpler pH recordings are amenable to high-throughput screens but lack specificity and rely on indirect means to control the voltage and pH-dependent channel opening. Patch-clamp recordings provide optimal control of the membrane voltage and yield a wealth of information on the channel biophysical properties and (when mutations are studied) on its ionic permeability profile. The approach also enables to impose a defined TM pH gradient by applying extracellular and intracellular solutions of known pH, using either the whole-cell or the inside-out configuration of the patch-clamp technique. Perfect control of the cytosolic pH is impossible to achieve in the whole-cell mode due to the large proton fluxes flowing across open H_V1 channels. However, due to its perfect proton selectivity, the H_V1 channel itself is a very accurate pH sensor and the reversal potential of the currents can be used to determine accurately the TM pH gradient (30). In contrast, pH recordings are performed in intact cells whose plasma membrane voltage and cytosolic pH varies dynamically. Regardless of whether channel activity is inferred from changes in extracellular pH (measured with electrodes) or cytosolic pH (measured with fluorescent dyes such as 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein [BCECF] or SNARF-1 or with genetically encoded pH indicators), indirect methods should be employed to impose the depolarizing voltage and/or an outwardly directed pH gradient required to activate H_V1 channels. Voltage control is usually achieved with ionophores such as gramicidin (a nonselective cation ionopore that will clamp the membrane voltage to 0 mV) or valinomycin (a potassium-selective ionophore which will clamp the membrane voltage to the K⁺ equilibrium potential). Control of the cytosolic pH is usually done by transiently exposing cells to ammonium (NH₄ ⁺), whose conjugated base NH₃ is membrane permeable and, thus, reaches the cytosol where it is protonated to NH₄ ⁺, thus acting as an effective proton sink. On external ammonium washout, the cytosolic NH₄ ⁺ that has accumulated is converted back to NH₃, which diffuses out of cells, with the released intracellular protons causing a rapid acidification of the cytosol to pH<6.0 depending on the concentration of NH₄ ⁺ and cell size (118). The outwardly directed pH gradient that is thus generated promotes the activation of H_V1 channels, provided the membrane potential is clamped at depolarizing voltages with ionophores. By following the recovery of the cytosolic pH, one can then obtain an indirect readout of the activity of plasma membrane voltage-gated proton channels, as long as the other plasma membrane acid extrusion systems are inhibited, either pharmacologically (e.g., with amiloride derivatives to inhibit sodium-proton exchangers) or by using specific ionic conditions (e.g., bicarbonate-free solutions to inhibit bicarbonate transporters). To ensure that the pH recovery is mediated by H_V1 channels, the kinetics of pH recovery is usually measured in the presence and absence of a known H_V1 channel inhibitor such as Zn²⁺. As can be appreciated, this approach relies heavily on pharmacological tools and on manipulations that are rather toxic for cells and does not provide an accurate voltage and pH control. However, despite their limitations, pH recordings have provided much insight into the physiological role of H_V1 channels and are now gaining renewed interest for high-throughput screens for H_V1 inhibitors. Inhibitor validation, however, will require electrophysiological recordings of either native proton currents or H_V1 channels exogenously expressed in mammalian transgene recipient cell lines (e.g., HEK293, chinese hamster ovary [CHO], etc.) or in other types of expression systems, such as Xenopus laevis oocytes.

Neutrophil and eosinophil granulocytes

Basophil granulocytes

Mast cells

Monocytes/macrophages

H⁺ currents with characteristics of H_V1 channels were recorded in mouse peritoneal macrophages (70); THP-1 monocyte cell line (34), and H_V1 protein and current were detected in mouse macrophages (107). The function of H_V1 in these cells is probably to sustain respiratory burst, but only a few experimental data are available (98).

Osteoclasts

Dendritic cells

B lymphocytes

T lymphocytes

Neurons

Microglia

Airway epithelia

Spermatozoa

Function

The cytoplasm of quiescent spermatozoa is acidic (pH<6.5), and its alkalinization is required for sperm motility, chemotaxis, and acrosome reaction. A seminal paper from Lishko et al. suggests that H_V1 activity is necessary for sperm activation: Functional assays showed that capacitation of spermatozoa occurred along with an increase in the amplitude of voltage-gated proton current. Interestingly, seminal fluid has an unusually high concentration of H_V1 blocker Zn²⁺ (2.2±1.1 mM), which might be a physiological mechanism of keeping the spermatozoa in the quiescent state. ROS production is also necessary for sperm maturation (27). One of the possible sources of ROS in spermatozoa is NOX5, a member of the NOX family that is highly expressed in human testis (4, 8, 99). The abundant expression of both H_V1 and NOX5 in human testis raises the question as to whether H_V1 activity in the testis is linked to electrogenic activity of an NOX. The relationship between H_V1 and NOX5 was addressed in a recent study (99). Using coimmunoprecipitation, they showed that NOX5 interacted with H_V1 in sperm cells, suggesting that these two molecules work as a functional unit. Furthermore, Zn²⁺ (1 mM) or RNAi-mediated inhibition of H_V1 in K562 leukemia cells overexpressing NOX5 resulted in a substantial decrease in O²⁻ production, supporting this hypothesis. Zn²⁺ treatment (1 mM) also significantly inhibited O²⁻ production in human spermatozoa stimulated by PMA, ionomycin, or H₂O₂. The authors also found a significant correlation between the percent of motile sperm and basal levels of O²⁻ production.

Another important factor in sperm activation is Ca²⁺ current mediated by pH-dependent flagellar Ca²⁺ channel, cation channels of sperm (CatSper) channel, which is inhibited by acidic intracellular pH (79), and can be activated by progesterone (129). Interestingly, spermatozoa from a CatSper2-deficient patient not only retained fully functional proton channel H_V1, but also its current density was up-regulated. This may indicate a possible linked developmental regulation between H_V1 and CatSper channels in human spermatozoa.

It should be noted that sperm activation in mice (and probably also in rats) must have some fundamental differences with the situation in humans. Not only do they lack H_V1 expression in sperm cells, but also these animals entirely lack NOX5 on a genomic level. In addition, progesterone activates CatSper only in humans and not in mice.

Figure 4 summarizes the role and potential interaction of three key players in human sperm activation: H_V1, the Ca²⁺-activated NOX5, and the sperm-specific Ca²⁺ channel CatSper. These proteins regulate three intracellular signaling systems with crucial importance for sperm activation: intracellular pH, redox potential, and cytosolic free Ca²⁺ concentration. Although details of their dynamic interaction still need to be elucidated, the following working hypothesis can be formulated. High Zn²⁺ concentrations in the male reproductive tract block spermatocyte H⁺ channels and lead to acidic sperm cytoplasm. It should be noted, however, that while the total Zn²⁺ concentrations in the male reproductive tract are indeed very high, little is known about the free Zn²⁺ concentration, which is the determinant for H_V1 inhibition. The fact that H⁺ channels are involved in cytosolic pH maintenance in resting spermatocytes suggests that these cells might have high basal H⁺ production and/or lack efficient other H⁺ extrusion mechanisms. On arrival of sperm cells in the female reproductive tract, H⁺ channels will be activated because of low Zn²⁺ concentration and possibly also through the endocannaboid Anandamide (78). It should be noted, however, that it is not clear as to what extent the Anandamide concentrations in the female genital tract are sufficiently high to activate Hv1. Simultaneously, CatSper is activated through progesterone, as well as through the H_V1-dependent cytosolic pH neutralization. CatSper-dependent Ca²⁺ influx activates the NOX5, which, due to its electrogenicity and cytosolic H⁺ generation, also requires H_V1 for optimal function. NOX5 further enhances the activity of H_V1 through depolarization and the activity of CatSper through H₂O₂. The conversion of O₂ ⁻ into H₂O₂ consumes H⁺ ions. This has two important functional consequences for the system: (i) H_V1 provides a driving force for dismutation of O²⁻ to the signaling molecule H₂O₂, and (ii) H⁺ ion consumption through O²⁻ dismutation supports H_V1 function by preventing excessive H⁺ ion accumulation at the extracellular site of the plasma membrane, which would inhibit H_V1. Thus, in a synergistic cooperation, the three sperm proteins abolish the inhibitory cytosolic acidification, increase the oxidative environment, and elevate cytosolic-free Ca²⁺ concentration. pH-, redox-, and Ca²⁺-signaling then collaborate and activate sperm effector functions, such as sperm movement, capacitation, sperm-zona pellucida interaction, acrosome reaction, and sperm-oocyte fusion.

OPEN IN VIEWER

FIG. 4.

The role of H_V1 in sperm. Sperm activation is related to H_V1 activity. (A) The conditions of male seminal fluid keep H_V1 inactive; (B) Passage of sperm into female genital tract leads to H_V1 and spermatozoid activation.

Cardiac fibroblasts

Zn²⁺ and polyvalent cations

Zn²⁺ and other polyvalent cations have been used for many years as proton current inhibitors for in vitro experiments. Half-maximal Zn²⁺ concentrations for H⁺ current inhibitions were generally in the range of 1 μM. It should be noted that this refers to free Zn²⁺ concentrations, and because of Zn²⁺ binding by proteins, metal chelators, and possibly other compounds, higher total Zn²⁺ concentrations were used depending on the buffer solutions (44). Before the identification of H_V1 and availability of H_V1-deficient mice, inhibition by Zn²⁺ was the gold standard to identify putative H⁺ currents. Of note, though, H_V1 channels from some species have been shown to be insensitive to Zn²⁺ inhibition, probably due to the lack of Zn²⁺ binding histidines at key locations (130, 136). However, since Zn²⁺ ions are implicated in many other physiological processes, the usefulness of Zn²⁺ as a specific H⁺ channel blocker is limited. Molecular mechanisms of Hv1 inhibition by Zn²⁺ have been extensively reviewed previously (30, 31, 35). In brief, Zn²⁺ appears to compete with H⁺ for binding to the external surface of the H_V1 channel. This changes the membrane potential perceived by the channel, and, therefore, stronger voltage needs to be applied to elicit the proton current.

Hanatoxin

Voltage sensors are known to be inhibited by venom toxins. Studies based on chimeras from different voltage sensors identified a modular unit composed of S3 and S4 helices, named paddle motif, which is highly conserved between different voltage-sensing proteins. The binding of hanatoxin, a toxin from venom of tarantula Grammostola spatulata (1) to the paddle motive, inhibits ion fluxes through various voltage-dependent ion channels. H_V1 channels also possess the paddle motif and were blocked by hanatoxin, which shifted the activation to more positive voltages, consistent with its predicted site of action (1). Interestingly, a mutation in the third TM segment of H_V1 (D185A; see Fig. 1) abolished this inhibitory effect. This work shows that the use of hanatoxin can be instrumental for further understanding the structure of H_V1 and could contribute to the development of new specific inhibitors of this ion channel.

Guanidine derivatives

Many inhibitors of ion channels are open channel blockers, including quaternary ammonium ions and cationic derivatives of local anesthetics. These compounds usually bind to the inside pore domain of an open channel and block the ion current. A recent study screened guanidine derivatives with different steric features in human H_V1 channels expressed in Xenopus oocytes. This work identified several molecules that are capable of blocking the depolarization-induced H⁺ current (59). The authors performed structure-activity analysis and found that even small differences in structure significantly affected the affinity. The most potent compounds were 2-guanidinobenzimidazole (2GBI) and 1-(1,3-benzothiazol-2-yl)guanidine, and their inhibitory effect depended on the presence of 2 guanidine groups conjugated on a five-membered aromatic ring. The inhibition by 2GBI had an IC₅₀ of 38 μM and was reversible. Further experiments with 2GBI allowed to obtain information on the binding site and to shed light on the location of the pore of H_V1: (i) The inhibitory effect of 2GBI was only observed on intracellular application; (ii) site-directed mutagenesis demonstrated that amino acid F150 in the S2 helix of H_V1 is important for the blocking by 2GBI (Fig. 1). The results of this study confirm that guanidine derivatives represent an interesting pharmacological tool to investigate the molecular mechanism of H_V1 inhibition. The discovery of this interaction is a promising basis for a further analysis of structure-activity relationship for H_V1 and might provide a molecular backbone for the development of more potent inhibitors. Interestingly, guanidine analogues showed neuroprotective potential in an in vitro model of ischemia (10), and it was recently shown that H_V1-deficient mice are protected in the model of stroke (146). None of the neuroprotective guanidine analogs has been tested for efficacy of H_V1 inhibition, but it will be of interest to see whether these compounds act—at least in part—through H_V1. In order to be of interest as for drug development, it will be necessary to develop cell-permeant guanidine analogs.

H⁺ current inhibitors with potentially indirect effects

Diethylpyrocarbonate (DEPC) is a histidine-modifying reagent that has been used as a biochemical tool to understand the role of histidines in the function of proteins. It was shown to inhibit NOX-dependent H⁺ current in human eosinophils and O²⁻ production on human neutrophils (5, 83). Indeed, histidines are known to be important for proton binding by H_V1: There is a competition between H⁺ and Zn²⁺ binding at His140 and His193. However, since the inhibitory effects of DEPC were only observed with an activated NOX and the effect was not found in eosinophils from CGD patients, it is more likely that DEPC acts through the modification of hemes in NOX2.

Three recent papers report the identification of compounds with inhibitory activity on H⁺ currents: tricyclic antidepressants (imipramine, amitryptiline, and desipramine), selective serotonin reuptake ihibitor fluoxetine, the morphine-derivative dextromethorphan (DM), and a tea catechin flavonoid EGCG. All of these compounds were found to inhibit H⁺ current in BV2 mouse microglial cells (67, 131, 132). The mechanism of H⁺ current inhibition by imipramine and DM was ascribed to the fact that they form weak organic bases, a mechanism of H⁺ channel inhibition previously discussed (30). With regard to EGCG, the authors argue that its inhibition of the H⁺ channel is pH independent. After a detailed analysis of the known properties of EGCG, they hypothesize that this compound might interfere with the activity of the H⁺ channel by modifying the structure of the lipid bilayer. Antidepressants, DM, and EGCG have been previously reported to have neuroprotective properties in multiple models of neurodegeneration and neuroinflammation (54, 80, 119, 147, 153). However, these drugs have multiple targets, including NOX2 (18, 153), and currently, there is no evidence that the neuroprotective effect is due to H⁺ channel inhibition.

Pharmacological modulation of H_V1 is not necessarily limited to inhibition. Indeed, as discussed next, in some disease situations, such as autoimmune disease and male infertility, H_V1 activation might be considered a relevant pharmacological approach. Currently, no drug-like molecules are known to activate H_V1. However, proton currents are pharmacologically enhanced by unsaturated long-chain fatty acids such as AA and others (32, 69, 78, 127). It appears that these fatty acid effects are due to a direct pharmacological activation of H_V1. A better understanding of the involved binding sites might lead to the discovery of drug-like H_V1 activators. In general, ion channel activators are more difficult to identify than ion channel inhibitors, as binding to the channel pore usually produces inhibition. However, it is likely that once large-scale screenings are performed, compounds which activate H_V1 by shifting the voltage- or pH dependency of channel activation could be identified.

Inflammatory diseases

No NOX2-independent function has been attributed to H_V1 in neutrophils and eosinophils. Thus, the impact of H_V1 inhibition in these cell types would be essentially a mild to moderate NOX2 inhibition, to the best of our knowledge without immunosuppression. The inhibition of neutrophil ROS generation through targeting H_V1 is obviously of interest in a variety of inflammatory pathologies, including septic shock, and other infectious diseases with an overshooting inflammatory component (e.g., bacterial meningitis, pulmonary infections, and hepatitis), ischemia/reperfusion injury, and stroke (microglia ROS generation probably is also important in this pathology, see section CNS diseases). The inhibition of H_V1 by bivalent cations should be carefully evaluated in prosthetics surgery: Cobalt ions released from metal prosthesis can promote bacterial infections in patients with metal-on-metal total hip arthroplasty (25). It should be noted that this list is similar to possible indications of inhibitors of the phagocyte NOX (76), and future research will need to determine which of both is the more promising therapeutic avenue.

Allergic diseases

Autoimmune disorders

Aging H_V1-deficient mice spontaneously develop mild autoimmunity [(123); see section T lymphocytes]. Along these lines, a study on gene expression in blood of patients with autoimmune inflammatory bowel disease detected an inverse correlation between the expression levels of H_V1 and the activity of Crohn's disease, with the expression of H_V1 being higher in the remission than in an active phase of disease (53). The mechanisms underlying the enhanced autoimmunity in the absence of H_V1 are not understood. Indeed, B-cell receptor signaling was decreased in H_V1-deficient mice (16). Thus, further studies to understand mechanisms will be necessary. However, one might speculate that H_V1 activation could be a useful therapeutic approach in certain types of autoimmune diseases. It should be noted that there is also an increased autoimmune phenotype in NOX2-deficient humans and mice. As shown in Table 4, at this point it is not clear whether there is a common denominator of the NOX2- and H_V1-associated autoimmunity.

B-cell-derived tumors

Certain lymphoid tumors are characterized by hyper-reactive B-lymphocyte receptor signaling. Therefore, H_V1 inhibition has been suggested as a therapeutic approach for this type of tumor (16).

Osteoporosis

Given the role of H_V1-dependent H⁺ secretion in acid-dependent calcium phosphate mobilization by osteoclasts, H_V1 inhibition might be an attractive concept for the treatment of osteoporosis. The underlying mechanism might be simply H⁺ secretion; however, in the case of osteoclasts, there is also an interesting link to NOX enzymes. NOX-dependent ROS generation has been implicated in osteoclast activity and, hence, in bone resorption (93, 121, 149, 150). However, while no bone alterations have been associated with NOX1 and NOX2 deficiency (9), NOX4-deficient mice appear to have increased bone density (Goettsch et al., 2013, in press). Thus, it is possible that H⁺ channels, in addition to their impact on extracellular acidity, may contribute to bone resorption through supporting NOX4 activity in osteoclasts. Thus, H_V1 inhibitors might target osteoporosis through the blocking of two osteolytic principles, namely acid secretion and ROS generation. However, no bone-related phenotype has been detected in H_V1-deficient mice so far.

CNS diseases

There are at least two mechanisms dealing with how H_V1 could be implicated in CNS diseases. H_V1 might modulate neuronal pH homeostasis. Indeed, it has been suggested that H⁺ channels might contribute to proton extrusion from metabolically active neurons (87). This H_V1 function would be neuroprotective rather than contributing to disease. However, inappropriate H_V1 function (e.g., dysregulated phosphorylation) or SNPs altering H_V1 activation properties might perturb normal neural pH homeostasis and, therefore, possibly also contribute to disease. To date, there is no experimental evidence for these scenarios and—as pointed out earlier—there is presently no convincing evidence for functional expression of H_V1 in mammalian neurons. However, H_V1 is abundantly expressed in microglia, where it might contribute to CNS disease through supporting NOX function. Indeed, oxidative stress, at least in part due to ROS generation by NOX, is a major contributor to the development of CNS diseases. A recent study by Wu et al. demonstrated that H_V1-deficient mice are protected in the pMCAO model of ischemic stroke (146). The infarct size in these mice was smaller, the neurological score was less severe as compared with the control mice, and molecular markers of cell death indicated the neuroprotective effect of H_V1 deficiency. Authors explain this effect by the reduced production of ROS by microglia. In the absence of H_V1, NOX2 causes excess depolarization of the membrane, which inhibits electron transfer by the oxidase, therefore abolishing ROS production. The authors also addressed the role of H_V1 in excitotoxicity. They observed that NMDA-induced neuronal death in primary cultures of neurons and microglia was not rescued by the absence of H_V1. This indicates that the beneficial effect of H_V1 deficiency in the model of stroke is dependent on its regulation of microglial activity rather than its contribution to excitotoxicity. These data suggest that H_V1 inhibition could be beneficial for the treatment of stroke. By extrapolation, on might notice that H_V1 inhibition could also be of interest for the treatment of other neurodegenerative processes accompanied by excessive production of ROS by microglia, such as, for example, stroke, Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis.

Pulmonary disease

No data on pulmonary disease models in H_V1-deficient mice are currently available, and no spontaneous airway phenotype has been detected in H_V1-deficient mice so far. However, based on the two documented physiological functions of H_V1 in airway epithelia, acidification of the ASL and facilitation of ROS generation by DUOX NOX (see section Airway epithelia), one may speculate on a possible role of H_V1 inhibitors in lung diseases. Maintaining the pH of ASL at a mildly acidic pH is important for normal physiology, and dysregulation might contribute to disease. In particular, overshooting acidification might be involved in the development of airway inflammation, as well as chronic obstructive lung disease, asthma, and cystic fibrosis [reviewed in Fischer and Widdicombe (45)]. The facilitation of ROS generation by DUOX NOX similarly might contribute to airway inflammation. Thus, H_V1 inhibitors might be of interest for the treatment of inflammatory airway diseases.

Male fertility

H_V1 is highly expressed in the testis, and its activity is crucial for sperm motility, chemotaxis, and acrosome reaction (78, 99) and therefore for male fertility. Interestingly, low levels of H_V1 expression are correlated with male infertility (111) and, more specifically, patients with teratozoospermia (i.e., abnormal sperm morphology) have reduced expression of H_V1 (Geo Profiles, NCBI ID: 38039902; ID: 38148335; ID: 38065002). Altogether, these findings suggest that H_V1 activation in human spermatozoa could be an interesting concept for male infertility. Conversely, H_V1 inhibition could be considered a tool for contraception.

Cancer

Otto Warburg described unique metabolic properties of cancer cells, namely a mostly nonoxidative breakdown of glucose (145). He also hypothesized that unique metabolic state leads to increased acidity in cancers. This should, indeed, make cancer cells particularly dependent on efficient proton extrusion pathways, in particular proton channels. Indeed, several recent publications point toward a privileged role of H_V1 in cancer cells (142 –144). For example, expression of H_V1 protein was associated with an invasive phenotype in breast cancer and in colorectal cancer (142 –144): H_V1 expression was increased on the surface of highly metastatic cells and correlated with tumor progression. Associations between H_V1 levels and tumor size, tumor classification, lymph node status, and clinical stage were detected. Tumor cells are highly metabolically active, resulting in low pH inside the cell. Voltage-gated proton channel expression is probably up-regulated in tumor cells in response to intracellular acidosis in order to regulate the pH. In fact, the inhibition of H_V1 expression by short-interfering ribonucleic acid (siRNA) in highly metastatic MDA-MB-231 cells reduced intracellular pH measured using BCECF pH-sensitive fluorescent probe and decreased the invasion and migration potential of these cells. Moreover, H_V1 silencing in grafted cells reduces tumor growth in nude mice injected sub-cutaneously with MDA-MB-231 cells. On the other hand, the expression of H_V1 in tumors might be linked to the role of NOX4 in breast cancer (152). Taken together, H_V1 inhibition might become an interesting concept to limit invasiveness of breast cancer, and possible other types of tumors. There is also an interest in H_V1 as a general biomarker for tumors. Indeed, several patents (WO 2010110346 A1, EP 1961825 A1, US 20110190157; EP 2253721 A1; EP 2115170 A2; US 20110217297 A1) suggest that H_V1 could be useful as a biomarker for leukemias and solid tumors.