Distinct Properties of Ca²⁺ Efflux from Brain, Heart and Liver Mitochondria: The Effects of Na⁺, Li⁺ and the Mitochondrial Na⁺/Ca²⁺ Exchange Inhibitor CGP37157

Abstract

Mitochondrial Ca²⁺ transport is essential for regulating cell bioenergetics, Ca²⁺ signaling and cell death. Mitochondria accumulate Ca²⁺ via the mitochondrial Ca²⁺ uniporter (MCU), whereas Ca²⁺ is extruded by the mitochondrial Na⁺/Ca²⁺ (mtNCX) and H⁺/Ca²⁺ exchangers. The balance between these processes is essential for preventing toxic mitochondrial Ca²⁺ overload. Recent work demonstrated that MCU activity varies significantly among tissues, likely reflecting tissue-specific Ca²⁺ signaling and energy needs. It is less clear whether this diversity in MCU activity is matched by tissue-specific diversity in mitochondrial Ca²⁺ extrusion. Here we compared properties of mitochondrial Ca²⁺ extrusion in three tissues with prominent mitochondria function: brain, heart and liver. At the transcript level, expression of the Na⁺/Ca²⁺/Li⁺ exchanger (NCLX), which has been proposed to mediate mtNCX transport, was significantly greater in liver than in brain or heart. At the functional level, Na⁺ robustly activated Ca²⁺ efflux from brain and heart mitochondria, but not from liver mitochondria. The mtNCX inhibitor CGP37157 blocked Ca²⁺ efflux from brain and heart mitochondria but had no effect in liver mitochondria. Replacement of Na⁺ with Li⁺ to test the involvement of NCLX, resulted in a slowing of mitochondrial Ca²⁺ efflux by ~70%. Collectively, our findings suggest that mtNCX is responsible for Ca²⁺ extrusion from the mitochondria of the brain and heart, but plays only a small, if any, role in mitochondria of the liver. They also reveal that Li⁺ is significantly less effective than Na⁺ in driving mitochondrial Ca²⁺ efflux.

Graphical Abstract

1. Introduction

Mitochondrial Ca²⁺ transport is essential for many cellular functions including the regulation of cell bioenergetics, Ca²⁺ signaling and cell death and survival [1–5]. Ca²⁺ uptake by mitochondria is mediated by the mitochondrial Ca²⁺ uniporter (MCU) complex, which functions as a Ca²⁺-selective and Ca²⁺-gated ion channel, whereas Ca²⁺ extrusion from mitochondria is controlled by Na⁺/Ca²⁺ and H⁺/Ca²⁺ exchangers (mtNCX and mtHCX, respectively) [3, 4, 6, 7]. Recent analysis in isolated mitochondria has demonstrated that MCU activity varies significantly among tissues, with some of the lowest activity reported for heart mitochondria, and higher activities found in mitochondria from liver and skeletal muscle [8, 9]. In addition, mitochondria from different tissues displayed a difference in the Ca²⁺ activation threshold and cooperativity [9]. This variability in mitochondrial Ca²⁺ uptake likely reflects attuning of the mitochondrial Ca²⁺ transport machinery to tissue-specific Ca²⁺ signaling and energy needs, and is fundamental for optimal performance of mitochondria in a given tissue [3, 8–10].

Ca²⁺ that accumulates in mitochondria during cellular activity is subsequently released back into the cytosol via mtNCX and mtHCX [3, 4, 6]. Earlier work in isolated mitochondria suggested that mtNCX plays a prominent role in removing Ca²⁺ from mitochondria in excitable tissues such as the brain, heart and skeletal muscle, whereas mtHCX is likely responsible for mitochondrial Ca²⁺ extrusion in non-excitable tissues such as liver, kidney and lung [6, 11, 12]. However recent identification of the protein NCLX (a product of gene Slc8b1) as a molecule likely responsible for the mtNCX transport has provided a new molecular perspective on the role of mtNCX in various tissues [13–19]. Notably, NCLX was found to be expressed in the liver, kidney and lung at levels similar or higher than those found in brain and heart [10], suggesting that the mtNCX mechanism could contribute to mitochondrial Ca²⁺ extrusion in both excitable and non-excitable tissues [20, 21]. The role of mtNCX is especially prominent in neurons, where it can significantly prolong elevation of the cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt) and regulate Ca²⁺-dependent processes such as synaptic transmission and gene expression [22–24]. Equally important, Ca²⁺ extrusion mediated by mtNCX is critical for preventing toxic Ca²⁺ build-up in mitochondria and protecting neurons and cardiomyocytes from cell death in the context of either excessive stimulation or ischemia [18, 25, 26]. Similarly, in non-excitable tissues, mtNCX was found to provide crucial support for the function and survival of cells, specifically for brown fat adipocytes in the context of adrenergic stimulation [27]. Thus, a precisely controlled balance between the activities of the mitochondrial Ca²⁺ uptake and efflux machinery is essential for preventing toxic mitochondrial Ca²⁺ overload and cell death. This concept also raises another key question of whether the tissue-specific variability in the activity of mitochondrial Ca²⁺ uptake is paralleled by a similarly diverse and tissue-specific activity of mitochondrial Ca²⁺ extrusion.

Thus, we sought to address this important question by carrying out a quantitative comparison of Ca²⁺ extrusion from isolated mitochondria obtained from the three tissues with prominent mitochondria function: brain, heart and liver. A second objective of this study was to re-assess the role of mtNCX in excitable and non-excitable tissues. To this end, we developed an assay based on the use of a low-affinity fluorescent Ca²⁺ indicator Ca²⁺-Green 5N (Kd~15 µM), for monitoring Ca²⁺ extrusion from isolated mitochondria under standardized and highly controlled recording conditions. A property unique to NCLX among the Na⁺/Ca²⁺ exchangers is that Li⁺ is nearly as effective as Na⁺ in driving its transmembrane transport of Ca²⁺ [28–31]. This unique feature of NCLX can be utilized to gain further insight into the molecular nature of the mitochondrial Ca²⁺ extrusion machinery. Thus, as a part of our assessment, we examined the sensitivities of mitochondrial Ca²⁺ extrusion to Li⁺ as well as to the mtNCX inhibitor CGP37157, which is a widely accepted pharmacological tool for studying mtNCX activity [32–34]. Our analysis revealed that Na⁺ strongly activated Ca²⁺ efflux from isolated brain and heart mitochondria, which was inhibited by CGP37157. In contrast, Na⁺ had no effect on Ca²⁺ efflux from isolated liver mitochondria, and the efflux was also insensitive to CGP37157. These findings were corroborated in intact primary neurons and hepatocytes by simultaneous measurement of the cytosolic and mitochondrial Ca²⁺ concentrations ([Ca²⁺]_cyt and [Ca²⁺]_mt, respectively). Although Li⁺ was able to promote Ca²⁺ efflux from both brain and heart mitochondria, the rate of Li⁺-driven Ca²⁺ efflux was only ~30% of that induced by Na⁺. Collectively, these findings suggest that mtNCX is the principal mechanism involved in extruding Ca²⁺ from mitochondria of the brain and heart, but that it does not play a significant role in mediating Ca²⁺ efflux from liver mitochondria. Our data also reveal that Li⁺ is significantly less effective than Na⁺ in its ability to drive mitochondrial Ca²⁺ efflux.

2. Materials and methods

2.1. Mitochondria isolation and measurement of Ca²⁺ efflux

All experiments involving mice were approved by the University of Iowa Institutional Animal Care and Use Committee and were carried out in strict accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Every effort was made to minimize the number of mice used and their suffering. Mitochondria were isolated from adult (6–10 weeks) C57BL/6J mice, using a protocol similar to one that was previously described, with some modifications [35]. Briefly, mice were anesthetized, swiftly decapitated and the brain was quickly removed. Brains were chopped and then homogenized in ice-chilled mitochondrial isolation buffer (MIB; 225 mM mannitol, 75 mM sucrose, 2 mM K₂HPO₄, 5 mM HEPES, 1 mM EGTA, 0.1% FAF BSA, pH 7.2 w/ KOH) using a tissue grinder (E2355, Eberbach; Ann Arbor MI). Homogenate was centrifuged at 1500 g (5 min) to pellet unlysed cells, and then the supernatant was additionally centrifuged at 21,000 g (10 min) to pellet crude mitochondria. Pelleted mitochondria were resuspended in 3.5 mL of 15% Percoll solution, which was layered on top of 3.7 mL 24% and 1 mL 40% Percoll solutions in MIB. The solutions were then centrifuged at 30,700 g for 8 min. Non-synaptosomal mitochondria were collected from a band between the 24% and 40% Percoll layers. Mitochondria were then resuspended in a 1:5 Percoll and MIB solution, and centrifuged at 6900 g for 10 min. The mitochondrial pellet was then resuspended in EGTA- and BSA-free MIB and then each sample was subjected to bicinchoninic acid assay (BCA) analysis.

For the isolation of heart and liver mitochondria, essentially the same protocol was used as for the brain mitochondria, but with slight modifications. Briefly, mice were anesthetized and swiftly decapitated, after which the tissue was quickly harvested. Tissue was chopped and homogenized (see above) in ice-chilled isolation buffer 1 (IB1; 225 mM mannitol, 75 mM sucrose, 10 mM HEPES, 1 mM EGTA, 0.1% FAF BSA, pH 7.4 with KOH). The homogenate was centrifuged at 700 g (10 min), and then the supernatant was centrifuged at 18,900 g (10 min). The pellet was resuspended in isolation buffer 2 (IB2; 225 mM mannitol, 75 mM sucrose, 10 mM HEPES, 0.1 mM EGTA, pH 7.4 w/ KOH) and then centrifuged at 18,900 g (10 min). Then, the pellet was resuspended in EGTA-free IB2 and BCA analysis was performed.

Mitochondrial Ca²⁺ flux was monitored using the low-affinity fluorescent Ca²⁺ indicator Calcium Green 5N (Kd~15 µM) and measuring changes in the extramitochondrial Ca²⁺ concentration. Specifically, mitochondria were suspended in a respiration buffer (Res; 150 mM KCl, 2 mM K₂HPO₄, 0.5 mM MgCl₂, 5 mM malate, 5 mM glutamate, 0.1 mM ADP, 10 mM HEPES, 10 μM EGTA, and 1 μM Calcium Green 5N) in a 96 well, black/clear, tissue culture treated plate (Falcon). Calcium Green 5N was excited at 506 nm and fluorescence collected at 532 nm using a SpectraMax M2 Microplate Reader (Molecular Devices). For a standard experiment, mitochondria were loaded with a bolus of Ca²⁺ (100 µM), followed by addition of the MCU inhibitor Ru360 (10 µM) to prevent mitochondrial Ca²⁺ uptake, thereby enabling to study mitochondrial Ca²⁺ efflux in isolation. To probe for the role of mtNCX, 20 mM of NaCl was added to mitochondria in the continuous presence of Ru360. This concentration of Na⁺ was chosen based on the reported mtNCX K_m~8–10 mM for Na⁺ and the findings that 20 mM Na⁺ produces maximal activation of mtNCX in various tissues [6]. The rate of Ca²⁺ efflux was quantified as nmol Ca²⁺ per minute per mg of mitochondrial protein during the first min of any given conditions (maximal slope). For calculating changes in extramitochondrial Ca²⁺ concentration, Calcium Green 5N fluorescence was calibrated by adding standard amounts of Ca²⁺ to mitochondria-free recording (respiration) buffer.

2.2. cDNA Plasmid Constructs and FIV Lentivirus

The plasmids encoding Flag-tagged mouse MCU and NCLX constructs were obtained from OriGene (Cat.#MR218926 for MCU-Flag, and Cat.#MR223948 for NCLX-Flag). The plasmid encoding the mitochondrial Ca²⁺ indicator mito-RGECO1 was obtained from Addgene (Cat.#46021). The Mito-V5-DsRed construct was modified from pDsRed2-Mito (Takara) by insertion of a V5 epitope tag. The GFP-KDEL plasmid construct was previously described [36] and kindly provided by Dr. Tom Rutkowski (University of Iowa).

For preparation of the mito-RGECO1 FIV (feline immunodeficiency virus) lentivirus, the corresponding cDNA sequence was PCR amplified (forward primer: 5’-GAGGTCTATATAAGCAGAGC-3’; reverse primer: 5’-GACGTCGACGAATTCGAGGCTGATCAGCGGTTTAAAC-3’), cut using EcoRI and NheI, and ligated into the EcoRI and SpeI sites of the lentiviral shuttle vector pFIV3.2-CAG-mcs. The lentivirus (1×10⁷ - 5×10⁸ transforming units/ml) was produced by the University of Iowa Viral Vector Core as previously described [23, 24].

2.3. Culture and transfection of mouse hippocampal neurons

Primary cultures of hippocampal neurons were prepared from C57BL/6J mouse neonates (P0–1) and transfected as previously described [37]. Briefly, the mice were swiftly decapitated, the brain was removed, the hippocampi were dissected in Neurobasal Plus medium supplemented with 20 mM HEPES (pH 7.3), and the tissue was digested in a trypsin solution (1 mg/mL) for 10 min at 25°C. Cells were washed in fresh medium and mechanically dissociated by sequential trituration with fire-polished Pasteur pipettes of increasingly smaller bore size. Cells were then plated onto 25 mm glass coverslips, pre-coated with poly-L-ornithine (0.2 mg/mL) and laminin (50 μg/mL) in a 6-well plate. The cultures were grown in Neurobasal Plus Medium with B27 Plus Supplement and penicillin-streptomycin (50 U/mL and 50 μg/mL, respectively) in a 5% CO₂ incubator at 37°C. Every 3–4 days, 50% of the media was replaced. After 6–7 DIV, hippocampal neurons were transfected with the mitochondrial Ca²⁺ indicator plasmid mito-RGECO1 using an FIV lentivirus [38]. Cells were used for imaging on DIV 10–14.

2.4. Culture and transfection of primary mouse hepatocytes

Hepatocytes were isolated from 6–8 weeks old C57BL/6J mice. Hepatocytes were isolated using a two-step collagenase perfusion as previously described [39] with slight modifications. Briefly, the abdomen of an anesthetized mouse was opened exposing the inferior vena cava (IVC). The IVC was cannulated with a 24-guage IV-catheter and perfusion was started with 37ºC pre-warmed perfusion buffer (HBSS -Ca²⁺, -Mg²⁺, - phenol red (Thermo Fisher 14175095); 25 mM HEPES (Thermo Fisher #15630080); 0.5 M EGTA; 1X penicillin/streptomycin (Thermo Fisher #10378016)) at 6 mL/min. The portal vein was quickly cut and the liver was observed to blanch immediately. After 5–8 minutes perfusion buffer was changed to 37ºC pre-warmed digestion buffer (HBSS +Ca²⁺, +Mg²⁺, - phenol red (Thermo Fisher #14025092); 25 mM HEPES; 0.2% BSA (Fraction V); 1X penicillin/streptomycin; 100 U/mL Collagenase IV (Worthington Biochemical #LS004188)) at 6 mL/min. Approximately 3–5 times during perfusion and digestions, the portal vein was gently clamped for 5 seconds using forceps and the liver was allowed to slightly swell. After 8 minutes, perfusion was halted, and the liver was carefully excised and placed in a 10 cm culture dish containing 10 mL of digestion buffer. The liver capsule was disrupted, and the liver was gently shaken liberating the hepatocytes into the excess digestion media. This hepatocyte suspension was filtered through a 100 µM cell strainer into 30 mL of washing media (DMEM low glucose (Thermo Fisher #11885076), 5% FBS, 10 mM HEPES, 1X penicillin/streptomycin). The hepatocyte suspension was centrifuged at 50 x g for 5 minutes, and the pellet was gently resuspended in 25 mL of washing media. This process was repeated 3 more times. The final hepatocyte pellet was resuspended in plating media (William’s E media (Thermo Fisher #12551032), 1X B27 supplement (Thermo Fischer #17504044), 10 nM dexamethasone, 1X Glutamax (Thermo Fischer #35050061), 1X penicillin/streptomycin), and cells were counted, and viability was assessed using trypan blue. Approximately 8.0×10⁵ cells/cm² hepatocytes were plated onto collagen coated coverslips. Hepatocytes were allowed to attach for 3.5 hours before plating media was changed to long-term maintenance media (William’s E media, 1X B27 supplement, 1X Glutamax, 1X penicillin/streptomycin, 20 µM forskolin (Tocris #1099), 10 µM SB431542 (Tocris #1614), 0.5 µM IWP2 (Tocris #3533), 1 µM DAPT (Tocris #2634), 0.1 µM LDN193189 (Tocris #6053) [40]. After 24 hrs in culture, hepatocytes were transfected with the mito-RGECO1 plasmid using the PromoFectin-Hepatocyte transfection kit (PromoCell; Cat.# PK-CT-2000-HEP) according to the manufacturer protocol, and used for imaging in 2–3 days after the transfection.

2.5. Ca²⁺ imaging in hippocampal neurons and primary hepatocytes

Simultaneous monitoring of the Ca²⁺ concentrations in the cytosol ([Ca²⁺]_cyt) and mitochondria ([Ca²⁺]_mt) of cultured hippocampal neurons and hepatocytes was performed as previously described [37]. On the day of imaging, transfected cell cultures were loaded with the cytosolic Ca²⁺ indicator Fura-2/AM (2 μM, 30 min). Cells were then placed in a flow-through perfusion chamber and mounted on an inverted IX-71 Olympus microscope (Olympus, Japan). Cells were perfused with a HEPES-buffered Hank’s salt solution of the following composition: 140 mM NaCl, 5 mM KCl, 1.3 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 10 mM glucose, pH 7.4 with KOH (310 mOsm/kg with sucrose). [Ca²⁺]_cyt and [Ca²⁺]_mt elevations in neurons were elicited by brief (15 s) depolarization using 30 mM KCl, and in hepatocytes, by application of 100 μM ATP (1 min).

Fura-2 and mito-RGECO1 were sequentially excited at 340 nm (12 nm bandpass), 380 nm (12 nm bandpass), and 550 nm (12 nm bandpass) using a Polychrome V monochromator (TILL Photonics, Germany) and a 40x oil-immersion objective (NA=1.35, Olympus). Excitation/emission fluorescence was separated using a dual fluorophore beamsplitter (FF493/574-Di01, Semrock; Rochester NY) while signal was collected using a dual emission filter (FF01–512/630, Semrock) and an IMAGO CCD camera (640×480 pixels; Till Photonics, Germany). Fluorescent images were acquired at a rate of 0.5 Hz. Fura-2 fluorescence ratio (R=340/380 nm) was converted to [Ca²⁺]_cyt using the equation: K_dβ((R-Rmin)/(R-Rmax)) as previously described [41, 42]. R_min, R_max and β were calculated by applying 10 μM ionomycin in either Ca²⁺-free buffer (1 mM EGTA) or Ca²⁺-containing (1.3 mM) HEPES-buffered Hank’s salt solution as previously described [38]. Using this approach, the calibration values were determined as follows: Rmin=0.19, Rmax=2.98, β=6.77. A K_d value of 275 nM was used for Fura-2 [41, 42]. Changes in [Ca²⁺]_mt were quantified as ΔF/F₀ = (F-F₀)/F₀, where F is current fluorescence intensity (λ_ex = 550 nm) and F₀ is the fluorescence intensity at baseline. At each wavelength, fluorescence was corrected for background as measured in an area free of cells.

2.6. Confocal microscopy and image analysis of NCLX and MCU localization in AML12 cells

Mouse liver AML12 cells [43] were purchased from ATCC (Cat.# CRL-2254), and after thawing and 3–6 passages, plated on collagen coated glass coverslip in Dulbecco’s modified Eagle’s medium/Ham’s F12 with 10 mM HPEPES (Gibco) supplemented with 10% fetal bovine serum (GIBCO), a 1x mixture of insulin, transferrin and selenium (ITS, GIBCO), 40 ng/ml dexamethasone (sigma) , and penicillin/streptomycin (1%) (GIBCO). The cells were co-transfected with plasmids encoding either MCU-Flag or NCLX-Flag together with the mitochondrial and endoplasmic reticulum (ER) reporters, Mito-V5-DsRed and GFP-KDEL, respectively, using Lipofectamine 2000 (Thermofisher) according to the manufacturer’s protocol. Two days after the transfection, cells were fixed with 4% paraformaldehyde in PBS, and subjected to immunofluorescence staining as previously described [23, 44, 45] using the following primary antibodies: rabbit polyclonal anti-GFP (Abcam Cat.# ab290; 1:500), mouse monoclonal anti-V5 (Invitrogen, Cat.# 14–6796-82; 1:1000), mouse monoclonal anti-Flag(DDK) (OriGene, Cat.#TA50011–100; 1:1000). Secondary antibodies conjugated to Alexa Fluor 488, 555, and 647 were targeted to each of the primary antibodies respectively (Thermofisher, 1:1000). After staining, the coverslips were mounted onto microscope slides using ProLong fluoromount with DAPI (Invitrogen Cat.# P36962). AML12 cells were imaged on an inverted Olympus FV-3000 confocal microscope using a 100x objective (UPLXAPO100X, Olympus, NA=1.45). Following image collection, all images were processed through a constrained iterative deconvolution program in the cellSens software (Olympus) designed for the FV-3000 microscope. The colocalization of the NCLX or MCU with the mitochondria and ER was quantified as the Pearson’s correlation coefficient using the JaCoP plug-in for ImageJ software [46] as previously described [45]. Analysis was performed on a single z-plane of each cell imaged, which was selected by finding the approximate center of the cells using Dapi.

2.7. qRT-PCR

RNA was isolated from adult (6–10 weeks) C57BL/6J mice, as previously described [24]. Briefly, mice were anesthetized and swiftly decapitated, after which the hippocampi, cortex, heart and liver were isolated. Tissue was placed in Trizol and homogenized, and then an additional 20% of total volume of chloroform was added. Homogenates were centrifuged at 13,200 rpm for 5 min, and then supernatant was removed and diluted 1:1 in 70% EtOH. RNA was extracted using the Qiagen RNeasy kit (Qiagen) and converted to cDNA using an Iscript-cDNA synthesis kit (Bio-Rad). mRNA expression was determined using the Step One Plus Real-time PCR System (Applied Biosystems) and the fast SYBR green master mix kit (ThermoFisher). The primer sequences were as follows. Mouse NCLX: 5’-ATACTGGAGACGGCGTCTGGGA-3’ (forward) and 5’-CTGCGGCAGTCCGGATTCCTC-3’ (reverse); mouse GAPDH: 5’-AGGTCGGTGTGAACGGATTTG-3’ (forward) and 5’-TGTAGACCATGTAGTTGAGGTCA-3’ (reverse). Data were normalized to GAPDH and analyzed using the 2ΔΔCt method.

2.8. Reagents

Ru360 was purchased from Calbiochem (Sigma-Millipore; 557440–1 SET). CGP37157, forskolin, SB431542, IWP2, DAPT and LDN193189 were purchased from Tocris (Biotechne; 1114–10 mg). Fura-2A/M (F1201) and Calcium Green 5-N (C3737) were purchased from Thermofisher. Percoll was purchased from GE Healthcare Bio-Sciences (17–0891-01). All other reagents were purchased from Sigma-Aldrich/RPI .

2.9. Statistical analysis.

Data were analyzed by Student’s t test (for comparing two groups of data; two-tailed with Welch’s correction) or one-way ANOVA with Bonferroni’s post hoc test (for comparing >2 groups of data), and for non-parametric testing (qRT-PCR data), Kruskal Wallis test was used followed by Dunn’s multiple-comparisons test (GraphPad Prism v7 or v8 for Windows). Means ± SEM plotted throughout, and significance levels are abbreviated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

3. Results

3.1. NCLX is expressed in the mouse brain, heart and liver

As the first step, we used quantitative RT-PCR to compare the expression of NCLX mRNA in the brain, heart and liver of the mouse. As shown in Figure 1, NCLX mRNA expression was similar in brain and heart, but significantly higher in the liver. A similar NCLX expression profile was recently reported for human tissues [10].

Figure 1. NCLX expression in excitable and non-excitable tissues.

Quantification of NCLX expression in adult mouse cortex, hippocampus (Hipp.), heart and liver using real-time PCR. Data are presented as mean ± SEM, n=12 mice, analyzed using Kruskal-Wallis with Dunn’s multiple comparison post-hoc test. **** p<0.001

3.2. Na⁺-dependent Ca²⁺ extrusion from isolated brain mitochondria is inhibited by CGP37157 and slowed by Na⁺ replacement with Li⁺

Next, we used isolated brain mitochondria to determine the properties of mitochondrial Ca²⁺ efflux (Fig. 2). In these experiments, isolated brain mitochondria were loaded with Ca²⁺ in Na⁺-free solution and then treated with the MCU inhibitor Ru360 (10 µM) to block mitochondrial Ca²⁺ uptake, enabling measurements of mitochondrial Ca²⁺ efflux in isolation. Addition of 20 mM NaCl to the extramitochondrial solution rapidly induced strong Ca²⁺ efflux from mitochondria. This Na⁺-induced Ca²⁺ efflux was blocked by application of the mtNCX inhibitor CGP37157 (Fig.2B-D). As shown in Fig. 2, the addition of 20 mM LiCl also stimulated mitochondrial Ca²⁺ efflux that was sensitive to CGP37157. However, mitochondrial Ca²⁺ efflux triggered by Li⁺ was significantly slower than that induced by Na⁺ (Fig. 2C and D).

Figure 2. Unlike Na⁺, Li⁺ fails to effectively stimulate Ca²⁺ efflux from mouse brain mitochondria.

(A) and (B), Representative traces of extracellular Ca²⁺ ([Ca²⁺]_ext) in a suspension of isolated brain mitochondria illustrating the protocol for measuring Na⁺ (Li⁺)-induced Ca²⁺ efflux from mitochondria in the absence (A) or presence (B) of the mtNCX inhibitor CGP37157 (30 µM). First, a pulse of Ca²⁺ (100 µM; arrow) was added to a suspension of mitochondria, followed by addition of the inhibitor of mitochondrial Ca²⁺ uniporter, Ru360 (10 µM; arrow) to block mitochondrial Ca²⁺ uptake; then either 20 mM Na⁺ (orange line) or Li⁺ (blue line) was added to induce Ca²⁺ efflux from mitochondria. This concentration of Na⁺ was chosen based on the reported mtNCX K_m~8–10 mM for Na⁺ and the findings that 20 mM Na⁺ produces maximal activation of mtNCX in various tissues [6]. (C) Summary of the maximal rates of mitochondrial Ca²⁺ efflux induced by Ru360 alone (dark blue circles), 20 mM Na⁺ without (orange circles) or with 30 µM CGP37157 (orange squares), 20 mM Li⁺ without (light blue circles) or with 30 µM CGP37157 (light blue squares), in the continuous presence of 10 µM Ru360. (D) The amount of Ca²⁺ extruded from mitochondria during a 30 min period (T₃₀ in (A), dotted vertical lines) after the addition of 20 mM Na⁺ (orange bar), Li⁺ (blue bar), without or with 30 µM CGP37157 (striped bars). Data are presented as mean ± SEM, n=6–8 independent experiments; ****p<0.0001, one-way ANOVA with Bonferroni’s post hoc test.

3.3. Mitochondrial Ca²⁺ efflux in hippocampal neurons is inhibited by Na⁺ replacement with Li⁺ and by CGP37157

Our findings that replacement of Na⁺ with Li⁺ significantly slowed mitochondrial Ca²⁺ efflux from isolated brain mitochondria (Fig. 2) was somewhat surprising given reports that NCLX-driven Ca²⁺ transport shows similar sensitivity to Na⁺ and Li⁺ [28–31]. Therefore, we decided to further test the effect of Li⁺ on mitochondrial Ca²⁺ extrusion in intact cells, using cultured primary mouse hippocampal neurons as a model. In these experiments, we simultaneously monitored [Ca²⁺]_mt and [Ca²⁺]_cyt using fluorescent Ca²⁺ indicators, mito-RGECO1 and Fura-2, respectively, as previously described [38, 47]. [Ca²⁺]_cyt and [Ca²⁺]_mt responses were induced by depolarization (30 mM KCl, 15 s), either under control conditions (140 mM NaCl in extracellular solution) or after extracellular Na⁺ had been completely replaced with Li⁺ (Fig. 3). Given that voltage-gated Na⁺ channels are highly permeable to Li⁺, this ion was expected to enter the cell upon depolarization [48, 49]. However unlike Na⁺, Li⁺ is negligibly transported out of the cell by the Na⁺/K⁺-ATPase [50, 51], which is expected to promote substitution of intracellular Na⁺ with Li⁺.

Figure 3. Substitution of Li⁺ for Na⁺ markedly inhibits mitochondrial Ca²⁺ efflux in hippocampal neurons.

(A) and (B), Representative traces from experiments with simultaneous monitoring of [Ca²⁺]_cyt (Fura-2; black) and [Ca²⁺]_mt (mito-RGECO1; red) in cultured hippocampal neurons. [Ca²⁺]_cyt and [Ca²⁺]_mt transients were induced by brief depolarization using 30 mM KCl (K⁺30, 15 s, arrows). Substitution of Li⁺ for Na⁺ in the extracellular solution (0 mM Na⁺, 140 mM Li⁺) is indicated by blue bar under the traces. In (B), extracellular Ca²⁺ was removed (0 mM Ca²⁺, 100 µM EGTA) immediately after K⁺30 stimulation (purple bars). (C), Mitochondrial Ca²⁺ efflux kinetics was quantified by measuring the duration of 70% [Ca²⁺]_mt recovery relative to its peak elevation (T₇₀), and compared between the control conditions (140 mM NaCl in extracellular solution) and following Na⁺ replacement with Li⁺ in the extracellular solution. Data are presented as mean ± SEM, n=36; ****p<0.0001, two-tailed paired Student’s t-test. (D) Representative traces of [Ca²⁺]_cyt (black) and [Ca²⁺]_mt (red) show the effect of the mtNCX inhibitor CGP37157 (10 μM; green bar) on mitochondrial Ca²⁺ efflux in hippocampal neurons. (E), Mitochondrial Ca²⁺ efflux kinetics was quantified by measuring the duration of 70% [Ca²⁺]_mt recovery relative to its peak elevation (T₇₀), and compared before and after treatment with CGP37157. Data are presented as mean ± SEM, n=9; ***p<0.001, two-tailed paired Student’s t-test.

Under the control conditions, depolarization induced rapid Ca²⁺ transients in both the cytosolic and mitochondrial compartments, followed by recovery to baseline levels. Recovery of the [Ca²⁺]_mt elevation was slower than the recovery of [Ca²⁺]_cyt signal, consistent with the previous reports [38, 47]. When extracellular Na⁺ was replaced with Li⁺, the inhibition of Ca²⁺ clearance was prominent in both mitochondria and the cytosol (Fig. 3A). Given that changes in [Ca²⁺]_cyt can affect mitochondrial Ca²⁺ transport, Li⁺-induced slowing of mitochondrial Ca²⁺ efflux could potentially have been a consequence of the prolonged [Ca²⁺]_cyt elevation, with the latter being caused by reversal of the plasma membrane Na⁺/Ca²⁺ exchanger in the absence of extracellular Na⁺. To rule out this potential effect of elevated [Ca²⁺]_cyt on mitochondrial Ca²⁺ efflux, we repeated the Na⁺ replacement experiments with extracellular Ca²⁺ removed immediately at the end of the depolarization. This approach indeed normalized recovery of [Ca²⁺]_cyt in Li⁺-containing extracellular solution (Fig. 3B). Notably, although recovery of [Ca²⁺]_cyt following depolarization was rapid under these conditions, the replacement of Na⁺ with Li⁺ resulted in marked slowing of recovery of [Ca²⁺]_mt (Fig. 3B and C). These findings in intact hippocampal neurons were similar to those in isolated brain mitochondria (Fig. 2). Specifically, T₇₀ of [Ca²⁺]_mt recovery increased from 4.1 ± 0.4 min in controls (Na⁺-containing solution) to 12.1 ± 1.1 min (p<0.0001; Student’s t-test) in the cells treated with Li⁺-containing extracellular solution (Fig. 3C). As positive controls, we tested the effect of the mtNCX inhibitor CGP37157, and found that in cells treated with the inhibitor [Ca²⁺]_mt recovery was significantly slowed (Fig. 3D and E). Overall, these data are consistent with our observations for isolated brain mitochondria (Fig.2), and they suggest that first, mtNCX is a major mechanism responsible for mitochondrial Ca²⁺ efflux in neurons, and second, that Li⁺ is significantly less effective than Na⁺ in driving this Ca²⁺ efflux.

3.4. Na⁺-dependent Ca²⁺ efflux from isolated heart mitochondria is inhibited by CGP37157 and slowed by Na⁺ replacement with Li⁺

We also examined the properties of mitochondrial Ca²⁺ extrusion in another excitable tissue, the heart (Fig. 4). As described for the experiments using mitochondria isolated from brain (Fig.2), isolated heart mitochondria were loaded with Ca²⁺ in a Na⁺-free solution and then treated with the MCU inhibitor Ru360 (10 µM) to block Ca²⁺ uptake. As with the brain mitochondria, 20 mM NaCl stimulated a strong Ca²⁺ efflux from heart mitochondria (Fig. 4A). However, the rate of Ca²⁺ efflux from heart mitochondria was significantly slower (12.3 ± 0.4 nmol Ca²⁺ min⁻¹ x mg protein⁻¹; n=6) than the rate of Ca²⁺ efflux from brain mitochondria (Fig.2C; 33.8 ± 2.2 nmol Ca²⁺ min⁻¹ x mg protein⁻¹; n=6; p<0.0001; ANOVA with Bonferroni’s post hoc test). Na⁺-induced Ca²⁺ efflux from heart mitochondria was blocked by the mtNCX inhibitor CGP37157 (30 µM; Fig. 4). Addition of 20 mM LiCl was also able to stimulate Ca²⁺ extrusion from heart mitochondria, although at a markedly slower rate than 20 mM NaCl (Fig. 4). Thus, similar with brain mitochondria, Ca²⁺ efflux from heart mitochondria is driven by Na⁺, blocked by CGP37157, and significantly less sensitive to Li⁺ than Na⁺ as an exchange partner.

Figure 4. Na⁺-induced Ca²⁺ extrusion from isolated heart mitochondria is inhibited by Li⁺ and CGP37157.

(A) Representative traces of extracellular Ca²⁺ ([Ca²⁺]_ext) in a suspension of mitochondria isolated from the heart, illustrating the protocol for measuring Na⁺ (Li⁺)-induced Ca²⁺ efflux from mitochondria. A pulse of Ca²⁺ (100 µM; arrow) was added to the suspension of mitochondria, followed by addition of the inhibitor of mitochondrial Ca²⁺ uniporter, Ru360 (10 µM; arrow) to block mitochondrial Ca²⁺ uptake, and in some instances mtNCX inhibitor CGP37157 (30 µM, gray line). Then either 20 mM Na⁺ (orange line) or Li⁺ (blue line) was added to induce Ca²⁺ efflux from mitochondria in the continuous presence of Ru360. (B) Quantification of the maximal rates of mitochondrial Ca²⁺ efflux induced by Ru360 alone (dark blue circles), 20 mM Na⁺ without (orange circles) or with 30 µM CGP37157 (orange squares), and 20 mM Li⁺ (light blue circles), in the continuous presence of 10 µM Ru360. (C) The amount of Ca²⁺ extruded from mitochondria during a 30 min period (T₃₀ in (A), dotted vertical lines) after the addition of 20 mM Na⁺ (orange bar), Li⁺ (blue bar), or Na⁺ plus 30 µM CGP37157 (striped bar). Data are presented as mean ± SEM, n=5–8 independent experiments; *p<0.05, **p<0.01, ****p<0.0001, one-way ANOVA with Bonferroni’s post hoc test.

3.5. Ca²⁺ efflux from isolated liver mitochondria is insensitive to Na⁺ or CGP37157

Whether or not mitochondria from non-excitable tissue, such as liver, could utilize an electrogenic mtNCX mechanism has been a matter of debate [6, 11, 12, 20]. Thus, we applied the same protocol that we established for brain and heart mitochondria (Figs. 2 and 4) to isolated liver mitochondria (Fig. 5A) and monitored Ca²⁺ efflux from this organelle. Similar to the results for brain and heart mitochondria, application of Ru360 caused a small but detectable Ca²⁺ efflux from liver mitochondria (1.8 ± 0.3 nmol Ca²⁺ min⁻¹ x mg protein⁻¹; n=7; Fig. 5B). However, in contrast to the findings for brain and heart mitochondria, the addition of 20 mM Na⁺ failed to significantly stimulate Ca²⁺ extrusion (1.7±0.1 nmol Ca²⁺ min⁻¹ x mg protein⁻¹; n=12; Fig. 5B) compared to Ru360 alone (p>0.05; one-way ANOVA with Bonferroni’s post hoc test), and the mtNCX inhibitor CGP37157 had no effect. Li⁺ also failed to stimulate Ca²⁺ extrusion from liver mitochondria (Fig. 5).

Figure 5. Effects of Na⁺, Li⁺ and CGP37157 on Ca²⁺ efflux from liver mitochondria.

(A) Representative traces of extracellular Ca²⁺ ([Ca²⁺]_ext) in a suspension of isolated liver mitochondria illustrating the protocol for measuring Na⁺ (Li⁺)-induced Ca²⁺ efflux from mitochondria. First a pulse of Ca²⁺ (100 µM; arrow) was added to a suspension of mitochondria, followed by addition of the MCU inhibitor, Ru360 (10 µM; arrow) to block mitochondrial Ca²⁺ uptake, and in some instances mtNCX inhibitor CGP37157 (30 µM, gray line); then either 20 mM Na⁺ (orange line) or Li⁺ (blue line) was added to induce Ca²⁺ efflux from mitochondria, all in continuous presence of Ru360. (B) Quantification of the maximal rates of mitochondrial Ca²⁺ efflux induced by Ru360 alone (dark blue circles), 20 mM Na⁺ without (orange circles) or with 30 µM CGP37157 (orange squares), and 20 mM Li⁺ (light blue circles), in the continuous presence of 10 µM Ru360. (C) The amount of Ca²⁺ extruded from mitochondria during a 30 min period (T₃₀ in (A), dotted vertical lines) after the addition of 20 mM Na⁺ (orange bar), Li⁺ (blue bar), or Na⁺ plus 30 µM CGP37157 (striped bar). Data are presented as mean ± SEM, n=7–12 independent experiments; one-way ANOVA with Bonferroni’s post hoc test.

These results suggest that in contrast to brain and heart mitochondria, mtNCX does not significantly contribute to mitochondrial Ca²⁺ efflux in the liver.

3.6. Mitochondrial Ca²⁺ efflux in primary mouse hepatocytes is slow and insensitive to CGP37157

To further examine the implications of distinctly slow Ca²⁺ efflux from liver mitochondria at a single cell level, we simultaneously monitored [Ca²⁺]_cyt and [Ca²⁺]_mt in primary mouse hepatocytes (Fig. 6), using the same approach as described for hippocampal neurons (Fig. 3). [Ca²⁺]_cyt and [Ca²⁺]_mt responses were induced by ATP (100 µM), which is a well-known activator of Ca²⁺ signaling in hepatocytes via P2Y and also likely P2X receptors [52–55]. As shown in Fig. 6A-B, ATP produced a large [Ca²⁺]_cyt elevation that rapidly recovered toward the baseline. In contrast, the recovery of [Ca²⁺]_mt was very slow, taking on average more than 20 min to return to the baseline (Fig. 6A). Notably, the process of [Ca²⁺]_mt recovery in hepatocytes was significantly slower (T₇₀=19.7 ± 2.7 min; n=9) than that in hippocampal neurons (T₇₀=4.1 ± 0.4 min; n=36; p<0.0001; Student’s t-test). The rate of mitochondrial Ca²⁺ efflux in hepatocytes was not significantly affected by the treatment with the mtNCX inhibitor CGP37157 (Fig. 6C).

Figure 6. Mitochondrial Ca²⁺ efflux in primary mouse hepatocytes is slow and insensitive to CGP37157.

(A, B), Examples of simultaneous measurements of [Ca²⁺]_cyt (Fura-2; black) and [Ca²⁺]_mt (mito-RGECO1; red) in primary mouse hepatocytes in response to 100 µM ATP (1 min) in the absence (A) or presence (B) of the mtNCX inhibitor CGP37157 (10 µM). (C) Comparison of mitochondrial Ca²⁺ efflux kinetics studied under control conditions (n=9; black) or in the presence of 10 µM CGP37157 (n=9; green), which was quantified by measuring the duration of 70% [Ca²⁺]_mt recovery relative to its peak elevation (T₇₀). The T₇₀ values were obtained from the experiments as those shown in (A) and (B). CGP37157 treatment did not significantly change the [Ca²⁺]_mt recovery kinetics compared to the control (p=0.77; unpaired Student’s t-test). Data are presented as mean ± SEM.

Overall, these results are consistent with the observations in isolated liver mitochondria and suggest that mtNCX does not play a significant role in mediating mitochondrial Ca²⁺ efflux in the liver.

3.7. Extramitochondrial localization of NCLX overexpressed in liver cells

It is thought that NCLX is responsible for the mtNCX transport in various tissues [20, 21]. Therefore, our observations that the mtNCX mechanism does not significantly contribute to mitochondrial Ca²⁺ efflux in the liver was surprising, given a relatively high expression of NCLX mRNA in the liver (Fig. 1). However, several groups have reported extramitochondrial localization and function of NCLX [28, 56–58], which could potentially help to explain our results. To test the possibility of extramitochondrial localization of NCLX in the liver, we co-expressed a Flag-tagged NCLX in the mouse liver cell line AML12 [43] together with mitochondria-targeted DsRed and ER-targeted GFP reporter plasmids. Co-localization analysis using confocal microscopy and Pearson’s correlation coefficient [46, 59] showed that a large fraction of NCLX was localized outside of mitochondria (Fig. 7). In contrast, Flag-tagged MCU demonstrated mitochondria-restricted subcellular localization in AML12 cells.

Figure 7. Extramitochondrial localization of NCLX overexpressed in liver AML12 cells.

(A, B) Subcellular localization of MCU-Flag (A, green) and NCLX-Flag (B, green) in mouse liver AML12 cells relative to mitochondria (mito-V5-DsRed; red) and ER (GFP-KEDL; magenta). (C) Summary of colocalization of MCU-Flag with mitochondria (blue; n=24 cells), NCLX-Flag with mitochondria (red; n=32 cells) and NCLX-Flag ER (green; n=32 cells) quantified by using the Pearson’s correlation coefficient [46, 59]. Data are presented as mean ± SEM; ****p<0.0001, one-way ANOVA with Bonferroni’s post hoc test.

4. Discussion

This study demonstrates that the properties of mitochondrial Ca²⁺ extrusion differ significantly between the brain, heart and liver. More specifically, our findings suggest that mtNCX is the principal mechanism responsible for Ca²⁺ extrusion from heart and brain mitochondria, but not in liver, where contribution of this mechanism was undetectable. This conclusion is supported by robust effects of Na⁺ and mtNCX inhibitor CGP37157 on mitochondrial Ca²⁺ efflux in both brain and heart mitochondria, and the absence of any effects of these treatments in liver mitochondria. We also found that Li⁺ was significantly less effective than Na⁺ in driving mitochondrial Ca²⁺ efflux, as apparent from a 70% difference in the rate of Ca²⁺ efflux triggered by these ions. Another key finding was that NCLX was expressed in all the tissues tested, including brain, heart and liver, with the highest level detected in the liver.

Our finding that mtNCX plays an important role in the brain and heart, but not in the liver, is consistent with earlier work by Carafoli and colleagues [11, 60]. Our conclusions are further strengthened by the distinctive effects of the mtNCX inhibitor CGP37157 [32–34] both in isolated mitochondria and intact cells (Fig.2–6). Quantitative comparison under similar conditions showed that the rate of Ca²⁺ efflux from brain mitochondria was nearly 3-fold greater than that found for heart mitochondria (33.8 ± 2.2 and 12.3 ± 0.4 nmol Ca²⁺ min⁻¹ x mg protein⁻¹, respectively). These values are consistent with rates previously reported for mitochondria of the rat heart [60], but exceed those for the mitochondria of rat and guinea pig brain [11, 61]. The latter discrepancy could potentially be explained by differences in mitochondrial Ca²⁺ load, ADP concentration and species used [62]. Importantly, our findings that Ca²⁺ efflux from liver mitochondria with or without Na⁺ was approximately 10-fold slower than Na⁺-dependent mitochondrial Ca²⁺ efflux from heart and brain mitochondria, are in good agreement with findings from previous studies [11, 21, 60].

Although consistent with previous work on isolated mitochondria [11, 21, 60], the lack of a detectable contribution of mtNCX in the liver was surprising in the context of high NCLX expression in this organ (Fig. 1). Our findings in mouse tissues are in good agreement with the NCLX expression profiling in human tissues, which also showed that NCLX expression in the liver is higher than that in the brain and heart [10] (Table 1 in this publication). In addition to the proposed role of NCLX in mitochondria, previous work suggested that this protein (also known as NCKX6) functions in the plasma membrane and possibly other cell membranes and organelles, including the endoplasmic reticulum [28, 56–58]. Thus, the apparent discrepancy between NCLX expression and a lack of its contribution to mitochondrial Ca²⁺ efflux in the liver might be potentially explained by the extra-mitochondrial localization of NCLX in the liver. Our finding that NCLX localization was mostly extramitochondrial in liver AML12 cells (Fig. 7) is consistent with this idea. The mechanisms that regulate NCLX trafficking within the cell remain largely unknown, and future work is needed to clarify some of the conflicting evidence about its cellular localization [28, 56–58].

In this study, we took advantage of the reported ability of NCLX to utilize either Li⁺ or Na⁺ to transport Ca²⁺ through an exchange mechanism [28–31], testing the consequences of Na⁺ substitution with Li⁺ on the rate of mitochondrial Ca²⁺ efflux. Our conclusion that Li⁺ is much less effective than Na⁺ in driving mitochondrial Ca²⁺ extrusion is supported by experiments using both isolated mitochondria and intact cells (Figs. 2–4). For the experiments in cultured hippocampal neurons, we simultaneously measured [Ca²⁺]_mt and [Ca²⁺]_cyt. Given that cytosolic Ca²⁺ changes can significantly impact mitochondrial Ca²⁺ transport, simultaneous monitoring of [Ca²⁺]_mt and [Ca²⁺]_cyt in the same cell provided an important control for potential effects of ion substitution on cytosolic Ca²⁺ signaling. This approach also enabled us to optimize experimental conditions for selective manipulation of mitochondrial Ca²⁺ transport while preserving the amplitude and kinetics of cytosolic Ca²⁺ responses. We found that under these controlled conditions, replacement of Na⁺ with Li⁺ led to marked slowing of Ca²⁺ efflux from mitochondria (Fig. 3). Importantly, these findings in intact cells were consistent with those from isolated brain and heart mitochondria (Figs. 2 and 4), with the latter allowing direct monitoring of Ca²⁺ fluxes in and out of mitochondria. As expected, the mtNCX inhibitor CGP37157 blocked mitochondrial Ca²⁺ efflux in both intact cells (Fig. 3) and isolated mitochondria (Figs. 2 and 4).

Our finding that Li⁺ was significantly less effective than Na⁺ in driving Ca²⁺ efflux from brain and heart mitochondria was somewhat unexpected given the proposed role of NCLX in mediating Na⁺-dependent Ca²⁺ extrusion from mitochondria [13–18]. Previous work using intact or permeabilized cells had suggested that NCLX is able to exchange Ca²⁺ for Li⁺ nearly as effectively as for Na⁺, as well as identified the amino acids within NCLX that are critical for this Li⁺ sensitivity [28–31]. However, our results regarding the effects of Li⁺ in isolated mitochondria are consistent with those from studies by others using isolated mitochondria obtained from the heart and vascular smooth muscles [60, 63]. Although the original, and commonly cited, article by the Carafoli group reported that both Na⁺ and Li⁺ can stimulate Ca²⁺ efflux from isolated heart mitochondria [64], a later publication by the same group provided quantitative analysis, which showed that the rate of Li⁺-driven mitochondrial Ca²⁺ efflux was only ~25–30% of that driven by Na⁺ [60] (Fig.4 in that publication). Similarly, measurements in mitochondria isolated from vascular smooth muscles demonstrated that the rate of Li⁺-driven mitochondrial Ca²⁺ efflux was only ~10% of that driven by Na⁺ [63]. Thus, Li⁺ is markedly less effective than Na⁺ in driving mtNCX-mediated Ca²⁺ efflux from mitochondria.

Utilization of the mtNCX mechanism in excitable cells such as neurons and cardiomyocytes is well suited to their patterns of physiological activity and Ca²⁺ signaling [9, 22, 65–69]. Indeed, high frequency [Ca²⁺]_cyt spikes in cardiomyocytes associated with every heart beat (up to 400–600 beats/min in rodents), and similarly frequent [Ca²⁺]_cyt transients produced by bursts of action potentials in neurons, necessitate rapid Ca²⁺ unloading from mitochondrial to prevent toxic accumulation of Ca²⁺ in the matrix. The mtNCX transport is also aided by Na⁺ entering through voltage-gated Na⁺ channels, which could potentially elevate the intracellular Na⁺ concentration to above 10 mM [70, 71], a value comparable to the mtNCX Km for Na⁺ (8–10 mM) [6]. In the case of liver, a much slower rate of mitochondrial Ca²⁺ extrusion in the liver might be optimal for the relatively infrequent [Ca²⁺]_cyt spikes that characterize hepatocytes [9, 72]. This is expected to prolong Ca²⁺ retention within the matrix after a single [Ca²⁺]_cyt transient (Fig. 6), allowing for more efficient use of Ca²⁺ signals to support stimulation of Ca²⁺-dependent dehydrogenases and ATP synthesis. Future studies will be needed to clarify the specific contributions of mtNCX and mtHCX mechanisms to Ca²⁺ signaling in excitable and non-excitable cells, and more generally, to identify molecular mechanisms that enable the coordination of mitochondrial Ca²⁺ uptake and release, processes that are tailored to unique characteristics of Ca²⁺ signaling and the metabolic needs of a particular cell type or tissue.

HIGHLIGHTS.

• Mitochondrial Na⁺/Ca²⁺ exchanger is active in brain and heart, but not in liver

• NCLX is expressed in brain, heart and liver

• Li⁺ is significantly less effective than Na⁺ in driving mitochondrial Ca²⁺ efflux

Abbreviations:

MCU

mitochondrial Ca²⁺ uniporter

mtNCX

mitochondrial Na⁺/Ca²⁺ exchanger

mtHCX

mitochondrial H⁺/Ca²⁺ exchanger

NCLX

Na⁺/Ca²⁺/Li⁺ exchanger