Mitochondrial calcium uniporter complex activation protects against calcium alternans in atrial myocytes

Abstract

Cardiac alternans, defined as beat-to-beat alternations in action potential duration, cytosolic Ca transient (CaT) amplitude, and cardiac contraction is associated with atrial fibrillation (AF) and sudden cardiac death. At the cellular level, cardiac alternans is linked to abnormal intracellular calcium handling during excitation-contraction coupling. We investigated how pharmacological activation or inhibition of cytosolic Ca sequestration via mitochondrial Ca uptake and mitochondrial Ca retention affects the occurrence of pacing-induced CaT alternans in isolated rabbit atrial myocytes. Cytosolic CaTs were recorded using Fluo-4 fluorescence microscopy. Alternans was quantified as the alternans ratio (AR = 1 − CaT_small/CaT_large, where CaT_small and CaT_large are the amplitudes of the small and large CaTs of a pair of alternating CaTs). Inhibition of mitochondrial Ca sequestration via mitochondrial Ca uniporter complex (MCUC) with Ru360 enhanced the severity of CaT alternans (AR increase) and lowered the pacing frequency threshold for alternans. In contrast, stimulation of MCUC mediated mitochondrial Ca uptake with spermine-rescued alternans (AR decrease) and increased the alternans pacing threshold. Direct measurement of mitochondrial [Ca] in membrane permeabilized myocytes with Fluo-4 loaded mitochondria revealed that spermine enhanced and accelerated mitochondrial Ca uptake. Stimulation of mitochondrial Ca retention by preventing mitochondrial Ca efflux through the mitochondrial permeability transition pore with cyclosporin A also protected from alternans and increased the alternans pacing threshold. Pharmacological manipulation of MCUC activity did not affect sarcoplasmic reticulum Ca load. Our results suggest that activation of Ca sequestration by mitochondria protects from CaT alternans and could be a potential therapeutic target for cardiac alternans and AF prevention.

NEW & NOTEWORTHY This study provides conclusive evidence that mitochondrial Ca uptake and retention protects from Ca alternans, whereas uptake inhibition enhances Ca alternans. The data suggest pharmacological mitochondrial Ca cycling modulation as a potential therapeutic strategy for alternans-related cardiac arrhythmia prevention.

INTRODUCTION

Cardiac alternans or pulsus alternans is a clinical condition observed in patients with heart failure (HF) and aortic valve disease and is presented as a beat-to-beat oscillation in cardiac contraction strength at a constant heart rate (17). Alternans is associated with cardiac arrhythmia, including atrial fibrillation, and sudden cardiac death (39, 69). At the cellular level, cardiac alternans is defined as beat-to-beat alternations in contraction amplitude (mechanical alternans), action potential duration (electrical alternans), and Ca transient amplitude (CaT alternans) at a constant heart rate (for reviews, see 16, 17, 30). Experimentally, in isolated myocytes or cardiac muscle preparations, alternans can be induced by increasing pacing frequency; however, an extensive plethora of conditions affects severity and pacing threshold of alternans, suggesting that cardiac alternans is a multifactorial process. Nonetheless, a growing body of evidence suggests that alternans is linked to disturbances of intracellular Ca handling during excitation-contraction coupling (ECC). Based on the analysis of experimental work and theoretical and modeling considerations (68, 69), cytosolic Ca sequestration during ECC has emerged as one of the key factors that determine vulnerability to alternans. Generally, factors that increase Ca sequestration protect against or rescue from alternans, whereas impairment of Ca sequestration generates conditions that are favorable to alternans. Here, Ca sequestration refers to the net efficiency of clearing the cytosolic compartment of excessive Ca and includes Ca reuptake into the sarcoplasmic reticulum (SR) via sarco(endo)plasmic reticulum Ca ATPase (SERCA), extrusion via Na/Ca exchange (NCX) and plasmalemmal Ca-ATPase, cytosolic buffering, diastolic SR Ca leak, and finally mitochondrial Ca uptake and storage. While it is well accepted that mitochondria have the capability to sequester substantial amounts of Ca, it is still a matter of debate whether and how mitochondria might play a [Ca]_i regulatory role on a beat-to-beat basis. There is indeed experimental evidence for two different models of mitochondrial Ca uptake (18, 55). One model proposes that oscillatory changes of mitochondrial [Ca] ([Ca]_m) mirror the repetitive cytosolic CaTs (19, 27, 59, 72), whereas the alternative model interprets the magnitude of mitochondrial Ca uptake and [Ca]_m changes as an integrative signal of cytosolic Ca spiking (2, 32, 46, 64). Mitochondrial Ca uptake serves to sustain crucial physiological processes and functions, including Ca-dependent mitochondrial ATP production to couple mechanical workload with ATP supply (4, 14). Ca transport into the mitochondrial matrix is accomplished by the highly selective mitochondrial calcium uniporter (MCU) residing in the inner mitochondrial membrane (IMM). The MCU was molecularly described (6, 17) as a 480 kDa multimeric complex [MCU complex, mitochondrial calcium uniporter complex (MCUC)], composed of the pore-forming MCU subunit and the regulatory elements mitochondrial calcium uptake-1 protein (MICU1), mitochondrial calcium uptake-2 protein (MICU2), essential MCU regulator (EMRE), and mitochondrial calcium uniporter subunit B (MCUB) (53). The driving force for mitochondrial Ca uptake is governed by the highly negative electrical gradient and differences in Ca concentrations across the IMM. Ca efflux from the mitochondrial matrix compartment occurs through two main pathways. The main avenue of Ca extrusion involves the mitochondrial Na(Li)/Ca exchange (NCLX) mechanism (56). In addition, there is growing evidence that the mitochondrial permeability transition pore (mPTP) can serve as a physiological Ca extrusion mechanism (27). The mPTP resides in the IMM, and long-lasting mPTP openings in response to mitochondrial Ca overload and oxidative stress permit permeation of compounds up to the size of 1,500 Da, including mitochondrial proapoptotic proteins responsible for induction of cell death (9, 18). However, through an alternative gating mode, mPTP can open and close transiently and repetitively (or flicker) in a subconductance state to release Ca under physiological conditions (14, 36, 37). This permeability transition pore (PTP) opening mode is considered a mitochondrial Ca efflux or Ca escape pathway, potentially protective against mitochondrial Ca overload.

The goal of the present study was to test the hypothesis that enhancement of mitochondrial Ca sequestration reduces CaT alternans, whereas mitochondrial Ca uptake inhibition enhances alternans. The study evolved from our previous findings that inhibition of various mitochondrial functions caused CaT alternans (23, 24). Here we explored MCUC activation, enhancement of mitochondrial Ca uptake, and retention through mPTP inhibition as pharmacological targets for alternans prevention and rescue.

METHODS

Chemicals, solutions, and experimental conditions.

Unless otherwise stated, all chemicals were obtained from Sigma-Aldrich (St. Louis, MO). The external standard Tyrode solution was composed of (in mM) 135 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 d-glucose (pH 7.4, adjusted with NaOH). All experiments were performed at room temperature (20°C–22°C).

Isolation of atrial myocytes.

Left atrial myocytes were enzymatically isolated from male New Zealand White rabbits (2.5–2.7 kg; 49 rabbits; Charles River Laboratories, Wilmington, MA) via Langendorff perfusion as previously described (38). Briefly, rabbits were anesthetized by intravenous injection of pentobarbital sodium (50 mg/kg) and heparin (1,000 U/kg). Hearts were excised, mounted on a Langendorff apparatus, and retrogradely perfused with a Ca-free Tyrode solution, followed by an enzyme solution containing 22.5 $\mu$g/mL Liberase TH (Research Grade; Roche, Basel, Switzerland), 20 µM Ca, and minimal essential medium Eagle (MEM) at 37°C. The left atrium was dissected from the heart and filtered and washed in MEM solution containing 50 µM Ca and 1% BSA. Isolated myocytes were washed and kept in MEM solution with 50 µM Ca until used. All aspects of animal husbandry, animal handling, anesthesia, surgery, and euthanasia were fully approved by the Institutional Animal Care and Use Committee (IACUC) of Rush University Chicago and comply with United States regulations on animal experimentation.

Cytosolic Ca measurements.

Cytosolic CaTs were recorded in intact cells with Fluo-4 AM using an Ionoptix epifluorescence microscopy setup (IonOptix, Milton, MA). Myocytes were electrically stimulated through a pair of platinum electrodes with electrical stimuli of voltage ~50% greater than the threshold for myocyte contraction (40). Left atrial myocytes were loaded with 10 µM Fluo-4 AM in standard Tyrode solution for 20 min. The cells were washed with standard Tyrode solution for 20 min to allow deesterification of the indicator. Fluo-4 fluorescence was excited at 488 nm (Xe arc lamp) and recorded at 515 nm. Fluorescence emission signals (F) were background subtracted and normalized to diastolic cytosolic Ca ([Ca]_i) (F₀), and changes of [Ca]_i are presented as F/F₀. In case of CaT alternans, diastolic [Ca]_i before the large amplitude CaT was chosen as F₀. CaT alternans was induced by increasing the pacing frequency until stable CaT alternans was observed. The degree of CaT alternans was quantified as alternans ratio (AR), defined as AR = 1 − CaT_small/CaT_large, where CaT_small and CaT_large are the amplitudes of the small and large CaTs of a pair of alternating CaTs, respectively (41, 71). Accordingly, AR values vary on a continuum between 0 and 1. AR = 0 indicates no alternans, and AR = 1 indicates the highest possible degree of Ca alternans, with every other stimulation failing to elicit a measurable CaT. CaTs were considered alternating when the beat-to-beat difference in CaT amplitude exceeded 10% (AR > 0.1) (38). CaT amplitude (∆F/F₀) was measured as the difference in F/F₀ measured immediately before the stimulation pulse and the peak of the CaT. To calculate average ARs, 7–10 consecutive alternans CaT pairs (14–20 individual CaTs) were analyzed.

SR Ca load measurements.

Rapid application of caffeine (10 mM) was used to estimate the SR Ca load (23). SR Ca load was quantified as the amplitude (ΔF/F₀) of the caffeine-induced cytosolic CaT.

Mitochondrial Ca uptake measurements.

[Ca]_m changes were measured in digitonin-permeabilized atrial cells previously loaded with 15 µM Fluo-4 AM for 40 min at room temperature in standard Tyrode solution as previously described (64). Briefly, myocytes were permeabilized in an intracellular solution containing 135 mM KCl, 0.5 mM MgCl₂, 0.5 mM KH₂PO₄, 5 mM succinate, 2 µg/mL rotenone, 20 mM HEPES, 15 mM 2,3-butanedione monoxime, 0.01 mM digitonin, 5 mM EGTA, and 1.86 mM CaCl₂ to obtain a free [Ca]_i of ~100 nM (pH 7.2, adjusted with KOH). After permeabilization, the bath solution was changed to the same intracellular solution but without digitonin. Myocytes were exposed to increasing concentrations of extramitochondrial free Ca ([Ca]_em) from 0.1 to 0.5 and 10 µM. Free Ca concentrations were calculated using the MaxChelator program (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/CaEGTA-NIST.htm). The mitochondrial Fluo-4 signal was background subtracted and normalized (%) by setting the fluorescence signal measured in [Ca]_em = 0.1 µM after digitonin treatment to 0 and the signal measured in saturating [Ca]_em (F_max; [Ca]_em = 10 µM) to 100%.

Statistical analysis.

Data were analyzed using Ion Wizard 6.2 (IonOptix), OriginPro2016 (OriginLab, Wellesley Hills, MA), and Prism software (GraphPad San Diego, CA). All summary data are presented as means ± SE for the indicated number (n) of cells. For two-group comparisons, Student’s t test for paired or unpaired data was used. In the case of multiple groups, one-way ANOVA for repeated or single measurements with Tukey’s post hoc test was used. Data were considered significant at P < 0.05.

RESULTS

Effects of MCUC activity on CaT alternans.

We tested the effect of pharmacological inhibition and activation of mitochondrial Ca uptake on the severity of pacing-induced CaT alternans in rabbit atrial myocytes (Fig. 1). In line with our previous observation, exposure to the specific MCUC inhibitor Ru360 (5 µM) enhanced CaT alternans (Fig. 1A). Ru360 is considered to be a rather specific MCUC inhibitor that does not affect Ca uptake and release from SR, NCX, L-type Ca current, and actomyosin ATPase activity (25, 48). In the presence of Ru360, the AR increased by 43% from 0.46 ± 0.06 to 0.66 ± 0.07 (P = 0.0052, n = 9). The opposite effect was observed when mitochondrial Ca uptake was stimulated pharmacologically (Fig. 1B). Activation of MCUC with the polyamine compound spermine (136 µM), a known enhancer of the MCUC (22, 29, 54), rescued CaT alternans. Spermine (Spm) is an allosteric MCUC agonist that decreases its apparent K_m for Ca uptake (54). As shown in Fig. 1B, Spm reduced CaT AR by 76% from 0.49 ± 0.03 in control to 0.12 ± 0.03 (P < 0.0001, n = 21). Similar results were found in supporting experiments using the flavonoid kaempferol (Kmpf) that has been reported to enhance mitochondrial Ca uptake (5, 7, 50) by increasing MCUC affinity for Ca and accelerating Ca uptake (5, 50). Kmpf (10 µM) suppressed CaT alternans and decreased AR by 79% from 0.53 ± 0.03 to 0.11 ± 0.02 (P < 0.0001, n = 14; data not shown). In an additional set of experiments (Fig. 1C), we sequentially exposed atrial myocytes to both drugs. Figure 1C shows that Ru360 enhanced CaT alternans (in this series of experiments, the AR increased by 86% from 0.28 ± 0.04 to 0.52 ± 0.07, P = 0.128, n = 3); however, subsequent exposure to Spm rescued CaT alternans and diminished AR by 79% from 0.52 ± 0.07 to 0.11 ± 0.07 (P = 0.028, n = 3). Overall, these data support the hypothesis that cytosolic clearance of Ca by mitochondrial Ca sequestration protects from CaT alternans, and thus MCUC activity modulation can exert opposite effects, resulting in abolishment or enhancement of CaT alternans.

Fig. 1.

Effect of mitochondrial calcium uniporter complex (MCUC) activity on calcium transient (CaT) alternans. In this set of experiments, pacing frequency was increased under control conditions until stable CaT alternans were established. Subsequently, pharmacological MCUC modulators were applied at the same stimulation frequency. A: CaT alternans recordings from atrial myocytes before [control (Ctrl)] and after application of the MCUC inhibitor Ru360 (5 μM). Right: average and individual cell CaT alternans ratios (ARs) in Ctrl and in Ru360 (n = 9, P = 0.0052, paired t test). B: CaT alternans recordings before (Ctrl) and after the application of the MCUC activator spermine (Spm) (136 µM). Right: average and individual cell CaT ARs in Ctrl and in Spm (n = 21, P = 0.0001, paired t test). C: interplay between MCUC inhibition and MCUC activation. CaTs recordings in Ctrl, followed by Ru360 and subsequent Spm exposure of the same cell. After AR increased in Ru360 (n = 3, P = 0.128, ANOVA), Spm led to a significant AR decrease (n = 3, P = 0.028, ANOVA).

In the next set of experiments, we determined whether mitochondrial Ca uptake affects the pacing threshold where CaTs alternans develops (Fig. 2). The CaT AR was assessed as a function of a stepwise increase of electrical stimulation frequency under control conditions and during exposure to Ru360 and spermine. Individual atrial myocytes were stimulated at increasing frequencies ranging from 0.5 to 2–3 Hz, in increments of 0.25–0.5 Hz. The CaT alternans pacing threshold was defined as AR exceeding a value of 0.1. Figure 2A shows the effect of MCUC inhibitor Ru360 on alternans pacing threshold and AR frequency dependence. Figure 2A, top, shows CaT recordings obtained from the same atrial myocytes under control conditions at three different frequencies. CaT recordings at identical frequencies in the presence of Ru360 (Fig. 2A, bottom), revealed that CaT alternans started to develop at a lower pacing frequency, and the degree of alternans (AR) was increased. Figure 2B summarizes the frequency dependency of AR in control and in the presence of Ru360. Ru360 increased AR across the entire stimulation range >1 Hz and shifted the pacing threshold from 1.82 to 1.1 Hz (P = 0.0250).

Fig. 2.

Mitochondrial calcium uniporter complex (MCUC) activity modulates calcium transient (CaT) alternans pacing threshold. A: effect of MCUC inhibition by Ru360. CaT recordings from atrial cells during electrical pacing at frequencies of 0.75, 1.25, and 2 Hz (n = 7) in control (Ctrl) (top) and in presence of Ru360 (5 µM, n = 8; bottom). Black circles indicate time of electrical stimulus. In the example shown, two pairs of alternans CaTs with an alternans ratio (AR) = 1 occurred. B: average ARs in control and Ru360 as a function of stimulation frequency (range 0.5–2.5 Hz). Arrow indicates AR pacing threshold shift; horizontal dashed line marks alternans threshold (AR > 0.1); analyzed by unpaired t test. C: effect of MCUC activation by spermine (Spm). CaTs recordings from atrial cells during electrical pacing at frequencies of 0.75, 1.25, and 1.75 Hz in control (top; n = 7) and in presence of Spm (bottom; 68 µM, n = 9). D: average ARs in control and Spm as a function of stimulation frequency (range 0.5–2.25 Hz). Arrow indicates AR pacing threshold shift; horizontal dashed line marks alternans threshold (AR > 0.1); analyzed by unpaired t test.

Opposite effects were observed with spermine. CaT traces reveal that Spm protects from alternans compared with control conditions (Fig. 2C). Spermine reduced AR across the entire frequency range tested (Fig. 2D) and shifted the alternans pacing threshold from 0.88 to 1.16 Hz (P = 0.0793).

Effect of spermine on mitochondrial Ca uptake.

To demonstrate the effects of Spm on mitochondrial Ca uptake directly, we measured [Ca]_m changes in single surface membrane permeabilized atrial myocytes with the Ca indicator Fluo-4 entrapped in mitochondria (64). After membrane permeabilization with digitonin, cells were exposed to 0.1 µM extramitochondrial Ca ([Ca]_em) simulating resting [Ca]_i in cardiac cells (Fig. 3A). Subsequently, myocytes were exposed to [Ca]_em = 0.5 µM, followed by an increase of [Ca]_em to 10 µM. [Ca]_m was normalized to F/F₀ level in 10 µM [Ca]_em. In [Ca]_em = 0.5 µM, [Ca]_m increased on average to 59 ± 3% (n = 7) of maximum F/F₀ in control (Fig. 3B). In the presence of Spm (136 µM), mitochondrial Ca uptake was enhanced and accelerated. [Ca]_m increased to 86 ± 2% (n = 7) of maximum in [Ca]_em = 10 µM, corresponding to a 46% enhancement of mitochondrial Ca uptake (P < 0.0001). Furthermore, Spm accelerated the rate of mitochondrial Ca uptake, expressed as the rate of [Ca]_m change to increase [Ca]_m from 10 to 90% of the level reached in [Ca]_em = 0.5 µM. The rate of [Ca]_m change increased from ∆%F/s = 0.58 ± 0.16 in control to 1.05 ± 0.10 in Spm (P = 0.0023, n = 7; Fig. 3C). In contrast, as shown previously (33), the MCUC inhibitor Ru360 slowed mitochondrial Ca uptake by nearly two orders of magnitude, demonstrating directly the inhibitory effect of Ru360 on mitochondrial Ca uptake.

Fig. 3.

Mitochondrial calcium (Ca) uptake stimulation by spermine (Spm) in permeabilized atrial myocytes. A: [Ca]_m in control (Ctrl) (n = 7) and in presence of Spm (136 µM, n = 7) at [Ca]_em = 0.1, 0.5, and 10 µM. [Ca]_m is normalized (%) to [Ca]_m measured in [Ca]_em = 10 µM. Dashed lines indicate [Ca]_m at [Ca]_em = 0.5 µM. B: average [Ca]_m at [Ca]_em = 0.5 µM, expressed as % of maximum fluorescence measured in [Ca]_em = 10 µM. C: mitochondrial Ca uptake rate (Δ%F/s) to increasse [Ca]_m from 10 to 90% in [Ca]_em = 0.5 µM. Statistical analysis by unpaired t test.

Effect of mPTP inhibition on CaT alternans.

As it has been shown that transient mPTP opening can serve as a physiological mitochondrial Ca efflux pathway (14, 36, 37), we hypothesized that mPTP inhibition enhanced mitochondrial Ca retention and therefore counteracted the development of CaT alternans. Mitochondrial PTP was inhibited with cyclosporine A (CsA) (8, 16, 30). CsA (10 µM) mitigated pacing-induced CaT alternans (Fig. 4A) by reducing the average AR by 70% from 0.46 ± 0.03 to 0.14 ± 0.02 (n = 19, P < 0.0001; Fig. 4B). Furthermore, CsA reduced AR (Fig. 4C) across the entire range of stimulation frequencies tested (0.5–3.0 Hz) and shifted the pacing frequency threshold for CaT alternans from 1.48 to 2.64 Hz (Fig. 4D). Taken together, these data indicate that mPTP opening is involved in the development of CaT alternans, and CaT alternans is attenuated by CsA-mediated mPTP inhibition.

Fig. 4.

Effect of mitochondrial permeability transition pore (mPTP) inhibition on calcium transient (CaT) alternans. A: CaT alternans recordings from atrial myocytes before [control (Ctrl)] and after application of the mPTP inhibitor cyclosporine A (CsA) (10 µM). B: average and individual cell CaT alternans ratios (ARs) in Ctrl and in CsA (n = 19). Statistical analysis by paired t test. C: CaT recordings from atrial cells during electrical pacing at frequencies of 1, 1.5, and 2.5 Hz in Ctrl (top; n = 7) and in presence of CsA (bottom; n = 5). D: average ARs in Ctrl and CsA as a function of stimulation frequency (0.5–3 Hz). Arrow indicates AR pacing threshold shift; horizontal dashed line marks alternans threshold (AR > 0.1); analyzed by unpaired t test.

Effect of [Ca]_m modulators on CaT amplitude.

To address potential off-target effects of [Ca]_m modulators on Ca release during ECC other than affecting mitochondrial Ca handling, we measured the average amplitude of two consecutive CaTs in control and in presence of Ru360, Spm, and CsA. We hypothesized that if such off-target drug effects were present, the average amplitude (ΔF/F₀) of two consecutive CaTs would potentially be affected by the degree of CaT alternans. The data show that for none of the three drugs, the average amplitudes were different from control [average CaT amplitude (∆F/F₀) control vs. drug; Ru360: 0.93 ± 0.11 and 0.74 ± 0.11, n = 9; Spm: 1.13 ± 0.16 and 0.97 ± 0.16, n = 21; and CsA: 1.33 ± 0.15 and 1.16 ± 0.13, n = 19]. None of these comparisons revealed statistically significant differences, suggesting absence of any major additional effects of these drugs on Ca release and ECC.

SR Ca load during enhanced mitochondrial Ca uptake.

SERCA activity is responsible for diastolic SR refilling, a process highly dependent on ATP availability. Mitochondria are the main source of cellular ATP, and mitochondrial ATP production is Ca dependent and occurs by oxidative metabolism. We therefore tested the hypothesis whether enhanced mitochondrial Ca uptake and retention resulted in SERCA stimulation and increased SR Ca load. SR Ca load was determined by measuring the CaT amplitude elicited by rapid application of caffeine (10 mM) during pacing-induced CaT alternans. Figure 5A shows caffeine CaTs during alternans in control conditions immediately after a small and after a large CaT. CaTs elicited by caffeine were identical in amplitude, irrespective of whether they were evoked after a small or a large alternans CaT (Fig. 5D). These results are consistent with our earlier observations that CaT alternans can occur in the absence of alternans of end-diastolic SR [Ca] (35, 58, 66). Furthermore, enhanced mitochondrial Ca uptake by Spm (Fig. 5B) or increased Ca retention by CsA (Fig. 5C) failed to affect SR Ca load, suggesting that a potential stimulation of mitochondrial ATP production did not precipitate in enhanced SR Ca uptake. These findings are reminiscent of our previous observations that also inhibition of mitochondrial Ca extrusion via NCLX, mitochondrial Ca-dependent dehydrogenases, or mitochondrial F₁/F₀-ATP synthase failed to affect SR Ca load (23).

Fig. 5.

Effect of mitochondrial calcium uniporter complex (MCUC) stimulation and mitochondrial permeability transition pore inhibition on sarcoplasmic reticulum (SR) calcium (Ca) load. Cytosolic calcium transients (CaTs) elicited with rapid application of 10 mM caffeine to estimate SR Ca load. A: caffeine-induced CaTs after a small [control (Ctrl) S] amplitude (n = 23) and a large (Ctrl L) amplitude (n = 12) alternans CaT. B: caffeine-induced CaT in spermine (Spm) (136 µM, n = 11). C: caffeine-induced CaT in cyclosporine A (CsA) (10 µM, n = 11). D: average amplitudes and individual cell amplitudes (ΔF/F₀) of caffeine-induced CaTs after small and large amplitude control alternans CaTs, Spm, and CsA. Averages were not significantly different among all groups (ANOVA).

DISCUSSION

Effects of mitochondrial Ca handling on CaT alternans.

In this study, we tested the hypothesis that mitochondrial Ca sequestration plays a critical role in the development of atrial calcium alternans. The main novel findings are as follows: 1) stimulation of mitochondrial Ca uptake with spermine abolished pacing-induced CaT alternans and reversed the effect of MCUC inhibitor Ru360, which through inhibition of mitochondrial Ca uptake, enhanced CaT alternans; 2) spermine shifted the alternans pacing threshold to higher frequencies, whereas Ru360 has the opposite effect on pacing threshold; 3) mPTP inhibition with CsA abolished CaT alternans and increased the alternans pacing threshold; and 4) spermine and CsA affected alternans propensity and threshold without any effect on SR Ca levels.

Our results are in line with the general notion that impaired cytosolic Ca sequestration efficiency during ECC increases the susceptibility for alternans, whereas efficient cytosolic Ca clearance is protective against alternans (68, 69). Our current and previous experiments demonstrated directly that spermine stimulated and Ru360 inhibited mitochondrial Ca uptake. In the current study, Spm increased mitochondrial Ca uptake by ~60% (Fig. 3) and doubled the rate of Ca uptake. In a previous study (33), we showed that Ru360 slowed atrial mitochondrial Ca uptake rate by nearly two orders of magnitude (46-fold).

It is well established that mitochondrial matrix Ca couples cardiac workload with ATP production, which in turn maintains ion gradients across the sarcolemma and SR membranes and fuels phosphorylation processes that modify the activity of Ca handling proteins. Mitochondrial Ca handling and Ca-dependent mitochondrial processes affect cardiac alternans in complex ways. We have shown previously that pharmacological inhibition of not only Ca uptake but also efflux via NCLX, Ca-dependent dehydrogenases of the Krebs cycle, several complexes of the electron transport chain, and the ATP synthesis machinery, as well as collapsing the mitochondrial membrane potential (ΔΨ_m), all enhanced CaT alternans (23, 24). As ATP is an activator of the SR ryanodine receptor (RyR) Ca release channel, and SERCA, impairment of ATP production is expected to compromise RyR and SERCA activity (75). In a comprehensive multi-parameter (75) study, we investigated the effect of mitochondrial uncoupling (with FCCP) on cellular Ca signaling and ECC. FCCP caused profound CaT alternans (23, 24). FCCP-induced mitochondrial depolarization resulted in a biphasic effect on SR Ca release, consisting of an initial inhibition followed by stimulation. The initial inhibition was the result of suppression of RyR channel activity by a decrease in [ATP], an increase of [Mg]_i, and cytoplasmic acidification and was accompanied by reactive oxygen species (ROS) formation. The subsequent stimulation resulted from reduced mitochondrial Ca buffering, cytosolic Na and Ca accumulation, SR Ca uptake (fueled by compensatory glycolytic ATP production), SR Ca overload, and eventually spontaneous Ca release in form of Ca waves. In this sequence of events initiated by mitochondrial uncoupling and ATP production impairment, the fact that SERCA activity can be maintained in part through activation of ATP production by glycolysis appears to be crucial for the regulation of CaT alternans. We have shown previously that selective inhibition of glycolysis caused alternans (35, 43), and the capability of β-adrenergic stimulation to rescue alternans was compromised when glycolysis was prevented from serving as a compensatory ATP source (24). These findings are consistent with the notion that the SERCA pump is closely associated with glycolytic enzymes (74, 75). This conclusion is supported by the observation that despite enhanced mitochondrial Ca uptake and Ca retention (Fig. 5), which would be expected to enhance ATP production, or despite direct inhibition of mitochondrial F₁/F₀-ATP synthase with oligomycin (23), there was no change in the SR Ca content, suggesting a degree of independence of SERCA from mitochondrial ATP production. There is also the possibility that enhanced mitochondrial Ca uptake deprives SERCA of Ca to be transported into the SR; thus, the net effect is essentially an unaltered SR Ca content. It is further noteworthy that Spm has been reported to enhance pyruvate dehydrogenase and citrate synthase activities by a Ca-independent mechanism (51, 62).

Unique features of atrial ECC add to the complexity of the relationship between alternans susceptibility and mitochondrial Ca cycling. Previous results from our group, as well as others, have shown that unlike in ventricular cells, in atrial myocytes, the transverse (T) tubule membrane system is poorly developed and even entirely lacking in rabbit atrial cells (11, 12, 34, 35, 41, 66, 67). Nonetheless, in the cell interior, nonjunctional SR is abundant and robustly endowed with RyR Ca release channels (11, 13, 42). As a consequence, the cell-wide CaT evoked by an action potential is spatially inhomogeneous (10). SR Ca release is initiated in the cell periphery from where Ca release from nonjunctional SR Ca release sites propagates in Ca wave-like fashion centripetally to the cell center. Because of the lack of a T-tubular system, the sarcolemmal Na/Ca exchanger, which in ventricular myocytes carries the main burden of beat-to-beat Ca removal, has no direct role in nonperipheral atrial cell regions, and thus beat-to-beat removal is reduced to Ca clearance by SERCA, mitochondrial Ca uptake activity, and to a minor extent cytosolic Ca buffering capacity (10, 33). Despite the ongoing controversy of how mitochondrial Ca uptake contributes to shaping the cytosolic CaT on a beat-to-beat basis, i.e., whether mitochondria show [Ca]_m oscillations mirroring cytosolic CaTs or whether beat-to-beat [Ca]_m changes are small and changes of [Ca]_m rather reflect an integrative response to cytosolic [Ca]_i (18, 55), we found in a previous study that centrally located mitochondria can modulate the CaT amplitude (33). Functional mitochondrial Ca uptake blunted the velocity of centripetal propagation of Ca release from nonjunctional SR, whereas MCUC inhibition with Ru360 accelerated propagation and increased the amplitude of central CaTs. The latter is indicative of an increase in SR Ca load and fractional Ca release, a condition known to be favorable for the development of alternans (20, 68). Thus, MCUC inhibition promotes alternans by a dual mechanism, predominantly by inhibition of cytosolic Ca clearance but also by enhancement of fractional SR Ca release (68, 69).

Effects of mPTP inhibition on CaT alternans.

We also investigated how mitochondrial Ca extrusion pathways affect alternans susceptibility. We showed previously that inhibition of the main mitochondrial Ca extrusion pathway, the mitochondrial NCLX, enhanced CaT alternans (23). We suggested that Ca taken up by mitochondria is not able to return to the cytosol, subsequently promoting long-lasting mPTP opening that leads to mitochondrial membrane depolarization and dysfunction (9). This notion was supported in our previous work by the experimental observation that inhibition of NCLX caused pronounced CaT alternans, but subsequent treatment with CsA to inhibit mPTP ameliorated CaT alternans (23). The mPTP has been described to be capable of opening transiently at a lower conductance level to serve as a Ca escape pathway that counteracts mitochondrial Ca overload under physiological conditions (27, 36, 37). We tested the hypothesis that mPTP inhibition with CsA increased mitochondrial Ca retention (thus assists cytosolic Ca clearance) and protects from alternans. CsA is an immunosuppressive drug that interacts with the mitochondrial protein cyclophilin-D (CypD) (8, 30), a structural mPTP component, resulting in mPTP inhibition (3, 16, 57). Also, recent observations suggest that CsA improves mitochondrial Ca sequestration by P_i-dependent mitochondrial Ca buffering that delays Ca-induced mPTP opening, thus maintaining mitochondrial function during Ca overload (49). In our experiments, CsA suppressed CaT alternans and increased the pacing threshold where CaT alternans developed, presumably by preserving ΔΨm and limiting Ca release to the cytosol. Moreover, the deletion of the Ppif gene that encodes cyclophilin-D (CypD knockout mice) protected from developing alternans in both single cells and at the whole heart level (28). There is also evidence that CsA is an inhibitor of calcineurin, a serine/threonine phosphatase that modulates the phosphorylation of phospholamban (15, 52), which would increase SERCA activity and protect from alternans by a mPTP-independent mechanism; however, in our experiments, we did not observe changes in the SR Ca load after CsA treatment (Fig. 5, C and D). Interestingly, spermine has been shown to reduce ROS production and inhibit mPTP (1, 62); thus, Spm has anti-alternans actions through MCUC activation and possibly also mPTP inhibition.

Mitochondrial Ca cycling and alternans in cardiac disease.

Ca alternans is associated with cardiac pathophysiologies such as HF, aortic valve disease, and myocardial ischemia, which are commonly accompanied by arrhythmias (21, 70). Ca alternans has been associated with the development of reentry phenomena (39, 69), providing an arrhythmogenic substrate that underlies atrial fibrillation, the most common cardiac arrhythmia that is prevalent in the elderly population and represents a leading cause of HF and stroke (41). In HF, the increase in Ca leak from the RyR, reduced SERCA expression, and mitochondrial dysfunction compromise cytosolic Ca sequestration, all conditions that enhance the propensity of Ca alternans. Several studies have suggested that improving mitochondrial function might be a strategy for ameliorating HF manifestations (44, 47, 60, 61). Recent work suggested that MCU overexpression in a pressure overload-induced HF model improved mitochondrial function, enhanced contractility, and inhibited arrhythmias (45). Along the same line, pharmacological MCUC activation had anti-arrhythmic effects in a murine model of RyR-mediated catecholaminergic polymorphic ventricular tachycardia (63). In a zebrafish model, stimulation of mitochondrial Ca uptake improved arrhythmogenic Ca waves and irregular contractions (65). While these studies document an improvement of arrhythmogenesis through enhancement of mitochondrial Ca uptake, other studies point in the opposite direction. Xie et al. (73) showed that action potential prolongations, early afterdepolarizations, and ventricular fibrillation in a hypertension-induced mouse HF model were abrogated with Ru360 or after genetic knockdown of MCU. Ru360 was also found to ameliorate mitochondrial damage and decrease arrhythmia incidence in a rat reperfusion injury model (26). Also, ventricular fibrillation was substantially reduced by MCUC inhibition with Ru360, in a pressure overload-induced hypertrophy model (31). Despite the differences in results and conclusions, these studies clearly indicate the central role played by MCU for cellular Ca signaling and the propensity for arrhythmia.

In conclusion, our data show that in atrial myocytes, functional mitochondrial Ca uptake and retention reduce cardiac alternans development. In cardiac disease, mitochondrial integrity is jeopardized and renders the heart more susceptible to alternans and arrhythmic disturbances. Therefore, interventions that increase mitochondrial Ca uptake and retention have the potential to be a therapeutic target for arrhythmia risk reduction.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-057832, HL-132871, and HL-134781.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.O.-A. and L.A.B. conceived and designed research; Y.O.-A. performed experiments; Y.O.-A. and L.A.B. analyzed data; Y.O.-A. and L.A.B. interpreted results of experiments; Y.O.-A. and L.A.B. prepared figures; Y.O.-A. and L.A.B. drafted manuscript; Y.O.-A. and L.A.B. edited and revised manuscript; Y.O.-A. and L.A.B. approved final version of manuscript.