Mitochondrial Ca²⁺ Influx and Efflux rates in Guinea Pig Cardiac Mitochondria: Low and High Affinity Effects of Cyclosporine A

Abstract

Ca²⁺ plays a central role in energy supply and demand matching in cardiomyocytes by transmitting changes in excitation-contraction coupling to mitochondrial oxidative phosphorylation. Matrix Ca²⁺ is controlled primarily by the mitochondrial Ca²⁺ uniporter (mCU) and the mitochondrial Na⁺/Ca²⁺ exchanger (mNCE), influencing NADH production through Ca²⁺-sensitive dehydrogenases in the Krebs cycle. In addition to the well-accepted role of the Ca²⁺-triggered mitochondrial permeability transition pore (PTP) in cell death, it has been proposed that the PTP might also contribute to physiological mitochondrial Ca²⁺ release. Here we selectively measure Ca²⁺ influx rate through the mCU and Ca²⁺ efflux rates through Na⁺-dependent and Na⁺-independent pathways in isolated guinea pig heart mitochondria in the presence or absence of inhibitors of mNCE (CGP 37157) or the PTP (CsA). CsA suppressed the negative bioenergetic consequences (ΔΨ_m loss, Ca²⁺ release, NADH oxidation, swelling) of high extramitochondrial Ca²⁺ additions, allowing mitochondria to tolerate total mitochondrial Ca²⁺ loads of >400 nmol/mg protein. For Ca²⁺ pulses up to 15µM, Na⁺-independent Ca²⁺ efflux through the PTP accounted for ~5% of the total Ca²⁺ efflux rate compared to that mediated by the mNCE (in 5mM Na⁺). Unexpectedly, we also observed that CsA inhibited mNCE-mediated Ca²⁺ efflux at higher concentrations (IC₅₀=2µM) than those required to inhibit the PTP, with a maximal inhibition of ~40% at 10µM CsA, while having no effect on the mCU. The results suggest a possible alternative mechanism by which CsA could affect mitochondrial Ca²⁺ load in cardiomyocytes, potentially explaining the paradoxical toxic effects of CsA at high concentrations.

Introduction

Ca²⁺ is the central player in excitation-contraction coupling in cardiac myocytes, but it also serves as an important signaling molecule between cytosol and mitochondria. In the heart, mitochondria take up Ca²⁺ from the cytosol through the calcium uniporter (mCU) and extrude it primarily through the mitochondrial Na⁺/Ca²⁺ exchanger (mNCE) [1, 2]. Mitochondrial calcium (Ca²⁺_m) regulates NADH production by activating Ca²⁺-sensitive dehydrogenases in the TCA cycle [3] and may stimulate oxidative phosphorylation at other sites as well [4, 5]. By changing the driving force for Ca²⁺ efflux through the mNCE, alterations in cytosolic Na⁺ (Na⁺_i) will significantly impact the rate of mitochondrial Ca²⁺ accumulation when the amplitude or frequency of the cytosolic Ca²⁺ transient changes. For example, Na⁺_i is elevated in chronic heart failure [6–8] and during ischemia-reperfusion injury [9–12], as well as during treatment with cardiac glycosides, and we have shown that high Na⁺_i blunts Ca²⁺_m accumulation in cardiac myocytes subjected to a rapid increase in work [13–16]. Moreover, the oxidation of NAD(P)H associated with inadequate mitochondrial Ca²⁺ signalling contributes to increased oxidative stress, arrhythmias and contractile dysfunction [15, 16]. These detrimental effects are prevented by treatment with an mNCE inhibitor, or by lowering Na⁺_i, emphasizing the importance of mitochondrial Ca²⁺ efflux as a modulator of excitation-contraction-bioenergetic coupling.

Na⁺-independent mitochondrial Ca²⁺ efflux can also play a role in mitochondrial Ca²⁺ regulation, particularly in certain non-cardiac tissues in which a H⁺/Ca²⁺ antiporter is active [17]. Alternatively, Gunter and Pfeiffer [18] proposed that the transient and reversible opening of the mitochondrial permeability transition pore (PTP) could be an energetically favorable way to release mitochondrial Ca²⁺ in a Na⁺-independent manner, albeit with the proviso that the concomitant loss of mitochondrial membrane potential (ΔΨ_m) would have to be rapidly reversed in order to maintain ATP production. In this light, Altschuld et al. [19] reported that cyclosporine A (CsA), an inhibitor of the PTP [20, 21], increased mitochondrial Ca²⁺ load in rat ventricular myocytes, possibly explaining the toxic effects of CsA on electrically-paced rat cardiomyocytes[22]. In a recent study, we too observed enhanced mitochondrial Ca²⁺ loading in response to pacing in the presence of CsA (see supplemental data Fig. S2 in ref.[14]). These observations could be interpreted as evidence for a physiological role for reversible PTP opening during Ca²⁺ cycling or that pathophysiological activation of the PTP is occurring in a fraction of mitochondria in the network of the isolated cardiomyocyte. However, based on the behavior of isolated mitochondria, unless there is significant metabolic stress [23, 24], mitochondria should be able to tolerate rather large Ca²⁺ loads without activation of the PTP because of the robust Ca²⁺ buffering capacity of the matrix due to the formation of calcium phosphate precipitates [25, 26].

In the context of the outstanding questions mentioned above, and in order to understand mitochondrial Ca²⁺ handling in both normal and disease states, quantitative measurements of unidirectional Ca²⁺ uptake and efflux rates are necessary. In this study, Ca²⁺ fluxes through mCU, mNCE and PTP were measured in normal isolated guinea pig heart mitochondria. The effects of variables such as Ca²⁺_i, Na⁺_i, and CsA on the mitochondrial Ca²⁺ transport were explored. The results indicate that the steady state extramitochondrial Ca²⁺ concentration is strongly influenced by Na⁺_i and that, unexpectedly, CsA, at concentrations higher than that required to inhibit the PTP, has an inhibitory effect on both PTP- and mNCE-mediated mitochondrial Ca²⁺ efflux. These kinetic measurements also provide essential information for refinement of computational models of mitochondrial Ca²⁺ handling, with the ultimate goal of interpreting the influence of mitochondria on cellular Ca²⁺ handling, redox potential and energetics.

Methods

Guinea pig heart mitochondria were isolated using a protocol described previously[27]. The extramitochondrial Ca²⁺ concentration ([Ca²⁺]_out) was measured using the Ca²⁺-sensitive fluorescent probe, CaGreen-5N, hexapotasssium salt (Molecular Probes, Invitrogen) in a fluorometer (Quantamaster, Photon Technologies International) at room temperature. Mitochondria (~0.5mg) were suspended in a potassium-based buffer solution consisting of: 137mM KCl, 2mM KH₂PO₄, 20µM EGTA, 20mM HEPES, and 5mM glutamate/malate (G/M) at pH 7.15. Calcium green-5N (0.1µM) fluorescence was recorded at excitation and emission wavelengths of 505nm and 535nm [28]. Mitochondrial 90° light scattering was monitored at 540nm with a second detector and NADH fluorescence was recorded with excitation at 350nm and emission at 450nm. Mitochondrial membrane potential was monitored by the ratiometric method of Scaduto and Grotyohann [29] using tetramethylrhodamine methyl ester (TMRM) at excitations of 546nm and 573nm and emission at 590nm. Mitochondrial protein concentrations were determined by the BCA assay (Thermo Scientific Pierce).

Free calcium in the buffer solution was calculated using MaxChelator (http://www.stanford.edu/~cpatton/maxc.html) and a standard curve was constructed in the presence of mitochondria, but with Ca²⁺ uptake blocked (see supplemental Figure S1) relating the CaGreen-5N signal to the free Ca2+ concentration in the buffer solution by fitting to the Grynkiewicz equation [30].

$${\left[{\mathit{Ca}}^{2+}\right]}_{\mathit{\text{free}}}=K_d\left[\frac{F-F_{\text{min}}}{F_{\text{max}}-F}\right]$$

, where F is the fluorescence intensity of calcium green at the experimental calcium level, F_min is the fluorescence intensity without calcium, F_max is the fluorescence intensity of CaGreen-5N saturated with calcium, and the K_d (14µM) was the solution dissociation constant of CaGreen-5N provided by the manufacturer.

To study CsA effects on Ca²⁺ transport by mitochondria, cyclosporine A (Sigma-Aldrich) was added directly to the mitochondrial suspension from a 4mM dimethyl sulfoxide (DMSO) stock solution. Amounts of DMSO alone, equivalent to those of the largest amount used in a given experiment (usually <1% of total volume), had no effect on the measured parameters.

Statistical Analysis

The results are presented as mean±SEM. An unpaired Student’s t-test was used to evaluate the significance of the differences between means of CsA-treated versus untreated mitochondrial Ca²⁺ fluxes using either Origin (Microcal) or Matlab statistical toolbox. Statistical significance was assumed at p<0.05.

Results

The capacity of guinea pig heart mitochondria to accumulate Ca²⁺ was tested by multiple additions of Ca²⁺ (first addition was 15µM and subsequent ones were 25µM each) while monitoring extramitochondrial Ca²⁺ concentration ([Ca²⁺]_out), NADH, light scattering and ΔΨ_m (Fig. 1). The first addition of Ca²⁺ evoked rapid Ca²⁺ uptake from the medium and a small (~5mV) depolarization of ΔΨ_m corresponding to the energetic cost of Ca²⁺ entry (that is, Ca²⁺ uptake, coupled to Na⁺/Ca²⁺ exchange, coupled to Na⁺/H⁺ exchange), a transient oxidation of NADH, and a decrease in light scattering corresponding to an increase in mitochondrial volume (Figs. 1A and 1B). In the absence of CsA, by the third or fourth Ca²⁺ addition, a secondary release of Ca²⁺ is evident, and is accompanied by a larger and sustained decrease in ΔΨ_m and NADH, but the ability of the suspension to regulate extramitochondrial Ca²⁺ is still partially maintained (Fig. 1A). After the sixth Ca²⁺ addition, however, corresponding to a total mitochondrial Ca²⁺ load of more than 400 nmol Ca²⁺/mg mitochondrial protein, ΔΨ_m collapses, NADH is oxidized, additional swelling occurs, and Ca²⁺ is released to the medium (Fig. 1A). In striking contrast, in the presence of CsA (4 µM), the mitochondria readily take up eight additions of Ca²⁺ while maintaining ΔΨ_m, NADH, and mitochondrial volume. Steady-state [Ca²⁺]_out was regulated at a constant setpoint (Fig. 1B) with no evidence of a permeability transition.

Figure 1.

Effects of sequential Ca²⁺ additions on mitochondrial Ca²⁺ uptake and energetic parameters. Mitochondria were suspended in KCl-based solution with 5mM NaCl and 5mM G/M and the mitochondrial membrane potential (ΔΨm), NADH, volume change(from light scattering) and extra-mitochondrial Ca²⁺ level were simultaneously monitored. (A) In the absence of cyclosporine A (CsA), the mitochondrial permeability transition pore (PTP) opens after a train of calcium pulses accompanied by mitochondrial swelling, membrane potential depolarization, and efflux of Ca²⁺ from the mitochondria. The first calcium pulse is 15µM free Ca²⁺ and each pulse afterward was 25µM. The total free Ca²⁺ load tolerated before full PTP opening in this experiment was 400nmol/mg. (B) 4µM cyclosporine A augmented mitochondrial calcium uptake capacity by preventing PTP opening. (C) After the Ca2+ addition, multiple pulses of Na⁺ were added from 5mM to 60mM to examine the effects of increased Na⁺-dependent efflux on steady state extramitochondrial Ca²⁺.

Since PTP activation was not a factor for the response to a single addition of Ca²⁺, we focused on measuring the mitochondrial Ca²⁺ influx and efflux rates under various conditions for additions of Ca²⁺ no greater than 20µM. For a single Ca²⁺ addition of 15µM, in the absence of Na⁺, mitochondrial Ca²⁺ uptake was rapid and [Ca²⁺]_out was lowered to < 0.32 µM (Fig. 1C; left panel) in less than 200 sec. The Ca²⁺ addition had minimal effects on ΔΨ_m (it decreased by only 2 mV)(Fig. 1C; left panel) and NADH (Fig. 1C; right panel). The effects of sequentially increasing Na⁺ on steady state [Ca²⁺]_out were then tested. Increasing extramitochondrial Na⁺ concentration to 5, 10 and 15 mM Na⁺ increased the [Ca²⁺]_out setpoints to 0.80µM, 1.28µM, and 1.34µM, respectively (Fig. 1C; left panel). For Na⁺ greater than 15mM, however, [Ca²⁺]_out paradoxically decreased to 1.12µM (30mM) and 0.76µM (60mM), presumably because the capacity of the mitochondria to regulate matrix Na⁺ through Na⁺/H⁺ exchange may have been saturated, thus allowing matrix Na⁺ levels to rise and decreasing the driving force for Ca²⁺ extrusion through the mNCE.

Mitochondrial calcium uptake and extrusion in cardiac mitochondria are mediated mainly through mCU and mNCE under normal conditions in the absence of PTP activation[18]. A protocol was developed to study the individual Ca²⁺ uptake and efflux rates by loading mitochondria with a single 15µM Ca²⁺ pulse and then selectively blocking mCU with Ru360 (mCU inhibitor)[31–33], mNCE with CGP-37157 (CGP)[34] or PTP with CsA[21]. With 5 mM Na⁺ in the buffer, Ca²⁺ was added to energized mitochondria pretreated with either CGP (10µM), CsA (10µM), or CGP+CsA. In all cases, mitochondrial Ca²⁺ uptake was rapid and a steady state representing the balance between Ca²⁺ uptake and efflux rates was achieved approximately 200 sec after the pulse (Fig. 2; phase i). The steady state [Ca²⁺]_out (Ca_SS) was modestly decreased by CsA and more significantly lowered by CGP or CGP+CsA (to ~0.5µM). The unidirectional Ca²⁺ efflux rate was then measured after application of Ru360 (Fig. 2; phase ii). CGP suppressed the majority of the Ca²⁺ efflux in 5mM Na⁺ (Fig. 2; blue trace). Interestingly, in addition to blocking the remaining small CGP-insensitive Ca²⁺ efflux (CGP+CsA; Fig. 2: green trace), CsA also inhibited a significant fraction of total Ca²⁺ efflux (Fig. 2; red trace), prompting further investigation of its effects.

Figure 2.

Selective block of mCU following a Ca²⁺ addition with Ca²⁺ uptake and efflux active. Phase (i): in the presence of 5mM NaCl and 5mM G/M, a 15µM Ca²⁺ pulse was given and there is a net uptake of Ca²⁺. A steady state extramitochondrial Ca²⁺ concentration is attained when the Ca²⁺ influx and efflux rates are equal. Phase (ii): after the addition of Ru360, influx through the mCU is blocked and net Ca²⁺ efflux occurs via the Na⁺-dependent pathway (mNCE) and the Na⁺-independent pathway (PTP). CsA partially inhibited efflux, which was mainly contributed by the mNCE, as indicated by it’s CGP-37157 sensitivity. Inhibitor concentrations: 5nM Ru360, 10µM CsA, and 10µM CGP.

The Na⁺ dependence of Ca_SS and the Ca²⁺ efflux rate in the presence and absence of CsA was analyzed by varying Na⁺ in the buffer from 0 to 60mM (Fig. 3A). Several parameters were measured for this protocol (Fig. 3B): the initial Ca²⁺ uptake rate (rate 1; Fig. 3C) after the addition of Ca²⁺; the net Ca²⁺ extrusion rate after Ru360 addition (rate 2; Fig. 3D); and the Ca_SS (at the end of phase i; Fig 3E). The average mitochondrial Ca²⁺ load after a single Ca²⁺ addition (15µM) was calculated as the total amount of Ca²⁺ added to the cuvette minus Ca_SS, normalized to the mitochondrial protein concentration (Fig. 3F).

Figure 3.

Effects of extramitochondrial Na⁺ on net Ca²⁺ uptake and extrusion. (A) Superimposed traces of [Ca²⁺]_out with the experimental protocol shown in Figure 2. Isolated mitochondria were equilibrated in buffer solutions containing 0 to 60mM Na⁺, with or without 10µM CsA. (B) Mitochondrial parameter measurements: Ca²⁺ uptake slope at 130 to 140sec (rate1), Ca²⁺ efflux slope at 380 to 390sec (rate2), steady state extramitochondrial Ca²⁺ at 350sec (Ca_SS). (C–E) Na⁺ effects on rate1, rate2 and Ca_SS. (F) The mitochondria Ca²⁺ load calculated from the difference between the Ca²⁺ added and the Ca_SS. Data presented as mean±SEM, n=3. * P<0.05, (−)CsA vs (+)CsA. Filled symbols: control data, Open symbols: CsA (10µM) treatment.

Ca_SS (Fig. 3E) increased from 0.6µM to 2µM as extra-mitochondrial Na⁺ concentration increased from 5mM to 15mM by increasing net Ca²⁺ extrusion (Fig. 3D) and reducing mitochondria Ca²⁺ load (Fig. 3F). When [Na⁺]_out was higher than 30mM, Na⁺ had the opposite effect and mitochondrial Ca²⁺ load increased (i.e., Ca_SS decreased). CsA (10µM) also decreased Ca_SS (Fig. 3E) and increased Ca²⁺ load in the mitochondria (Fig. 3F). The net unidirectional Ca²⁺ efflux rate was significantly reduced by CsA (Fig. 3D) but there was no effect of CsA on the Ca²⁺ uptake rate (Fig. 3C).

A second protocol was employed to more selectively measure individual fluxes from mCU, mNCE and PTP (Fig. 4A). A single Ca²⁺ addition was made under zero-Na⁺ conditions and then Ru360 was added to block the mCU, leaving active only the Na⁺-independent Ca²⁺ efflux pathway [35]. Na⁺ was then added to activate mNCE [17, 36]. The maximum uptake rate of mCU measured in guinea pig heart mitochondria with a 15µM free Ca²⁺ addition was 0.49±0.04 nmol/mg/sec (Fig. 4B). The Na⁺-independent Ca²⁺ extrusion rate was 0.003±0.0004 nmol/mg/sec and the Na⁺-dependent (plus Na⁺-independent) extrusion rate with 5mM Na⁺ was 0.018±0.003 nmol/mg/sec (Fig. 4B). CsA (10µM) reduced Na⁺-independent Ca²⁺ efflux measured in phase ii from 0.003 to 0.001 nmol/mg/sec, consistent with a small Ca²⁺ leak mediated by the PTP [37, 38]. However, in the presence of CsA, a 40% inhibition of the Na⁺-dependent Ca²⁺ efflux measured in phase iii, was also observed (CsA decreased the Na⁺-dependent efflux rate to 0.011±0.001 nmol/mg/sec; Fig, 4B).

Figure 4.

Unidirectional Ca²⁺ flux rates through the mCU, PTP, and mNCE (A) A single Ca²⁺ pulse (15µM) is given in 0mM Na⁺ solution (i: mCU-mediated Ca²⁺ uptake) and then Ru360 (5nM) is added to block the Ca²⁺ uptake pathway, leaving only Na⁺-independent Ca²⁺ extrusion (ii: includes PTP-mediated leak). 5mM Na⁺ is then added to activate Na⁺-dependent Ca²⁺ extrusion (iii: mNCE). (B) Summary data for mitochondrial Ca²⁺ transport rates in the presence and absence of 10µM CsA. Data presented as mean±SEM, n=11. *P<0.05. (C) Ca²⁺ dependence of the Ca²⁺ influx and efflux rates for Ca²⁺ additions between 2 and 20µM using the same protocol. Data presented as mean±SEM, n=5. (D) Na⁺-dependence of the initial mNCE rate with additions of different concentrations of Na⁺ (2.5mM to 60mM). Data presented as mean±SEM, n=5. Filled symbols: control data, Open symbols: CsA (10µM) treatment.

The dependence of the Ca²⁺ transport rates on the size of the extramitochondrial Ca²⁺ addition ([Ca²⁺]_added), ranging from 2µM to 20µM, using the second protocol was determined (Fig. 4C). The maximal Ca²⁺ uptake rate increased linearly as a function of [Ca²⁺]_added from 0.05nmol/sec/mg to 0.6nmol/sec/mg and was not significantly altered by CsA. The mNCE flux (with 5mM NaCl) also increased from ~0.015 to 0.03 nmol/sec/mg as [Ca²⁺]_added was increased (Fig. 4C), corresponding to mitochondrial Ca²⁺ loads after the Ca²⁺ uptake phase of 4.5, 19, 37, 57, and 77 nmol/mg for [Ca²⁺]_added levels of 2, 5, 10, 15, and 20µM, respectively. In the presence of 10µM CsA, the mNCE rate was approximately 0.012 nmol/sec/mg and, interestingly, this rate did not increase as a function of [Ca²⁺]_added (Fig. 4C). The Na⁺-independent Ca²⁺ efflux was minimally increased by raising [Ca²⁺]_added over the range of 2–15µM (from 0.001 to 0.002 nmol/mg/sec), but a larger CsA-sensitive Na⁺-independent Ca²⁺ efflux (0.006nmol/mg/sec) was evoked after a 20µM Ca²⁺ addition (Fig, 4C).

The effects of Na⁺, in the range of 2.5mM to 60mM (covering both the physiological and pathophysiological range of Na⁺_i [14, 16, 39]), on the mNCE rate (phase iii) were investigated after Ca²⁺ loading with 15µM Ca²⁺ (Fig. 4D). The mNCE rate increased from 0.01 nmol/mg/sec to 0.07 nmol/mg/sec as [Na⁺]_out increased from 2.5mM to 15mM then decreased at higher [Na⁺]_out from 30mM to 60mM. Again, CsA partially inhibited mNCE flux, reducing the rates by about 0.01nmol/mg/sec over the full range of [Na⁺]_out.

The instantaneous mCU rate as a function of [Ca²⁺]_out could also be determined during active mitochondrial Ca²⁺ uptake from single Ca²⁺ additions (Fig. 5). A relationship similar to that determined above from the maximal Ca²⁺ uptake rates for various Ca²⁺ additions (Fig 4C) was obtained.

Figure 5.

Instantaneous relationship between mCU-mediated Ca²⁺ influx rate dCa²⁺/dt and extramitochondrial Ca²⁺.

The concentration dependence of the CsA inhibition of Ca²⁺ efflux was investigated for a range of [CsA] from 0.05µM to 40µM (Fig. 6). Maximal inhibition of mNCE flux (15µM Ca²⁺ loading pulse; 5mM Na⁺ addition) by CsA was 40% (reduced from 0.020 nmol/mg/sec to 0.012 nmol/mg/sec) for [CsA] ≥10µM (10–40µM range equivalent to 40 to 160nmol CsA/mg). The half-maximal inhibitory concentration (IC₅₀) for inhibition of the mNCE by CsA was 2µM. Inhibition of the Na⁺-independent Ca²⁺ efflux, presumably mediated by the PTP, was maximal at a lower CsA concentration (0.05µM; 200pmol/mg), which corresponds to the effective inhibitory concentration of CsA for the PTP reported previously [20, 21, 40].

Figure 6.

Concentration-dependence of CsA inhibition of PTP- or mNCE-mediated Ca²⁺ efflux (left panel). Lack of effect of CsA on mCU rate (right panel). Mitochondria were incubated with different concentrations of cyclosporine A (0, 0.05, 2, 10, 20, 40µM) in a KCl-based buffer solution with 5mM G/M. Mitochondria then were given a 15µM Ca²⁺ loading pulse and a 5mM Na⁺ addition after a Ru360 addition. Data presented as mean±SEM, n=5~6.

Discussion

The present work provides the first quantitative measurements of the unidirectional Ca²⁺ uptake and extrusion rates of mitochondria from the guinea pig heart and analyzes the influence of [Ca²⁺]_out, [Na⁺]_out, CsA, CGP-37157 and Ru360 on Ca²⁺ transport. Maximal Ca²⁺ uptake rates through the Ru360-sensitive mCU for Ca²⁺ additions of 2 to 20µM ranged from 0.05 to 0.6 nmol/sec/mg and the steady-state extramitochondrial Ca²⁺ level was dependent on concomitant Ca²⁺ efflux, primarily through the mNCE. Both Na⁺-independent and Na⁺-dependent Ca²⁺ efflux pathways were present, with the mNCE rate predominating (roughly 10-fold higher than the Na⁺-independent rate). The mNCE had a biphasic dependence on Na⁺; its rate increasing over the range of 2.5 to 15mM and then decreasing at 30–60mM. In addition to preventing PTP activation for large mitochondrial Ca²⁺ loads exceeding 400 nmoles/mg, CsA increased the mitochondrial Ca²⁺ load for single lower Ca²⁺ pulses by inhibiting the Na⁺-independent Ca²⁺ efflux pathway with high affinity (pmol/mg range) and partially inhibiting Na⁺-dependent Ca²⁺ efflux with a lower affinity (~2µM IC₅₀; nmol/mg range).

Mitochondrial Ca²⁺ transport in the heart is a topic of increasing interest and frequent controversy [41–43]. In terms of physiological regulation, Ca²⁺ uptake during EC coupling provides a crucial feedforward signal to mitochondrial oxidative phosphorylation to increase NADH production and ATP supply to meet the demands of contractile activation and ion transport. On the other hand, excessive mitochondrial Ca²⁺ loading, by activating the PTP, is a key event leading to necrotic or apoptotic cell death [24, 44]. Understanding the balance between the positive and negative effects of mitochondrial Ca²⁺ requires a detailed understanding of the factors modulating the Ca²⁺ uptake and efflux pathways under normal and pathophysiological conditions. Early work by Chance [5] showed that Ca²⁺ additions to mitochondria evoked rapid (<100 msec) changes in the redox potential of the respiratory chain carrier cytochrome b and transient oxidation followed by reduction of the pyridine nucleotide (NADH) pool, as well as transient stimulation of mitochondrial respiration. A decrease in light scattering (mitochondrial swelling) was also observed upon Ca²⁺ addition, and when multiple Ca²⁺ additions were made, a threshold was reached at which NADH completely oxidized and swelling was maximal. These early observations of large amplitude mitochondrial swelling are quite similar to what is observed using the “standard” PTP assay, as shown in Figure 1 for multiple Ca²⁺ additions. Notably, the CsA-inhibitable Ca²⁺ release and reuptake evident after the third, fourth, and fifth Ca²⁺ additions in Figure 1A are followed by a complete release of Ca²⁺ from the mitochondrial matrix after the sixth addition (total Ca²⁺ added was 140µM, or 400 nmoles/mg protein), which also corresponded to the collapse of ΔΨ_m, NADH oxidation, and large amplitude swelling. Prior to the sustained activation of the PTP, the CsA-sensitive partial release and reuptake of Ca²⁺ could be interpreted in several ways. First, the response could be due to a reversible or transient opening of the PTP, which has been previously proposed [37]. Alternatively, the transient Ca²⁺ release could represent irreversible opening of the PTP in a fraction of the mitochondrial population and reuptake of Ca²⁺ by the remaining intact mitochondria. Third, a combination of the two effects could be occurring, i.e., Ca²⁺ release from mitochondria in which the PTP is open, followed by PTP closure as their Ca²⁺ load is decreased [45]. Arguing against the idea that the PTP opening is reversed is the observation that depolarization of ΔΨ_m and oxidation of the NADH pool is sustained even after the released Ca²⁺ is taken back up, supporting the interpretation that a fraction of the mitochondria have undergone irreversible PTP activation. Interestingly, in the presence of CsA, the initial swelling evoked by Ca²⁺ is almost completely reversed in a stepwise manner by subsequent Ca²⁺ additions, indicating that some type of Ca²⁺-mediated volume regulation has been activated (see light scattering recording in Fig. 1B). Further investigation will be required to characterize the mechanism responsible for this effect.

For smaller single Ca²⁺ additions ≤15µM, corresponding to mitochondrial Ca²⁺ loads of 57 nmol/mg, PTP activation was not a significant factor, and the maximal Ca²⁺ uptake rate in the absence of Na⁺ was not altered by CsA. This indicates that we were truly measuring the unidirectional flux through the mCU, which was completely inhibited by Ru360. The Ca²⁺ uptake rates that we measured were ~10-fold higher than the Na⁺-dependent Ca²⁺ efflux rate, and were similar in magnitude to, but slightly higher than, those reported previously in guinea pig heart mitochondria under less selective conditions [46]. The relatively low affinity of the mCU for Ca²⁺ (10’s of µM[46]) relative to the range of diastolic (0.1µM) and systolic (1µM) [Ca²⁺] evident during the cytosolic transient has always raised the question of the relevance of mitochondrial Ca²⁺ uptake in the beat-to-beat regulation of Ca²⁺[43]. Nevertheless, there is strong evidence that the close juxtaposition of mitochondria and the sites of SR Ca²⁺ release (as close as 37nm [47]) creates a local Ca²⁺ microdomain that could support significant Ca²⁺ uptake by the mitochondria surrounding the diad[13, 48, 49]. Computational studies of the local Ca²⁺ in the dyadic cleft during EC coupling indicate that, at the peak of triggered SR Ca²⁺ release, [Ca²⁺] at the center of the cleft may approach 600µM, while 200nm away at the periphery of the cleft [Ca²⁺] may be as high as 100µM, declining back to the diastolic level over 150ms [50]. This would certainly be sufficient to support fast Ca²⁺ uptake through the mCU, although other cytosolic factors, such as Mg²⁺,[36, 51], adenine nucleotides [52], or endogenous polyamines [53] could shift the apparent K_m for Ca²⁺ uptake (see supplemental Fig. S3 for uptake rates in the presence of MgATP). The kinetic data obtained in the present work will be vital for refining such models of local Ca²⁺ transport to include mitochondria near the junctional microdomain.

The effect of extramitochondrial Na⁺ on the rate of Ca²⁺ efflux through the mNCE was shown to be biphasic, increasing for Na⁺ up to 15mM and then decreasing for Na⁺ in the range of 30–60mM. The latter was true in the experiments where both influx and efflux were active, in which case the extramitochondrial Ca²⁺ setpoint was affected (Fig. 1C and Fig. 3), or when only Ca²⁺ efflux was active (Fig. 4D). The steady state [Ca²⁺]_out levels varied from 1 to 2µM as [Na⁺]_out was increased from 5mM to 15mM and then decreased to ~1.25µM at 60mM [Na⁺]_out. The suppressive effect on Ca²⁺ efflux at higher Na⁺ was not due to generalized degradation of energetic functions, since ΔΨ_m, NADH, and volume were not significantly altered (Fig. 1C and 1D). We hypothesize that the suppression of the rate at Na⁺ > 30mM may be due to Na⁺ loading of the mitochondrial matrix, which would decrease the driving force for Ca²⁺ extrusion. This could occur if Na⁺ influx exceeds the capacity of the proton pumps to support Na⁺ extrusion coupled to Na⁺/H⁺ exchange. A sigmoidal fit of the mNCE rates as Na⁺ was varied over the range of 5–30 mM yielded a K_m for Na⁺ of 7.5 mM, close to the previously reported value of 8mM obtained for mNCE protein reconstituted in lipid vesicles[54].

The normal range of Na⁺_i reported in unstimulated guinea pig cardiomyocytes is 5–8mM and it increases to 15–20 mM in heart failure [14]; therefore, based on the present findings, approximately a 4-fold increase in the mitochondrial Ca²⁺ efflux rate would be expected. Consistent with this effect, we have previously reported that elevated Na⁺_i (15mM) accelerates the decay rate and amplitude of mitochondrial Ca²⁺ transients and decreases the overall mitochondrial Ca²⁺ loading during a train of electrically-evoked Ca²⁺ transients in intact guinea pig myocytes - this results in insufficient activation of NADH production to compensate for increased energy demand [13]. A similar impairment of energy supply and demand exists during the development of chronic heart failure [14], or after inhibition of the sarcolemmal Na⁺ pump with the cardiac glycoside ouabain, which also elevates Na⁺_i to greater than 15mM [16]. During the course of these earlier studies, we also noted that mitochondrial Ca²⁺ accumulation during a train of Ca²⁺ transients was moderately potentiated by CsA (see supplementary Fig. S2 in ref [14]), which motivated the present investigation of the effects of CsA on unidirectional mitochondrial Ca²⁺ influx and efflux rates.

For a single Ca²⁺ addition, CsA decreased steady state extramitochondrial Ca²⁺, indicative of an increase in mitochondria Ca²⁺ accumulation. Hence, we examined whether CsA influenced mCU-mediated Ca²⁺ influx, Na⁺-independent Ca²⁺ efflux, and Na⁺-dependent Ca²⁺ efflux. The maximal Ca²⁺ uptake rate through the mCU was not affected by CsA (up to 10µM). The rate of Na⁺-independent Ca²⁺ efflux was less than 5% of the Na⁺-dependent flux for Ca²⁺ additions up to 15µM and this pathway was inhibited by just 50nM CsA. This suggests that there may be some small role for PTP opening as a Ca²⁺ efflux pathway under “normal” conditions; however, one would also expect that other factors present in the cytoplasm, such as ATP[55], would further suppress this residual activation of the PTP. The larger effect of CsA on Ca²⁺ flux, in the absence of PTP activation, was to inhibit the Na⁺-dependent Ca²⁺ efflux rate, with an IC₅₀ of 2µM. Alterations in the energy state of the mitochondria by CsA could not account for this effect, as no significant differences in ΔΨ_m, NADH, or volume were observed in the presence of CsA. The possibility that CsA could have been acting by suppressing a Na⁺-dependent potentiation of PTP opening could also be ruled out, based on the observation that all of the Na⁺-dependent Ca²⁺ efflux could be blocked by the mNCE inhibitor 1µM CGP 37157, which has an IC₅₀ of approximately 0.36uM[34]. Thus, we conclude that the inhibitory effect of CsA on Na⁺-dependent mitochondrial Ca²⁺ efflux was due to inhibition of the mNCE. The nature of this inhibition is presently unknown, but it does not appear to be a direct competitive effect of CsA on the transport site, since the fractional inhbition of mNCE by CsA decreased at higher concentrations of Na⁺. A kinetic analysis of the data shows that the CsA effect is consistent with a “mixed type” inhibition (see Fig. S2). Notably, the CsA inhibition of Na⁺-dependent Ca²⁺ efflux was larger in the presence of higher mitochondrial Ca²⁺ loads (Fig. 4C), which might indicate that CsA is interfering with an intramitochondrial mNCE regulatory process. By analogy with the sarcolemmal NCX, CsA could be inhibiting internal Ca²⁺ activation of the transporter (this occurs via a distinct module in the intracellular regulatory loop of the sarcolemmal NCX). Alternatively, there may be multiple isoforms of mNCE present, one of which is activated at higher mitochondrial Ca²⁺ loads and is inhibited by CsA, with the other forms insensitive to CsA inhibition. These speculative hypotheses can be tested in future studies, aided by the recent discovery of a molecular candidate for mNCE (NCLX)[56].

While the kinetics of mitochondrial Ca²⁺ fluxes for the mCU, mNCE, and PTP have been previously been characterized in mitochondria from various tissues and species [2, 32, 35, 36, 54, 56–58], a systemic and quantitative assessment of the unidirectional influx and efflux rates obtained under identical conditions is required to improve our understanding of mitochondrial Ca²⁺ transport as a whole. These data will therefore be invaluable for integrated model development in the future. From the perspective of pathophysiology, there is an emerging view that inhibition of PTP opening protects against acute ischemic injury [24, 59], as well as progressive degenerative diseases such as muscular dystrophy [60, 61] and Alzheimer’s [62]. The present findings showing an additional effect of CsA on mNCE suggest that high concentrations of CsA can influence mitochondrial Ca²⁺ dynamics by a PTP-independent mechanism. Inhibition of mNCE could be beneficial in instances where impaired mitochondrial Ca²⁺ loading is a problem, such as in heart failure [14, 15]; however, increased mitochondrial Ca²⁺ loading could also have toxic effects, such as those reported by Olbrich et al [22] for CsA concentrations ranging from 2–8µM in electrically-paced rat cardiomyocytes. In addition, previous studies of the concentration dependence of CsA-mediated effects on ischemia-reperfusion injury indicated that 0.2µM CsA, but not 2µM CsA, was protective [63]. Thus, increased mitochondrial Ca²⁺ loading could potentially contribute to detrimental effects of CsA at concentrations above that required to inhibit the PTP. Further studies will be required to examine the mechanism of the CsA effect on mNCE and whether it requires the presence of cyclophilin D, as does PTP inhibition [64].