Na⁺ Transport in Cardiac Myocytes; Implications for Excitation-Contraction Coupling

Abstract

Intracellular Na⁺ concentration ([Na⁺]_i) is very important in modulating the contractile and electrical activity of the heart. Upon electrical excitation of the myocardium, voltage-dependent Na⁺ channels open, triggering the upstroke of the action potential (AP). During the AP, Ca²⁺ enters the myocytes via L-type Ca²⁺ channels. This triggers Ca²⁺ release from the sarcoplasmic reticulum (SR) and thus activates contraction. Relaxation occurs when cytosolic Ca²⁺ declines, mainly due to re-uptake into the SR via SR Ca²⁺-ATPase and extrusion from the cell via the Na⁺/Ca²⁺ exchanger (NCX). NCX extrudes one Ca²⁺ ion in exchange for three Na⁺ ions and its activity is critically regulated by [Na⁺]_i. Thus, via NCX, [Na⁺]_i is centrally involved in the regulation of intracellular [Ca²⁺] and contractility. Na⁺ brought in by Na⁺ channels, NCX and other Na⁺ entry pathways is extruded by the Na⁺/K⁺ pump (NKA) to keep [Na⁺]_i low. NKA is regulated by phospholemman, a small sarcolemmal protein that associates with NKA. Unphosphorylated phospholemman inhibits NKA by decreasing the pump affinity for internal Na⁺ and this inhibition is relieved upon phosphorylation. Here we discuss the main characteristics of the Na⁺ transport pathways in cardiac myocytes and their physiological and pathophysiological relevance.

[Na⁺]_i during cardiac cycle and its role in excitation-contraction coupling

Mammalian cells maintain a large electrochemical gradient of Na⁺ across the plasma membrane. Intracellular Na⁺ concentration ([Na⁺]_i) is kept low (4-14 mM) by the Na⁺/K⁺-ATPase (NKA) at the expense of metabolic energy, while extracellular [Na⁺] is ∼140 mM. In neurons and muscles, the energy stored in the transmembrane [Na⁺] gradient provides the basis for fast electrical signaling, i.e. propagation of action potentials. Many cells utilize the electrochemical Na⁺ gradient to couple energetically unfavorable transmembrane solute flow to Na⁺ transport (secondary active transport). Examples include Na⁺/neurotransmitter, Na⁺/glucose and Na⁺/amino acids cotransporters as well as various Na⁺/ion co- and counter-transporters.

In the heart, [Na⁺]_i is very important in modulating intracellular Ca²⁺, and thus contractility, through the Na⁺/Ca²⁺ exchanger (NCX). The exchanger is the main pathway for Ca²⁺ extrusion from cardiac myocytes and therefore is essential in excitation-contraction coupling. NCX can operate in both Ca²⁺ efflux and Ca²⁺ influx (or reverse) mode, depending on the internal and external concentration of both Na⁺ and Ca²⁺, as well as on the membrane potential. High [Ca²⁺]_i favors Ca²⁺ efflux whereas positive membrane potential and high [Na⁺]_i favor Ca²⁺ influx. In normal myocytes, under physiological conditions NCX works almost exclusively in the Ca²⁺ extrusion mode, driven mostly by the high subsarcolemmal Ca²⁺ transient (1). However, elevated [Na⁺]_i would shift the balance of fluxes through NCX to favor more Ca²⁺ influx and less Ca²⁺ efflux, resulting in larger Ca²⁺ transients and therefore enhanced contractility. Indeed, the inotropic effect of increasing [Na⁺]_i is well known and is the basic mechanism behind cardiac-glycosides induced inotropy. In cardiac Purkinje fibers, the developed tension doubles with a 1 mM increase in Na⁺ activity (2). The effect is smaller but still prominent in ventricular myocytes (3).

During electrical excitation of the heart, voltage-dependent Na⁺ channels open, triggering the upstroke of the action potential (AP; Fig. 1). Na⁺ channels inactivate rapidly (within a few milliseconds) at positive potentials, which limits the gain in intracellular Na⁺ ([Na⁺]_i) to 6-15 μmol/L (4). During the AP, Ca²⁺ enters the myocytes via L-type Ca²⁺ channels, which triggers Ca²⁺ release from the sarcoplasmic reticulum (SR). This raises the free intracellular [Ca²⁺] ([Ca²⁺]_i), allowing Ca²⁺ to bind to the myofilaments and trigger contraction. For relaxation to occur, Ca²⁺ must be taken out of the cytosol, mainly via the SR Ca²⁺-ATPase (SERCA), which takes Ca²⁺ back into the SR, and NCX (Fig. 1). NCX transports one Ca²⁺ ion in exchange for three Na⁺ ions, thus Ca²⁺ extrusion leads to an increase in [Na⁺]_i by ∼32 μmol/L during each AP (Fig. 1). This is required to extrude the ∼10 μmol/L Ca²⁺ that enters via L-type Ca²⁺ channels. The Na⁺/H⁺ exchanger brings in ∼2 μmol/L Na⁺ at physiological intracellular pH and ∼16 μmol/L during intracellular acidosis (see 4). Other transporters, including the Na⁺/HCO₃^- cotransporter, the Na⁺/K⁺/2Cl cotransporter and the Na⁺/Mg²⁺ exchanger bring in smaller amounts of Na⁺. At steady-state, the excess Na⁺ (∼40-45 μmol/L) is then extruded by NKA to keep [Na⁺]_i constant (Fig. 1).

Figure 1.

Na⁺ and Ca²⁺ transport during cardiac cycle in ventricular myocytes. Inset shows the integrated Na flux through NCX, voltage-gated Na⁺ channels and NKA during an action potential (AP). The corresponding Ca²⁺ transient is also shown. NCX - Na⁺/Ca²⁺ exchanger; ATP - ATPase; PLB - phospholamban; PLM – phospholemman; SR - sarcoplasmic reticulum; NHE – Na/H exchanger. Modified from Bers (2001).

Resting [Na⁺]_i in cardiac myocytes is in the range of 4-8 mM for most mammalian species, including human, and higher (10-15 mM) in rat and mouse (4). [Na⁺]_i increases during stimulation in a frequency-dependent manner, because, as detailed above, Na⁺ influx is larger due to more frequent activation of Na⁺ channels and NCX. [Na⁺]_i is elevated in hypertrophy and heart failure (HF) (3,5,6) and this may limit the contractile dysfunction observed in these pathophysiological conditions.

Voltage-gated Na⁺ channels

The cardiac voltage-gated Na⁺ channel is composed of a pore-forming α subunit Na_v1.5 (∼240 kDa) and auxiliary β subunits (∼30-35 kDa, β1-β4 subunits; (7)). The α subunit consists of four repeat domains (I-IV), each containing six transmembrane segments (S1-S6) and one membrane reentrant domain, connected by internal and external polypeptides loops (7). The S4 segments contain 4-8 positively charged residues, which serve as voltage sensors. An outward movement of these charges under depolarization underlies the activation of the channel (8). Inactivation is mediated by the short intracellular loop connecting domains III and IV. One α subunit is associated with one or two auxiliary β subunits. These auxiliary subunits modulate channel gating, interact with extracellular matrix and play a role as cell adhesion molecules (9). Na_v1.5 is much less sensitive to inhibition by tetrodotoxin (K_0.5 ∼ 2 μM) and is more sensitive to Cd²⁺ block than neuronal or skeletal muscle Na⁺ channels. There may also be neuronal Na⁺ channels in cardiac myocytes (10) but their contribution to the total Na⁺ current is small (<15%) and their role remains controversial.

Most cardiac Na⁺ channels inactivate within a few milliseconds. However, a small percentage of channels either do not close, or close and then reopen, generating a slowly inactivating, persistent Na⁺ current, I_Na,slow (11,12). This persistent Na⁺ current is more sensitive to block by tetrodotoxin, is activated at more negative potentials and has an amplitude up to 0.5% of the peak transient I_Na. Despite the small amplitude, I_Na,slow may prolong the AP and generate a substantial Na⁺ influx. Some mutations in the Na⁺ channel gene SCN5A result in an increased I_Na,slow, which leads to long QT syndrome (LQT3). An increased I_Na,slow was found in ventricular myocytes from a canine ischemic model of HF (13) and more recently in dogs with pacing-induced HF (14). The latter study also reports an increased slowly inactivating Na⁺ current in human HF. Thus, I_Na,slow might be partly responsible for the prolongation of the AP and the rise in [Na⁺]_i that occur in HF. We have found that a tetrodotoxin-sensitive pathway is responsible for the increased [Na⁺]_i in a rabbit HF model (5). Genetic (15) or pharmacologically-induced (16) increases in I_Na,slow were shown to result in delayed afterdepolarizations and triggered arrhythmias. Moreover, ranolazine, an I_Na,slow inhibitor, reduced [Na⁺]_i and diastolic Ca²⁺ overload and improved diastolic function in failing human hearts (17). Interestingly, Wagner et al. (18) found that overexpression of CaMKII slows fast I_Na inactivation, enhances I_Na,slow and increases [Na⁺]_i in cardiac myocytes. CaMKII also slows recovery from inactivation and reduces steady-state channel availability at physiological diastolic membrane potential (18). Thus, CaMKII causes both gain and loss of Na⁺ channels function in a manner similar to that produced by a human mutant Na⁺ channel (Asp insertion at 1795 in the C-terminus), which causes both long-QT (at low heart rates) and Brugada syndrome (at high heart rates) (19). CaMKII expression and activity is augmented in HF (20,21). Thus, the acquired Na⁺ channel dysfunction may affect millions of patients with heart failure.

In addition to normal biophysical properties, Na⁺ channels activity requires proper sarcolemmal localization. Immunofluorescence measurements revealed that Na_v1.5 channels are located both in the T-tubules (TT) and external sarcolemma (ESL) (22,23), although in some studies they were found mostly at the intercalated disks (23). Functional studies indicated a relatively uniform Na_v1.5 distribution between the TT and ESL (24). It has been recently shown that ankyrin-G, an adapter protein that links membrane proteins to the cytoskeleton, plays an important role in the proper membrane targeting of Na⁺ channels. Lowe et al. (25) showed that reduced ankyrin-G expression decreases total cellular Na_v1.5 expression and its efficient membrane localization, as well as I_Na. Moreover, a mutation in the ankyrin-binding motif of Na_v1.5 that abolishes the binding of ankyrin-G has been shown to induce Brugada syndrome (26).

Na⁺/Ca²⁺ exchanger

NCX1 is the only NCX isoform present in the heart and consists of 970 amino acids (∼110 kDa). Topological data suggest that NCX1 has nine transmembrane segments (27,28) and a large intracellular loop (∼550 amino acids) between TMS 5-6, with the N- and C-termini located on the external and internal sides, respectively. There are also two regions of internal repeats (α1 and α2) facing opposite sides of the membrane. The α1 repeat comprises part of the TMS 2 and 3 and the reentrant loop between them while the α2 repeat contains part of the TMS 7, the cytosolic loop connecting it to the TMS 8 and part of the TMS 8. Cross-linking experiments indicate that the α repeats are near one another in the folded protein, suggesting that they may contribute to the ion conduction pathway (29,30). The large cytosolic loop contains two separate Ca²⁺ binding domains CBD1 and CBD2 (also known as β-repeats) that are essential for the allosteric regulation by Ca²⁺ (see below and Refs. (31-33)). There is also a charged segment of 20 residues (219-238) at the N-end of the cytosolic loop (XIP) which resembles a calmodulin binding domain and may be autoinhibitory (since a peptide based on this sequence inhibits NCX activity).

Besides being transport substrates, intracellular Ca²⁺ and Na⁺ exert important modulatory effects on NCX activity. Increases in cytosolic Ca²⁺ allow Ca²⁺ binding to the CBD domains, which increases the activity of the exchanger (allosteric [Ca²⁺]_i-dependent activation). Weber et al. (34) found a physiologically relevant K_m(Ca) of 125 nM in intact ferret ventricular myocyte during dynamic [Ca²⁺]_i changes. Allosteric Ca²⁺-activation develops rapidly (within tens of ms). However, once activated, NCX does not revert to an inactivated state on return of [Ca²⁺]_i to rest levels for tens of seconds (35,36). This suggests that Ca²⁺-dependent activation may only gradually change during alterations in heart rate, but it is undoubtedly substantially activated during basal heart rates. High [Na⁺]_i inhibits NCX (Na⁺-dependent inactivation) in a time and [Na⁺]_i-dependent manner. Because of the high [Na⁺]_i (> 30 mM) required, Na⁺-dependent inactivation may not be very important under normal physiological conditions. However, it could prevent excess Ca²⁺ influx and cellular Ca²⁺ overload under pathophysiological conditions of high [Na⁺]_i, such as ischemia/ reperfusion, where net Ca²⁺ influx via NCX might be strongly favored. [Na⁺]_i-dependent inactivation is eliminated by an increase in membrane PIP₂, which interacts with the XIP region of the exchanger (37). Recent evidence indicates that PIP₂ levels may also modulate NCX by affecting its membrane trafficking (38).

NCX is concentrated at the T-Tubules (39-41), however it is not known how much is at the junctions with the SR. Functional data indicate that Ca²⁺ entry via reverse-mode NCX can trigger Ca²⁺ release from the SR at least under certain conditions (42-44). However, only a small NCX fraction seems to co-localize with proteins specific to the dyadic cleft, i.e. L-type Ca²⁺ channels and RyRs (45).

Na⁺/K⁺-ATPase

NKA extrudes three Na⁺ ions in exchange for two K⁺ ions using the energy derived from the hydrolysis of one ATP molecule, and thus moves out one net charge per cycle. NKA has two essential subunits: α and β. The NKA-α subunit (∼ 110 kDa) contains the binding sites for Na⁺, K⁺, ATP and cardiac glycosides and its crystal structure has been recently uncovered (46). NKA-α has 10 transmembrane domains with the N- and C- termini located on the cytosolic side of the membrane. The extracellular loops are short, with the exception of the TM7 - TM8 loop. NKA-α has a large intracellular loop between TM4 and TM5, composed of about 430 amino acid residues, a long N-terminal tail of about 90 amino acid residues, and an intracellular loop of about 120 residues between TM2 and TM3. The smaller β subunit (∼50 kDa, one transmembrane domain) has a role in the processing and in the proper membrane insertion of the pump.

There are three α (α1-α3) and three β (β1-β3) NKA subunit isoforms in the heart, with any αβ combination resulting in a functional pump. NKA-α1 is present in the heart of all species studied, while the expression of NKA-α2 and NKA-α3 differs significantly between species (47). For instance, the fetal and neonatal rodent hearts express NKA-α3, and this is replaced by NKA-α2 early in development (48). Interestingly, the reverse switch occurs in hyperthrophied or failing rat heart (49,50). All three NKA-α isoforms can be detected in the human heart (51).

It has been suggested that different NKA isoforms may function differently in the cell, depending on specific localization in the membrane. Thus, it was proposed that NKA-α2 and NKA-α3 are located mainly in the T-tubules, at the junctions with the SR, where they could regulate local NCX and [Ca²⁺]_i. There is rather convincing evidence supporting such a model in smooth muscle (52). However, things are less clear in the heart. James et al. (53) showed that mouse hearts with genetically reduced levels of NKA-α2 were hypercontractile as a result of increased Ca²⁺ transients. In contrast, hearts with reduced levels of NKA-α1 were hypocontractile. Moreover, inhibition of NKA-α2 with ouabain increased the contractility of heterozygous NKA-α1 hearts. These data indirectly implicate a selective involvement of NKA-α2 in cardiac myocyte Ca²⁺ regulation. Further support for this idea came from the observation that increased expression of NKA-α2 (at the expense of NKA-α1) decreased the reverse-mode Na⁺/Ca²⁺ exchange current and Ca²⁺ transients in mouse ventricular myocytes (54). Dostanic et al. (55) showed that reducing the ouabain affinity of mouse NKA-α2 leads to a loss of glycoside-induced inotropy, which suggests that NKA-α2 regulates local [Ca²⁺]_i near the Na⁺/Ca²⁺ exchange, as in smooth muscle. However, the same authors later found that NKA-α1 is also functionally and physically coupled with NCX in cardiac myocytes (56).

The differential NKA-α1/NKA-α2 localization (TT vs. ESL) in cardiac myocytes is controversial. Using immunofluorescence, McDonough et al. (47) found that NKA-α1 is preferentially distributed in the TT, while NKA-α2 was uniformly distributed in the TT and ESL in the rat. Conversely, Silverman et al. (57) showed that in guinea-pig ventricular myocytes NKA-α2 is mainly in TT and NKA-α1 is predominantly on the ESL. Recent studies using formamide-induced detubulation of ventricular myocytes and measurements of dose-dependent inhibition of NKA current by ouabain showed that the functional density of NKA-α2 is significantly higher (about 4-fold) in the TT (vs. ESL) while NKA-α1 is more uniformly distributed in cardiac myocytes from both rats (58,59) and mice (60). Moreover, we (58) found that the amount of NKA-α1 and NKA-α2 in the TT is comparable. This raises the intriguing possibility that NKA-α2 is indeed localized to sarcolemmal junctions with the SR (48% of the TT area and ∼8% of ESL in rat ventricle, as measured by electron microscopy; Ref (61)) and could be important in regulating local cleft [Na⁺]_i and [Ca²⁺]_i and therefore excitation-contraction coupling as in smooth muscle. However, as discussed above this hypothesis requires further structural and functional tests in adult ventricular myocytes.

NKA regulation by phospholemman

Phospholemman (PLM) is a small (72 amino acids), single membrane-spanning sarcolemmal protein that has long been known as a major phosphorylation target for PKA and PKC in the heart (62,63). However, its physiological role was poorly understood. Initially it was thought that PLM forms ion channels selective for taurine and thus might be involved in cell volume regulation (64). More insight came with the finding that PLM belongs to a family of proteins that associate with and modulate NKA in various tissues (65). The protein family, named FXYD after a conserved Pro-Phe-X-Tyr-Asp motif in the extracellular N-terminus domain, also includes the NKA γ-subunit found in the kidney. PLM is the only FXYD protein highly expressed in the heart and is also unique in the family in having multiple phosphorylation sites at its cytosolic carboxyl terminus.

Recent evidence shows that PLM modulates NKA the same way phospholamban (PLB) modulates SERCA, a P-type pump closely related to NKA. That is, unphosphorylated PLM inhibits NKA, mostly by reducing its affinity for internal Na⁺ (57,66-68) and PLM phosphorylation relieves this inhibition (67-68). β-adrenergic agonists had no effect on NKA function in myocytes from PLM-KO mice (67). This indicates that phosphorylation of PLM, rather than direct NKA phosphorylation, mediates the enhancement of NKA function during sympathetic stimulation of the heart. This is in agreement with data showing that NKA-α subunit can be phosphorylated by PKA, but only in the presence of detergents, while in situ the phosphorylation site may be inaccessible to the kinase (69).

Interestingly, the physical association between PLM and NKA-α, as determined by co-immunoprecipitation experiments, appears to be unaffected by PKA phosphorylation (70,71). However, when fluorescence resonance energy transfer (FRET) was used to look for more subtle changes in the NKA-PLM association, robust FRET was observed at baseline and the FRET was dramatically decreased upon PLM phosphorylation by either PKA or PKC (72). Thus, PLM phosphorylation changes the PLM-NKA interaction but does not result in a complete dissociation. In the analogous PLB-SERCA system it was long thought that PLB phosphorylation caused PLB to dissociate from SERCA, but recently this has been challenged (73).

We have shown recently (74) that NKA activation, mediated by phosphorylation of PLM, limits the rise in [Na⁺]_i and Ca²⁺ transient amplitude during β-adrenergic stimulation in cardiac myocytes. Enhancement of NKA activity may thus be an integral part of the sympathetic fight or flight response of the heart, by enhancing Na⁺ extrusion to better keep up with the higher level of Na⁺ influx (that is caused by the combined inotropic and chronotropic effects on the heart). We found (74) that the inability of NKA to be activated by β-adrenergic stimulation in the PLM-KO mouse leads to excessive elevation of [Na⁺]_i during sympathetic activation, which has both the benefits and the risks associated with NKA inhibition by cardiac glycosides (inotropy, but enhanced arrhythmogenesis). Thus, the physiological role of PLM may be to prevent Ca²⁺ overload and triggered arrhythmias by limiting the rise of [Na⁺]_i during sympathetic stimulation of the heart.

NKA and NCX regulation by ankyrin-B

Ankyrin-B is another member of the ankyrin family of adaptor proteins responsible for the proper localization of various membrane proteins at specialized membrane domains. In humans, loss-of-function mutations in ankyrin-B are associated with cardiac phenotypes including QT interval prolongation (LQT4 syndrome), bradycardia, idiopathic ventricular fibrillation and catecholaminergic polymorphic ventricular tachycardia (for a review, see Ref 75). Ankyrin-B has long been known to bind and target NKA to the membrane (76). More recently, it has been shown that the proper membrane localization and post-translational stability of NCX also require direct interaction with ankyrin-B (77). A decrease in ankyrin-B function resulted in reduced NKA and NCX targeting to the T-Tubules as well as reduced overall NKA and NCX protein level (78) in cardiac myocytes. This led to altered Ca²⁺ signaling and aftercontractions, which may cause the arrhythmias observed in vivo. However, whether ankyrin-B brings NCX and NKA close to each other in the sarcolemma, how alterations in ankyrin-B level and activity affect [Na⁺]_i, cellular and SR Ca²⁺ cycling and the molecular mechanism leading to arrhythmias are still largely unknown.

Conclusion

The numerous Na⁺ transport pathways are critical individually in regulating cardiac [Na⁺]_i. However, the integrated cellular handling of Na⁺ must be kept in mind in order to understand and predict how alterations in these pathways affect [Ca²⁺]_i, contractility and electrical activity of cardiac myocytes.