Mitochondrial Matrix Ca²⁺ Accumulation Regulates Cytosolic NAD⁺/NADH Metabolism, Protein Acetylation, and Sirtuin Expression

Abstract

Mitochondrial calcium uptake stimulates bioenergetics and drives energy production in metabolic tissue. It is unknown how a calcium-mediated acceleration in matrix bioenergetics would influence cellular metabolism in glycolytic cells that do not require mitochondria for ATP production. Using primary human endothelial cells (ECs), we discovered that repetitive cytosolic calcium signals (oscillations) chronically loaded into the mitochondrial matrix. Mitochondrial calcium loading in turn stimulated bioenergetics and a persistent elevation in NADH. Rather than serving as an impetus for mitochondrial ATP generation, matrix NADH rapidly transmitted to the cytosol to influence the activity and expression of cytosolic sirtuins, resulting in global changes in protein acetylation. In endothelial cells, the mitochondrion-driven reduction in both the cytosolic and mitochondrial NAD⁺/NADH ratio stimulated a compensatory increase in SIRT1 protein levels that had an anti-inflammatory effect. Our studies reveal the physiologic importance of mitochondrial bioenergetics in the metabolic regulation of sirtuins and cytosolic signaling cascades.

INTRODUCTION

Cells respond to environmental cues largely through receptor-mediated pathways that facilitate Ca²⁺ mobilization from intracellular endoplasmic reticulum stores (1). Once released into the cytosol, Ca²⁺ drives energy-intensive signaling pathways that evoke a cellular response. In parallel, Ca²⁺ is sequestered by functional mitochondria to speed ATP generation (2). Thus, a Ca²⁺-mediated rise in bioenergetics increases mitochondrial energy production to match cellular energy demand. Imbalances in mitochondrial Ca²⁺ handling disrupt this delicate balance and over time, manifest as pathogenic processes that contribute to human disease (3).

Mitochondria have a unique ability to decode and transduce cytosolic Ca²⁺ signals into an energetic output (ATP) to match cellular energetic demand. Specifically, mitochondrial Ca²⁺ uptake regulates bioenergetics both by increasing the supply of reducing equivalents (NADH production) to the electron transport chain and by increasing F₁-F_o ATP synthase activity (NADH consumption) (4, 5). The balance between NADH production and NADH oxidation is therefore exquisitely dependent upon both the shape and duration of the mitochondrial Ca²⁺ signal to activate NADH-generating matrix enzymes, as well as the rate of NADH consumption by the respiratory chain (6). Indeed, a single cytosolic Ca²⁺ event or sustained Ca²⁺ increase stimulates a mitochondrial NADH transient that diminishes as NADH is consumed by the respiratory chain (7). Repetitive Ca²⁺ events conversely trigger a persistent elevation in the NADH signature that reflects a dynamic relationship between Ca²⁺-stimulated NADH production and respiratory chain activity (8).

As opposed to NADH production, the consumption of NADH by the respiratory chain is related to the cell's energetic demand and depends on metabolic preference (glycolytic or oxidative) and substrate supply to mitochondria (6), as well as sufficient quantities of oxygen and ADP (9). However, the coordination between NADH production and NADH consumption was delineated largely in metabolic cells with robust mitochondrial activity. Many cell types, such as leukocytes (10) and vascular endothelial cells (ECs) (11), have low basal energy requirements and generate ATP primarily through anaerobic mechanisms outside the mitochondria. Nonetheless, endothelial mitochondria rapidly sequester cytosolic Ca²⁺ transients to influence vascular function (12). Repetitive Ca²⁺ oscillations are commonly observed in ECs (13–15) and influence NF-κB transcriptional activity (16) and expression of vascular adhesion molecules and interleukins (17, 18). Whether Ca²⁺-mediated alterations in mitochondrial bioenergetics can influence Ca²⁺-regulated cytosolic signaling cascades is unclear.

In the present study, we delineate a fundamental mechanism by which repetitive Ca²⁺ signals influence mitochondrial bioenergetics and intracellular signaling independent of ATP generation. Continual Ca²⁺ transients enhance mitochondrial Ca²⁺ loading and result in a persistent elevation in matrix NADH. Using primary human ECs as a model system, we demonstrate that Ca²⁺-stimulated mitochondrial NADH generation is shuttled to the cytosol as a bioenergetic cue to influence signaling. Specifically, mitochondrion-to-cytosol NADH transmission influences the expression of sirtuins, resulting in altered protein acetylation and cellular activation. Thus, matrix NADH metabolism is a key rheostat by which mitochondria integrate and decode cytosolic Ca²⁺ signals to influence cytosolic signaling cascades and nuclear gene expression.

MATERIALS AND METHODS

Cell culture.

Primary human pulmonary artery endothelial cells (HPAECs) (Invitrogen) were cultured according to the manufacturer's instructions and used between passages 5 and 10. AD-293 cells (Stratagene) were cultured in Dulbecco modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS), 50 U/ml penicillin, and 50 μg/ml streptomycin. J774 macrophages (ATCC) were cultured in RPMI 1640 with 10% FBS, 50 U/ml penicillin, and 50 μg/ml streptomycin. For plasmid expression experiments, HPAECs were transfected with 0.25 μg DNA using the Neon electroporation system (Invitrogen). The AD-293 calcium-sensing receptor (CaSR) stable clone was generated by transfection (Lipofectamine 2000) with a Flag-CaSR plasmid followed by G418 selection. To reduce confluence-dependent variability in the NAD⁺/NADH ratio and sirtuin activities, all experiments were performed on confluent cells plated 2 to 3 days before assays.

Western blotting and immunoprecipitation.

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (Sigma) containing a protease inhibitor cocktail (Thermo Scientific) and phosphatase inhibitor cocktail (Roche) and 10 mM nicotinamide and 10 μM trichostatin A for acetylation experiments. For Western blotting, 25 μg proteins was electrophoresed on NuPAGE 4 to 12% Bis-Tris acrylamide gels (Invitrogen), transferred onto polyvinylidene difluoride (PVDF) membranes, and probed with primary antibodies at 4°C overnight. For immunoprecipitation, 2 mg cell lysate was incubated with primary antibody overnight at 4°C followed by incubation with A/G Plus agarose beads (Santa Cruz). Horseradish peroxidase (HRP)-conjugated secondary antibodies (Thermo Scientific) were used for detection.

Cellular respiration.

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured using the XF24 analyzer (Seahorse Bioscience). HPAECs were plated in Seahorse 24-well assay plates, 30,000 cells/well, in M200 growth medium 48 h before the assay. OCR and ECAR measurements were performed in XF assay medium (Seahorse Bioscience) at 10-min intervals.

Intracellular calcium concentration ([Ca²⁺]_i) imaging.

Cells plated on MatTek dishes were loaded with 5 μM Fura-2-acetoxymethyl ester (Fura-2 AM; Molecular Probes) in ECM buffer (120 mM NaCl, 5 mM NaHCO₃, 10 mM Na-HEPES, 4.7 mM KCl, 1 mM KH₂PO₄, 1.2 mM MgSO₄, 2 mM CaCl₂, 10 mM glucose, and 2.0% bovine serum albumin [BSA] [pH 7.4]), with 100 μM sulfinpyrazone and 0.003% pluronic acid for 30 min at room temperature. After dye loading, cells were washed and imaged in ECM buffer with 0.25% BSA and 100 μM sulfinpyrazone using a Nikon Eclipse Ti fluorescence microscope calibrated for Fura-2 fluorescence (Molecular Probes). Spectral analysis of Ca²⁺ oscillations was measured as described previously (19). Mitochondrial Ca²⁺ concentration ([Ca²⁺]_m) changes were measured with the fluorescence resonance energy transfer (FRET)-based genetically encoded mitochondrial Ca²⁺ indicator Cameleon D3cpv (Addgene plasmid 36324) (20). Images were acquired every 10 s with an LSM510 META Zeiss confocal microscope using a Fluar 40×/1.3 oil objective at 405/488 nm excitation/emission. Ratio images were obtained by dividing the intensity of the FRET channel to the intensity of the cyan fluorescent protein (CFP) channel. The sensor response was calibrated at the end of the experiment for each cell by measuring R_min (5 μM ionomycin and 5 mM EGTA) and R_max (5 μM ionomycin and 5 mM CaCl₂) (21).

NADH and NAD⁺/NADH ratio measurements.

Mitochondrial NADH fluorescence was measured with a Nikon Eclipse Ti microscope equipped with a xenon arc lamp and DeltaRamX monochromator (Photon Technology International) and an Evolve 512 electron-modifying charge-coupled-device (EMCCD) camera (Photometrics) with the assistance of EasyRatioPro software using a UV filter. Specificity for mitochondrial NADH was determined by colocalization with the mitochondrial dye MitoTracker Green (Invitrogen).

Cytosolic NAD⁺/NADH ratio was measured using the genetically encoded ratiometric fluorescence indicator Peredox (Addgene plasmid 32383) (22). Green and red fluorescence images were acquired every 20 s with an LSM510 META Zeiss confocal microscope using a Fluar 40×/1.3 oil objective at excitation wavelengths of 405 nm and 543 nm. Images were background corrected, and green-to-red ratio images were obtained using ImageJ software. For each cell, ratio data were normalized to the minimal green-to-red ratio signal obtained with 10 mM pyruvate. For quantitative measurements, the Peredox sensor was calibrated in ECM buffer without glucose by changing extracellular concentrations of lactate and pyruvate as previously described (23). Total intracellular NAD⁺/NADH ratio was measured using the EnzyChrom NAD⁺/NADH assay kit (BioAssay Systems). Intracellular lactate and pyruvate concentrations were measured using the lactate assay kit and pyruvate assay kit (Cayman Chemical) and used to calculate the cytosolic NAD⁺/NADH ratio (24).

Leukocyte adhesion assay.

HPAECs were plated into 96-well plates and cultured until confluent. J774 macrophages were stained with 1 μM Cell Tracker Green (Molecular Probes) for 0.5 h at 37°C. A total of 3 × 10⁵ macrophages were placed on top of HPAECs and allowed to adhere to endothelial cells for 1 h at 37°C. Cells were washed vigorously with HBSS, and the fluorescence of adherent J774 macrophages was measured using a Synergy 4 fluorescence plate reader (BioTek) at 490-nm excitation and 525-nm emission wavelengths. Images of J774 macrophages adherent to HPAECs were acquired using a Nikon Eclipse Ti microscope with a 10× objective and a fluorescein isothiocyanate (FITC) filter.

qRT-PCR.

Quantitative reverse transcription-PCR (qRT-PCR) analysis of mRNA expression was performed using a Fast real-time PCR system (Applied Biosystems) on cDNAs synthesized using the Omniscript RT kit (Qiagen) from previously purified mRNA (RNeasy minikit; Qiagen).

ATP content.

ATP levels were measured using the EnzyLight assay kit (BioAssay Systems).

Statistical analysis.

The data are shown as means ± standard errors of the means (SEMs) from 3 or more independent experiments. Statistical significance was assessed using Student's t test, and P values of <0.05 were considered statistically significant.

Reagents.

Pan-acetylated lysine, SIRT2 (sirtuin 2), SIRT3, SIRT4, SIRT5, SIRT6, p53, phosphatase and tensin homolog (PTEN), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), actin, histone H3, and acetyl-histone H3 antibodies were from Cell Signaling. SIRT1 and manganese superoxide dismutase (MnSOD) antibodies were from Invitrogen.

RESULTS

Receptor-mediated Ca²⁺ oscillations stimulate mitochondrial Ca²⁺ loading in primary human pulmonary artery endothelial cells.

Oscillatory Ca²⁺ waves are a common event in both excitable and nonexcitable cells and control processes ranging from fertilization to cell death (25). The inflammatory agonist histamine activates ECs by initiating repetitive Ca²⁺ signals (26) that regulate vascular permeability, leukocyte recruitment, and vasomotor tone (27). In primary human pulmonary artery endothelial cells (HPAECs), histamine (100 nM) triggered a rapid mobilization of intracellular Ca²⁺ ([Ca²⁺]_i) that presented as an oscillatory pattern for as long as histamine was present in the extracellular environment (Fig. 1A). Kinetically, [Ca²⁺]_i oscillations rose rapidly from 0.171 ± 0.006 μM at rest to an initial amplitude of 1.155 ± 0.05 μM that decreased slightly but remained consistent during subsequent transients (Fig. 1B). While the biophysical patterning of the histamine-induced Ca²⁺ response was heterogeneous over a wide range of concentrations, a majority of cells between 100 nM and 500 nM histamine exhibited oscillations with a mean frequency of ∼2 mHz (Table 1). At histamine concentrations below 100 nM, the [Ca²⁺]_i response was comprised primarily of a single transient. Conversely, histamine at 1 μM resulted in a single [Ca²⁺]_i event that decreased but remained elevated ∼50% over baseline (“plateau”). Mechanistically, histamine triggered an initial release of Ca²⁺ from the endoplasmic reticulum that required sufficient extracellular Ca²⁺ to refill the intracellular stores and sustain the oscillatory phenotype (Fig. 1C and Table 1). Mitochondria are important buffers that uptake and release cytosolic Ca²⁺ and in doing so, dramatically shape intracellular Ca²⁺ signaling patterns (28). To evaluate the effects of [Ca²⁺]_i oscillations on mitochondrial Ca²⁺ handling, the mitochondrial Ca²⁺ concentration ([Ca²⁺]_m) was measured via confocal microscopy using the FRET-based genetically encoded mitochondrial Ca²⁺ indicator Cameleon D3cpv (20). Cameleon D3cpv measurements of [Ca²⁺]_m showed a rapid and significant increase in matrix Ca²⁺ that coincided with the first histamine-induced cytosolic Ca²⁺ transient (i.e., initial release of Ca²⁺ from intracellular stores) (Fig. 1D). Although the simultaneous measurement of both cytosolic and mitochondrial Ca²⁺ was not possible using the Cameleon D3cpv indicator, [Ca²⁺]_m exhibited a heterogeneous response that presented as repetitive [Ca²⁺]_i transients (i.e., oscillations). While these individual transients are masked when averaging all cells in the field, mean [Ca²⁺]_m was higher at any time point after histamine stimulation (Fig. 1E and F). Biophysically, [Ca²⁺]_m rose from 0.126 ± 0.022 μM at baseline to 0.839 ± 0.186 μM immediately after histamine treatment and then decreased to 0.224 ± 0.035 μM and remained statistically elevated at later time points. This finding suggests that repetitive mitochondrial Ca²⁺ loading serves to persistently elevate Ca²⁺ over baseline. Inhibition of mitochondrial Ca²⁺ uptake with the mitochondrial Ca²⁺ uniporter blocker Ruthenium Red significantly decreased the amplitude of the first [Ca²⁺]_m transient in a dose-dependent manner and abolished the subsequent elevation in [Ca²⁺]_m following histamine stimulation (Fig. 1E and F) with no effect on [Ca²⁺]_i oscillatory pattern (data not shown). Together, these results show synchronized mitochondrial Ca²⁺ loading during histamine-induced cytosolic Ca²⁺ oscillations that effectively serve to elevate matrix Ca²⁺ content.

FIG 1.

Receptor-mediated Ca²⁺ oscillations stimulate mitochondrial Ca²⁺ loading in primary endothelial cells. (A) Single-cell oscillatory [Ca²⁺]_i mobilization pattern in a representative HPAEC stimulated with histamine (100 nM) as demonstrated by the raw Fura-2 ratio. The addition of histamine is indicated by the arrow. Removal of histamine (washout) after 15 s eliminates the development of Ca²⁺ oscillations. (B) Fura-2 AM quantification of [Ca²⁺]_i in HPAECs at baseline and after treatment with histamine (100 nM) (during the first, second, and third [Ca²⁺]_i transients) (mean plus standard error of the mean [SEM] [error bar]; n = 109). (C) Representative single-HPAEC [Ca²⁺]_i mobilization response to histamine in the absence of extracellular Ca²⁺ and in which the endoplasmic reticulum Ca²⁺ store was predepleted with thapsigargin (Tg) (1 μM) for 20 min prior to the histamine addition. (D) Oscillatory [Ca²⁺]_m transients in a representative single cell as assessed by the genetically encoded Ca²⁺ sensor Cameleon D3cpv and represented as the CFP/FRET ratio after histamine stimulation (100 nM). (E) Mean calibrated [Ca²⁺]_m of HPAECs during histamine addition (100 nM). Data are shown as means ± SEMs from control cells (n = 31) and those pretreated with Ruthenium Red (RR) (20 μM, n = 26; 100 μM, n = 37). (F) Calibrated [Ca²⁺]_m at the indicated time intervals in control cells and those pretreated with 20 μM and 100 μM Ruthenium Red. The values that are significantly different by Student's t test are indicated by asterisks as follows: ****, P < 0.0001; *, P < 0.05.

TABLE 1.

Percentage of responding HPAECs and corresponding mean frequency of calcium oscillations in response to increasing doses of histamine

Buffer and histamine concn (nM)	% HPAECs (mean ±SEM) showing:			Frequency (mHz) (mean ± SEM)
	Single Ca2+ transient	Ca2+ oscillations	Ca2+ plateau
Ca2+-containing buffer
20	0	0	0	0
50	30 ± 10	45 ± 10	2.5 ± 2.5	1.48 ± 0.18
100	21.56 ± 8.44	67.78 ± 4	8.9 ± 5.88	2.74 ± 0.18
300	24.17 ± 19.61	45.31 ± 16.55	27.07 ± 10.9	2.44 ± 0.25
500	14.8 ± 7.78	45.94 ± 7.85	39.26 ± 7.6	2.15 ± 0.13
1,000	16.71 ± 3.78	17.34 ± 3.36	65.95 ± 7.13	1.45 ± 0.09
Ca2+-free buffer
50	0	0	0	0
100	27.82 ± 2.62	2.16 ± 1.26	0	0.24 ± 0.08
300	79.67 ± 3.67	2.67 ± 1.45	0	0.64 ± 0.01
500	91.3 ± 8.7	0	0	1.06 ± 0.05
1,000	89.6 ± 10.4	4.1 ± 2.4	0	1.09 ± 0.05

Sustained mitochondrial Ca²⁺ loading regulates mitochondrial NADH accumulation.

Mitochondrial matrix Ca²⁺ activates the Ca²⁺-sensitive enzymes pyruvate dehydrogenase, isocitrate dehydrogenase, and oxoglutarate dehydrogenase (2) to speed mitochondrial NADH production (7, 29). Indeed, the histamine-induced elevation in matrix Ca²⁺ triggered a significant and sustained mitochondrial NADH accumulation (+11.4% ± 0.02% with respect to baseline) as measured by time course fluorescence microscopy (Fig. 2A and B) that was maximized by the mitochondrial complex I inhibitor rotenone. In energetic tissue, NADH funnels electrons into the respiratory chain to drive oxidative phosphorylation and ATP synthesis (30). However, endothelial cells are largely glycolytic and do not rely on mitochondria as primary energy generators (31). We estimated that approximately 60% of intracellular ATP production was derived via glycolytic mechanisms and that histamine challenge did not change the relative contribution of glycolysis and oxidative phosphorylation (data not shown). In support, histamine treatment did not increase total cellular ATP production in HPAECs (Fig. 2C). Similarly, HPAECs exhibited a low basal oxygen consumption rate (OCR) that was unaffected by histamine stimulation as measured with a Seahorse analyzer (Fig. 2D). Likewise, Seahorse measurements of the extracellular acidification rate (ECAR) showed no change in the glycolytic flux after histamine (Fig. 2E). The accumulation of NADH in response to histamine did, however, coincide with an enhanced spare respiratory capacity measured upon trifluorocarbonylcyanide phenylhydrazine (FCCP)-mediated uncoupling as the difference between carbonilcyanide FCCP -stimulated and basal OCR (Fig. 2F).

FIG 2.

Matrix Ca²⁺ regulates mitochondrial bioenergetics. (A) Time course fluorescence microscopy of mitochondrial NADH autofluorescence upon stimulation of HPAECs with histamine (100 nM) and rotenone (2 μM). NADH autofluorescence was normalized to baseline. (B) Quantification of changes in NADH autofluorescence of HPAECs upon the addition of histamine (+Hist) (100 nM) and rotenone (+Rot) (2 μM) (n = 147 cells). (C) ATP content of HPAECs at 0.5, 1, 2, and 4 h after histamine stimulation (100 nM). Values are normalized to baseline ATP content. (D) Representative traces of oxygen consumption rate (OCR) of HPAECs upon sequential additions of histamine (100 nM or 500 nM), oligomycin (Oligo) (1 μM), FCCP (1 μM), and rotenone and antimycin (Rot/AA) (1 μM each) using a Seahorse XF24 analyzer. (E) Seahorse X24 measurement of the extracellular acidification rate (ECAR) of HPAECs upon sequential addition of histamine (100 nM or 500 nM), oligomycin (1 μM), FCCP (1 μM), and rotenone and antimycin A (1 μM each). (F) Spare respiratory capacity of control and histamine-treated HPAECs as measured upon addition of FCCP (1 μM) with a Seahorse XF24 analyzer. Values represent the means from three independent experiments ± SEMs. The values that are significantly different by Student's t test are indicated by asterisks as follows: ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Ca²⁺-dependent mitochondrial NADH metabolism transmits to the cytosol.

While not coupled to energy production, histamine-stimulated NADH production did manifest as a significant and persistent reduction in the total cellular NAD⁺/NADH ratio (53% ± 0.096%, up to 8 h after histamine treatment) (Fig. 3A). The depressed NAD⁺/NADH ratio was mostly due to an increase in NADH, since NAD⁺ levels did not significantly change (data not shown). The ratio between reduced NADH and oxidized NAD⁺ regulates the intracellular redox state and controls the activity of several enzymes and transcription factors (32). However, the relative impermeability of the mitochondrial inner membrane dictates that NAD⁺ and NADH are highly compartmentalized, with approximately 75% of cellular NAD⁺/NADH confined within the mitochondria (33). To our knowledge, there is no evidence that Ca²⁺-mediated increases in mitochondrial NADH influence cytosolic NAD⁺/NADH homeostasis.

FIG 3.

Mitochondrial NADH metabolism transmits to the cytosol. (A) Total intracellular NAD⁺/NADH ratio measured in HPAECs at baseline and at 0.5, 2, and 8 h after stimulation with histamine (100 nM). (B) Cytosolic NAD⁺/NADH ratio calculated from the measured lactate/pyruvate ratio in HPAECs at baseline and at 0.5 and 1 h after histamine treatment (100 nM). Values in panels A and B are mean from 3 independent experiments plus SEMs (error bars). (C) Representative pseudocolored green-to-red fluorescence ratio images of HPAEC expressing the Peredox sensor at baseline and after treatment with histamine (100 nM) and pyruvate (10 mM). (D) Representative time course of green-to-red fluorescence ratio of HPAECs expressing Peredox after treatment with 100 nM histamine and 10 mM pyruvate. Fluorescence ratio was normalized to the values obtained in the presence of pyruvate alone. Values are means ± SEMs from six cells in the field. (E) Time course of cytosolic NAD⁺/NADH ratio in HPAECs treated with histamine (100 nM) calculated from the Peredox green-to-red fluorescence ratio after sensor calibration. (F) Cytosolic NAD⁺/NADH ratio values measured with Peredox in HPAECs at baseline and at 10 and 15 min after histamine treatment. Values are means ± SEMs from 26 cells from three independent experiments. (G) Dose-dependent decrease in Peredox green-to-red fluorescence in response to Ruthenium Red (RR) (20 μM and 100 μM). Fluorescence ratio was normalized to baseline values for comparison. CON, control. (H) Peredox green-to-red fluorescence 10 min after histamine stimulation in the presence of RR (20 μM and 100 μM) and AOAA (500 μM), and when Ca²⁺ oscillations were ceased by histamine washout and by the absence of Ca²⁺ in the experimental buffer. Fluorescence ratio was normalized to baseline values for comparison. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

While NADH is not actively exported from the mitochondrial matrix, it is possible that persistent NADH buildup facilitates the transmission of reducing equivalents from the matrix to the cytosol through redox shuttles (34). We therefore calculated the cytosolic NAD⁺/NADH ratio ([NAD⁺/NADH]_cyt) by measuring the concentrations of lactate and pyruvate, which are linked by NADH-dependent lactate dehydrogenase (24). Similar to the total NAD⁺/NADH ratio, histamine treatment resulted in a 30% ± 0.019% reduction in [NAD⁺/NADH]_cyt, decreasing from 248.7 ± 6.482 to 175 ± 4.638 at 0.5 h after histamine stimulation (Fig. 3B). To directly measure the dynamics of [NAD⁺/NADH]_cyt, HPAECs were transfected with the genetically encoded ratiometric indicator Peredox (22) and visualized via confocal microscopy. Changes in the green-to-red fluorescence ratio of the sensor are directly related to changes in the cytosolic NADH/NAD⁺ ratio, and quantitative measurements of the ratio can be obtained after sensor calibration with exogenous lactate and pyruvate (23). Histamine treatment resulted in a rapid increase in the green-to-red ratio of Peredox that remained elevated throughout the experiment and was consistent with an increase in the cytosolic NADH/NAD⁺ (decrease in the cytosolic NAD⁺/NADH) (Fig. 3C and D). Quantitation of [NAD⁺/NADH]_cyt showed significantly decreased [NAD⁺/NADH]_cyt from 235 ± 38.27 at baseline to 129 ± 21.08 15 min after histamine challenge similar to values estimated from lactate and pyruvate measurement (Fig. 3E and F). Importantly, the changes in [NAD⁺/NADH]_cyt were related to the Ca²⁺-induced mitochondrial NADH increase, since blocking mitochondrial Ca²⁺ uptake with Ruthenium Red abolished the decrease in [NAD⁺/NADH]_cyt upon histamine treatment (Fig. 3G and H) in a dose-dependent manner consistent with mitochondrial Ca²⁺ uptake (Fig. 1E and F). Likewise, limiting the histamine response to a single Ca²⁺ transient either by histamine washout or by removing extracellular Ca²⁺ resulted in a normalization of the [NAD⁺/NADH]_cyt, conclusively demonstrating that repetitive transients are required to maintain a reduced [NAD⁺/NADH]_cyt. Further, while Ruthenium Red inhibited the ability of mitochondrial bioenergetics to alter the [NAD⁺/NADH]_cyt, it did not appreciably alter the histamine oscillatory response (data not shown). These data are important in that they prove that the cytosolic alterations in NADH originate within the mitochondria and are not simply occurring in parallel with mitochondrial bioenergetics. Inhibition of the malate-aspartate shuttle using aminooxyacetic acid (AOAA) also prevented the persistent alteration in the [NAD⁺/NADH]_cyt, suggesting that mitochondrial redox shuttles relay acute NADH changes to the cytosol (Fig. 3H). We observed that inhibition of both the mitochondrial Ca²⁺ uniporter and malate/aspartate shuttle resulted in a more reduced cytosolic NADH-NAD⁺ redox state at baseline, consistent with a decrease in basal mitochondrial metabolism. Indeed, a reduction in mitochondrial Ca²⁺ loading was detected at baseline in the presence of Ruthenium Red (Fig. 1E and F) that might relate to lessened mitochondrial function. Similar to ECs, an AOAA-mediated increase in the cytosolic NADH/NAD ratio has been reported in vascular smooth muscle cells due to impaired removal of cytosolic reducing equivalents through the malate/aspartate shuttle and reduced oxidative metabolism (35).

Oscillating 293AD CaSR cells show changes in [NAD⁺/NADH]_cyt.

To verify that the communication between mitochondrial and cytosolic NAD⁺/NADH metabolism was not restricted to endothelial cells, we employed an alternative cell model of inducible Ca²⁺ oscillations. 293AD cells were transfected with the calcium-sensing receptor (CaSR) (Fig. 4A), a G-protein-coupled receptor normally absent in 293AD cells (36). When cultured in media containing less than 0.4 mM Ca²⁺, 293AD CaSR cells do not exhibit baseline Ca²⁺ signaling. However, bolus addition of extracellular CaCl₂ triggers rapid and repetitive Ca²⁺ oscillations in CaSR-expressing cells (Fig. 4B). Concomitant with Ca²⁺ oscillations, the 293AD CaSR line exhibited a significant decrease in the [NAD⁺/NADH]_cyt that was absent in nontransfected 293AD cells (Fig. 4C). As the cessation of oscillations restored the [NAD⁺/NADH]_cyt, it is likely that the reversible malate-aspartate shuttle is activated by Ca²⁺-stimulated mitochondrial bioenergetics. Indeed, similar to HPAECs, AOAA preincubation in 293AD CaSR cells rendered the [NAD⁺/NADH]_cyt insensitive to oscillations. Interestingly, stimulation of mitochondrial NADH accumulation by the respiratory chain inhibitor rotenone also lowered the [NAD⁺/NADH]_cyt in nonoscillating 293AD CaSR, suggesting that a Ca²⁺-independent rise in mitochondrial NADH can also influence cytosolic NAD⁺/NADH metabolism. In total, these findings are the first to reveal that Ca²⁺-induced alterations in mitochondrial bioenergetics are actively communicated to the cytosol, likely by the reversible malate-aspartate shuttle.

FIG 4.

Ca²⁺-induced Ca²⁺ oscillations in 293AD CaSR cells result in decreased cytosolic NAD⁺/NADH ratio. (A) Immunoblot analysis of CaSR overexpression in 293AD cells transfected with Flag-CaSR DNA. The positions of molecular mass markers (M) are indicated. (B) [Ca²⁺]_i measured by Fura-2 AM in 293AD CaSR under extracellular Ca²⁺ concentrations. (C) Cytosolic NAD⁺/NADH ratio calculated from the measured lactate/pyruvate ratio in 293AD CaSR cells in the presence of high (1.8 to 1.4 mM) and low (0.4 mM) CaCl₂ concentrations in the presence and absence of aminooxyacetic acid (AOAA) (2 mM) and after 30-min treatment with rotenone (2 μM). Values are normalized to low CaCl₂ 293AD CaSR cells and are the means from three independent experiments plus SEMs. ***, P < 0.001.

Mitochondrial regulation of NAD⁺/NADH metabolism alters protein acetylation and sirtuin expression.

In both the mitochondrial and cytosolic compartments, NAD⁺/NADH metabolism coincides with alterations in the activity of NAD⁺-dependent deacetylase enzymes known as sirtuins (37). In general, a reduced [NAD⁺/NADH]_cyt regulates the activity of SIRT1 to deacetylate protein lysine residues (38, 39). Indeed, in vitro deacetylase activity of cytosolic and nuclear localized SIRT1 exhibited an inverse relationship with NADH concentrations that corresponded to ratios measured in histamine-stimulated HPAECs (Fig. 5A). Since the activity of sirtuins influences protein acetylation, total lysine acetylation was evaluated in HPAECs and 293AD CaSR cells under reduced [NAD⁺/NADH]_cyt conditions. Both histamine challenge (HPAECs) and CaCl₂ stimulation (293AD CaSR) resulted in an overall increase in protein acetylation (Fig. 5B). To determine whether acetylation was confined to the mitochondria, prominent cytosolic and mitochondrial protein targets of sirtuins were selected for individual acetylation analysis: phosphatase and tensin homolog (PTEN), p53, and MnSOD (40). Calcium oscillations were initiated in 293AD CaSR cells for 0.5 h, and endogenous PTEN and p53 (nuclear-cytosolic localization) and MnSOD (mitochondrial) were immunoprecipitated and probed for lysine acetylation. At high CaCl₂ (decreased [NAD⁺/NADH]_cyt), PTEN, p53, and MnSOD demonstrated increased acetylation (Fig. 5C and D). These data are the first to demonstrate that mitochondrial Ca²⁺ loading rapidly influences protein acetylation, likely via a buildup of NADH and the inhibition of sirtuin enzymatic activity.

FIG 5.

Mitochondrial regulation of NAD⁺/NADH metabolism alters protein acetylation and sirtuin expression. (A) In vitro measurement of SIRT1 deacetylase activity in the presence of 2 mM NAD⁺ and increasing concentrations of NADH (0 to 0.6 mM). (B) Representative Western blot of extracts of 293AD CaSR cells grown in CaCl₂ (final concentration of 1.4 mM for 0.5 h) and HPAECs in the presence or absence of histamine (Hist) (100 nM, 0.5 h). The extracts were probed with a pan-acetyllysine (Ac-Lysine) antibody and normalized to GAPDH protein levels. Con, control. (C) Western blot (WB) analysis of PTEN, p53, and MnSOD (tetramer) acetylation immunoprecipitated (IP) from 293AD CaSR cells at low (0.4 mM) (−) and high (1.4 mM) (+) CaCl₂ concentrations in the presence of trichostatin A (TSA) (250 nM). (D) Quantitation of protein acetylation in panel C as calculated by densitometry of Ac-lysine/total protein. (E) Western blot analysis of SIRT1 and SIRT3 protein levels in 293AD CaSR cells at 0.5, 1, 2, and 4 h after treatment with high CaCl₂ (final concentration of 1.4 mM). Cells were incubated for 0.5 h in low CaCl₂ (0.4 mM) before the CaCl₂ concentration was increased. GAPDH was used as a loading control. (F) Western blot analysis of SIRT1, SIRT2, SIRT3, SIRT4, SIRT5, and SIRT6 protein levels in HPAECs treated with histamine (100 nM) at 0.5, 1, and 2 h after histamine treatment, using GAPDH as a loading control. (G) Quantitation of protein levels in panel F as calculated by densitometry relative to GAPDH levels. (H) Relative expression of SIRT1, SIRT2, SIRT3, SIRT4, SIRT5, and SIRT6 mRNA in HPAECs treated with histamine at 0.5, 1, and 2 h. 18S mRNA was used for normalization. (I and J) Western blot analysis of acetylated histone H3 protein levels in HPAECs treated with histamine (100 nM) (I) and 293AD CaSR cells treated with 1.4 mM CaCl₂ in the presence of TSA (250 nM) (J). GAPDH and histone H3 were used as loading controls. The positions of molecular mass markers are indicated to the right of the blots. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Lysine acetylation is broadly associated with cellular metabolism, and increased acetylation is commonly observed in metabolic diseases such as diabetes and cancer (41). Thus, it is likely that long-term histamine-triggered accumulation of NADH (Fig. 3A) would initiate a compensatory response to restore cytosolic and mitochondrial protein acetylation and cellular metabolic homeostasis. We therefore measured nucleocytosolic SIRT1 and mitochondrial SIRT3 expression at different time points following the initiation of Ca²⁺ oscillations in both 293AD CaSR cells and HPAECs. At the 30-min time point that corresponded to increased protein acetylation, we detected no alteration in the expression of SIRT1, which is consistent with diminished enzymatic activity. However, as oscillations persisted, we discovered a gradual increase in SIRT1 protein expression (in 293AD CaSR cells [Fig. 5E] and HPAECs [Fig. 5F and G]). SIRT3 exhibited a variable expression pattern that fluctuated over time. Both SIRT1 and SIRT3 mRNA levels corresponded to increased protein expression in HPAECs (Fig. 5G and H). Along with SIRT1 and SIRT3, HPAECs exhibited an overall increase in the expression of nuclear SIRT6, as well mitochondrion-localized SIRT4 and SIRT5. Cytosolic SIRT2 expression was unaltered by histamine (Fig. 5F and G). Aside from SIRT1 and SIRT3, only elevated SIRT4 and SIRT6 expression was a result of increased transcription, suggesting a complex regulatory process (Fig. 5H). Histone H3 is one of the five core histones and a nuclear target of nucleocytosolic SIRT1 (42), SIRT2 (43), and SIRT6 (44). Concurrent with sirtuin expression, changes in histone H3 acetylation (Ac-H3) were also noted (in HPAECs [Fig. 5I] and in 293AD CaSR cells [Fig. 5J]): Ac-H3 increased at early time points after induction of Ca²⁺ oscillations and decreased when SIRT1 and SIRT6 expression were at the highest levels.

To establish a causal relationship between protein acetylation and sirtuin activity, the [NAD⁺/NADH]_cyt was restored using the cell-permeable NAD⁺ precursor β-nicotinamide mononucleotide (NMN) (45) and Ac-H3 evaluated at the early time point after CaCl₂ challenge in 293AD CaSR cells. NMN completely prevented histone H3 hyperacetylation at high CaCl₂ concentrations (Fig. 6A). While not as effective as NMN, inhibition of the malate-aspartate shuttle (AOAA) also reduced histone H3 acetylation at early time points of CaCl₂ stimulation. Since SIRT1 is the most well-understood member of the sirtuin family in the vascular endothelium, we next measured how manipulating mitochondrial and cytosolic Ca²⁺ loading influences SIRT1 expression. In HPAECs, diminishing mitochondrial Ca²⁺ uptake with Ruthenium Red completely abolished the histamine-mediated increase in SIRT1 expression at 2 h (Fig. 6B). Further, eliminating Ca²⁺ oscillations by either washing out histamine after the initial transient (Fig. 6C) or by removing extracellular Ca²⁺ uptake also blocked the increase in SIRT1 (Fig. 6D). In total, our data demonstrates that mitochondrial Ca²⁺ loading can acutely influence protein acetylation by modulating both the activity and expression of sirtuins not only in mitochondria but also in the cytosolic and nuclear compartments.

FIG 6.

Protein acetylation and SIRT1 expression are dependent upon sustained Ca²⁺ oscillations and mitochondrial Ca²⁺ loading. (A) Western blot analysis of histone H3 acetylation in 293AD CaSR cells kept at low CaCl₂ (0.4 mM) in the presence of TSA (250 nM) at different time points after extracellular CaCl₂ was increased to 1.4 mM in the presence of NMN (200 μM) or AOAA (2 mM). GAPDH was used as a loading control. (B) Western blot analysis of SIRT1 protein levels in HPAECs treated with histamine (100 nM) in the presence of Ruthenium Red (20 μM) at 0.5, 1, and 2 h after histamine treatment, using GAPDH as a loading control. (C and D) Western blot analysis of SIRT1 protein levels in HPAECs treated with histamine (100 nM) in which Ca²⁺ oscillations ceased due to histamine washout (C) and removal of extracellular Ca²⁺ (D). GAPDH was used as a loading control. M, molecular mass.

Mitochondrion-directed SIRT1 upregulation is anti-inflammatory in HPAECs.

In the vascular endothelium, SIRT1 is an important contributor to angiogenesis, vascular tone, and inflammation (46, 47). During inflammation, activated endothelial cells upregulate the expression of adhesion molecules (intercellular adhesion molecules [ICAMs], vascular cell adhesion molecules [VCAMs], and selectins) that facilitate the arrest and recruitment of circulating leukocytes (48–50). To evaluate whether mitochondrion-dependent SIRT1 expression influences endothelial/leukocyte adhesion, J774 macrophages were incubated with HPAECs after histamine challenge. As expected, histamine stimulated an acute increase in J774 macrophage adherence (0.5 and 1 h). Conversely, J774 adhesion was significantly diminished 2 h after histamine challenge (Fig. 7A), coincident with the demonstrated increase in SIRT1 expression by mitochondrion-to-cytosol NADH transmission. To mechanistically link SIRT1 to macrophage adherence, HPAECs were pretreated with the SIRT1 enzymatic inhibitors nicotinamide and suramin (51, 52) and then incubated with J774 macrophages at the 2-h time point. Both suramin and nicotinamide significantly increased J774 adhesion to histamine-treated HPAECs relative to controls (Fig. 7B), demonstrating that sirtuin inhibition increases leukocyte adhesion. Pretreatment of HPAECs with Ruthenium Red, which removed the ability of the mitochondria to upregulate SIRT1 expression, had an even more pronounced effect on J774 adhesion (Fig. 7B) and suggests that mitochondrial Ca²⁺ import is the key step required to prevent chronic inflammation and leukocyte recruitment. In support, Ruthenium Red treatment significantly increased the mRNA levels of E-selectin 2 h after histamine treatment (Fig. 7C). To evaluate whether the SIRT1-mediated inhibition of leukocyte adhesion was exclusive to histamine, additional studies were conducted by adding the inflammatory agent tumor necrosis factor alpha (TNF-α) to HPAECs at 0.5, 1, and 2 h after histamine challenge. Short-term histamine pretreatment (1 h) enhanced TNF-α-mediated leukocyte adhesion over TNF-α alone (Fig. 7D). Conversely, TNF-α-stimulated adhesion was diminished following histamine pretreatment for 2 h, the time point at which SIRT1 was elevated. Inhibition of SIRT1 activity by nicotinamide significantly increased J774 adhesion at both time points, indicating that increased SIRT1 expression is anti-inflammatory in HPAECs. In total, these findings suggest that Ca²⁺-stimulated mitochondrial bioenergetics are a critical modulator of intracellular signaling cascades via SIRT1 in general and may play a prominent role in vascular inflammation specifically.

FIG 7.

Mitochondrion-directed SIRT1 upregulation is anti-inflammatory in HPAECs. (A) J774 macrophage adhesion to HPAECs treated with histamine (100 nM) at 0.5, 1, and 2 h after histamine treatment. J774 macrophages were stained with Cell Tracker Green, and the fluorescence of macrophages adherent to HPAECs was measured with a fluorescence plate reader. ns, not significant. (B) J774 macrophage adhesion to HPAECs treated with histamine (100 nM) for 2 h in the presence of the sirtuin enzymatic inhibitors suramin (50 μM) and nicotinamide (NAM) (10 mM) and in the presence of the mitochondrial uniporter blocker Ruthenium Red (20 μM). (C) Relative mRNA levels of E-selectin in HPAECs treated with histamine (100 nM) in the presence or absence of Ruthenium Red (20 μM) at 0.5, 1, and 2 h after histamine treatment. 18S mRNA was used for normalization. (D) J774 macrophage adhesion to HPAECs treated with TNF (40 ng/ml, 2 h) and histamine (hist) (100 nM) for 1 or 2 h in the presence or absence of nicotinamide (10 mM). Values are means from 3 independent experiments plus SEMs. (E) Schematic representation of results. Ca²⁺ oscillations effectively raise Ca²⁺ levels within the mitochondrial matrix, resulting in NADH production by the tricarboxylic acid (TCA) cycle. NADH not used for energy production will activate the malate-aspartate (MAL-ASP) shuttle to transmit excess reducing equivalents to the cytosol, resulting in a diminished cytosolic NAD⁺/NADH ratio ([NAD⁺/NADH]_cyto). A lowered [NAD⁺/NADH]_cyto influences cytosolic/nuclear sirtuin activity and increases protein acetylation. To compensate for this increase in protein acetylation, a mitochondrial Ca²⁺-directed increase in SIRT1 expression occurs, which in the endothelium reduces inflammation over time. MCU, mitochondrial calcium uniporter. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

DISCUSSION

Aside from their obvious importance in energy production, mitochondria have emerged as signaling organelles that integrate and coordinate metabolic responses. While the nuclear control of mitochondrial function (anterograde signaling) is well established (53), the mitochondrial modulation of nuclear gene expression (retrograde response) is largely unknown. Examples of mitochondrial retrograde signaling include the mitochondrial unfolded protein response and mitochondrial release of oxidants, nitric oxide, and Ca²⁺ (54). Our data introduce an additional mode of retrograde signaling, in which mitochondria integrate and translate cytosolic Ca²⁺ transients into a genetic response via the transmission of mitochondrial NADH to the cytosol and nucleus. We demonstrate the biological significance of Ca²⁺-mediated mitochondrial retrograde signaling by using primary ECs and provide evidence that the mitochondrial regulation of [NAD⁺/NADH]_cyt can influence the inflammatory responses of these cells.

Cytosolic Ca²⁺ oscillations are universal signals that affect diverse cellular processes (55) and are rapidly buffered by mitochondria. While the cytosolic consequences of Ca²⁺ oscillations have been extensively studied, how repetitive Ca²⁺ waves affect mitochondrial function is still unclear, as mitochondrial Ca²⁺ uptake is largely cell type and agonist dependent. For example, in pancreatic islets and HeLa cells, repetitive [Ca²⁺]_i oscillations desensitize the mitochondrial Ca²⁺ uniporter and reduce long-term matrix Ca²⁺ levels (56, 57). Conversely, vasopressin-induced [Ca²⁺]_i oscillations in hepatocytes rapidly transmit to the mitochondrial matrix and result in parallel [Ca²⁺]_m oscillations (7). In RBL-2H3 mast cells, adenosine receptor activation (carboxamidoadenosine) effectively increased mitochondrial Ca²⁺ uniporter activity to elicit a stepwise increase in mitochondrial Ca²⁺ with each successive transient (58). In our studies, endothelial mitochondrial Ca²⁺ loading was more efficient during the first cytosolic Ca²⁺ transient that is released from the intracellular stores. However, while mean mitochondrial Ca²⁺ loading never reached the magnitude of the original transient, we observed repetitive mitochondrial Ca²⁺ transients in response to histamine that effectively resulted in an overall mean rise in matrix Ca²⁺ concentration above the baseline, triggering a chronic activation of matrix dehydrogenases and NADH production. In highly metabolic cardiomyocytes, the Ca²⁺-stimulated increase in mitochondrial NADH is rapidly consumed to enhance ATP production and provide the energy necessary for cardiac contraction (59, 60). Conversely, our experiments revealed a persistent buildup of mitochondrial NADH in HPAECs in response to cytosolic Ca²⁺ oscillations that was the result of enhanced NADH production without a requisite increase in ATP synthesis. Ca²⁺-dependent mitochondrial NADH accumulation has been documented in numerous cell types, including neurons during membrane depolarization (30), hormone-stimulated hepatocytes (7, 61), and pancreatic islets of Langerhans during glucose-induced insulin release (8). In concert with the present findings, we surmise that cytosolic oscillations are likely an efficient means to effect an overall increase in [Ca²⁺]_m, and in turn, initiate a persistent elevation in NADH production within the mitochondrial matrix.

A major consequence of mitochondrial NADH accumulation in HPAECs was the transmission of mitochondrial NADH changes to the cytosolic and nuclear compartments. An estimation of the [NAD⁺/NADH]_cyt revealed a diminished ratio that corresponded to increased mitochondrial NADH production during [Ca²⁺]_i oscillations. With the inner mitochondrial membrane impermeable to NAD⁺ and NADH, the cytosolic and mitochondrial pools of pyridine nucleotides are well separated and present as distinct free NAD⁺/NADH ratios (62). However, while there is likely no direct export of NADH out of mitochondria, reducing equivalents can be transported from the matrix to the cytosol via the malate-aspartate (reversible) and glycerol phosphate (irreversible) shuttles (63). In our studies, we detected both a normalization of the NAD⁺/NADH ratio when Ca²⁺ oscillations ceased in AD293 CaSR cells and an elevated NAD⁺/NADH ratio in cells pretreated with AOAA, indicating that the reversible malate-aspartate shuttle is likely the prominent mechanism through which Ca²⁺-induced mitochondrial bioenergetics are relayed to the cytosol. To compensate, the irreversible glycerol phosphate shuttle could reoxidize NADH to NAD⁺ in the cytosol (64) and restore the [NAD⁺/NADH]_cyt over time. However, optimal glycerol phosphate activity requires equimolar concentrations of both the cytosolic and mitochondrial components of the cycle (65), which appears to occur only in highly metabolic tissue (66). Indeed, we observed a steady reduction in the [NAD⁺/NADH]_cyt for up to 8 h after histamine stimulation for as long as Ca²⁺ oscillations were present (data not shown), suggesting that metabolically quiescent ECs may have little glycerol phosphate shuttle activity. In support, the glycerol phosphate shuttle was unable to restore the [NAD⁺/NADH]_cyt in porcine aortic tissue in which malate-aspartate activity was inhibited by AOAA (35), indicating lowered glycerol phosphate shuttle activity in vascular cells. Further study is required to define the relative contributions of both the malate-aspartate and glycerol phosphate shuttle in the transmission of mitochondrial NADH to the cytosol. Interestingly, recent work has identified the malate-aspartate and glycerol phosphate NADH shuttles as regulators of life span in calorie-restricted Saccharomyces cerevisiae via Sir2 (67). In this study, the authors propose that a chronic increase in mitochondrial respiration triggers a rise in the NAD⁺/NADH ratio, which then transmits to the cytosol through redox shuttles to activate Sir2 (the yeast homolog of SIRT1). Our data demonstrate that mitochondrial redox shuttles not only sense chronic alterations in nutrient availability but that they also allow mitochondria to integrate and convert receptor-mediated Ca²⁺ transients into an acute cytosolic metabolic signal.

In our studies, the biological effect of mitochondrion-to-cytosol NADH shuttling coincided with alterations in protein acetylation in both the mitochondrial and cytosolic-nuclear compartments. Sirtuins are highly conserved proteins that utilize NAD⁺ as a cofactor and are therefore exquisitely responsive to changes in the NAD⁺/NADH ratio (38, 68). Concordant with the notion that NAD⁺ and NADH are immobile within cells, sirtuins have multiple isoforms localized to different cellular compartments and respond to discrete alterations in their particular microenvironment (SIRT1 and SIRT2 are nucleocytosolic, SIRT3, SIRT4, and SIRT5 are mitochondrial, SIRT6 and SIRT7 are nuclear) (37). We originally surmised that Ca²⁺-mediated NADH accumulation would predominately affect proteins within the mitochondria (i.e., MnSOD). However, we were surprised at the extent to which alterations in PTEN and p53 acetylation occurred in the cytosolic compartment. Reversible protein acetylation is an evolutionarily conserved posttranslational regulatory mechanism that alters the activity of metabolic enzymes involved in glycolysis, gluconeogenesis, tricarboxylic acid cycle, urea cycle, and fatty acid metabolism (69). Thus, the ability of the mitochondria to influence cytosolic metabolic proteins via acetylation represents a convenient mechanism by which mitochondria can coordinate metabolism between adjoining cellular compartments. In addition to cytosolic proteins, we also detected mitochondrial Ca²⁺-dependent alterations in the acetylation of nuclear histone H3. Changes in histone acetylation concomitant with the mitochondrial NADH homeostasis suggest the possibility that mitochondrial metabolism influences not only cytosolic and mitochondrial processes but may also initiate cellular metabolic reprogramming through epigenetic mechanisms.

The transcriptional regulation of sirtuin expression is complex and has been studied predominantly during dietary modifications by nutrient starvation (caloric restriction) or nutrient excess (high-fat diet) (37). Several pathways have been proposed to explain SIRT1 upregulation upon nutrient withdrawal, including Forkhead box O3a (FOXO3a)-p53 (70), peroxisome proliferator-activated receptor α (PPARα) (71), cyclic AMP response element-binding protein (CREB) (72), and PPARδ (73). Our studies demonstrate an inverse relationship between protein acetylation and sirtuin expression. The rapid cellular dynamics of this interaction suggests a compensatory mechanism in which expression of nucleocytosolic and mitochondrial sirtuins (SIRT1 and SIRT3) rises to counter a reduction in enzyme activity. An interesting potential link between changes in NADH levels and variations in SIRT1 expression is the SIRT1 transcriptional regulation by the NADH-sensitive transcriptional corepressor C-terminus-binding protein (CtBP) (74). However, this study noted an inverse relationship between free NADH levels and SIRT1 transcription. While we did not investigate the exact transcriptional machinery engaged by mitochondrial NADH shuttling, we demonstrate using two different cell types that increased SIRT1 expression occurs during physiologic Ca²⁺ signaling and SIRT1 upregulation is linked to a mitochondrion-directed decrease in the cytosolic NAD⁺/NADH ratio.

Endothelial SIRT1 targets multiple proteins to effectively modulate nitric oxide production (75, 76), inflammation (47, 77, 78), and vessel growth (79, 80). Thus, SIRT1 expression is widely considered a positive effector of EC function. Indeed, our study demonstrates that mitochondrial upregulation of SIRT1 expression has an anti-inflammatory effect by reducing the expression of adhesion molecules (E-selectin) and preventing the recruitment of circulating leukocytes. At the molecular level, the inhibitory effect of increased SIRT1 on adhesion molecule expression may be mediated by the NF-κB pathway. Indeed, SIRT1 inhibits NF-κB transcription by directly interacting and deacetylating RelA/p65 (81). The vasoprotective effect SIRT1 as well as its inhibitory effect on the expression of adhesion molecules has been previously documented. In TNF-α-stimulated endothelial cells, endogenous SIRT1 levels prevented superoxide production and expression of ICAM-1 and VCAM-1 by suppressing NF-κB signaling (82). Similarly, resveratrol-induced SIRT1 expression directed the expression of vasoprotective transcription factor Kruppel-like factor KLF2 (83) and protected endothelial cells against the effects of cigarette smoking-induced oxidative stress by invoking antioxidant, anti-inflammatory, and antiapoptotic mechanisms (84).

In summary, our studies demonstrate that enhanced mitochondrial Ca²⁺ loading results in a persistent elevation of NADH within the matrix. When not used as a substrate for ATP generation, NADH transmits to the cytosol to influence protein acetylation and gene expression. The biological significance of this finding was demonstrated using primary human ECs, in which mitochondrial Ca²⁺-driven bioenergetics participates as a negative regulator of inflammation and leukocyte adhesion. Taken together, our results reveal a fundamental mechanism by which mitochondria regulate cytosolic NAD⁺/NADH metabolism to initiate retrograde signaling, epigenetic regulation, and metabolic adaptation.