Late Na⁺ current produced by human cardiac Na⁺ channel isoform Na_v1.5 is modulated by its β₁ subunit

Abstract

Experimental data accumulated over the past decade show the emerging importance of the late sodium current (I _NaL) for the function of both normal and, especially, failing myocardium, in which I _NaL is reportedly increased. While recent molecular studies identified the cardiac Na⁺ channel (NaCh) α subunit isoform (Na_v1.5) as a major contributor to I _NaL, the molecular mechanisms underlying alterations of I _NaL in heart failure (HF) are still unknown. Here we tested the hypothesis that I _NaL is modulated by the NaCh auxiliary β subunits. tsA201 cells were transfected simultaneously with human Na_v1.5 (former hH1a) and cardiac β₁ or β₂ subunits, and whole-cell patch-clamp experiments were performed. We found that I _NaL decay kinetics were significantly slower in cells expressing α + β₁ (time constant τ = 0.73 ± 0.16 s, n = 14, mean ± SEM, P < 0.05) but remained unchanged in cells expressing α + β₂ (τ = 0.52 ± 0.09 s, n = 5), compared with cells expressing Na_v1.5 alone (τ = 0.54 ± 0.09 s, n = 20). Also, β₁, but not β₂, dramatically increased I _NaL relative to the maximum peak current, I _NaT (2.3 ± 0.48%, n = 14 vs. 0.48 ± 0.07%, n = 6, P < 0.05, respectively) and produced a rightward shift of the steady-state availability curve. We conclude that the auxiliary β₁ subunit modulates I _NaL, produced by the human cardiac Na⁺ channel Na_v1.5 by slowing its decay and increasing I _NaL amplitude relative to I _NaT. Because expression of Na_v1.5 reportedly decreases but β₁ remains unchanged in chronic HF, the relatively higher expression of β₁ may contribute to the known I _NaL increase in HF via the modulation mechanism found in this study.

Introduction

Experimental data accumulated over the past decade show the emerging importance of the late sodium current (I _NaL) for the function of both normal and, especially, failing myocardium, in which I _NaL is reportedly increased [1–3] The importance of the contribution of I _NaL to heart failure (HF) mechanisms has been demonstrated in experiments in which “correction” of I _NaL in failing cardiomyocytes resulted in:

• 1. rescue of normal repolarization;

• 2. reduced beat-to-beat action potential duration (APD) variability; and

• 3. improvement of Ca²⁺ handling and contractility [1, 3–5].

Accordingly I _NaL has recently emerged as a novel possible target for cardioprotection to treat the failing heart [6, 7].

Voltage-clamp studies have identified several types of single Na⁺ channel activity and whole-cell Na⁺ currents that could contribute to APD in cardiomyocytes. The variety of Na⁺ channel activities identified so far has been classified (for review see Ref. [6]) in terms of the late (or persistent) Na⁺ current i.e. I _NaL (or I _pNa), and background Na⁺ currents. In contrast with I _NaL, background Na⁺ currents have been poorly characterized and have no clear molecular identity.

Major biophysical and pharmacological characteristics of the whole-cell I _NaL have been studied in great detail in human cardiomyocytes by our research group [3, 8, 9] and can be summarized as follows:

• 1. potential-independent slow inactivation and re-activation (~0.5 s);

• 2. steady-state activation and inactivation similar to that for I _NaT; and

• 3. low sensitivity to the specific toxins TTX and STX similar to the cardiac Na⁺ channel isoform Na_v1.5.

A slowly inactivating I _NaL with aforementioned biophysical characteristics has been identified in ventricular cardiomyocytes and cardiac Purkinje cells of dogs [1, 3, 5, 10–12], guinea pigs [13–15], rabbits [16], rats [17] and mice [18]. I _NaL is also produced by the heterologously expressed cardiac Na⁺ channel isoform main α-subunit Na_v1.5 [7, 19].

Despite explosive interest in this new component of the Na⁺ current (for recent reviews see Refs. [6, 7, 14, 20]) the mechanisms of I _NaL regulation in normal heart and its alterations in HF are not yet understood and are likely to need further collective efforts based on different approaches including detailed biophysical and molecular biology examinations in addition to traditional pharmacological studies. Utilizing antisense inhibition and siRNA technologies our most recent studies explored the molecular identity of I _NaL in ventricular cardiomyocytes [7, 21]. These studies suggested the cardiac Na⁺ channel α-subunit isoform (Na_v1.5) was a major contributor to I _NaL.

Although most recent studies have shown that I _NaL is strongly and differently modulated by intracellular Ca²⁺ in the cardiomyocytes of normal and failing hearts [18, 22], Na⁺ channels operate not in isolation but within macromolecular complexes [23, 24], which are critical attributes of Na⁺ channel function (in addition to membrane voltage and ion concentrations). The macromolecular complexes include auxiliary β-subunits, phospholipids and elements of the cytoskeleton, each of which can modulate Na⁺ channel function including I _NaL (for review see Ref. [7]). The β-subunit gene family has four members—β₁ (SCN1B), β₂ (SCN2B), β₃ (SCN3B), and β₄ (SCN4B) (for review see Ref. [24]). Despite high homology between β₁ and β₃, and β₂ and β₄ the different functional role of these newly discovered isoforms (β₃ and β₄) could not be ruled out. In addition there is a splice variant β₁A of SCN1B that is expressed in embryonic brain and adult heart in rat [25]. All five β-subunits are expressed in rodent heart and are differently localized to specific sub-cellular domains and cell types. The β₁ subunit is non-covalently attached to the α subunit, and the β₂ subunit is covalently linked to the α subunit by a disulfide bond [26]. Numerous studies indicate a possible role of β auxiliary subunits in modulating Na⁺ channel expression and function (for review see Ref. [24]), but the possible implication of β-subunits in I _NaL modulation has not been studied in detail, especially in HF. Our previous studies using the canine chronic HF model showed that in the state of HF the protein level of Na_v1.5 is reduced but remains unchanged for β₁ and β₂ subunits, making these β subunits relatively upregulated [27].

Thus, an intriguing possibility could be that differential expression of α- and β-subunits in normal and failing hearts can contribute, at least in part, to I _NaL alterations observed in HF. Accordingly, in this study, using a heterologous expression system, we specifically tested whether I _NaL is modulated by β₁ or β₂ co-expression with Na_v1.5. Our experiments show that β₁ substantially and significantly affects I _NaL, whereas the effects of β₂ were insignificant. More specifically, the β₁ subunit modulates I _NaL by two mechanisms one of which slows I _NaL decay and the other of which increases I _NaL amplitude relative to the peak transient (I _NaT) current.

Methods

Human kidney epithelial cells tsA201 were transiently transfected by Na_v1.5 alone and/or simultaneously with human cardiac β₁ or β₂ subunits (tagged by GFP or fivefold excess of β-encoding cDNA, as previously suggested [28]). We chose tsA201 cells because there was no expression of endogenous α or β₁ and β₂ subunits compared with that of some HEK293 cell lines (our unpublished data). Whole-cell recordings were made using the conventional patch-clamp technique. The heterologous expression system provides an unique possibility of measuring the peak transient current (I _NaT) and I _NaL at the same physiological [Na⁺]_o, = 140 mM. This is impossible in cardiomyocytes because of voltage-control problems. The I _NaL amplitude was measured as averaged current within 200–220 ms after the onset of depolarization (2 s duration) and was normalized to the peak I _NaT (see examples in Fig. 3a, b). The I _NaL decay was evaluated by the single-exponential fit starting 200 ms after the onset of depolarization, as previously suggested [8]. The data points of the peak current in the current–voltage relationships were fitted to the function [29]:

$$ I_{\text{Na}} \left( {V_{\text{t}} } \right) = G_{ \max } * \left( {V_{\text{r}} - V_{\text{t}} } \right)/\left( { 1+ { \exp }\left( {\left( {V_{{\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{G}}}} - V_{\text{t}} } \right)/k_{\text{G}} } \right)} \right) $$

(1)

where G _max is a normalized maximum Na⁺ conductance, V _r is a reversal potential, V _t is the testing voltage, and V _1/2G and k_G are the midpoint and slope of the respective Boltzmann functions underlying the NaCh steady-state activation (SSA). The data points were fitted to model equations by use of nonlinear regression (Origin 7.0 software, Microcal Software, MA, USA).

The steady-state availability (or inactivation, SSI) terms (V _1/2A, the midpoint and k _A, the slope of the relationship) were measured by a standard double-pulse procedure with the 2-s-duration pre-pulse (V _p) ranging from −130 to −40 mV. The data points of I _NaT normalized to I _NaT measured at −30 and −130 mV prepulse were fitted to a Boltzmann function A (V _p):

$$ A\left( {V_{\text{p}} } \right) = 1/\left( { 1+ { \exp }\left( {\left( {V_{\text{p}} - V_{{\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{A}}}} } \right)/k_{\text{A}} } \right)} \right). $$

(2)

Full details can be found in an extended “Materials and methods” section in the online supplement linked with the online version of this article.

Results

The reporter GFP gene expression monitored the transient transfection of the β-subunits (Fig. 1). Evidently both β₁ (Fig. 1a) and β₂ (Fig. 1b) can be detected both perinuclear and, importantly, in membrane compartments of the cells in the optical slices under the confocal microscope. Our confocal imaging thus validated the transfection procedure, ensuring that β subunits are located in the cell membrane together with functional Na channel α subunits. It is also important to note that we did not make any attempt to quantify the β-subunits expression level based on these results because our objective was to study β subunit effects on I _NaL in all-or-none fashion.

Fig. 1

Confocal microscopy of the live tsA201 cells transfected with Na_v1.5 + β₁ (a) or Na_v1.5 + β₂ (b). Fluorescence of GFP linked to β subunit C-terminus at the cell membrane is evident for both β subunits. Optical slices were 0.5 μm (Zeiss Axiovert 100, Bio-Rad MRC 1024, excitation/emission wavelength 488/522 nm, laser power 10%)

To carefully evaluate I _NaL, we averaged 20–50 original current traces, and the “zero” current (obtained after TTX 25 μM application) was subtracted from the current traces (typical examples are shown in Fig. 2). We determined effects of β subunits on I _NaL kinetics and I _NaL relative amplitude compared with that of I _NaT (Fig. 3). We found that co-expression of α with the β₁ subunit but not the β₂ subunit significantly slows I _NaL decay (Fig. 3a) and substantially increases the relative amplitude of I _NaL (Fig. 3b). Accordingly α + β₂ can be interpreted as a mock control for these studies. The effect of the β₁ subunit on these I _NaL values was independent of the reporter GFP gene. This is evident from comparison of the β₁-GFP construct with β₁ alone (Fig 3b, c). Furthermore, the β₁ subunit caused a significant rightward shift of the SSI curves for both peak I _NaT and I _NaL (measured at 200 ms; not shown). Representative examples of I _NaT are shown in Fig. 4, statistics for SSI data are given in Table 1. The current–voltage relationship and the steady-sate activation values remained unchanged (Fig. 5, Table 1).

Fig. 2

Experimental approach used to elucidate I _NaL of heterologously expressed Na_v1.5. a, b Typical examples of the I _Na currents recorded in tsA201 cells transiently transfected by α + β₁-GFP. Shown are averaged traces from 50 sweeps before and after TTX (25 μM) application to assess “zero” current. c Difference I _NaL current obtained by subtraction of “zero” current. Note different currents and time scales in a, b, and c to demonstrate peak (I _NaT) and I _NaL. V _h = −120 mV, V _m = −30 mV, 23°C

Fig. 3

Effects of the β subunits on the I _NaL produced by the heterologously expressed human cardiac NaCh isoform α subunit (Na_v1.5). a Representative examples of superimposed current traces recorded in tsA201 cells transfected with cDNA encoding human cardiac NaCh α-subunit alone or together with β₁ (α + β₁) or β₂ (α + β₂). Shown are averaged (10–20) currents with a single-exponential fit to I _NaL decay (solid lines) starting 200 ms after the onset of depolarization. The time constant (τ) values are given in the panel. The current amplitudes are relative to the peak I _NaT. The voltage-clamp procedure given in the inset. b, c Statistical analysis of the I _NaL/I _NaT ratio (b) and the decay time course (c) changes in response to coexpression of α with β subunits. The statistically significant difference (P) in panels b and c was evaluated by ANOVA followed by the Bonferroni’s post hoc test. Bars in panels b and c represent means ± SE, n number of cells. There was no significant difference between α + β₁-GFP compared with α + β₁ (1:5 cDNA ratio) or for α alone compared with α + β₂-GFP for both b and c panels

Fig. 4

Effect of β₁ subunit on the steady-state inactivation (SSI) voltage-dependency of the heterologous expressed Na_v1.5. a, b Representative raw current recordings at −30 mV in response to a different 2-s-long prepulse potential (voltage procedure is shown in c). c Data points of the relative (normalized to maximum) peak I _NaT measured at −30 mV and plotted against prepulse voltage (V _p). Solid lines represent the Bolzmann fit (Eq. 2 in “Methods”), and values are given on the plots. Statistics for values evaluated in numerous cells are given in Table 1. V _h = −140 mV, [Na]₀ = 140 mM, 24°C

Table 1 Steady-state activation (SSA) and inactivation (SSI) data for I _Na in the heterologously expressed cardiac Na_v1.5 (α) without or with its auxiliary β subunits

Fig 5

Voltage–current relationship for heterologous expressed Na_v1.5 (α alone a, or with the β₁ subunit b). a, b Left panels raw current recordings at different membrane potentials, right panels data point plots, with the fits (solid lines) to Eq. 1 (“Methods”). The equation is given in a, and fit values for the steady-state activation are indicated on the plots in a and b. The statistical data for the steady-state activation are given in Table 1

Discussion

For the first time we demonstrate that the β₁ subunit can significantly modulate I _NaL produced by the heterologously expressed Na_v1.5. The modulation includes slowed inactivation, augmented amplitude relative to I _NaT, and rightward SSI shift.

Although β subunits do not form ion-conducting pores, they are important modulators of Na_v function, expression levels at the plasma membrane (trafficking), and cell adhesion [23, 24]. Recent studies support the emerging significance of the β₁ auxiliary subunit in modulation of Na_v1.5 function. It has been shown that the β₁-subunit:

• 1. is involved in abnormal NaCh activity associated with the LQT3 mutation [30];

• 2. aggravates NaCh dysfunction in Brugada syndrome [31];

• 3. modifies block of NaCh by fatty acids [32] and lidocaine [33]; and

• 4. modulates trafficking of Na_v1.5 [34].

As to modulation of the late Na⁺ channel activity by the β subunits, there are only a few controversial reports. Both β₁ and β₃ subunits exhibit dual and opposite effects on the decay time course of the Na⁺ current produced by heterologously expressed rat brain αIIA (Na_v1.2) [35]. The current decay was assessed by use of two exponential models and the fast time constant, τ ₁, was accelerated whereas the slow time constant, τ ₂, was increased more than twofold (10.9 vs. 25 ms) by β₁ and β₃, respectively. In line with this finding it has been shown that expression of the β₁ subunit increased slowly, inactivating current (called persistent current by the authors) produced by the new epilepsy-related Na_v1.1 mutant D1866Y [36]. These data are in line with our findings reported here.

Transient expression of Na_v1.5 into the HEK293 cell line stably expressing β₁ subunit reduced a non-inactivating current component measured 750 ms after depolarization onset [37]. A similar non-inactivating Na⁺ current has been also reported in rat cardiomyocytes; it was highly TTX-sensitive and could be augmented by hypoxia or cyanides [38, 39]. In both cases, when the non-inactivated current was measured, the authors used intracellular solutions deprived of ATP and containing artificial (non-physiological) anion fluoride. It has long been known that fluoride can retard Na⁺ channel inactivation significantly, as was shown in internally perfused axons [40], and in cardiac cells [41]. Metabolic regulation of Na⁺ channel inactivation and its rundown was demonstrated in neonatal cultured rat cardiomyocytes [42]. Therefore, this non-inactivating I _Na component is likely to be related to these non-physiological experimental conditions. In contrast, our current recordings were performed in close-to-physiological conditions (physiological [Na⁺] and with no fluoride) and they did not reveal any non-inactivating current either in heterologously expressed Na_v1.5 clone (see Fig. 2, and Ref. [19]) or in cardiomyocytes from human hearts [3, 8]. In addition to the non-inactivating current discussed above, a background Na⁺ current that could be recorded at very negative membrane potentials (−120 mV) has been reported in rabbit cardiac Purkinje cells and ventricular cardiomyocytes [43]. In contrast, recent studies [11, 12] show that canine cardiac Purkinje cells exhibit slowly inactivating I _NaL, similar to that described in ventricular cardiomyocytes of humans and dogs [3, 8]. Unlike background Na⁺ currents, this I _NaL in Purkinje cells possesses steady state inactivation and is not activated at resting membrane potentials. Furthermore, non-inactivating Na⁺ current was not present in human cardiomyocytes [3, 8]. Our single-channel data, in fact, excluded the presence of the non-inactivating component under close-to-physiological conditions for both Na_v1.5 clone and for human cardiomyocytes [2, 19]. Thus, the non-inactivating or background Na⁺ currents are more likely to be species-dependent and/or were recorded in the presence of artificial anion fluoride and absence of ATP in the intracellular milieu. The molecular and genetic origins of background currents in cardiac cells remain unknown, but Denis Noble in his recent review [6] suggested that they could result from a leak form of Na⁺–K⁺ ATPase [44] or from NCX [45]. Accordingly, it is important to emphasize that we report here for the first time the modulatory effect of the β₁ subunit on cardiac-type late Na⁺ current I _NaL (also known as persistent Na⁺ current I _pNa) rather than on background non-inactivated currents of yet unknown nature reported in some previous studies.

The potency of the β₁ subunit to modulate I _NaL shown herein has been confirmed in the native cell environment by our preliminary study in normal dog cardiomyocytes in which antisense inhibition of SCN1B significantly accelerated I _NaL decay ([46], see Fig. 8A in Ref. [7]). It has been also shown that at the protein level Na_v1.5 is downregulated whereas β₁ remained unchanged in HF, pointing toward relative higher membrane content of β₁ [27], but the SSI shift was not found in HF [2, 3, 29]. The SSI shift is dependent on variety of factors, including intracellular [Ca²⁺], cytoskeleton, and membrane lipid content, that may affect SSI in different ways negating the β₁-related effect ([22], see review [7]). These data together with the findings of our study suggest a potential mechanism for the contribution of β₁ to HF-related I _NaL alterations [1, 3]. Furthermore, in addition to the I _NaL decay slowing, β₁ can also change I _NaL via its well-known effect of SSI shift. This β₁-induced SSI shift under physiological conditions (i.e. at a resting potential of ~−80 mV) may have profound effect on I _NaL enhancement during action potentials, as more Na⁺ channels operating in late modes become available.

The next important question is how, specifically, the β₁ subunit interacts with Na⁺ channel to produce the observed I _NaL changes. One possibility, however, could be related to the C-terminus (CT). The role of the CT in regulating Na_v1.5 inactivation via the Ca²⁺-calmodulin-dependent interaction with the III–IV linker, responsible for the initial fast inactivation, has recently been suggested [47, 48]. Direct interaction between the cytoplasmic CT domain of Na_v1.1 with β₁ and β₃ has recently been demonstrated [36] and thus provides a possible molecular mechanism for the I _NaL modulation found here.

Single-channel studies in heterologously expressed Na_v1.5 and human cardiomyocytes show that late NaCh activity is arranged in two major gating modes—late scattered mode (LSM) and ‘burst” mode (BM) [19]. Numerical evaluation based on the Markovian chain model revealed that BM + LSM is responsible for the intermediate phase (40–300 ms) whereas LSM is responsible for the ultra-late (>300 ms) phase on I _NaL inactivation [2]. In this study we analyzed the amplitude and decay of I _NaL after 200 ms, thus presumably the function of LSM mode. Therefore, the β₁ subunit may increase the probability of occurrence of these modes (i.e. make modal switch more probable), thus increasing relative I _NaL/I _NaT. Also β₁ affects the gating kinetics (inactivation) of LSM. The idea of β₁ subunit involvement in the modal switch has previously been suggested for muscular isoform Na_v1.4 [49].

We did not find β₂-related effects on I _NaL in our experimental setting. The role of this subunit in physiological function of NaCh is not clear. Initially the β₂ subunit was implicated in intercellular adhesion and recruitment of the cytoskeletal protein ankyrin to the plasma membrane at sites of cell-to-cell contact [50]. Accordingly the direct role of the β₂ subunit on NaCh gating was not suggested, because NaCh protein has direct attachment to the sub-membrane cytoskeleton via ankyrin (for a review see Ref. [48]) and may directly be related to the cytoskeleton-dependent effects on Na⁺ channel gating [51, 52]. It has been recently demonstrated that β₁/β₂ chimeras may cause the additional closed state for Na_v1.5 that is accessible at hyperpolarized potentials, although this effect was not produced by the wild type β₂ subunit [53]. Given the multi-modal origin of Na_v1.5-related I _NaL it is not obvious how this subunit can play a role in the kinetic transition constants assuming the new closed state [2]. Although our experiments do not reveal apparent effects on the I _NaL properties studied, we cannot exclude the possibility that the β₂ subunit can still affect late Na⁺ channel openings taking into account a possible role of neuronal isoforms (for a review see Ref. [7]).

We conclude that the auxiliary subunit β₁ modulates I _NaL, produced by the human cardiac Na⁺ channel Na_v1.5 by slowing its decay and increasing I _NaL amplitude relative to I _NaT. Because expression of Na_v1.5 reportedly decreases but β₁ remains unchanged in chronic HF, the relatively higher expression of β₁ may contribute to known I _NaL increase in HF via the modulation mechanism found in this study.