Calcium Channels and Calcium-Regulated Channels in Human Red Blood Cells

Abstract

Free Calcium (Ca²⁺) is an important and universal signalling entity in all cells, red blood cells included. Although mature mammalian red blood cells are believed to not contain organelles as Ca²⁺ stores such as the endoplasmic reticulum or mitochondria, a 20,000-fold gradient based on a intracellular Ca²⁺ concentration of approximately 60 nM vs. an extracellular concentration of 1.2 mM makes Ca²⁺-permeable channels a major signalling tool of red blood cells. However, the internal Ca²⁺ concentration is tightly controlled, regulated and maintained primarily by the Ca²⁺ pumps PMCA1 and PMCA4. Within the last two decades it became evident that an increased intracellular Ca²⁺ is associated with red blood cell clearance in the spleen and promotes red blood cell aggregability and clot formation. In contrast to this rather uncontrolled deadly Ca²⁺ signals only recently it became evident, that a temporal increase in intracellular Ca²⁺ can also have positive effects such as the modulation of the red blood cells O₂ binding properties or even be vital for brief transient cellular volume adaptation when passing constrictions like small capillaries or slits in the spleen. Here we give an overview of Ca²⁺ channels and Ca²⁺-regulated channels in red blood cells, namely the Gárdos channel, the non-selective voltage dependent cation channel, Piezo1, the NMDA receptor, VDAC, TRPC channels, Ca_V2.1, a Ca²⁺-inhibited channel novel to red blood cells and i.a. relate these channels to the molecular unknown sickle cell disease conductance P_sickle. Particular attention is given to correlation of functional measurements with molecular entities as well as the physiological and pathophysiological function of these channels. This view is in constant progress and in particular the understanding of the interaction of several ion channels in a physiological context just started. This includes on the one hand channelopathies, where a mutation of the ion channel is the direct cause of the disease, like Hereditary Xerocytosis and the Gárdos Channelopathy. On the other hand it applies to red blood cell related diseases where an altered channel activity is a secondary effect like in sickle cell disease or thalassemia. Also these secondary effects should receive medical and pharmacologic attention because they can be crucial when it comes to the life-threatening symptoms of the disease.

25.1 Introduction to Calcium in Red Blood Cells

Free Calcium (Ca²⁺) is an important and universal second messenger in all cells [1, 2], red blood cells (RBCs) included [3,4,5]. This results in the abundance of Ca²⁺-binding proteins in RBCs with differing Ca²⁺ sensitivities as outlined in Fig. 25.1. Mature mammalian RBCs are believed to not contain organelles as Ca²⁺ stores such as the endoplasmic reticulum or mitochondria [6]. Compared to other cell types, where the Ca²⁺ liberated from stores within intracellular organelles can be used in the regulation of free cytosolic Ca²⁺ concentration and thereby Ca²⁺ signalling, in mammalian erythrocytes the control of free intracellular Ca²⁺ concentration must be done by regulation of membrane transport. A 20,000-fold gradient based on an intracellular Ca²⁺ concentration of approximately 60 nM vs. an extracellular concentration of 1.2 mM makes Ca²⁺-permeable channels a major signalling tool of RBCs. As historically RBCs served as the model cell to investigate membrane transport, it is well known that the internal Ca²⁺ concentration is tightly controlled, regulated and maintained primarily by the Ca²⁺ pumps PMCA1 and PMCA4 [4, 7]. The Ca²⁺ pumping in turn is regulated by multiple factors, such as the Ca²⁺ concentration itself [8], calmodulin [9], calpain [10], phospholipids and various kinases [11] or even self-association [12]. Within the last two decades it became evident that an increased intracellular Ca²⁺ is associated with RBC clearance in the spleen and promotes RBCs aggregability and clot formation [3, 13,14,15,16]. There was a long debate within the community whether this process should be called eryptosis [17], which is no longer recommended [18]. In contrast to this rather uncontrolled deadly Ca²⁺ signals (resulting in Ca²⁺ overload), within the recent years it became evident that a temporal increase in intracellular Ca²⁺ can also have positive effects such as the modulation of the RBCs O₂ binding properties [19] or even be vital for brief transient cellular volume adaptation when passing constrictions like small capillaries or slits in the spleen [5, 20, 21] as depicted in Fig. 25.2. The perilous balance of Ca²⁺ in RBCs was recently reviewed [3].

Fig. 25.1

Overview of concentration dependence of Ca²⁺ activated proteins in RBCs. The yellow column indicates the estimated range of RBCs‘resting free Ca²⁺ [96]. The gray/black bars indicate the activation of the proteins with the intensity of darkness related to the activation level (details see below). The red lines depict the half activation concentration. For orientation the green line provides the in vivo k_D for Fluo-4 [101], probably the most appropriate Ca²⁺ fluorophore to be used in RBCs [102]. The universal intermediate messenger calmodulin has a dissociation constant for Ca²⁺ of 920 nM [103], which can be shifted down to 100 nM, indicated by the red arrow. The Gárdos channel has an open probability of EC₅₀ of 4.7 μM with a Hill slope factor of approximately 1 [104]. Values were measured in excised patches at a membrane potential of 0 mV. The curve of the opening frequency is almost superimposable (EC₅₀ of 4.3 μM) [104] keeping the values given in the figure valid also for whole cell and hence population based investigations. The values for half maximal activation of the scramblase was determined by different studies with varying methodologies and a slightly different result. Values varied between approximately 30 μM determined in liposomes [105] and 70 μM measured in RBC ghosts [106]. The flippase displays almost full inhibition already at a Ca²⁺ concentration of 400 nM [107]. μ-Calpain, a protein that cleaves cytoskeleton and membrane proteins depicts an half activation at 40 μM Ca²⁺ [108] but can be activated and then shifting half-maximal activation down to 200 nM [109]. Transglutaminase mediating polymerisation of RBC membrane proteins in its native form has a dissociation constant for Ca²⁺ of 190 μM [110]. Adenosine monophosphate (AMP) deaminase is an enzyme that converts AMP into inosin monophosphate and is directly stimulated by Ca²⁺ at a half maximal concentration of 50 μM free Ca²⁺ [111]. The binding of Ca²⁺ to the C2-domain of PKCα was determined in vitro to be 35 μM with a Hill coefficient of 0.9 [112]. Although the Ca²⁺ dependence of the membrane binding was measured to be one order of magnitude lower [112], the initial Ca²⁺ binding is the crucial step for PKCα activation and therefore the relevant number in this compilation. (This figure is reproduced from Bogdanova et al. 2013 [3])

Fig. 25.2

In vivo Ca²⁺-signalling of mouse RBCs when passing through capillaries. Mouse RBCs were ex vivo stained with Fluo-4 and then re-injected into the mouse circulation. Fluorescence imaging of capillaries was performed in the dorsal skinfold chamber. (A) shows representative snapshots of RBCs passing a bifurcation. For the 30 ms sequence only every second recorded image is presented. For a better orientation the vessel walls are indicated by red dashed lines in the leftmost image and further annotations (grey) are added in the other images. (B) depicts the positions where fluorescence intensity (F/F₀) was analysed for a decreasing vessel calibre (Ba) and for a constant vessel calibre (Bb) of the same example section as in (A). The dashed red lines mark the vessel walls, the red arrow indicates the blood flow direction and the yellow circles depict the analysis positions which are plotted in the following diagrams. Example fluorescence traces of the two cells analysed as pointed out in (Ba) and (Bb) are shown in (Bc). (C) depicts the analysis of 30 cells passing through a capillary with decreasing vessel calibre and 28 cells passing through a capillary with constant vessel calibre. Analysis was performed at 3 vessel-bifurcations in 2 mice. The fluorescence intensity (F/F₀) traces of all measured RBCs passing through a capillary with decreasing vessel calibre is plotted in (Ca), while the traces of all measured RBCs passing a capillary with constant vessel calibre is plotted in (Cb). The statistical analysis of the maximal fluorescence intensity (F/F₀) of RBCs from both groups is depicted in (Cc). The increase in Ca²⁺, while passing through a vessel with decreasing calibre is significant (p = 0.014; *). (This figure is reproduced from Danielczok et al. 2017 [5])

Fig. 25.3

Milestones in Gárdos channel research. (a) Reprint from Gárdos 1958 [25]; a one page article mentioning a dependence of K⁺-efflux from Ca²⁺ in the category “preliminary notes” was retrospectively sufficient for the naming of an ion channel. IA stands for iodoacedic acid. (b) Reprint from Hamill 1981 [23]; the first published single channel current traces from the Gárdos channel appeared in a poster Abstract for the Physiological Society (London) meeting. Numerous patch-clamp based characterisations followed this initial recording [22, 28, 104, 103,104,105,106,107]. (c) Reprint from Hoffman et al. 2003 [26] providing the molecular identification of the Gárdos channel to be KCNN4 (K_Ca3.1, IK1, SK4). (d) Reprint from Rapetti-Maus et al. 2015 [29] providing the first report of a Gárdos channel mutation and its link to a pathophysiological setting. Upper panel: Blood film smears for two patients with the Gárdos channel mutation p.Arg352His (mother and son). Lower panel KCNN4 transcript sequencing; upper line presents the wild type sequence and the lower line the transcript with mutation c.1055G.A (p.Arg352His)

25.2 The Gárdos Channel – A Calcium-Activated Potassium Channel

The Gárdos channel is one of the Ca²⁺ sensors in RBCs transferring Ca²⁺ uptake into K⁺ and water loss and mediating thereby Ca²⁺-dependent volume regulation and changes in RBC rheology. It is also annotated as KCNN4, K_Ca3.1, IK1 or SK4. It is the first channel we describe in this chapter because it was the first channel found in RBCs [22, 23], i.e. utilising the patch-clamp technique as a direct read-out of channel activity. However, its name goes back to the effect of Ca²⁺ dependent K⁺ efflux found in RBCs by G. Gárdos once metabolic pathways are poisoned [24, 25]. As the molecular identity of this transport in RBCs was not known, it was referred to as Gárdos effect and later, when it turned out to be ion channel mediated, the involved transport protein was called Gárdos channel. Even after its molecular identification in 2003 [26] in the RBC field it remained to be referred to as Gárdos channel. Figure 25.3 provides the milestones in the Gárdos channel research. A comprehensive review of the Gárdos channel structure and function was published some 15 years ago [27]. An interesting property of the channel is its temperature dependence. With decreasing temperature, a continuous decrease of Gárdos channel conductance is observed. The Arrhenius plot of the unitary channel conductance between 0 °C and 47 °C is strictly linear and has a slope which corresponds to an activation energy of 29.6 ± 0.4 kJ/mol. Nevertheless, simultaneously, altered gating kinetics results in an increase of channel opening probability at reduced temperatures. At saturating concentration of intracellular Ca²⁺ (10 μM), reducing the temperature from 35 to 30 °C results in a change of the opening and closing kinetic of the Gárdos channel. Brief channel openings and closing are progressively replaced by longer openings and shorter closing states [28]. More importantly, reducing the Ca²⁺ concentration at the intracellular face of the channel at half the EC₅₀ for Ca²⁺ and temperature close to 0 °C drastically increases open probability indicating that even at very low Ca²⁺ concentrations the Gárdos channel may be activated. Knowing that blood samples for analyses and RBC concentrates for blood transfusion are kept refrigerated, one has to keep in mind this peculiar property of the Gárdos channel that may have deleterious effects respective to the cell volume.

The physiological function of the Gárdos channel was a speculative topic for many decades although a link to RBC cell volume threatening was clear, since activation of the Gárdos channel results in cellular K⁺ loss associated with Cl⁻ and osmotically obliged water loss may lead to rapid cell shrinkage. In this context the channel was believed to be a ‘suicide mechanism’ triggered by the intracellular increase in Ca²⁺. This was proposed to happen in the process of clot formation, in thrombotic events as well as during RBC clearance. With the finding of the first mutations in the Gárdos channel [29, 30] (see also Fig. 25.3D) and the associated pathophysiology, a physiological function of the Gárdos channel was doubtless proven. Although the initial reports link the mutation of the Gárdos channel to Hereditary Xerocytosis [29, 30], further studies revealed that ‘Gárdos channelopathy’ is its own disease or at least an own variant of Hereditary Stomatocytosis [31]. Any prolonged Gárdos channel activation lead to changes in cell volume which eventually affect rheological, stiffness and rigidity properties that compromise their survival within the circulation especially during their passage within the slits of the spleen. The mutations reported so far resulted in a gain of function that could be treated with a Gárdos channel inhibitor. There are numerous Gárdos channel inhibitors available some of them already clinically tested for other diseases, like clotrimazole for topical applications or senicapoc [32]. The Gárdos channel also shows an increased activity in other haemolytic diseases, such as sickle cell disease. Therefore one has tried to use the Gárdos channel as a pharmacological target to treat sickle cell disease [33] overlooking the fact that upstream of the signalling cascade is an increase in intracellular Ca²⁺ through a pathway named P_sickle for which we are still seeking molecular identity (see below) and triggers numerous other pathophysiological processes in the RBCs (compare Fig. 25.1) leaving the Gárdos channel only a minor portion to account for the cellular symptoms of sickle cell disease [34]. However, what failed in sickle cell disease may still work out well in Gárdos Channelopathy [31, 35].

25.3 Non-selective Cation Channels Permeable for Calcium

25.3.1 Non-selective Voltage Dependent Cation Channel and Piezo1

RBCs contain a variety of non-selective ion channels that are permeable to Ca²⁺. First there is a non-selective voltage dependent cation channel initially described by Christophersen and Bennekou [36, 37] and later to be reported to be Ca²⁺ permeable [38]. However, the molecular identity of this channel is still not quite clear [39, 40]. Although it was proposed to be a conductive state of the voltage-dependent anion channel (VDAC) [41], recent reports rather make a link to the Piezo1 [42, 43]. Figure 25.4 provides a comparison of the non-selective voltage dependent cation channel and Piezo1. The unique hysteresis-like open probability was also modelled successfully [44].

Fig. 25.4

Comparison of the non-selective voltage activated cation channel recorded in RBCs and Piezo1 recorded in overexpressing Neuro2A cells. (a) Current traces of the non-selective voltage activated cation channel in inside-out patches of RBCs in symmetrical KCl-solution in mM (500 KCl, 5 MOPS, 4 NMDG, 1 EGTA, 0.02 Ca²⁺, pH = 7.4). (b) Current traces of Piezo1 in outside-out patches of overexpressing Neuro2A cells in symmetrical NaCl-solution in mM (140 NaCl, 10 HEPES, 5 EGTA, pH = 7.4). The difference in ion strength between panels a and b could explain the different current amplitudes shown in the single channel openings. (c) The open state probability as function of the membrane potential. Filled symbols represent data from the experiment shown panel a. Open symbols represent data from an identical experiment performed on the same patch after 15 min. In both series, the open probability was calculated from 3 min of continuous recording at each potential. The curves were drawn by eye. (d) Tail currents from individual cells were normalized to their maximum and fitted to a Boltzmann relationship. Pooled data are shown as mean ± SEM. (Panels a and c are reproduced from Kaestner et al. 2000 [38] and panels b and d from Moroni et al. 2018 [42])

PIEZO1 and in particular mutations of the channel have been associated with the RBC-related disease Hereditary Xerocytosis [45, 46]. Therefore it seems obvious that this channel, originally described as a mechanosensitive channel, is present in the RBC membrane. Furthermore, knock-out approaches in zebrafish [21] and mice [20] gave further evidence for the conserved abundance of Piezo1 in RBCs as well as for its function (see below). Piezo1 and its mutations were mainly characterised in heterologous overexpressing cell lines and initial measurements of Piezo1 in RBC have been rather episodic [47]. However, the discovery of the pharmacological activation of Piezo1 by Yoda1 [48] lead to the development of high throughput patch-clamp assays as potential diagnostic tools that were recently implemented [49].

The interplay of Piezo1 (in particular its property to mediate Ca²⁺ entry) with the above mentioned Gardos channel [5, 20, 21] provides additional evidence for a functional Piezo1 in RBCs. Furthermore, the interplay between the channels was proposed to be vital for the RBCs to maintain their ion homeostasis [50]. In pathophysiology, Piezo1 seems to play a mayor role in an increased RBC Ca²⁺-homeostasis. The reported mutations of Piezo1 [45, 46, 51,52,53,54] are mostly gain of function mutations suggesting an easier and more pronounced Ca²⁺ entry. The consequent increase in intracellular Ca²⁺ is most likely the trigger for the early removal of the RBC from the circulation and hence the reason for the anaemic symptoms [55]. One of the clinical treatments to handle severe anaemias is splenectomy. Interestingly splenectomy introduces comorbidity, namely thrombotic events, in a subpopulation of Hereditary Xerocytosis patients. This could be explained by an active participation of RBCs in the thrombus formation due to increased intracellular Ca²⁺ [13, 14, 56,57,58]. Surprisingly such thrombotic events were not reported for splenectomised Gárdos Channelopathy patients [31].

It is worthwhile to mention that in sickle cells an increased conductance carrying also Ca²⁺ was reported and named P_sickle [59, 60]. It is likely that P_sickle resembles the superposition of several ion channel entities. The sensitivity of P_sickle to GsMTx-4 [61], a toxin that inhibits Piezo1 [62] points to this mechanosensitive channel, while the increased abundance of N-methyl-D-aspartate (NMDA) receptors in sickle cells [63] is a very strong indicator for these ionotrope glutamate receptors (see below).

Fig. 25.5

NMDA receptors structure and function. (A) Erythroid NMDA receptors are expressed early on in differentiating erythroid progenitor cells starting from proerythroblasts (CD34 + -derived cells on day 5 in culture) to orthochromatic erythroblasts (day 16 in culture). (B) Electrophysiological recordings of currents mediated by treatment of these cells with agonists NMDA and glycine reveal the change in subunit composition of the receptor during differentiation. Whereas proerythroblasts and polychromatic erythroblasts are equipped with receptors built by glycine-binding GluN1 and 3A/B and glutamate/NMDA-binding Glun2A and 2D (see cartoons), late orthochromatic erythroblasts and reticulocytes contain receptors in which GluN2A is replaced by GluN2C. As a result channels with high current amplitudes and short times to inactivation in early progenitors turn into slowly inactivating channels with lower amplitude in late progenitors and circulating RBCs. For details see [67]. (C) Repeated activation of eNMDARs triggers inactivation of the channels. (D) Changes in the intracellular Ca²⁺ concentration in RBCs upon systemic administration of memantine in patients with sickle cell disease. (Da) Dot plot showing the heterogeneity of free Ca²⁺ levels in RBCs (gates F, E, and H) and RBC-derived vesicles (gate D). (Db) shows a histogram of Ca²⁺-dependent fluorescence (all cells in population, gate H) in unstained cells (negative control, red), at baseline before the onset of treatment with 5 mg memantine a day (green) after 24 h (blue) and 1 week (pink) of therapy with 5 mg Memantin Mepha daily. The treatment was performed within the MemSID clinical trial (ClinicalTrials.gov Identifier: NCT02615847 approved by SWISSMEDIC (# 2015DR2096) and Cantonal ethic committee of canton Zurich (#2015–0297)). (Panels A–C are reproduced from Hänggi et al. 2015 [67] and panel D from Makhro et al. 2017 [66])

25.3.2 N-Methyl D-Aspartate Receptor

There is clear evidence for the abundance of erythroid N-methyl D-aspartate (NMDA)-receptors in RBCs in particular in the young cell population [19, 64, 65] and its abnormally high prevalence and activity in sickle cell patients [34, 63]. Inhibition of these receptors by oral administration of memantine, the pore-targeting antagonist of NMDARs, results in a decrease of the intracellular Ca²⁺ (Fig. 25.5). In general, cells stemming from myeloid lineage are expressing particular type of ionotropic glutamate receptors making immune responses, clotting and RBC function sensitive to the changes in ambient glutamate levels [66]. Subunit composition of erythroid NMDA receptors (eNMDARs) as well as the number of receptor copies changes in the course of differentiation (Fig. 25.5A–C). Receptor abundance declines from thousand copies in proerythroblasts to about 30 in reticulocytes and about 5 (on average) in mature RBCs [19, 67]. At the same time high amplitude currents with short inactivation time carried by the GluN2A/2D-containing receptors in proerythroblasts are replaced by the currents mediated by the receptors built by the GluN2C/2D subunits with much smaller amplitude and prolonged inactivation time in ortho/polychromatic erythroblasts and reticulocytes [67], (Fig. 25.5A–C). eNMDARs are highly permeable for Ca²⁺ [19, 64] and are actively involved in Ca²⁺-driven signaling during differentiation and maintenance of intracellular Ca²⁺ in mature RBCs [19, 67]. Clearance of eNMDARs in reticulocytes released into the circulation most likely occurs by way of ‘shedding’ when receptors are released together with other membrane proteins from the membrane in the form of vesicles. Whereas no direct measurements of eNMDARs in vesicles were performed so far, inability to clear eNMDARs from membranes of RBCs of patients with sickle cell disease is an indirect evidence for this hypothesis. In erythroid precursor cells obtained by differentiation of peripheral CD34+ cells of sickle cell disease patients the number of eNMDAR copies was like that in cells of healthy patients [67]. However, circulating RBCs of patients were presented with abnormally high abundance and activity levels of eNMDARs. As a result, basal Ca²⁺ levels in RBCs of sickle cell disease patients were exceeding that in cells of healthy subjects. Pharmacological inhibition of the receptors decreased Ca²⁺ levels and resulted in rehydration and reduction in oxidative load [67]. First pilot clinical trial MemSID in which patients with sickle cell disease were treated with the antagonist of NMDA receptors, memantine, revealed that these receptors may be an attractive pharmacological target for this group of patients [68, 69] (Fig. 25.5D). Among physiological factors that may control eNMDARs are endurance exercises that are associated with glutamate release into the circulation [70].

Fig. 25.6

VDAC, ANT and TSPO detection in human red blood cell ghosts. (a) Samples (15 μg of protein) of whole lysates were subjected to SDS-PAGE (10% acrylamide) and analysed by Western blotting using polyclonal anti-ANT (1:1000 dilution), polyclonal goat anti-TSPO raised against the C terminus of human TSPO (1:1000) or rabbit polyclonal anti VDAC 1–2-3 (1:100 dilution). Multiple bands at different molecular weights is consistent with the oligomerisation of VDAC proteins. ANT and TSPO proteins are also clearly visible. (b) Immunofluorescence experiments were performed on smears Dilution were 1/5 for primary and 1/20 for secondary antibodies. Scale bars represent 10 μm. (This figure is reproduced from Bouyer et al. 2011 [41])

25.3.3 Voltage-Dependent Anion Channel

Another multifunctional channel with a clear molecular identity in RBCs that also conducts Ca²⁺ is the Voltage-Dependent Anion Channel (VDAC) [41, 71]. VDACs have originally been characterized as mitochondrial porins [72]. Three different isoforms of VDAC have been identified so far: VDAC1, VDAC2 and VDAC3. Showing the expression of ‘Porin 31HL’ in the plasmalemma of human B lymphocytes, gave first evidence on the multitopological localisation of VDAC [73]. The existence in the membrane of RBCs of a 32 kDa associated voltage dependent anion channel (VDAC) in a peripheral benzodiazepine receptor-like PBR protein complex to 18 kDa protein TSPO ‘translocator proteins’ and 30 kDa ANT ‘adenine nucleotide transporter’ proteins has been demonstrated [41, 74]. It has a nanomolar affinity for PK11195, Ro5–4864 and Diazepam ligands [75, 76]. All blood cells have a population of receptors with micromolar affinity for PK11195 ranging from approximately 750,000 sites for lymphocytes to over one hundred sites for RBCs [77, 78]. These indications are corroborated by analysis of messenger RNA expression data provided by GeneAtlas U133A where 3 isoforms of VDAC, 2 isoforms of ANT and 2 isoforms of TSPO were found in erythroid progenitors from CD34+ to CD71+ (Fig. 25.6). VDAC is a protein that has remarkably well-preserved structural and functional characteristics, despite major variations in the sequence [79]. Although it is also present in the plasma membrane, most of the information we have on its structure function comes from studies on mitochondrial proteins [80, 81]. The maximum conductances reach 4–5 nS in the presence of 1 M NaCl or KCl, 350–450 pS for more physiological concentrations (NaCl or 150 mM KCl). Conductance and selectivity are voltage dependent; at low voltages, close to −10 mV, the channel is stable and remains open, whereas at positive or negative potentials higher than 40 mV, VDAC has multiple sub states of different permeabilities and selectivities, as well as closing episodes of which frequency increases with voltage [82, 83]. The highest levels are permeable to small ions (Na⁺, K⁺, Cl⁻, etc.) but also to large anions (glutamate, ATP) and large cations (acetylcholine, dopamine, Tris, etc.). They have a preference for anions (2:1) when saline solutions are composed of ions of equal mobility such as NaCl or KCl. More importantly, at low conductances, VDAC is more permeable to small ions with, apparently, a marked preference for cations and higher permeability to Ca²⁺ ions than in large conductances [82,83,84]. VDAC may have different oligomerization states: mono-, di-, tri, tetra-, hexamers or even more. Indeed, atomic force microscopy revealed the presence of VDAC1 monomers as well as dimers and larger oligomers showcasing the interaction of the pore with itself, however, dimers are more frequent. Very little is known about the activation and regulation mechanisms of the channel. Nevertheless, when the pores dimerize, the selectivity for Ca²⁺ increases. Various studies support the function of VDAC (more precisely VDAC1 the most studied yet) in the transport of Ca²⁺ and in cellular Ca²⁺ homeostasis. Lipids-reconstituted bilayer incorporating VDAC1 in the presence of different CaCl₂ concentration gradients showed well-defined voltage-dependent channel conductance, as observed with either NaCl or KCl solution, with higher permeability to Ca²⁺ once VDAC is in the low conductance state. It is obvious that the permeability ratios of VDAC1 for Ca²⁺ is very low compared to Cl⁻ (P_Ca2+/P_Cl− is 0.02–0.38) [83] but considering the tremendous electrochemical gradient for Ca²⁺ between intra- and extracellular face of RBCs (see Introduction) a short activation may represent a significant input of Ca²⁺ into the cell.

25.3.4 Transient Receptor Potential Channels of Canonical Type

Yet another type of non-selective cation channels that are believed to be abundant in RBCs are Transient Receptor Potential channels of Canonical type (TRPC channels). Indications point to a different expression pattern of isoforms in precursor cells compared to mature RBCs and also differences between mammalian species seem likely [85,86,87,88]. In humans it is believed that TRPC6 is abundant in RBCs [39, 89, 90]. So far a dedicated physiological function of TRPC6 in mature RBCs remains elusive.

25.4 Voltage-Activated Calcium Channels and Their Regulation

Evidence for the existence of a number of voltage-activated Ca²⁺ channels in RBCs has been reported [91, 92], and the most convincing evidence is for Ca_V2.1, based on molecular biology data (Western blot) [93] and, presumably, Ca_V2.1-specific pharmacological interactions (ω-agatoxinTK) [93, 94] – both are shown in Fig. 25.7A, B. Nevertheless, so far, we and others have failed to obtain direct functional evidence for the existence of Ca_V2.1 or other voltage-activated Ca²⁺ channels in RBCs by patch-clamp techniques [95]. However, also RBCs although non-excitable cells meet the condition of voltage jumps necessary to activate voltage-activated channels such as Ca_V2.1 [95]. In particular when the Gárdos channel (see Sect. 25.2) is activated, the resting membrane potential changes from approximately −10 mV to approximately −70 mV [96]. Not hyperpolarisation but depolarisation is required to activate Ca_V2.1 [97]. Nevertheless, hyperpolarisation is a requirement to switch Ca_V2.1 channels from the inactivated state to the closed state, which is a prerequisite to subsequently transition to the open state [98] (Fig. 25.7C). Closing of the Gárdos channels after their initial activation could well provide the necessary conditions for subsequent depolarisation to activate Ca_V2.1 [95]. Such a proposed mechanism is sensible also in the context of other voltage-activated channels in the RBC membrane (compare Sect. 25.3).

Fig. 25.7

Ca_V 2.1 in human RBCs. (A) Western blot analysis of the 1A subunit of voltage-gated calcium channels in human erythrocyte ghosts from 3 different donors. RBC membrane proteins were separated electrophoretically at high protein loads (80 μg). After electrophoretic transfer to nitrocellulose paper, blots were stained with antibodies directed against residues 865 to 881 of the α_1A subunit of the rat brain voltage-gated calcium channel (lanes 1A, 2A, 3A). Because several nonspecific bands were also visualized, competition of Ca_V2.1 antibody staining using its specific peptide was also performed. For this purpose, 10 μg of the Ca_V2.1 antibody was preincubated with 10 μg of antibody-specific peptide for 1 h at 22 °C and then further incubated with the blot for 2 h at 22 °C (lanes 1B, 2B, 3B). Polypeptides with M_r of approximately 190,000 and 220,000 are characteristic of the major splicing variants of the α_1A subunit of Ca_V2.1 in the brain. An SDS-PAGE Coomassie blue–stained gel of RBC membranes (left lane; “RBC Ghost”) serves as an approximate molecular weight marker. (B) Kinetics of lysophosphatidic acid (LPA; phospholipid released by activated platelets)-induced (Ba) and Phorbol 12-myristate 13-acetate (PMA; protein kinase C activator)-induced (Bb) Ca²⁺ entry in the presence and absence of ω-agatoxin TK. Average traces of single cells derived from live cell imaging experiments are presented as self-ratio values. Labelled lines above the traces indicate the stimulation regime. The traces are the mean values of 3 independent experiments, and the numbers in brackets at the end of the colour legend refer to the number of cells measured. (C) Activation scheme for the Ca_V2.1 channel modulated by underlying Gardos channel activity. Closing of the Gardos channels after their initial activation could provide the necessary conditions for subsequent depolarisation to activate Ca_V2.1. Since the hypothetical switching behaviour of the Gardos channel would be crucial for the activation of Ca_V2.1, we propose three principle modes by which this switching could occur: (i) Because channel activity is a stochastic event and because the number of Gardos channels per RBC is rather low (in the single digit numbers [114, 118]), depolarisation could be the result of stochastic Gardos channel closures. This hypothesis is supported by the rather sparse whole cell patch-clamp recordings of Gardos channel activity in human RBCs [31, 119,120,121]. Whole cell current traces do not show a smooth appearance but rather a flickering pattern similar to that observed with single channel recordings, especially at higher (positive and negative) membrane potentials. (ii) When looking at Gardos channel-induced changes in the membrane potential of cell populations, a gradual Ca²⁺ concentration-dependent effect can be seen [122], i.e., the hyperpolarisation observed in RBC suspensions is a gradual Ca²⁺ concentration-dependent effect. However, the abovementioned study [122] as well as another report [123] showed that the activation of the Gardos channel at the cellular level is an all-or-none response. This means that the gradual change in membrane potential would be the result of the summation of cells with open or closed Gardos channels. Taking into consideration that the Ca²⁺ pump [124] continuously operates in response to any increase in intracellular Ca²⁺ levels, one would imagine that the state of the Gardos channels is exclusively modulated by variations in intracellular Ca²⁺ concentrations. Hence, the switching behaviour of the Gardos channel would be the direct consequence of continuous variations in RBC intracellular Ca²⁺ concentrations. (iii) Localized interactions between the Gárdos channel and Ca_V2.1 in RBCs could occur in lipid rafts or nanodomains, as is the case with closely related ion transporters in other cell types, for example, within the fuzzy space or dyadic cleft in myocytes [125]. Although RBCs do not possess membrane-constricted subspaces, there are indications for functional compartments in the immediate vicinity of the plasma membrane [126]. Colocalization of ion channels is common in excitable cells [127, 128]. For RBCs, it is still unknown if the different ion channels colocalize or cluster to allow their interaction in nanodomains. However, in support of this idea is the observation that local activation of mechanosensitive channels (most likely Piezo 1) by patch-clamp micropipettes resulted in local activation (single-channel recordings) of the Gardos channel [124]. (Panels A, B and C are reproduced from Andrews et al. 2000 [93], Wagner-Britz et al. 2013 [94] and Kaestner et al. 2018 [95], respectively)

Fig. 25.8

Evidence for non-selective cation channel activated by the removal of Ca²⁺ . (A) Whole-cell patch clamp recordings in a Cs⁺-based internal and a TEACl-based external solutions (Aa) I/V curves with 2 mM CaCl₂ (blue) and 0 mM CaCl₂ (red) in the external solution (n = 5). (Ab) I/V curve of the Ca²⁺ blocked current – the current recorded in 2 mM CaCl₂-external solution was subtracted from the current recorded in 0 mM CaCl₂ -external solution. Currents were elicited by voltage steps from −100 mV to 100 mV for 500 ms in 20 mV increments at V_h = −30 mV. Detailed solutions composition is given next to the graphs. Data are presented as mean ± SEM, with n being the number of cells. Significance is assessed with a paired Student’s t test and set at p < 0.05. For better visualization, a significance anywhere below p < 0.05 is denoted with one star. (B) Whole-cell patch clamp recordings in physiological (a K⁺-based internal and a Na⁺-based external) solutions (Ba) Raw current traces from a representative RBC in an external solution containing 2 mM CaCl₂ at t₁ (dark blue) and 0 mM CaCl₂ at t₂ (green). Detailed solutions composition is given next to the current traces. (Bb) I/V curves in 2 mM CaCl₂ (t₁) (dark blue) and 0 mM CaCl₂ (t₂) (green) external solutions (n = 7). (C) Blood plasma ion content of healthy adults in heparin and EDTA. (Ca) Ca²⁺ content of a control aqueous non-buffered 1.8 mM CaCl₂ solution filled in heparin and EDTA vacutainers. The abbreviation b.d.l. denotes ‘below detection limit’. (Cb) Na⁺ plasma content, (Cc) K⁺ plasma content of blood anticoagulated with heparin and EDTA. Measurements in heparin and EDTA were performed on blood of healthy adults (n = 3) collected in heparin and EDTA vacutainers, respectively and reference values were taken from Liappis, 1972 [130]. Error bars represent SEM and stars denote significances as follows: n.s for not significant, ** for p < 0.01 and *** for p < 0.001. (This figure is reproduced from Petkova-Kirowa et al. 2018 [99])

25.5 Evidence for a Calcium-Inhibited Channel

There is also evidence for a non-selective cation channel in RBCs that is activated when extracellular Ca²⁺ is removed [99]. Original recordings and an I–V curve are shown in Fig. 25.8A, B. There are two conceptual question related to this recent report (a) if the channel is abundant in almost all RBCs why it was not reported before (in four decades of patch-clamping RBCs) and (b) since divalent cations in general and Ca²⁺ in particular support seal formation, a removal of Ca²⁺ could impair the seal quality/tightness. Under these circumstances it is almost impossible to discriminate a leak in the seal from an ion channel. However, here are also two arguments in favour of the existence of this channel: (A) The suspicion of the phenomenon described in (b) could have prevented scientists to report about the channel (a). The non-ohmic behaviour of the I-V curve (Fig. 25.8B) is in favour of a channel rather than a leak. (B) A channel activated by the removal of Ca²⁺ is an ideal explanation of the dissipation of the monovalent cation gradients when cells are placed in tubes containing Ca²⁺-chelating anticoagulants as exemplified in Fig. 25.8C. This experimental result is a showcase of the cation gradient dissipation associated with RBC storage lesions [100].

Fig. 25.9

Proposed mechanisms leading to increased intracellular Ca²⁺ levels in diseased RBCs and accordingly to accelerated clearance of cells from the blood stream. Alternative or cumulating Ca²⁺ entry pathways are highlighted with grey background: increased abundance of NMDA-receptors (NMDAR), e.g., in sickle cell disease, altered activity of Piezo1, e.g. in Hereditary Xerocytosis, increased activity of Gárdos Channel, e.g. in Gárdos Channelopathy, or unspecified Ca²⁺ transport mechanisms. Additionally, ineffective extrusion of Ca²⁺ due to disruption of ATP pools fueling the plasma membrane Ca²⁺ ATPase (PMCA) can contribute. Several downstream processes follow Ca²⁺ overload in RBCs, e.g.: activation of calmodulin by formation of the Ca²⁺-calmodulin complex (Ca-CaM) and activation of calpain, thereby loosening the cytoskeletal structure; activation of the scramblase (Scr) leading to exposure of phosphatidylserine on the outer leaflet of the membrane; activation of the Gárdos channel followed by the efflux of K⁺, Cl⁻ and H₂O and consecutive cell shrinkage. (This figure is reproduced from Hertz et al. 2017 [55])

25.6 Summary

The importance of Ca²⁺ in the membrane transport regulation and mediation of RBCs was early recognised. However, only in the recent years it became evident how this ion transport is related to ion channels and a correlation to molecular entities could be performed. This process is everything but finished and in particular the understanding of the interaction of several ion channels in a physiological context just started. This includes also pathophysiological conditions, on the one hand channelopathies, where a mutation of the ion channel is the direct cause of the disease, like the above described Hereditary Xerocytosis [45] and the Gárdos Channelopathy [31]. On the other hand it applies to RBC related diseases where an altered channel activity is a secondary effect like in sickle cell disease [34, 63] or thalassemia. Also these secondary effects should receive medical and pharmacologic attention because they can be crucial when it comes to the life-threatening symptoms of the disease [55]. An overview of the involvement of Ca²⁺ and Ca²⁺-conducting channels as general components in anaemias is summarised in Fig. 25.9. However, this scheme can only be regarded as a current snapshot of our knowledge about Ca²⁺ and Ca²⁺-conducting channels in RBCs. Further investigations on a better match between functional and molecular knowledge will arise as well as a better understanding of the activity of Ca²⁺ and Ca²⁺-conducting channels within the signalling networks in RBCs.