Protein Kinase B (PKB/AKT1) Formed Signaling Complexes with Mitochondrial Proteins and Prevented Glycolytic Energy Dysfunction in Cultured Cardiomyocytes During Ischemia-Reperfusion Injury

Abstract

Our previous studies showed that insulin stimulated AKT1 translocation into mitochondria and modulated oxidative phosphorylation complex V in cardiac muscle. This raised the possibility that mitochondrial AKT1 may regulate glycolytic oxidative phosphorylation and mitochondrial function in cardiac muscle cells. The aims of this project were to study the effects of mitochondrial AKT1 signaling on cell survival in stressed cardiomyocytes, to define the effect of mitochondrial AKT1 signaling on glycolytic bioenergetics, and to identify mitochondrial targets of AKT1 signaling in cardiomyocytes. Mitochondrial AKT1 signaling played a protective role against apoptosis and necrosis during ischemia-reperfusion stress, suppressed mitochondrial calcium overload, and alleviated mitochondrial membrane depolarization. Activation of AKT1 signaling in mitochondria increased glucose uptake, enhanced respiration efficiency, reduced superoxide generation, and increased ATP production in the cardiomyocytes. Inhibition of mitochondrial AKT attenuated insulin response, indicating that insulin regulation of ATP production required mitochondrial AKT1 signaling. A proteomic approach was used to reveal 15 novel targets of AKT1 signaling in mitochondria, including pyruvate dehydrogenase complex (PDC). We have confirmed and characterized the association of AKT1 and PDC subunits and verified a stimulatory effect of mitochondrial AKT1 on the enzymatic activity of PDC. These findings suggested that AKT1 formed protein complexes with multiple mitochondrial proteins and improved mitochondrial function in stressed cardiomyocytes. The novel AKT1 signaling targets in mitochondria may become a resource for future metabolism research.

Heart failure and cardiomyopathy are major public health issues in the United States and developed countries. Although there are several contributing factors leading to the development of heart failure, metabolic dysregulation plays a critical role in the development of cardiomyopathy, and diabetes aggravates the morbidity and mortality associated with heart failure (1–4). Diabetes mellitus is characterized by insulin deficiency and/or insulin resistance, which leads to hyperglycemia, defective glucose uptake, reduced glycolytic oxidative phosphorylation, inefficient energy production, and oxidative stress (3).

Mitochondria are the major source of ATP production in the heart via glycolysis and β-oxidation. Oxidative phosphorylation through glycolysis is significantly reduced in the diabetic myocardium. This creates a metabolic switch to selectively rely on β-oxidation, which is associated with lipotoxicity and inefficient energy conversion (3). Reversing metabolic abnormalities and restoring myocardial energetics prevented subsequent development of cardiac dysfunction in persons with diabetes (5, 6). Metabolic dysfunction increased mitochondrial reactive oxygen species and induced cardiac muscle apoptosis and subsequent development of myocardial fibrosis (7). Although mitochondria play a pivotal role in the dysregulation of bioenergetics and oxidative stress, the exact molecular mechanisms underlying mitochondria dysfunction in diabetic cardiomyopathy are not completely understood.

Insulin deficiency or insulin resistance decreased insulin receptor signaling and glycolysis. Reduced glucose transporter type 4 expression and impaired insulin-induced glucose transporter type 4 translocation may lead to less pyruvate available for glycolytic oxidative phosphorylation (8). The mechanisms linking impaired insulin receptor signaling to mitochondrial dysfunction in cardiac muscle cells have begun to unfold in recent years. Our laboratory has identified AKT1 (protein kinase B), a serine/threonine-specific protein kinase, translocation to mitochondria as a critical link between insulin receptor signaling and mitochondria oxidative phosphorylation. AKT1 was the primary AKT isoform that relayed insulin receptor signaling to mitochondria and modulated oxidative phosphorylation complex V activity (9). Insulin activation of mitochondrial AKT1 was hampered in an animal model of type 2 diabetes (10). We hypothesized that impaired mitochondrial AKT1 signaling plays a role in the dysregulation of bioenergetics in cardiac muscle cells.

There are several unanswered questions regarding mitochondrial AKT1 actions in the heart. The effect of mitochondrial AKT1 signaling on the efficiency of ATP production and reactive oxygen species formation during cardiac stress has not been investigated, and the targets of mitochondrial AKT1 signaling within mitochondria are not known. In this study, we investigated the effect of mitochondrial AKT1 on mitochondrial function in an experimental model of ischemia-reperfusion (I/R) injury, dissected the effects of mitochondrial AKT1 activation on glycolytic bioenergetics, and identified novel mitochondrial targets of AKT1 signaling. The results suggest that activation of mitochondrial AKT1 enhanced glucose uptake and glycolysis, reduced oxidative stress, and improved cardiomyocyte survival during cardiomyocyte I/R injury. These findings suggest that the efficiency of respiration was improved, because more ATP was generated with less oxygen consumption. Using a proteomic approach, we have identified 15 protein targets of AKT1 signaling in cardiac mitochondria. Among these proteins, we have confirmed and characterized the association of AKT1 with pyruvate dehydrogenase complex (PDC) subunits and verified stimulatory effects of mitochondrial AKT1 on the enzymatic activity of PDC. These novel findings suggested that AKT1 formed signaling complexes with multiple mitochondrial proteins, improved mitochondrial function in stressed cardiomyocytes, and thereby shed new insight into the mechanisms of AKT1 actions in mitochondria.

Research Design and Methods

Cell culture and transduction of viral constructs

Primary cultures of neonatal cardiomyocytes were prepared from Sprague-Dawley rats according to a protocol we previously described (11). The animal experimental procedures were approved by the Institutional Animal Care and Use Committee at University of California, Irvine. Cardiomyocytes were maintained with DMEM containing 10% fetal bovine serum and 1% penicillin/streptomycin and incubated at 37°C, 5% CO₂. To transduce adenoviral or lentiviral vectors, cardiomyocytes were incubated with equal amounts of viral vectors in DMEM containing 10% fetal bovine serum for 48–72 hours. Two recombinant adenoviruses were used in this study. An adenoviral vector expressed a mitochondria-targeting constitutively active AKT (Ad-Mito-Akt) (9, 12). Mitochondrial targeting was achieved by fusing a mitochondria-targeting sequence (MSVLTPLLLRGLTGSARRLPVPRAKIHSL) to the N terminus. The control adenovirus-expressed green-fluorescent protein (GFP). A His-tagged dominant-negative AKT (substitutions at K179A, T308A, and S473A) with mitochondria-targeting sequence at the N terminus in a Tet-on inducible lentiviral vector with and without red fluorescent protein (RFP) was used (9, 12). The unmodified lentiviral vector with RFP served as a control. Ad-Mito-AKT and the dominant-negative mitochondria-targeting AKT virus, respectively, achieved more than 95% and more than 90% transduction efficiency in the primary cardiomyocytes. Rabbit antiphospho-T308 AKT1 and ERK2 antibodies were purchased from Cell Signaling. Rabbit antipyruvate dehydrogenase subunits E1α, E1β, E2, E3, component X of pyruvate dehydrogenase complex (PDHX), His-tag, and AKT1(E45) were purchased from Genetex.

In vitro model of I/R

Cardiomyocytes were incubated in a low oxygen chamber with mock ischemia media (modified Krebs-Henseleit buffer: [pH 6.5] 0.9% O₂ and 0mM glucose, 37°C) for 2 hours and then incubated in a normal oxygen chamber with a mock reperfusion media for 1 hour ([pH 7.4] 21% O₂ and 11mM glucose, 37°C).

Immunofluorescence staining

To visualize the effect of I/R on AKT1 subcellular localization, the cardiomyocytes were fixed with 4% formaldehyde for 30 minutes at room temperature. After washing with PBS, cells were treated with 0.05% saponin in ddH₂O for 20 minutes and blocked with 10% normal goat sera for 30 minutes. The fixed cells were then incubated with specific primary antibodies overnight at 4°C, conjugated to the appropriate fluorochrome-conjugated secondary antibodies, stained with MitoTracker Green (Invitrogen) for 1 hour, counterstained with 4′,6-diamidino-2-phenylindole, and then analyzed with a Zeiss AxioPlan2 Fluorescence microscope.

Apoptosis/necrosis assay

An Annexin V Apoptosis kit (MitoTracker Red and Alexa Fluor 488 Annexin V) was purchased from Invitrogen. Cardiomyocytes were stained according to the instructional manual, and their fluorescence was measured with a BD LSR II Flow Cytometer (excitation at 488 nm, emission at 525 and 575 nm). Although apoptotic cells showed green fluorescence with reduced red fluorescence, live cells showed minimal green fluorescence and bright red fluorescence, and necrotic cells showed neither. The fluorescence activated cell sorting (FACS) measurement was gated according to the positive (doxorubicin and H₂O₂ treated) and negative controls (no death-inducing agent). The analysis was presented as the percentages of apoptotic and necrotic cells.

FACS measurements of ΔΨm and superoxide

For mitochondrial membrane potential (ΔΨm), the cells were incubated with JC-1 for 25 minutes before they were analyzed at green and red emission wavelengths. Intact mitochondrial membranes retained the JC-1 dye, a cationic carbocyanine dye that accumulates in intact mitochondria, and formed J-aggregate (orange-red fluorescence), whereas depolarization of the membranes decreased the ability to retain dye that remains as a monomer (green fluorescence). ΔΨm was determined by the ratio of green to red fluorescence in the cells. To measure mitochondrial superoxide, the cells were loaded with 1μM MitoSOX Red (Molecular Probes) at 37°C for 30 minutes and then washed 3 times with PBS. The fluorescence was measured with a BD LSR II Flow Cytometer (excitation at 514 nm and emission at 575 nm). BD FACSDiVa software was used to analyze the data.

Analysis of mitochondrial O₂ respiration by extracellular flux measurements

To measure mitochondrial function in cardiomyocytes, we used a Seahorse Bioscience XF24 Extracellular Flux Analyzer (Seahorse Bioscience) and followed the manufacturer's protocol. Cells were briefly plated in a 0.2% gelatin coated 24-well Seahorse XF24 assay plate at 5 × 10⁴ cells/well and then incubated for 16 hours before analysis. Before the metabolic flux analysis, cells were washed once with unbuffered, serum-free DMEM (4.5 g/L glucose [pH 7.4]) and incubated with buffered media at 37°C in a non-CO₂ incubator for 1 hour. Three baseline measurements of oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were taken before sequential injection of the following mitochondrial inhibitors (final concentration): oligomycin (1 μg/mL), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) (3μM), and rotenone (0.1μM). Three measurements were taken after the addition of each inhibitor. OCR and ECAR values were automatically calculated and recorded by the Seahorse XF24 software. The basal respiration was calculated by averaging the 3 measurements of OCR before injecting the inhibitors. The spare respiration capacity was calculated by using the OCR measurement after FCCP and subtracting the basal respiration. The proton leak was calculated by using the OCR measurement after oligomycin injection and subtracting the OCR measurement after rotenone injection.

Mitochondria abundance

The content of mitochondria in cardiomyocytes was determined by the ratio of mtDNA to nDNA. Quantitative real-time PCR was used to determine the copy number of mitochondrial 12S rDNA and nuclear 18S rDNA. The mitochondria 12S and nuclear 18S primer sets were used for quantitative polymerase chain reaction. Mitochondria 12S, forward 5′-ACCGCGGTCATACGATTAAC-3′ and reverse 5′-CCCAGTTTGGGTCTTAGCTG-3; and nuclear 18S, forward 5′-CGCGGTTCTATTTTGTTGGT-3′ and reverse 5′-AGTCGGCATCGTTTATGGTC-3′. PCR was performed with an ABI 7900 real-time thermocycler coupled with SYBR Green using the following conditions: stage 1, 50°C for 2 minutes; stage 2, 95°C for 10 minutes; and stage 3, 40 cycles of 95°C for 15 seconds, then 60°C for 60 seconds. Each sample was analyzed in triplicates, and the copy numbers were determined by the Comparative Threshold Cycle Method (ABI User Bulletin 2). Standard curves were obtained by serial dilutions of known 12S and 18S DNA fragments corresponding to the primer sets.

Glucose uptake assay

Glucose uptake was measured with 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) (Invitrogen). The cells were serum-deprived overnight and stimulated with insulin or vehicles for 30 minutes in the presence of 200μM 2-NBDG at 37°C and then washed 3 times with ice-cold PBS. The content of 2-NBDG was measured with a BD LSR II Flow Cytometer (excitation at 488 nm and emission at 560 nm).

ATP assay

ENLITEN ATP assay kits were obtained from Promega. Cultured neonatal cardiomyocytes were trypsinized and counted before use. For each assay, 1 × 10⁶ cells were used. Cells were digested with 2.5% trichloroacetic acid (TCA) and then neutralized with Tris-acetate (pH 7.75). The luminescence was measured immediately using a Synergy HT multidetection microplate reader. A 2-second delay time after 100-μL rLuciferase/Luciferin reagent injection and a 10-second relative light unit signal integration time were used according to the ENLITEN ATP assay manual.

Proteomic identification of mitochondrial proteins interacted with Akt1

To identify the proteins bound to AKT1 in mitochondria, His-tagged mitochondria-targeting constitutively active AKT1 was expressed in 293 cells. The mitochondria-targeting effect was confirmed with fluorescence microscopy, and the mitochondria were isolated by differential centrifugation as we previously reported (9). The proteins associated with AKT1 were cross-linked with crosslinker, dithiobissuccinimidyl propionate in their native form, and the protein complexes were extracted. Two purification steps were performed as outlined in figure 4 below. After elution from the His-binding column, the protein complexes were treated with 200mM dithiothreitol to reverse cross-linking and then resolved with SDS-PAGE. Specific bands were cut and digested with trypsin, and the resulting peptides were analyzed with an AB Sciex 5800 MALDI TOF/TOF (matrix-assisted laser desorption/ionization time-of-flight/time-of-flight). Data processing was performed with the ProteinPilot 4.0 software, which incorporated a Paragon search algorithm to finalize the protein identification process.

Pyruvate dehydrogenase activity assay

A pyruvate dehydrogenase activity kit was purchased from the Biomedical Research Service Center at SUNY Buffalo. For each assay, 1 × 10⁶ cardiomyocytes were washed twice with ice-cold PBS, 100-μL ice-cold 1× cell lysis solution was added to cell pellets after removing PBS. For each well of a 96-well microplate, 10-μL sample and 50-μL assay solution were added and mixed via gentle agitation. The plate was incubated at 37°C for 60 minutes, and the reaction was stopped by adding 50-μL 3% acetic acid per well. The OD at 492 nm was measured using a Synergy HT multidetection microplate reader, and the readings were normalized by the protein contents in each sample.

Statistical analysis

Data are presented as mean ± SEM, unless noted otherwise. FACS data were analyzed with BD FACSDiVa software. Student's t test and one-way ANOVA with Holm-Sidak post hoc analysis were performed with the SigmaStat 3.11 software. The statistical significance level was set at P < .05.

Results

Alterations of mitochondrial AKT1 signaling during simulated I/R in primary cardiomyocytes

To study mitochondrial AKT1 signaling and its impact on respiration during cardiac stress, we have established an in vitro model of I/R injury. Antibodies against activated AKT1 (p-AKT1) were used in this series of experiments. When primary cardiomyocytes were exposed to low oxygen, low nutrient environment (ischemia), p-AKT1 were nearly undetectable in the mitochondria (Figure 1). Upon reperfusion, significant amounts of p-AKT1 can be detected in the mitochondria within 5 minutes. These data indicated dynamic changes of mitochondrial AKT1 translocation and activation during cardiac muscle I/R. During I/R, Caspase 3 was activated, and the percentage of cardiomyocytes with activated Caspase 3 was significantly increased (Figure 1C). A mitochondria-targeting constitutively active AKT1 reduced activation of Caspase 3 during I/R, whereas inhibition of mitochondrial AKT signaling with a mitochondria-targeting dominant-negative AKT (mdnAKT) exacerbated the activation of Caspase 3.

Figure 1.

Dynamic activation of AKT1 in the mitochondria during simulated I/R injury. A, Activation of mitochondrial AKT1 during I/R in cardiomyocytes. Primary cardiomyocytes were subjected to in vitro ischemia and reperfusion as outlined in Research Design and Methods. The cells were harvested and subfractionated into cytosolic and mitochondrial fractions. A total of 100 μg of cytosolic proteins or 10 μg of mitochondrial proteins was used for immunoblotting of phospho-AKT1 (p-AKT1). β-Tubulin served as a marker for cytosolic proteins, and VDAC (voltage dependent anionic channel) was used as a marker for mitochondrial subfractionation. U, untreated cells grown in regular growth media. B, Cardiomyocytes were subjected to ischemia and reperfusion. Then, p-AKT1 (red) and mitochondria (green) were analyzed with fluorescence microscopy. The merged immunofluorescence images confirmed the activation of p-AKT1 in mitochondria. C, The effect of mitochondrial Akt1 signaling on Caspase 3 activation. Cardiomyocytes were transduced with either the mitochondria-targeting constitutively active AKT1 (mAKT), mdnAKT, or the control viruses. Caspase 3 activation was analyzed by immunostaining with antibodies against activated Caspase 3. The bar graph represented the percentage of cardiomyocytes stained positive for activated Caspase 3 in each group. A representative immunofluorescence photo is shown on the right panel.

The effect of mitochondrial AKT1 activation on cell death during simulated I/R

To define how activation of AKT1 in mitochondria modulates mitochondrial function, we first studied the occurrence of cell death. Apoptosis and necrosis of cardiac muscle cells are pathological features of myocardial injury during I/R. Specific activation of mitochondrial AKT1 was achieved by using a mitochondria-targeting constitutively active AKT1, and inhibition of mitochondrial AKT1 was achieved by using a mdnAKT as we previously reported (9). Activation of mitochondrial AKT1 inhibited both apoptosis and necrosis of cardiomyocytes induced by I/R (P < .01) (Figure 2A). Interestingly, the mdnAKT did not alter I/R-induced necrosis (P > .3). However, it significantly increased basal apoptosis and further increased apoptosis during I/R. These results indicate that mitochondrial AKT1 signaling protected against cell death in this experimental model. We had previously reported that overexpression of the mitochondria-targeting constitutively active AKT1 in primary cardiomyocytes did not alter cytosolic or nuclear AKT activity (12). Therefore, the endogenous AKT signaling was not altered in this experimental design.

Figure 2.

The effect of mitochondrial AKT1 signaling on cardiomyocytes survival and mitochondrial function during simulated I/R injury. Each figure represents data summarized from 3–5 independent experiments. A, Necrosis and apoptosis of cardiomyocytes during I/R. Cardiomyocytes were transduced with either the mitochondria-targeting constitutively active AKT1 (mAKT), mdnAKT, or the control viruses. Necrosis and apoptosis were analyzed with MitoTracker/Annexin V double labeling. B, Mitochondrial AKT1 signaling and ΔΨm depolarization during I/R. ΔΨm was analyzed with JC-1 staining and FACS. GFP served as a control for mAKT and RFP as control for mdnAKT. The ratio between green and red fluorescence for control cells was set as 1. C, The effect of AKT1 signaling on mitochondria ROS during I/R. Mitochondrial superoxide production was measured by MitoSOX and FACS. mAKT had no significant effect on basal mitochondrial superoxide but inhibited I/R-induced mitochondrial superoxide production (P < .01). mdnAKT significantly increased both basal and I/R- induced superoxide (P < .01). D, Mitochondrial AKT1 inhibited angiotensin II-induced mitochondrial superoxide production. The time-course effect of angiotensin-II (ATII) on cardiomyocyte superoxide was attenuated by mAKT and exacerbated by mdnAKT (P < .01). The fluorescence intensity of control cells was set as 1. E, Mitochondrial AKT1 regulated intracellular calcium overload in cardiomyocytes during I/R. Intracellular calcium was analyzed with fura 3-AM and FACS. mAKT inhibited calcium rise and mdnAKT exacerbated calcium overload (P < .01). GFP served as control for mAKT and RFP as control for mdnAKT. F, Mitochondrial AKT1 regulated mitochondrial calcium overload in cardiomyocytes during I/R. Mitochondrial calcium sequestration was detected with and rhod-2 and FACS. I/R-induced calcium influx was alleviated by mAKT and aggravated by mdnAKT (P < .01). *, P < .01, vs normal control; #, P < .05, vs control/GFP with I/R; &, P < .01, vs control/RFP; $, P < .05, vs control/RFP with I/R. Ctrl, control.

The effect of mitochondrial Akt1 activation on ΔΨm, calcium loading, and oxidative stress during simulated I/R, basal redox homeostasis

Loss of cross-membrane electrochemical gradient (ΔΨm), increased oxidative stress, and increased calcium overload are 3 critical factors that had been mechanistically linked to the induction of cardiomyocyte apoptosis during I/R injury. JC-1 is a mitochondrial dye that exhibits potential-dependent accumulation in mitochondria, which allows detection of mitochondrial ΔΨm by measuring a potential-sensitive color shift. I/R reduced ΔΨm-induced oxidative stress and increased intracellular and mitochondria calcium levels (Figure 2, B, C, E, and F). Inhibition of mitochondrial AKT reduced basal ΔΨm, increased basal oxidative stress, and increased basal mitochondrial calcium influx. The regulatory effect of AKT1 signaling on ΔΨm during I/R was similar to its effect on apoptosis.

Superoxide production was detected with MitoSOX red, which specifically targets mitochondria in live cells and produces red fluorescence after reacting with superoxide. Activation of mitochondrial AKT1 had minimal effect on basal mitochondrial superoxide but significantly inhibited I/R-induced mitochondrial superoxide production. On the other hand, inhibition of mitochondrial AKT significantly increased both basal and I/R-induced mitochondrial superoxide generation, suggesting that mitochondrial AKT signaling is essential for the maintenance of basal redox hemostasis and the prevention of mitochondrial oxidative stress during I/R. To further determine whether mitochondrial AKT1 signaling modulated oxidative stress beyond I/R, angiotensin II was used as a stress inducer (Figure 2D). Accumulation of superoxides was suppressed by the active mitochondrial AKT1 and exacerbated by the mdnAKT.

The intracellular Ca²⁺ concentration and mitochondrial calcium sequestration were detected with Fura 3-AM and [2-[2-[2-(bis(carboxymethyl)amino)-5-methyl-phenoxy]ethoxy]-4-[3,6-bis(dimethylamino)xanthen-9-ylidene]-1-cyclohexa-2,5-dienylidene]-bis(carboxymethyl)ammonium (Rhod-2), respectively (Figure 2, E and F). I/R induced a 4.84 ± 0.75-fold increase in fura 3-AM fluorescence and a 33.3 ± 2.6% increase in Rhod-2 fluorescence. I/R-induced calcium overload was alleviated by the active mitochondrial AKT1 and exaggerated by the dominant-negative mitochondrial AKT. Although basal intracellular Ca²⁺ was not significantly affected by mdnAKT, basal mitochondrial Ca²⁺ was increased by mdnAKT. This suggests that mitochondrial AKT1 has a role in regulating basal mitochondrial Ca²⁺ handling. Taken together, these data suggests that activation of mitochondrial AKT1 signaling reduced oxidative stress, prevented cross-membrane electrochemical gradient alterations, and stopped calcium overload in cardiac mitochondria.

Mitochondrial Akt1 enhanced the efficiency of O₂ respiration in cardiomyocytes

To identify the direct effect of mitochondrial AKT1 on O₂ respiration, we performed a Seahorse Bioscience XF24 extracellular flux analysis in cardiomyocytes transduced with various viral constructs (Figure 3). Activation of mitochondrial AKT1 reduced the OCRs for basal mitochondrial respiration, whereas inhibition of mitochondrial AKT1 increased the basal OCR in cardiomyocytes (Figure 3A). Mitochondrial AKT1 also significantly modulated the oligomycin-suppressible ATP turnover, proton leak, and maximal respiration capacity in cardiomyocytes (Figure 3A). During I/R injury, spare respiration increased, whereas activation of mitochondrial AKT1 inhibited spare respiration rise (Figure 3B). To verify that mitochondrial respiration was modulated independently of cytosolic metabolic changes, we isolated intact mitochondria from cardiomyocytes and measured respiration with the Seahorse analyzer (Figure 3, C and D). The OCR was evaluated in 2 sets of isolated mitochondria, using succinate as the substrate. The basal respiration did not differ in mitochondria with activated AKT1 in comparison with the control. However, the maximal respiration was lower in the AKT1-activated mitochondria compared with the control. Inhibition of mitochondrial AKT significantly increased the OCR (Figure 3C). The effects of insulin on the OCR were similar to mitochondrial AKT1 activation in isolated mitochondria (Figure 3D). The data confirmed the effect of mitochondrial AKT1 signaling, independent of cytosolic factors, on cardiomyocyte respiration. Because respiration can be modulated by the abundance of mitochondria in the cells, we analyzed the mitochondrial content in cardiomyocytes (Figure 3E). Although the mitochondrial content was slightly increased by the activation of mitochondrial Akt1, it did not explain the reduced OCRs. Thus, we concluded that the effect of AKT1 on mitochondrial function was unrelated to the changes in mitochondria abundance.

Figure 3.

The effect of mitochondrial AKT1 signaling on glycolytic respiration and ATP production in cardiomyocytes. A, OCR in cardiomyocytes. Measurements in a Seahorse XF24 analyzer were carried out in cardiomyocytes transduced with mitochondrial activated AKT1 (mAKT), mdnAKT, or control viruses. The data represented mean ± SD of quintuplicate measurements. FCCP is carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone. B, Mitochondrial AKT1 inhibited spare respiration spikes during I/R. OCR was measured in an identical set of cells with or without I/R treatment. C, OCR in isolated mitochondria. Mitochondria were isolated from cardiomyocytes transduced with mAKT, mdnAKT, or control viruses. Succinate was used as the substrate in Seahorse XF24 measurements to assess glycolytic respiration. The data represented mean ± SD of quintuplicate measurements. A, ADP. B, Oligomycin. C, FCCP. D, Antimycin A. D, Insulin effect on mitochondrial respiration. Cardiomyocytes were stimulated with insulin (10⁻⁷M) or vehicles and mitochondria were isolated for Seahorse analysis using succinate as substrate. The data represented mean ± SD of quintuplicate measurements. Insulin stimulation led to lower OCR. mAKT represents mitochondria from cardiomyocytes with mAKT. E, Mitochondrial contents in cardiomyocytes. The abundance of mitochondria was determined by real-time PCR of the copy number of 12S and 18S rDNA. This was used for the ratio of mitochondria/nuclear DNA in these cells. The index was increased slightly in the cardiomyocytes with mAKT (P = .049). F, ECAR in cardiomyocytes. These data represent glycolysis-dependent respiration. Measurements in a Seahorse XF24 analyzer were carried out in cardiomyocytes transduced with mAKT, mdnAKT, or control viruses. The data represented the mean ± SD of quintuplicate measurements. P < .001. G, Mitochondrial AKT1 effect on glucose uptake. 2-NBDG glucose uptake was measured as described in Research Design and Methods. As a positive control, insulin (10⁻⁷M) increased glucose uptake in cardiomyocytes. Mitochondrial AKT1 positively regulated glucose uptake. H, Mitochondrial AKT1 increased ATP production. The ATP level in the cells with mAKT increased 2.05 ± 0.30-fold at basal, and insulin stimulation further increased uptake (P < .01). Basal and insulin-induced ATP production was lower in cells transfected with mdnAKT (P < .001). For panels B, E, F, G, and H: *, P < .01, vs control; #, P < .01, vs control/GFP/RFP; &, P < .01, mAKT or mdnAKT vs control/GFP/RFP; $, P < .05, mdnAKT vs control/RFP. Ctrl, control; WT, wild type.

In addition to mitochondrial signaling, we also investigated the ECAR, which represents the result of glycolysis. Mitochondrial AKT1 signaling positively regulated the ECAR in cardiomyocytes (Figure 3F), which raised the possibility that mitochondrial AKT1 signaling might have modulated glucose transport into cardiomyocytes. To address this question, we measured 2-NBDG-glucose uptake, and the results showed that mitochondrial AKT1 activation increased glucose uptake in cardiomyocytes (Figure 3G). The effect of mitochondrial AKT1 on glucose uptake was on par with insulin stimulation. Insulin could not further increase glucose uptake in the cells transduced with active AKT1. In the last series of mitochondrial function analysis, we analyzed ATP production (Figure 3H). As expected, activation of mitochondrial AKT1 increased ATP, whereas inhibition of mitochondrial AKT reduced ATP in cardiomyocytes. Insulin stimulation further enhanced ATP production when mitochondrial AKT1 was activated, suggesting synergistic energy production. These results suggested that mitochondrial AKT1 signaling increased the efficiency of cell respiration, by increasing glucose uptake and ATP production while minimizing oxygen consumption and reactive oxygen species (ROS) production.

Protein targets of mitochondrial AKT1 signaling

To identify the targets of AKT1 signaling in mitochondria, we overexpressed His-tagged mitochondria-targeting active AKT1 in the cells and then cross-linked AKT1, using crosslinker, dithiobissuccinimidyl propionate, with its interactive mitochondrial proteins in their native forms. The protein complexes were isolated, 2-step purified, resolved with SDS-PAGE, and trypsinized for protein identification with MALDI TOF/TOF (Figure 4A). With this approach, we have identified 15 AKT1-interactive proteins (Figure 4B). Some of these proteins are recognized mitochondrial proteins, whereas others are known targets of phosphatidylinositol-4,5-bisphosphate 3-kinase/AKT signaling (Supplemental Data, Table 1). Among these proteins, PDHX belongs to the PDC. PDC is the gate keeper for the mitochondrial glycolytic pathway, linking pyruvate to the TCA cycle and oxidative phosphorylation. Additional experiments confirmed that the constitutively active AKT1 and the dominant-negative AKT that we used for this study are indeed associated with PDHX in cardiomyocytes (Figure 4C). To verify an interaction of AKT1 with PDC subunits in mitochondria, coimmunoprecipitation experiments were performed with cardiomyocyte mitochondria prep. The results showed coimmunoprecipitation of AKT1 with PDHX or other PDC subunits (Figure 4, D and E). Insulin stimulation acutely enhanced association of AKT1 with PDHX (Figure 4E). To explore whether AKT1 can modulate PDC enzymatic activities, we analyzed PDC activities in either cardiomyocytes transduced with mitochondria-targeting active AKT1 or dominant-negative AKT. PDC activity was increased by the active AKT1 and inhibited by the dominant-negative AKT (Figure 4F). Insulin stimulation could not further increase PDC activity in the cells transduced with active AKT1. The dominant-negative AKT prevented insulin from reaching a maximal stimulatory response. These findings suggested that mitochondrial AKT1 signaling is a key regulator of PDC activity in cardiomyocytes.

Figure 4.

Potential substrates of AKT1 signaling in mitochondria. A, Schematic illustration of the proteomic approach to identify the mitochondrial proteins interacting with AKT1. B, Mitochondrial proteins associated with AKT1. C, Association of mitochondria-targeting constitutively active AKT1 (mAKT)/mitochondria-targeting dominant negative (mdnAKT) with PDHX. mAKT or mdnAKT was, respectively, overexpressed in primary cardiomyocytes, and proteins were extracted for these experiments. For the association of mAKT-His with PDHX, the protein complexes were captured with NAP-Sepharose column, and the proteins were eluted and immunoblotted with anti-His antibodies or anti-PDHX antibodies (upper panel). For the association of mdnAKT-His with PDHX, the protein complexes were immunoprecipitated with anti-His antibodies or anti-PDHX antibodies and then immunoblotted with anti-PDHX antibodies (lower panel). D, Coimmunoprecipitation of AKT1 and PDHX. Mitochondria were isolated from cardiomyocytes, and the proteins were extracted and immunoprecipitated with anti-AKT1 antibody or anti-PDHX antibody. Then, they were, respectively, blotted with anti-AKT1 antibody and anti-PDHX antibody. IP, immunoprecipitation. WB, Western blotting. E, AKT1 coimmunoprecipitated with PDC subunits in mitochondria. Mitochondrial proteins were immunoprecipitated with anti-antibodies against PDC subunits E1α, E1β, E2, and E3, and blotted with antibody against Akt1. F, Insulin stimulation formed the mitochondrial AKT1-PDC complex. Cardiomyocytes were first treated with insulin or vehicles. Then, the mitochondria were isolated for the coimmunoprecipitation experiment. G, The effect of mitochondrial AKT1 on pyruvate dehydrogenase activity in cardiomyocytes. Activation of mitochondrial AKT1 (mAKT) increased the basal PDH activity by 33.4 ± 6.2% (P < .02), and insulin treatment caused further stimulation (P < .01). Inhibition of mitochondrial AKT1 (mdnAKT) suppressed basal activity by 39.7 ± 4.1% (P < .01) and the effect of insulin by 50%. *P < .01, control vs insulin; #, P < .01, mAKT vs control/GFP; &, P < .01, mdnAKT vs control/GFP; $, P < .01 mdnAKT vs control/GFP with insulin.

Discussion

Mitochondria play a key role in maintaining normal function of cardiac muscle and during the development of cardiomyopathy. Our current study suggests that mitochondrial AKT1 signaling positively regulates glycolytic ATP production by increasing glucose uptake and enhancing the efficiency of mitochondrial respiration in cardiomyocytes. Because the endogenous cytosolic and nuclear AKT activities were not altered in this experimental model (8), the cardioprotective effects of mitochondrial AKT1 are likely mediated through signaling proteins in the mitochondrial proteome that regulate a coordinated signaling network consisting of interlinking biochemical pathways (Supplemental Figure 1).

The mitochondrial electron transport chain is the principal machinery for aerobic ATP production through oxidative phosphorylation and the major source of ROS generation. Mitochondria are also the target for the damaging effect of ROS and excessive Ca²⁺ (13, 14). The electron transport chain is largely inhibited during ischemia due to lack of oxygen, the terminal acceptor of electrons. The accumulation of electrons alters the ΔΨm, which serves as a driving force for the electron leakage and superoxide generation from complex I and III during reperfusion (14). Eventually, the excessive ROS increased mitochondrial Ca²⁺ and altered intracellular pH. This could have induced the mitochondrial permeability transition pore to open and consequently cause a loss of ΔΨm (15–17). Our findings provided compelling evidence that mitochondrial AKT1 inhibited I/R-induced superoxide production, cytoplasmic and mitochondrial Ca²⁺ overload as well as prevented a loss of ΔΨm. In the cardiomyocytes that overexpressed mdnAKT, we found an increased rate of superoxide production, ΔΨm depolarization, and mitochondrial calcium influx under basal (nonstressed) conditions. This suggests a role of mitochondrial AKT1 in the maintenance of basal mitochondrial redox and Ca²⁺ homeostasis.

AKT has been identified as one of the common kinases that is critical for cardioprotection against I/R injury (18, 19). Our laboratory had shown that insulin induced translocation of activated AKT1 into mitochondria stimulated oxidative phosphorylation complex V's activity (9). The current study showed that mitochondrial AKT1 significantly increased both basal and insulin-stimulated ATP production. Moreover, mitochondrial AKT1 signaling enhanced glucose uptake and stimulated PDC activity, which is the regulatory enzyme that connects pyruvate to the TCA cycle and (Supplemental (Data, Table 1). The effect of insulin on PDC was not augmented by mitochondrial AKT1, but the effect of insulin on ATP production was significantly augmented (Figure 3). This suggests that insulin has additional effects on ATP production through pathways independent of mitochondrial AKT1. However, insulin regulation of ATP production required mitochondrial AKT1 signaling, because inhibition of mitochondrial AKT suppressed insulin response (Figure 3). Oxidative stress may inhibit oxidative phosphorylation and ATP production. It is possible that inhibition of mitochondrial AKT signaling attenuated bioenergetics because of the altered redox state and uncoupled respiration. The stimulatory effect of mitochondrial AKT1 on glucose uptake was somewhat unexpected, albeit it fit the results of the Seahorse analyzer data. Although it is tempting to speculate potential implications of mitochondrial signaling on glucose homeostasis, how it modulates glucose uptake in vivo waits to be further studied. The mdnAKT construct provided an effective tool to study the role of insulin mediated through mitochondrial AKT1. A limitation in our experimental design is the possibility that the observed regulatory events by the constitutively active AKT1 may be different from what occurs with rapid phosphorylation of physiologic protein levels of AKT in vivo.

Evidence has emerged that energy deficiency can be a cause of heart failure (20). Cardiac metabolic regulatory events may also be adaptive in disease states (21). The multiple targets of AKT1 signaling in mitochondria suggest that AKT1 can remodel the mitochondrial proteome and exert coordinated actions to maintain normal mitochondrial function and help mitochondria adapt during myocardial injury (Supplemental Table 1). We have only just begun to scratch the surface of AKT1 signaling in mitochondria. PDC is a gate keeper for glycolytic oxidative phosphorylation in mitochondria and plays a role in the regulation of myocardial metabolism and function. Inhibition of PDC in the heart shifted mitochondrial metabolism from glycolysis to fatty acid oxidation, which is similar to the insulin-resistance state (21). Reversing PDC deficiency in the hyperthyroid heart reduced ventricular hypertrophy and improved cardiac output (22). Knockout of the cardiac pyruvate dehydrogenase α-subunit caused left ventricular hypertrophy and systolic dysfunction (23). Our coimmunoprecipitation experiment suggested that AKT1 preferentially bound to the E3 subunit (Figure 4). The exact interactive domains on PDC and AKT1 cannot be determined with these experiments. However, further studies are underway in our laboratory to investigate the mechanisms of their binding and interaction.

In conclusion, activation of AKT1 signaling in mitochondria increased glucose uptake, enhanced respiration efficiency through better respiration coupling, and increased ATP production in the cardiomyocytes. Mitochondrial AKT1 signaling plays a protective role against apoptosis and necrosis in nonstressed cardiomyocytes and during I/R injury. We have identified 15 novel targets of AKT1 signaling in mitochondria. Disrupting AKT1's interaction with its target substrates in mitochondria impaired mitochondrial function and amplified the occurrence of cardiomyocytes death during stress. These new findings raise the possibility to develop future therapeutic interventions by enhancing mitochondrial AKT1 signaling to its target substrates.