Can the Sodium-Calcium Exchanger Initiate or Suppress Calcium Sparks in Cardiac Myocytes?

Abstract

Positive feedback of Calcium (Ca)-induced Ca release is the mechanism of Ca spark formation in cardiac myocytes. To initiate this process, a certain amount of Ca in the cleft space is necessary. When the membrane potential becomes higher during excitation-contraction coupling, Ca can enter through both Ca current (I_CaL) and sodium-calcium exchanger (NCX) and may activate ryanodine receptors to initiate a Ca spark. On the other hand, at the resting membrane potential (V_m ∼–80 mV), NCX removes Ca from the cell (forward mode). If Ca released from the sarcoplasmic reticulum is quickly removed via forward mode NCX before Ca-induced Ca release starts, the Ca release becomes nonspark Ca leak. This would also be influenced by the cleft/noncleft distribution of NCX, which is unknown. Using a physiologically detailed mathematical model of subcellular Ca cycling, we analyze how NCX strength and distribution alter Ca spark formation. During excitation-contraction coupling, most Ca sparks are induced by I_CaL with very few due to NCX current. At the resting membrane potential if most NCX is localized to the cleft, spontaneous Ca sparks are significantly reduced.

Sodium-calcium exchange (NCX) is an important regulator of the cardiac action potential, intracellular calcium (Ca), and sodium (Na) concentrations ([Ca]_i and [Na]_i) (1–4). At the resting membrane potential (V_m ∼−80 mV), three Na ions enter through the NCX and one Ca ion leaves the cell (forward mode). However, at more positive V_m, less Ca is removed and, above the reversal potential, Ca enters in exchange for Na (reverse mode). Ryanodine receptors (RyR) are Ca-sensitive channels that form clusters called Ca release units (CRUs) containing several to several hundred RyRs. If a few RyRs open in a CRU, it can trigger more sarcoplasmic reticulum (SR) Ca release from that CRU via positive feedback, forming a Ca spark.

Our recent study (5) showed that a stochastic opening of a single RyR effectively initiates Ca sparks when [Ca]_SR is relatively high. Furthermore, we showed that one or a few L-type Ca channel openings effectively initiate Ca sparks during excitation-contraction coupling (ECC). Reverse mode NCX current is relatively small compared to a single RyR current or L-type Ca current (I_CaL).

Here, we first investigate how effectively reverse mode NCX current initiates Ca sparks during ECC. Second, at resting V_m, if spontaneous Ca released from the SR is quickly removed via forward mode NCX, it could interrupt the positive feedback process, such that the Ca release becomes a nonspark SR Ca leak, thus reducing the propensity for Ca waves and triggered arrhythmias via delayed afterdepolarization. We show how NCX can suppress Ca sparks in this manner. Third, these effects are influenced by the distribution of NCX (cleft versus noncleft), which is unknown. We investigate how the distribution of NCXs can alter Ca spark formation. Fourth, we also show how [Na]_i and NCX strength can alter Ca spark formation at a range of values of intra-SR [Ca] ([Ca]_SR).

In this study, we analyze Ca flux through RyRs, I_CaL, and NCX using a physiologically detailed mathematical model of junctional SR Ca release in which RyR gating is regulated by [Ca]_SR and cleft [Ca] ([Ca]_Cleft). The details of the model are in our previous work (5) and described briefly as follows. Our base model is a recently developed Ca cycling model by Restrepo et al. (6). In this model, there are 100 RyR channels and four functional L-type Ca channels in each CRU. The RyR model is a four-state Markovian model. Each RyR opens stochastically depending on [Ca]_Cleft and [Ca]_SR. They adopted the NCX model from Shannon at al. (7) and the seven-state Markovian model of the L-type Ca channel from Mahajan et al. (8). Although the open probability of the L-type Ca channel at the resting membrane potential is small, to investigate pure NCX effects, I_CaL is removed for diastolic spontaneous Ca spark tests (i.e., see Fig. 3, B–D and F).

Figure 3.

Effect of NCX distribution and [Na]_i on diastolic spontaneous Ca sparks. (A–D) Black curve: all NCX is located in submembrane space. Blue curve: 50% NCX is in submembrane and 50% NCX in cleft. Red curve: all NCX is located in cleft [Na]_i = 12 mM. (A) [Ca]_cleft accumulation when single RyR channel opens at [Ca]_SR = 800 μM (solid) and 600 μM (dashed). (B) Ca spark event ratio (Ca spark events/total leak events) versus [Ca]_SR. (C) Ca spark flux-SR load relationship. (D) Ca spark flux-SR load relationship when NCX conductance is doubled. (E) Forward mode NCX versus [Na]_i at the indicated [Ca]_i values. V_m is –80 mV. (F) Ca spark-load relationship at [Na]_i = 6–16 mM. 11% of NCX is located in the cleft space.

Experimentally it is known that reverse mode NCX current is much smaller than I_CaL and less effective as a trigger of SR Ca release (9). Because Ca accumulation at the cleft space depends on the Ca flux rate, the speed of accumulation and the plateau [Ca]_Cleft is much smaller for NCX (Fig. 1). At a given [Ca]_SR, the probability of opening of a closed RyR depends on [Ca]_Cleft and the time. Typically, each L-type Ca channel opens only briefly (∼1 ms) but the current strength is large. On the other hand, NCX current can continue until it reaches the equilibrium set by [Ca], [Na], and V_m. How do these currents contribute ECC?

Figure 1.

(A) Ca accumulation when L-type Ca channels open at V_m = 0 (red) and 40 mV (black). Solid and broken lines show opening of one or four L-type Ca channels, with all NCX in the submembrane (noncleft) space. (B) Ca accumulation by NCX current at V_m = 40 (black) and 0 mV (red). Lines are for all 100% of NCX either in submembrane space (solid) or cleft (broken).

To address this, we examine fractional SR Ca release (Fig. 2). Initially V_m is held at −80 mV and then changed to either 0 or +40 mV for 200 ms. For each case, we do four simulations: 1), Ca sparks are induced by I_CaL, NCX, or spontaneously (red curves). 2), Ca sparks are induced by NCX or spontaneously (black). 3), Only spontaneous Ca sparks are allowed (green; to estimate the impact of I_CaL without NCX). 4), NCX is blocked for the first 10 ms after the V_m step (blue dashed curves). At V_m = 0 mV, most Ca sparks are induced by I_CaL and only a tiny fraction of Ca sparks can be recruited by NCX (and only at high [Ca]_SR). At higher V_m (40 mV), because I_CaL becomes smaller and NCX becomes larger, the fraction of Ca sparks induced by NCX is increased. However, the absolute number of Ca sparks is still limited (only 10% via NCX at [Ca]_SR = 800 μM). If all NCXs locate at cleft space, NCXs initiate Ca sparks more efficiently (Fig. 2, C–D). A uniform sarcolemmal distribution would give ∼11% of NCX in clefts (10,11). Therefore, during ECC, most of the Ca sparks are induced by I_CaL. Regardless of the V_m and NCX distribution, blocking NCX for 10 ms (or 30 ms; not shown) does not alter the result (blue dashed lines). This implies that NCX does not appreciably help to initiate Ca sparks by I_CaL. Moreover, once an L-type Ca channel opens at a cleft, the local rise in [Ca]_Cleft will prevent net Ca entry via NCX.

Figure 2.

Fractional SR Ca release versus [Ca]_SR. Red curve: fractional release. Black curve: fractional release without I_CaL. Green curve: fractional release without I_CaL and NCX (i.e., spontaneous Ca release only). Blue dashed curve: NCX is blocked for first 10 ms after V_m step. [Na]_i = 12 mM. Insets are magnified plots from 0 to 0.1 (y axis). (A) Holding V_m is 0 mV and all NCX is in submembrane space (noncleft). (B) V_m = 40 mV and all NCX is in submembrane space. (C) V_m = 0 mV and all NCX is in the cleft. (D) V_m = +40 mV and all NCX is in the cleft. (E) Reverse mode NCX versus [Na]_i at V_m = 0, +20, and +40 mV ([Ca]_i = 0.1 μM). (F) Fractional release versus [Ca]_SR at [Na]_i = 6∼16 mM without I_CaL. V_m = 0 mV. 11% NCXs locate at cleft space.

In addition to V_m and [Ca]_i, NCX also depends on [Na]_i. For reverse mode NCX, higher [Na]_i increases Ca influx via NCX (Fig. 2 E). Therefore, at higher [Na]_i, more Ca sparks are induced by NCX at the same [Ca]_SR (Fig. 2 F).

At the resting V_m of cardiac myocytes (∼ –80 mV), NCX removes Ca from the cell. Because NCX extrudes Ca better as [Ca] rises, any NCX close to the cleft space where [Ca] is high, will remove Ca more efficiently during spontaneous SR Ca release. Once a single RyR opens in a CaRU, [Ca]_Cleft rises quickly. Depending on the NCX distribution, the level of [Ca]_Cleft achieved is different (Fig. 3 A). If all NCX is at the cleft (red), the plateau [Ca]_Cleft is ∼60% to ∼70% of the plateau [Ca]_Cleft for the case where all NCX is outside of the cleft (black). Due to lower [Ca]_Cleft, Ca spark events are also significantly reduced (Fig. 3 B). At [Ca]_SR = 700 μM, 12% of total Ca leak events are sparks (88% are nonspark leak) when NCX is excluded from cleft (black). On the other hand, when all NCX is in cleft (red) only 7% of the Ca leak events are Ca sparks. Thus, in Ca spark flux there is a 40% reduction versus the case where NCX is excluded from the cleft (Fig. 3 C). When NCX conductance is doubled, the effect of NCX localization is significantly increased (Fig. 3 D). Alternatively, at V_m = –80 mV, as [Na]_i rises, less Ca is extruded by NCX (Fig. 3 E), therefore [Ca]_Cleft gets higher and more Ca sparks occur at lower [Ca]_SR (Fig. 3 F).

Ca sparks are elementary events for ECC induced by Ca-induced Ca release (12). Ca-induced Ca release can be initiated by I_CaL, NCX, Ca diffusion from the neighbors, or spontaneous opening of RyRs. We focus here on the contribution of NCX with respect to Ca spark formation. NCX plays opposite roles depending on V_m. Reverse mode NCX can trigger Ca sparks, but it is not efficient (9,13–15). On the other hand, forward mode NCX may suppress Ca sparks by removing [Ca]_Cleft. In addition, regardless of the importance of NCX, the distribution of NCX in cardiac myocytes is still unclear. Mathematical modeling is a useful tool to analyze and estimate experimentally unknown values. Using our recently developed mathematical model, we show that during ECC, most Ca sparks are induced by I_CaL and only a very small fraction are due to NCX current (except at very high V_m and [Na]_i and if NCX are highly localized at the cleft space). This agrees with previous observations (9,13–15). At the resting membrane potential, NCXs suppress Ca sparks especially if NCXs are localized to the cleft. These findings underscore the criticality of NCX localization regarding functional impact on SR Ca release. The strength of NCX also substantially impacts our results. High [Na]_i decreases the forward mode NCX (Ca efflux) and increases the reverse mode NCX (Ca influx). Within the physiological range of [Na]_i, the amplitude of NCX changes severalfold (Fig. 2 E, Fig. 3 E). Both initiation and suppression of Ca sparks are significantly affected by [Na]_i. This implies that changing NCX strength by NCX overexpression, NCX knockout or NCX blockers also affects Ca spark formation.