The transcriptional coactivator PGC-1α is essential for maximal and efficient cardiac mitochondrial fatty acid oxidation and lipid homeostasis

John J. Lehman, Sihem Boudina, Natasha Hausler Banke, Nandakumar Sambandam, Xianlin Han, Deanna M. Young, Teresa C. Leone, Richard W. Gross, E. Douglas Lewandowski, E. Dale Abel, Daniel P. Kelly

Abstract

High-capacity mitochondrial ATP production is essential for normal function of the adult heart, and evidence is emerging that mitochondrial derangements occur in common myocardial diseases. Previous overexpression studies have shown that the inducible transcriptional coactivator peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α is capable of activating postnatal cardiac myocyte mitochondrial biogenesis. Recently, we generated mice deficient in PGC-1α (PGC-1α−/− mice), which survive with modestly blunted postnatal cardiac growth. To determine if PGC-1α is essential for normal cardiac energy metabolic capacity, mitochondrial function experiments were performed on saponin-permeabilized myocardial fibers from PGC-1α−/− mice. These experiments demonstrated reduced maximal (state 3) palmitoyl-l-carnitine respiration and increased maximal (state 3) pyruvate respiration in PGC-1α−/− mice compared with PGC-1α+/+ controls. ATP synthesis rates obtained during maximal (state 3) respiration in permeabilized myocardial fibers were reduced for PGC-1α−/− mice, whereas ATP produced per oxygen consumed (ATP/O), a measure of metabolic efficiency, was decreased by 58% for PGC-1α−/− fibers. Ex vivo isolated working heart experiments demonstrated that PGC-1α−/− mice exhibited lower cardiac power, reduced palmitate oxidation, and increased reliance on glucose oxidation, with the latter likely a compensatory response. 13C NMR revealed that hearts from PGC-1α−/− mice exhibited a limited capacity to recruit triglyceride as a source for lipid oxidation during β-adrenergic challenge. Consistent with reduced mitochondrial fatty acid oxidative enzyme gene expression, the total triglyceride content was greater in hearts of PGC-1α−/− mice relative to PGC-1α+/+ following a fast. Overall, these results demonstrate that PGC-1α is essential for the maintenance of maximal, efficient cardiac mitochondrial fatty acid oxidation, ATP synthesis, and myocardial lipid homeostasis.

  • nuclear receptors
  • ATP
  • cardiac energetics
  • heart failure
  • left ventricular hypertrophy
  • peroxisome proliferator-activated receptor-γ coactivator-1α

a flurry of recent studies has implicated the transcriptional coactivator peroxisome proliferator-activated receptor (PPAR)-γ coactivator (PGC)-1α as a regulator of genes responsible for the maintenance of high-level cardiac mitochondrial oxidative capacity and energy transduction in the postnatal heart. Originally cloned as a cold-inducible, transcriptional coactivator of PPAR-γ in brown adipose tissue (44), PGC-1α is also cardiac enriched (28, 44). PGC-1α is induced during postnatal development and with fasting (33) and exercise (20), conditions known to increase cardiac mitochondrial energy transduction via the fatty acid β-oxidation pathway. Forced expression of PGC-1α in cardiac myocytes in culture or in hearts of transgenic mice increases mitochondrial numbers and stimulates respiration (33, 47).

To accomplish its pleiotropic effects on mitochondrial biogenesis and respiratory function, PGC-1α boosts the activity of downstream transcriptional regulatory circuits, including those involving nuclear respiratory factor (NRF)-1 and NRF-2 (60). PGC-1α activation of NRF-1 and NRF-2 expression, as well as its direct coactivation of NRF-1, induces the expression of mitochondrial transcription factor A, thus facilitating mitochondrial DNA replication and coordinating the expression of mitochondrial and nuclear genes central to mitochondrial biogenesis (15, 16, 31, 51, 58, 60). In addition, PGC-1α coactivates the estrogen-related receptor (ERR) (23) and PPAR (57), nuclear receptor transcription factors that regulate genes central for the cellular uptake and mitochondrial β-oxidation of fatty acids. PPAR-α regulates virtually every step of cardiac fatty acid utilization (12). Given that ERR-α in cardiac myocytes activates the expression of many known PPAR-α and NRF target genes (25), the ability of PGC-1α to coactivate ERR-α (24, 52) allows for further amplification of cardiac mitochondrial fatty acid oxidation (23). In addition, pyruvate dehydrogenase kinase-4, a negative regulator of glucose oxidation, was recently identified as a PGC-1α/ERR-α target in skeletal muscle, providing a mechanism whereby PGC-1α exerts reciprocal inhibition of glucose catabolism while increasing fatty acid oxidation pathways (59).

Abnormalities in energy transduction mediated by reduced PGC-1α activity have been implicated in the evolution of important disease states, including diabetes mellitus, pathologic cardiac hypertrophy, and heart failure. In human Type 2 diabetes mellitus, PGC-1α and its target genes involved in oxidative phosphorylation are coordinately downregulated in skeletal muscle (38). PGC-1α expression is reduced in skeletal muscle of Zucker diabetic fatty rats and is restored by troglitazone (26). Activation of cardiac Cdk9 in mice represses PGC-1α and confers a predisposition to heart failure (50). PGC-1α expression is reduced in pressure overload-induced cardiac hypertrophy (34), and mice with generalized loss of PGC-1α develop accelerated profound cardiac dysfunction 2 mo following pressure overload by transverse aortic constriction (2). The level of cardiac PGC-1α gene expression is correlated with mitochondrial oxidative capacity in both healthy rat hearts and failing rat hearts subjected to chronic pressure overload (17).

This study employed PGC-1α-deficient mice to define the role of PGC-1α in the coordinate regulation of both cardiac fuel preference for mitochondrial fatty acid oxidation and efficient mitochondrial ATP-producing capacity in vivo. Recently, we and others have generated mice with a generalized deficiency of PGC-1α (PGC-1α−/− mice), which survive with modestly blunted postnatal cardiac growth (1, 35). Mitochondrial numbers and respiratory capacity are diminished in skeletal muscle of PGC-1α−/− mice, leading to reduced muscle performance and exercise capacity (35). The PGC-1α−/− line developed by our group does not exhibit reduced cardiac function at rest (35), in contrast with an independently generated model of global PGC-1α gene disruption in which PGC-1α-null mice demonstrated a significant reduction in fractional shortening by 7–8 mo of age (1). Our results demonstrate an essential role for PGC-1α in the maintenance of maximal and efficient cardiac mitochondrial fatty acid oxidation and ATP synthesis, implicating reduced mitochondrial fatty acid oxidative capacity and inefficiency in mitochondrial energy transduction as key mechanisms of pathology in PGC-1α-deficient states.

MATERIALS AND METHODS

Animal experiments.

All animal experiments conducted were approved by the Animal Studies Committee of Washington University School of Medicine. Mice were kept in separate cages with identical light-dark cycles and free access to water. For 24-h fasting experiments, standard chow was removed, and animals were housed in separate cages with wood chip bedding.

Mitochondrial function experiments.

The mitochondrial respiratory function of myocardial left ventricular (LV) fibers from 3.5-mo-old female mice was measured in the presence of saponin, which perforates the sarcolemma while leaving the mitochondria morphologically and functionally intact (49). The respiration of saponin-permeabilized LV fibers was measured at 25°C using an optical probe (Oxygen FOXY Probe, Ocean Optics, Dunedin, FL) in a 2-ml sealed, continuously stirred respiration chamber. The previously described respiration buffer (49) contained the following (in mM): 125 KCl, 20 HEPES, 3 Mg-acetate, 5 KH2PO4, 0.4 EGTA, and 0.3 DTT, pH 7.1 at 25°C, with 2 mg/ml BSA added. For palmitoyl-l-carnitine (PC) respiration, the respiration buffer also contained 20 μM PC and 5 mM malate. For pyruvate respiration, the respiration buffer contained 10 mM pyruvate and 5 mM malate. Following the measurement of basal respiration, state 3 (maximal ADP-stimulated) respiration was determined by exposing fibers to 1 mM ADP. For each sample, the integrity of the outer mitochondrial membrane was assessed by adding 8 μM exogenous cytochrome c to ADP-stimulated mitochondria. Postoligomycin (uncoupled) respiration was evaluated 10 min following the addition of oligomycin (1 μg/ml) to inhibit ATP synthase. The solubility of oxygen in the respiration buffer at 25°C was taken as 246.87 nmol O2/ml. Respiration rates were expressed as nanomoles of O2 per minute per milligram dry weight of fibers. State 3 (maximal ADP-stimulated) ATP synthesis rates were assessed employing previously described methods (43). The ATP produced per oxygen consumed (ATP/O) ratio reflects the ratio of state 3 ATP synthesis rates to state 3 O2 consumption. State 3 ATP synthesis was initiated by the addition of 1 mM ADP, with serial removal of 10-μl aliquots to define the rate of maximal ATP synthesis. These serial aliquots were quenched in DMSO and frozen prior to quantitation of ATP content by bioluminescence using an ATP-monitoring reagent (Promega Enlighten FF2000) with known quantities of ATP measured under the same conditions. We have previously found in saponin-permeabilized cardiac fibers that extramitochondrial ATP generation and ATP hydrolysis is minimal (9), as assessed using EDTA to inhibit ATP hydrolysis (by ATPases), iodoacetate to inhibit hexokinase and thus ATP generation via glycolysis, and di(adenosine-5′)pentaphosphate to inhibit adenylate kinase and thus ATP generation in the cytosol. Consistent with this prior work, residual ATP generation by extramitochondrial pathways in state 4 respiration (in the presence of oligomycin to inhibit mitochondrial ATP synthesis) was measured for a subset of samples in the present study and again found to be minimal in saponin-permeabilized cardiac fibers (data not shown).

Respiratory function was also assessed for mitochondria isolated from combined left and right cardiac ventricles using a previously described isolation protocol (7). Hearts from 3- to 4-mo-old sex-matched mice were pooled per sample. The final washed pellet was suspended in isolation medium (pH 7.2) containing (in mM) 300 sucrose, 10 Na-HEPES, and 0.2 EDTA. The protein concentration of the mitochondrial isolate was determined using Micro BCA (Pierce, Rockford, IL). The respiration of the mitochondrial isolate containing 0.5 mg protein was measured at 30°C using an optical probe (Oxygen FOXY Probe, Ocean Optics) in a 2-ml sealed, continuously stirred respiration chamber. The respiration buffer employed for the saponin-permeabilized fiber respiration analysis (described above) was also used for the isolated mitochondrial respiration experiments except that for succinate respiration, 5 mM succinate was present in addition to 10 μM rotenone (to inhibit complex I). Following the measurement of basal respiration, state 3 (maximal ADP-stimulated) respiration was determined by exposing mitochondria to 350 μM ADP. Postoligomycin (uncoupled) respiration was evaluated following the addition of oligomycin (1 μg/ml) to inhibit ATP synthase. The solubility of oxygen in the respiration buffer at 30°C was taken as 230 nmol O2/ml. Respiration rates were expressed as nanomoles of O2 per minute per milligram of mitochondrial protein.

Isolated working mouse heart perfusion.

The isolated mouse working heart perfusion was based on previously described procedures (4). Adult sex-matched mice (7–8 mo old) were heparinized (100 units ip) 10 min prior to anesthesia. Animals were then deeply anesthetized with 5–10 mg pentobarbitol sodium (ip). Hearts were excised and placed in an ice-cold Krebs-Henseleit bicarbonate (KHB) solution (118 mM NaCl, 25 mM NaHCO3, 4.7 mM KCl, 1.2 mM KH2PO4, 2.5 mM CaCl2, and 5.0 mM glucose, and 10 μU/ml insulin; pH 7.4). Hearts were cannulated first via the aorta and perfused retrogradely by the Langendorff method. Following left atrial cannulation, perfusion was switched to the working heart perfusion with KHB solution containing 1.2 mM palmitate bound to 3% fatty acid-free BSA with a preload pressure of 11.5 mmHg and an afterload pressure of 50 mmHg for 60 min with oxygenated buffer solution. To determine palmitate and glucose oxidation rates, trace amounts of [3H]palmitate (0.1 μCi/ml) and [U-14C]glucose (0.1 μCi/ml) were used, respectively. Samples were collected every 10 min for 14CO2 trapped in 1 M hyamine hydroxide solution as a result of glucose oxidation and 3H2O released into the buffer due to palmitate oxidation, and the radioactivity was counted. Functional measurements, such as cardiac output, aortic flow, peak systolic pressure, and heart rate (HR), were acquired every 10 min for 10 s using inline flow probes (Transonic Systems), an MP100 system (AcqKnowledge, BioPac Systems), and a pressure transducer (TSD 104A, BioPac Systems). Cardiac hydraulic work (expressed in J·s−1·g wet wt−1) was calculated as the product of cardiac output and peak systolic pressure, normalized to heart wet weight. At the end of each perfusion, hearts were frozen immediately in liquid nitrogen. A small piece of heart tissue was also used for determining the dry-to-wet weight ratio.

RNA analysis.

Cardiac ventricles were isolated from 1- to 2-mo-old, littermate-matched PGC-1α+/+ and PGC-1α−/− mice. Gene expression comparisons were made with sex-matched littermates fed ad libitum. The protocols for RNA isolation and RT-PCR as well as mouse-specific primer and probe sets have been previously described (10, 33).

Electron microscopy.

In brief, LV tissue was fixed in modified Karnofsky's fixative as previously described (35). Thin sections were obtained and viewed with a Japan Electronic Optics Laboratory 1200 transmission electron microscope.

Electrospray ionization mass spectrometry of myocardial lipid extracts.

Electrospray ionization mass spectrometry (ESI/MS) analysis was performed as previously described (21). Briefly, mice were killed by inhalation of CO2 prior to tissue collection. The hearts were excised quickly and immersed in ice-cold diluted PBS buffer. After extraneous tissue and epicardial fat had been removed, each heart was quickly dried and immediately freeze clamped at the temperature of liquid nitrogen. Myocardial wafers (25 mg) were pulverized into fine powder with a stainless steel mortar and pestle, with a subsequent homogenization in 0.5 ml ice-cold LiCl solution (50 mM) using a Potter-Elvehjem tissue grinder for 2 min. A small volume of homogenate containing 2–5 mg protein was transferred to a glass test tube. Methanol and chloroform (2 ml each), as well as an additional volume of LiCl solution to make a final volume of 1.8 ml with a final LiCl concentration of 50 mM, were added to the test tube containing the heart homogenate for lipid extraction by the Bligh and Dyer procedure (6). The protein concentration of homogenates was then determined using a BCA protein assay kit (Pierce). At this point, internal standards including T17:1 triglyceride (TAG; 10 nmol/mg protein) were added to each homogenate based on the experimentally determined protein concentrations for normalization to the protein content. Next, the extraction mixture was centrifuged, and the chloroform layer was carefully removed and saved. To the MeOH/aqueous layer of each test tube, an additional 2 ml chloroform was added; the mixture was vortexed and centrifuged, and the chloroform layers were combined and subsequently dried under a nitrogen stream. Each residue was then resuspended in 4 ml chloroform-methanol [1:1 (vol/vol)] and reextracted against 1.8 ml of 20 mM LiCl aqueous solution, and the extract was dried as described above. Each residue was resuspended in 1 ml chloroform and filtered with a PFTE syringe filter into a 5-ml glass centrifuge tube (this step was repeated twice). The chloroform filtrate was subsequently dried under a nitrogen stream and resuspended with a volume of 500 μl/mg protein in 1:1 (vol/vol) chloroform-methanol, and lipid extracts were finally flushed with nitrogen, capped, and stored at −20°C for ESI/MS analyses (typically within 1 wk). ESI/MS analyses of lipids directly from lipid extracts of biological samples were performed using a triple-quadrupole mass spectrometer (ThermoFinnigan TSQ Quantum, San Jose, CA) operating under Xcalibur software as described in detail previously (22). Briefly, each prepared lipid solution was diluted ∼50-fold with 1:1 (vol/vol) chloroform-methanol just prior to infusion for anionic lipid analyses, which contained <10 pmol/μl total lipids. A small amount of LiOH (50 nmol/mg protein) was added to the diluted lipid solution just before other lipid analyses were performed in both negative and positive ion modes. The diluted lipid extract solution was directly infused into the ESI source at a flow rate of 2 μl/min with a syringe pump using an orthogonal injection.

Triacylglyceride turnover and isolated heart protocols for 13C NMR.

Twelve-week-old animals were heparinized (50 U/10 g ip) and anesthetized with ketamine (80 mg/kg ip) plus xylazine (12 mg/kg ip). Hearts were excised and perfused in retrograde fashion with modified Krebs-Henseleit buffer (118.5 mM NaCl, 4.7 mM KCl, 1.5 mM CaCl2, 1.2 mM MgSO4, and 1.2 mM KH2PO4) equilibrated with 95% O2-5% CO2 and containing 0.4 mM unlabeled palmitate-fatty acid-free albumin complex (3:1 molar ratio) and 10 mM glucose. A water-filled latex balloon was fitted into the LV and set to a diastolic pressure of 5 mmHg. The LV developed pressure (LVDP) and HR were continuously recorded with a pressure transducer and digital recording system (Powerlab, AD Instruments, Colorado Springs, CO). The rate-pressure product (RPP) was calculated as the product of HR and developed pressure. Temperature was maintained at 37°C. Hearts were perfused in a 14.1-T NMR magnet at baseline workload or with adrenergic challenge (0.1 μM isoproterenol). At the start of each protocol, the hearts continued to be supplied buffer containing unlabeled palmitate-albumin complex and glucose for 10 min to ensure metabolic equilibrium and allow for the collection of 13C NMR background signals of naturally abundant 13C (1.1%). At the start of each enrichment protocol, the perfusate was then switched to a supply of 13C-enriched buffer containing 0.4 mM [2,4,6,8,10,12,14,16-13C8] palmitate (Isotec, Miamisburg, OH) plus 10 mM unlabeled glucose. Perfusion with 13C-enriched media continued for 20 min at baseline workload or for 10 min with adrenergic challenge. Sequential 13C NMR spectra (2-min time blocks) were collected throughout these perfusion periods for the determination of metabolic flux. Separate perfusions were performed under each protocol for either measurements of oxidative flux or TAG turnover due to the limited tissue sample size of the mouse hearts for assays. Additional hearts were perfused for 120 min to ensure the stability of TAG turnover and content over time. At the end of each protocol, hearts were freeze clamped in liquid nitrogen-cooled clamps. 13C enrichment of TAG in the intact heart was monitored from the NMR signal at 30.5 ppm from the TAG methylene groups, and TAG turnover was calculated from total TAG content and enrichment over time (42, 48).

NMR spectroscopy and tissue chemistry.

Measurements of oxidative flux and TAG turnover were performed on intact beating hearts situated within a 10-mm NMR probe inside a 14.1-T NMR magnet. Sequential, proton-decoupled 13C NMR spectra were acquired (2 min each) with a natural 13C abundance correction using previously reported NMR methods (40, 42). Kinetic analysis of dynamic 13C spectra from intact beating hearts was performed as previously reported by our laboratory (40, 42, 61). In vitro 13C NMR was also performed on acid extracts of the myocardium at 14.1 T to determine the fractional enrichment of [2-13C]acetyl-CoA (37, 42). Tissue concentrations of glutamate, aspartate, citrate, malate, and α-ketoglutarate from perchloric acid extracts of frozen LVs were determined spectrophotometrically and fluorometrically using previously described assays (61). Lipid extracts were obtained from heart samples and triacylglerides quantified by colorimetric assay, as previously described (Wako Pure Chemical Industries) (42). TAG was isolated and saponified, and the fractional 13C enrichment of the fatty acids was assessed by MS analysis (Waters X-terra C18MS column, MS:scan m/z 100–600 Fragmentor 75V Negative ESI).

Statistical analysis.

Statistical comparisons were made using Student's t-test, with a statistically significant difference defined as a P value of <0.05. ANOVA with Newman-Keul's test was employed for multiple group comparisons. Error bars represent SEs.

RESULTS

Fibers from hearts of PGC-1α−/− mice exhibit decreased maximal capacity for mitochondrial fatty acid β-oxidation and ATP synthesis.

To characterize the mitochondrial functional phenotype of hearts from PGC-1α−/− mice in a manner that would detect potential differences in mitochondrial volume density or function, respiration (O2 consumption) experiments were performed on saponin-permeabilized myocardial fibers. This technique allowed for the selective use of different metabolic substrates to define the maximal respiratory capacity of specific mitochondrial oxidative pathways. Experiments were performed with LV fibers isolated from hearts of 3.5-mo-old mice using PC, pyruvate, or glutamate as substrates. Basal respiration was assessed, followed by maximal ADP-stimulated state 3 respiration. Finally, the proportion of uncoupled respiration was assessed by measuring oxygen consumption following the addition of the ATP synthase inhibitor oligomycin, allowing for the calculation of the respiratory control (RC) quotient (state 3 respiration/postoligomycin respiration). Maximal (state 3) PC respiration was reduced in cardiac muscle fibers from PGC-1α−/− mice (PGC-1α+/+ vs. PGC-1α−/−: 12.7 ± 1.0 vs. 9.2 ± 0.9 nmol O2·min−1·mg dry wt−1, P < 0.05; Fig. 1A). Conversely, state 3 pyruvate respiration in PGC-1α−/− cardiac fibers was increased (PGC-1α+/+ vs. PGC-1α−/−: 12.6 ± 1.6 vs. 17.4 ± 1.3 nmol O2·min−1·mg dry wt−1, P < 0.05). There were no differences in state 3 respiration of glutamate, a substrate converted to enter the tricarboxylic acid (TCA) cycle (Fig. 1A), although this does not exclude defects in upstream pathways including glucose oxidation or fatty acid β-oxidation. A reduction in the RC quotient of PGC-1α−/− fibers was noted for PC respiration only, suggesting that the proportion of respiration coupled to ATP synthesis relative to uncoupled respiration is reduced in PGC-1α−/− hearts in the presence of fatty acid substrate (Fig. 1A). Overall, these data demonstrate that the maximal capacity for mitochondrial fatty acid β-oxidation is decreased in the myocardium of PGC-1α−/− mice, with an increased maximal capacity for glucose oxidation, which is consistent with a shift away from the normal postnatal heart's preference for fatty acid oxidation.

Fig. 1.

Respiration and ATP synthesis rates for permeabilized myocardial fibers and isolated mitochondria. A: respiration of saponin-permeabilized left ventricular (LV) fibers from hearts of 3.5-mo-old female peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α+/+ and PGC-1α−/− mice was measured as described in materials and methods. The respiration buffer contained 20 μM palmitoyl-l-carnitine (PC) and 5 mM malate for PC respiration, 10 mM pyruvate and 5 mM malate for pyruvate respiration, and 5 mM glutamate and 2 mM malate for glutamate respiration. Following measurements of basal respiration, state 3 (maximal ADP-stimulated) respiration was determined by exposing fibers to 1 mM ADP, with the subsequent determination of postoligomycin (uncoupled) respiration. The respiratory control (RC) quotient represents the following ratio: (state 3 respiration/postoligomycin respiration). Results are means ± SE; n = 10 for PC and pyruvate experiments and 6 for glutamate experiments. *P < 0.05 and **P < 0.01 compared with corresponding PGC-1α+/+ values. B: maximal ATP synthesis rates and efficiency of ATP synthesis [ATP produced per oxygen consumed (ATP/O)] in permeabilized LV fibers prepared in parallel from the same hearts used for respiration analysis. ATP/O represents the following ratio: (state 3 ATP synthesis rate/state 3 respiration rate). Results are means ± SE; n ≥ 8 for PC and pyruvate experiments and 5 for glutamate experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with corresponding PGC-1α+/+ values. C: respiration of mitochondria isolated from combined left and right cardiac ventricles of 4-mo-old sex-matched PGC-1α+/+ and PGC-1α−/− mice was measured in buffer containing 5 mM succinate and 10 μM rotenone as described in materials and methods. Following the assessment of basal respiration, state 3 (maximal ADP-stimulated) respiration was determined by exposing mitochondria to 350 μM ADP with the subsequent determination of postoligomycin (uncoupled) respiration. The RC quotient represents the following ratio: (state 3 respiration/postoligomycin respiration). Results are means ± SE; n = 6. *P < 0.05 compared with corresponding PGC-1α+/+ values.

Additionally, maximal rates of ATP synthesis from ADP were assessed during state 3 respiration of saponin-permeabilized myocardial fibers using a fluorometric assay. For all three substrates tested (PC, pyruvate, or glutamate), maximal ATP synthesis rates were significantly reduced in PGC-1α−/− cardiac muscle fibers (Fig. 1B; pyruvate respiration, PGC-1α+/+ vs. PGC-1α−/−: 36.2 ± 3.7 vs. 23.0 ± 2.9 nmol·min−1·mg dry wt−1, P < 0.05). Thus, despite increased state 3 pyruvate respiration (i.e., O2 consumption) for PGC-1α−/− cardiac fibers, the maximal ATP synthesis rate was reduced with pyruvate respiration. The ratio of state 3 ATP synthesis to state 3 O2 consumption allowed for the determination of the efficiency measure ATP/O (Fig. 1B). ATP/O was decreased by 58% for PGC-1α−/− saponin-permeabilized myocardial fibers (pyruvate respiration, PGC-1α+/+ vs. PGC-1α−/−: 3.3 ± 0.4 vs. 1.4 ± 0.2 nmol·min−1·mg dry wt−1, P < 0.001). For PC and glutamate respiration, ATP/O was also significantly decreased for PGC-1α−/− cardiac fibers (Fig. 1B). Thus, for all substrates tested, maximal ATP synthetic rates were reduced and metabolic efficiency of ATP synthesis (ATP/O) was impaired in hearts of PGC-1α−/− mice.

Given that the saponin-permeabilized myocardial fiber respiration data suggested reduced mitochondrial coupling (i.e., increased uncoupled respiration) in PGC-1α−/− cardiac fibers, we performed respiration analysis on mitochondria isolated from cardiac ventricles of PGC-1α+/+ and PGC-1α−/− mice to further assess the extent of mitochondrial coupling. There were no differences in basal or state 3 respiration of succinate, a substrate for the TCA cycle and electron transport chain complex II (Fig. 1C). Postoligomycin (uncoupled) respiration was increased in mitochondrial isolates of PGC-1α−/− hearts (PGC-1α+/+ vs. PGC-1α−/−: 68 ± 6 vs. 103 ± 13 nmol O2·min−1·mg mitochondrial protein−1, P < 0.05; Fig. 1C). There was an associated reduction in the RC quotient for PGC-1α−/− mitochondria, indicating that the proportion of respiration coupled to ATP synthesis relative to uncoupled respiration was reduced in mitochondria of PGC-1α−/− hearts (Fig. 1C). Overall, the saponin-permeabilized myocardial fiber data and isolated mitochondrial data are consistent with reduced maximal ATP synthetic capacity and reduced mitochondrial coupling in hearts of PGC-1α−/− mice.

Isolated working hearts deficient in PGC-1α exhibit decreased palmitate oxidation and diminished cardiac power.

The mechanical and fuel metabolic implications of these mitochondrial alterations were investigated using an ex vivo isolated working heart model. Hearts from sex-matched 7- to 8-mo-old mice were perfused in conditions of relatively high fatty acid and low insulin concentrations (1.2 mM palmitate, 5 mM glucose, and 10 μU/ml insulin) to allow the assessment of functional capacity in the setting of high fatty acid oxidation rates. Hearts from PGC-1α−/− mice demonstrated a significant reduction (19%) in cardiac hydraulic power, the product of cardiac output and peak systolic pressure normalized to heart wet weight (Fig. 2A). This reduction in cardiac hydraulic power represented the net effect in PGC-1α-deficient hearts of reduced heart rate, reduced stroke volume, and reduced peak systolic pressure, none of which were statistically significantly different independently (data not shown). [3H]palmitate oxidation, a measure of mitochondrial fatty acid β-oxidation, was decreased by 25% in PGC-1α−/− hearts (Fig. 2B). Conversely, glucose oxidation, as assessed with [U-14C]glucose, was increased by 91% in PGC-1α−/− hearts (Fig. 2B). Given that the molar contribution of palmitate to TCA cycle flux is four times greater than that of glucose (given the fewer carbon atoms in glucose), the observed relative enhancement in glucose oxidation rates in PGC-1α−/− hearts is consistent with a balanced shift toward increased reliance on glucose oxidation in the setting of decreased fatty acid oxidation. Taken together, these results are consistent with the permeabilized myocardial fiber respiration analysis in that working hearts isolated from PGC-1α−/− mice exhibited reduced palmitate oxidation, with an increased reliance on glucose oxidation, likely a compensatory response.

Fig. 2.

Alterations in substrate utilization and cardiac hydraulic work in isolated working hearts from PGC-1α−/− mice. A: an isolated mouse working heart perfusion was performed with hearts from 7- to 8-mo-old sex-matched male and female mice (as described in materials and methods). Results for cardiac hydraulic work are means ± SE; n ≥ 6. *P < 0.05 compared with corresponding PGC-1α+/+ values. B: to determine palmitate and glucose oxidation rates, trace amounts of [3H]palmitate (0.1 μCi/ml) and [U-14C]glucose (0.1 μCi/ml) were used in the isolated working heart perfusate. Results represent mean oxidation rates per gram ventricular dry weight ± SE; n ≥ 5. *P < 0.05 compared with corresponding PGC-1α+/+ values.

Reduced expression of genes central to mitochondrial fatty acid oxidation and oxidative phosphorylation in PGC-1α−/− hearts.

To characterize the molecular mechanisms responsible for the observed metabolic reprogramming in PGC-1α-deficient hearts, we employed quantitative real-time PCR to examine the expression of relevant energy metabolic genes in cardiac ventricles from 1- to 2-mo-old PGC-1α+/+ and PGC-1α−/− mice (Table 1). Specifically, genes encoding the mitochondrial fatty acid oxidation enzymes, muscle-type carnitine palmitoyltransferase I (M-CPT-I) and very-long-chain acyl-CoA dehydrogenase (VLCAD), were modestly but significantly reduced in PGC-1α−/− hearts by 22 ± 6% and 20 ± 6%, respectively (relative to PGC-1α+/+, P < 0.01). The alteration of M-CPT-I levels is important because this enzyme catalyzes a rate-limiting step in fatty acid β-oxidation. Gene expression of the TCA cycle enzyme citrate synthase was also reduced by 21 ± 7%. The gene expression of the catalytic β-subunit of ATP synthase, which is essential for oxidative phosphorylation, was reduced by 17 ± 6% in hearts of PGC-1α−/− mice compared with PGC-1α+/+ mice (P < 0.05). Additionally, the expression of genes encoding proteins central to the electron transport chain, including cytochrome b, cytochrome c, cytochrome c oxidase subunit II (mitochondrial encoded), and cytochrome c oxidase subunit IV (nuclear encoded) was decreased in hearts from PGC-1α−/− mice. The expression of candidate genes involved in glucose utilization and cardiac hypertrophy or heart failure was not altered. Interestingly, the expression of the gene encoding microsomal TAG transfer protein (MTP), which facilitates the cellular export of excess accumulated TAG, was upregulated in PGC-1α-deficient hearts. Consistent with the above, electron microscopy of cardiac ventricle from 2-mo-old PGC-1α−/− mice demonstrated a generalized mild decrease in the density of mitochondrial cristae (Fig. 3) in the setting of preserved overall mitochondrial volume density. Overall, these results are in agreement with the metabolic experiments and provide a potential mechanism for the decreased capacity for mitochondrial fatty acid β-oxidation and oxidative phosphorylation.

Fig. 3.

Altered mitochondrial cristal density in PGC-1α−/− cardiac ventricles. Representative electron micrographs of the cardiac LV apex obtained from normally fed 2-mo-old PGC-1α+/+ and PGC-1α−/− female mice are shown (representative of n ≥ 6 comparisons). Scale bars are shown to the right for the low (×30,000; top) and high (×120,000, bottom) magnifications.

View this table:
Table 1.

Cardiac gene expression fold change by RT-PCR for PGC-1α−/− mice relative to PGC-1α+/+ mice (normalized to 18S)

13C NMR experiments demonstrate an accentuated shift toward glucose oxidation relative to palmitate in PGC-1α−/− hearts in the context of a metabolic stress conferred by adrenergic challenge.

The data shown above indicate that PGC-1α-deficient hearts have a decreased capacity for fatty acid oxidation. Such a defect could be manifest in conditions of stress that demand increased absolute rates of both fatty acid oxidation and glucose oxidation, such as with increased work in response to β-adrenergic challenge. To evaluate the capacity of hearts from PGC-1α−/− mice to augment fatty acid oxidative metabolism to meet the energy requirements of stress, 13C NMR-derived TAG turnover rates and fractional enrichment of acetyl-CoA from [13C]palmitate were obtained for isolated retrograde perfused hearts from 4-mo-old mice at baseline and with infusion of the β-adrenergic agonist isoproterenol (0.1 μM). In this nonworking mode, the RPP was not significantly different between groups at baseline work (PGC-1α+/+ vs. PGC-1α−/−: 47,000 ± 8,000 vs. 39,000 ± 4,000 beats·min−1·mmHg). During the administration of isoproterenol, the RPP in both groups initially increased within 1 min (PGC-1α+/+ vs. PGC-1α−/−: 54,000 ± 4,000 vs. 56,000 ± 7,000 beats·min−1·mmHg, P < 0.05) and remained elevated throughout the perfusion period. In both PGC-1α+/+ and PGC-1α−/− hearts, adrenergic stimulation significantly decreased 13C enrichment of TAG (66% and 84% decrease, respectively), indicating a shift away from [13C]palmitate storage (data not shown). At baseline, there was a trend toward elevated TAG content in PGC-1α−/− hearts compared with PGC-1α+/+ hearts (Table 2). TAG turnover at baseline was elevated by 49% in PGC-1α−/− hearts compared with PGC-1α+/+ hearts (Table 2). In the PGC-1α+/+ group, TAG content and turnover remained unchanged during isoproterenol challenge (Table 2). In contrast, adrenergic stimulation of PGC-1α−/− hearts significantly decreased TAG turnover and content (Table 2). Increased workload resulted in a significant decrease in the fractional enrichment of acetyl-CoA from [13C]palmitate in PGC-1α−/− hearts (Table 2), consistent with a general shift toward glucose oxidation relative to palmitate to meet the additional energy requirements of adrenergic challenge in PGC-1α−/− hearts.

View this table:
Table 2.

13C NMR-derived TAG turnover rates and fractional enrichment of acetyl-CoA from [13C]palmitate for isolated retrograde perfused hearts at baseline and with infusion of 0.1 μM isoproterenol

Hearts deficient in PGC-1α accumulate excess TAG with fasting and exhibit alterations in TAG molecular species profiles.

Decreased capacity for fatty acid oxidation in PGC-1α-deficient hearts could also result in an impaired response to other conditions of stress, such as fasting, which necessitate an increased reliance of the heart on fat. With fasting, there is increased delivery of fatty acid to the heart mediated by increased peripheral lipolysis. Given that PGC-1α-deficient hearts would have a decreased capacity for the oxidation of fatty acid substrate, hearts deficient in PGC-1α would be expected to accumulate lipid substrate with fasting. The response of 4-mo-old PGC-1α−/− mice to the stress of a 24-h fast was assessed by performing oil red O (ORO) staining for neutral lipid in the myocardium. ORO staining was increased in hearts of PGC-1α−/− mice following a 24-h fast (Fig. 4). To quantitate and extend these findings, two-dimensional ESI/MS analysis was performed on myocardial lipid extracts from both fed and 24-h fasted PGC-1α+/+ and PGC-1α−/− mice. Quantitation of TAG molecular species of myocardial lipid extracts demonstrated a nonsignificant trend toward increased TAG in hearts from fed PGC-1α−/− mice; however, following a 24-h fast, there was a significant (120%) increase in TAG in hearts from PGC-1α−/− mice relative to similarly fasted PGC-1α+/+ mice (Fig. 4). Thus, consistent with the reduction in mitochondrial fatty acid oxidative capacity in PGC-1α−/− hearts, TAG homeostasis is perturbed following the stress of a 24-h fast in hearts of PGC-1α−/− mice.

Fig. 4.

Assessment of cardiac ventricular neutral lipid and triglyceride (TAG) content. Representative histological sections stained with oil red O of the cardiac left ventricle from 4-mo-old PGC-1α+/+ and PGC-1α−/− mice following a 24-h fast are shown (left). The red droplets are indicative of neutral lipid accumulation. Mean cardiac ventricular TAG levels (right) were determined by two-dimensional electrospray ionization mass spectrometric (ESI/MS) analysis performed on myocardial lipid extracts from both fed and 24-h fasted PGC-1α+/+ and PGC-1α−/− mice (n ≥ 5). For the fast, 4-mo-old female mice were individually housed on wood chip bedding. Results are means ± SE. *P < 0.0001 compared with fasted PGC-1α+/+ values.

ESI/MS analysis also allowed the determination of the molecular fingerprint of the increased TAG species present in hearts of 24-h fasted PGC-1α−/− mice relative to fasted PGC-1α+/+ mice. As shown in representative myocardial lipid extracts, ESI/MS demonstrated greater TAG content in hearts of PGC-1α−/− mice for 837.7, 863.7, and 889.7 m/z TAG species relative to the TAG internal standard (Fig. 5, A and B, top “relative intensity” spectra). These increased TAG species contained acyl groups derived from linoleate (18:2), palmitate (16:0), and oleate (18:1) (Fig. 5, A and B). Indeed, quantitation of the relative abundance of these fatty acyl species in TAG from hearts of fasted mice demonstrated an increased relative abundance of 16:0, 18:1, and 18:2 fatty acyl groups in TAG of myocardial lipid extracts derived from PGC-1α−/− mice compared with PGC-1α+/+ mice (Fig. 5C). Conversely, there was a relative decreased abundance of 20- and 22-carbon fatty acyl species in TAG of myocardial extracts derived from fasted PGC-1α−/− mice compared with PGC-1α+/+ mice (Fig. 5C). ESI/MS analysis allowed quantitation not only of the relative fatty acyl composition of TAG but also of the total mass of specific TAG species. Relative to a fasted PGC-1α+/+ mouse heart value of 1.0, hearts of fasted PGC-1α−/− mice demonstrated a 3.6-fold change in 16:0/16:0/16:0 TAG, a 1.9-fold change in 16:0/16:0/18:1 TAG, and a 2.6-fold change in 16:0/18:1/18:2 TAG (Fig. 5D).

Fig. 5.

Two-dimensional ESI/MS fingerprint and quantitation of TAG molecular species of cardiac ventricular lipid extracts from fasted PGC-1α+/+ and PGC-1α−/− mice. A and B: two-dimensional ESI/MS analyses were performed on cardiac ventricular lipid extracts from 4-mo-old female PGC-1α+/+ (A) and PGC-1α−/− (B) mice following a 24-h fast as described in materials and methods. The top spectra depict the relative intensity of each TAG species compared with the T17:1 internal standard (IS) peak for TAGs. Below the relative intensity TAG spectrum, the subsequent horizontal rows depict the detailed analyses of TAG molecular species that contain the specific fatty acyl chain noted along the left axis. For example, the “NL 256.2 (16:0)” row indicates all the TAG molecular species containing at least one 16:0 fatty acyl chain. C: quantitation by ESI/MS of the relative abundance of fatty acyl species in TAG of cardiac ventricular lipid extracts from 24-h fasted PGC-1α+/+ and PGC-1α−/− mice. Results are means ± SE; n ≥ 5. *P < 0.05 and **P < 0.01 compared with fasted PGC-1α+/+ values. D: quantitation by ESI/MS of representative increased TAG species in cardiac ventricular lipid extracts from 24-h fasted PGC-1α+/+ and PGC-1α−/− mice. The increased TAG species in hearts of fasted PGC-1α−/− mice relative to PGC-1α+/+ mice included 16:0/16:0/16:0 TAG, 16:0/16:0/18:1 TAG, and 16:0/18:1/18:2 TAG. Results are means ± SE; n ≥ 5. *P < 0.05, **P < 0.01, and ***P < 0.0001 compared with corresponding PGC-1α+/+ values.

DISCUSSION

The results of our study involving the assessment of respiration in permeabilized myocardial strips and isolated mitochondria, coupled with isolated working heart metabolic experiments, demonstrate that PGC-1α is essential for the maintenance of maximal, efficient cardiac mitochondrial fatty acid oxidation and ATP synthesis. Additionally, the energy metabolic alterations in hearts deficient in PGC-1α contribute to a decrement in mechanical function in an isolated working model and impaired metabolic response to the stresses of fasting and β-adrenergic challenge. Indeed, diminished cardiac mitochondrial fatty acid oxidative capacity is associated with pathological cardiac hypertrophy and heart failure, and PGC-1α gene expression is reduced in the setting of pathological pressure overload-induced hypertrophy, potentially contributing to the altered energy metabolic balance (34). Studies in humans with hypertensive LV hypertrophy have demonstrated not only a reduction in myocardial fatty acid uptake and oxidation but also a reduction in myocardial efficiency (i.e., the ratio of myocardial work to myocardial O2 consumption) (11, 30, 53). In the present study, independent of the specific metabolic substrate, maximal ATP synthesis rates were reduced and metabolic efficiency (ATP/O) was impaired in hearts of PGC-1α−/− mice, consistent with inefficiencies in the electron transport chain and overall oxidative phosphorylation. The observed reduction in ATP synthetic capacity and mitochondrial coupling in hearts deficient in PGC-1α is consistent with prior work employing 31P NMR, which revealed that Langendorff-perfused hearts from mice deficient in PGC-1α exhibit a reduction in their concentrations of ATP both at rest and with maximal β-adrenergic stimulation (1). To expand upon the present metabolic efficiency (ATP/O) data obtained with permeabilized myocardial strips, future multisubstrate studies assessing the ATP synthetic capacity of isolated mitochondria from hearts of PGC-1α−/− mice are warranted.

Mechanisms contributing to reduced mitochondrial fatty acid oxidation and impaired efficiency in cardiac energy transduction with PGC-1α deficiency are likely multifactorial. Impaired stoichiometry of key components of electron transport and oxidative phosphorylation with PGC-1α deficiency may contribute to the present observed reduction in metabolic efficiency, as could the possible increased activity of uncoupling proteins facilitated by accumulation of lipid species. Hearts of PGC-1α−/− mice exhibit a mild reduction in the expression of genes central to fatty acid β-oxidation, the TCA cycle, electron transport, and oxidative phosphorylation. Decreased gene expression of the catalytic β-subunit of ATP synthase as well as the mitochondrial creatine kinase 2 isoform in hearts of PGC-1α−/− mice could contribute to decreased mitochondrial efficiency through a defect in the terminal event of oxidative phosphorylation and impaired mitochondrial ATP/phosphocreatine exchange of high-energy phosphoryl groups. Recently, increased expression of PGC-1α and mitochondrial genes has been shown in skeletal muscle to constitute an essential homeostatic control mechanism for the maintenance of cellular ATP levels in response to chemical uncoupling of mitochondria (46). Impediments to efficient electron transport (e.g., decreased cytochrome c in the setting of PGC-1α deficiency) may contribute to an increase in the net driving force for the reduction of O2 at a given reaction site along the electron transport chain and thus to increased ROS production (3). A central role for PGC-1α in the coordinated regulation of mitochondrial oxidative metabolism and ROS protection mechanisms has previously been described, with PGC-1α playing a significant role in the control of expression of genes central to protection from ROS-mediated injury (8, 27, 29, 5456). PGC-1α-null fibroblasts have reduced expression of ROS-detoxifying enzymes, a higher steady-state level of ROS, and greater sensitivity to oxidative stress, which itself is capable of inducing PGC-1α gene expression (54).

Consistent with the reduction in gene expression of electron transport and oxidative phosphorylation enzymes and the decrement in overall ATP synthetic capacity, cardiac ventricles from PGC-1α−/− mice exhibited a mild decrease in the density of mitochondrial cristae. Other investigators using a different animal model of PGC-1α loss of function also observed no change in cardiac mitochondrial volume density but did describe occasional subtle defects in mitochondrial packing and slight dilation of cristae (1). Tissues with high mitochondrial energy transduction rates such as the heart, brown adipose, and skeletal muscle have highly dense intramitochondrial spanning of cristae, the principal site of oxidative phosphorylation (18). PGC-1α gain of function has been shown to promote mitochondrial biogenesis in the heart, skeletal muscle, and brown adipose, with recent work demonstrating that a growth hormone-releasing peptide promotes an increase in the density of 3T3-L1 white adipocyte mitochondrial cristae in the setting of an associated increase in PGC-1α (45).

Challenging the PGC-1α-deficient mouse with fasting or with isoproterenol provided important information about the role of this factor in maintaining the dynamic balance between myocyte lipid storage and oxidation. Indeed, there was a significant increase in TAG content following a 24-h fast in hearts from PGC-1α−/− mice compared with similarly fasted wild-type mice, consistent with the observed deficit in mitochondrial fatty acid oxidative capacity with PGC-1α deficiency and consistent with other models of impaired mitochondrial fatty acid oxidative capacity, such as the PPAR-α−/− mouse (13). Interestingly, hearts from PGC-1α−/− mice had significantly greater expression of MTP, which facilitates the export of excess accumulated TAG from the heart (39). Specific TAG species that accumulated with fasting of PGC-1α−/− mice contain an abundance of long-chain fatty acids such as palmitate and oleate, prominent substrates for the mitochondrial fatty acid β-oxidation pathway. Conversely, the relative decrease in 20-carbon fatty acyl species suggests a possible compensatory upregulation of peroxisomal oxidation of these longer-chain fatty acids in the setting of mitochondrial dysfunction. Consistent with these findings of excess TAG, an inverse relationship between PGC-1α protein expression and TAG accumulation has been described in rodent skeletal muscle (5).

13C NMR revealed that hearts from PGC-1α−/− mice exhibited a limited capacity to recruit TAG as a source for lipid oxidation during β-adrenergic challenge. At baseline, PGC-1α−/− hearts exhibited a 49% higher TAG turnover rate than wild-type hearts, possibly attributable to a trend toward elevated storage of long-chain fatty acids in the fed state with an associated induction of MTP to facilitate TAG export. During β-adrenergic stimulation, energy-consuming pathways, such as TAG synthesis, would be inhibited. Whereas adrenergic challenge with isoproterenol resulted in a drop in the 13C enrichment of TAG in both groups, only hearts from PGC-1α−/− mice showed a decrease in TAG turnover. At the same time, a decrease in TAG content in PGC-1α−/− hearts with isoproterenol indicates depletion of the TAG pool. This observation is consistent with the increased expression of MTP in PGC-1α−/− hearts, which would allow for the increased export and consequent depletion of the TAG pool in times of high energy demand. During a 10-min perfusion with isoproterenol, a decrease in expression of MTP would not be seen, thus not affecting TAG export. With isoproterenol, the relative contribution of glucose oxidation increased more than the relative increase in fatty acid oxidation in hearts from PGC-1α−/− mice, resulting in a relative reduction in acetyl-CoA 13C enrichment from [13C]palmitate. The relative increase in unlabeled substrates oxidized at high workload in PGC-1α−/− hearts was presumably from glucose and not endogenous fat, as indicated by previous work on the recruitment of glucose and glycogen oxidation in the stressed rodent heart (19). Interestingly, a recent study (41) of hypertrophic rat hearts showed a decrease in palmitate oxidation and a decrease in TAG turnover. Hearts from PGC-1α−/− mice are less reliant on mitochondrial fatty acid oxidation and more reliant on a boost in glucose oxidation to meet the energy demands of adrenergic challenge. This impairment of β-adrenergic metabolic responsiveness may contribute to the previously described blunted mechanical response to the β-adrenergic agonist dobutamine in PGC-1α-deficient mice (1, 35). Future studies in which the PGC-1α−/− heart is perfused solely with glucose substrate may yield further insight into the defects of the PGC-1α-deficient state, in that “glucose-only” perfusion may preserve function by preventing the accumulation of potentially toxic lipid intermediates and reducing the generation of reactive species in peroxisomal and mitochondrial compartments.

PGC-1α integrates input from developmental cues, fasting, and exercise as well as calcium- and stress-activated pathways to facilitate its interaction with downstream targets including NRFs, PPARs, and ERRs, boosting fatty acid oxidation enzyme gene expression and mitochondrial biogenesis (14). The complexity of these interactions yields insight into the mechanisms that may compensate for chronic loss of PGC-1α function. These mechanisms may involve increased activity of PGC-1β or of partners such as ERR-α, potentially through events including posttranslational modification.

The PGC-1α-deficient heart's increased reliance on glucose oxidation in the setting of diminished mitochondrial fatty acid oxidative capacity is strikingly reminiscent of the metabolic remodeling of both pathological cardiac hypertrophy and the aged heart (32, 34, 36). It will be interesting to evaluate the cardiac phenotype of aged PGC-1α-deficient mice to understand how PGC-1α potentially antagonizes age-associated mitochondrial dysfunction by inducible promotion of efficient mitochondria with a robust capacity for mitochondrial fatty acid oxidation. Also, conditional, cardiac-specific PGC-1α- and PGC-1β-null mice will allow evaluation of the impact of acute, midlife loss of PGC-1 function: a situation that may more closely mimic the evolution of adult disease states. By facilitating an energy metabolic balance, the appropriate promotion of PGC-1α activity may represent a novel potential therapy for the pathological hypertrophied, failing, or senescent heart.

GRANTS

This work was supported by National Institutes of Health Grants KO8-AG-024844 (to J. J. Lehman), RO1-HL-3749244 (to E. D. Lewandowski), R01-HL-73167 (to E. D. Abel), and RO1-HL-058493 (to D. P. Kelly). S. Boudina was supported by a postdoctoral fellowship from the Juvenile Diabetes Research Foundation. Additional assistance was provided by the morphology core supported by Washington University Digestive Diseases Research Core Center Grant P30-DK-052574 and Washington University Clinical Nutrition Research Center Grant P30-DK-056341.

DISCLOSURES

D. P. Kelly is a scientific consultant for Novartis and is on the Scientific Advisory Boards of Phrixus and Eli Lilly.

Acknowledgments

The authors thank Bill Kraft for expert technical assistance with electron microscopy and Mary Wingate for assistance with manuscript preparation.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract