Mitochondrial health is enhanced in rats with higher vs. lower intrinsic exercise capacity and extended lifespan

Autophagy/mitophagy and mitochondrial fitness are better preserved in cardiomyocytes from HCR vs. LCR

Next, we investigated whether a higher Rres was associated with mitochondrial fitness and turnover linked to autophagy/mitophagy. These metrics were assessed in parallel with Rres and VO₂max employing live-cell imaging and electron microscopy (EM) of cardiac myocytes from all age groups. In addition, accumulation of lipofuscin, a biomarker of molecular/organellar recycling impairment, was also measured.

Autophagy was quantified as accumulation of autophagic vesicles, visualized by combining vital-dye staining with confocal microscopy imaging of cardiac myocytes subsets from LCR and HCR in which Rres and VO₂max were measured in parallel. At a young age, autophagy was already significantly higher in HCR vs. LCR (Fig. ​(Fig.1d1d and Supplementary Fig. 7a) and was resistant to oxidative stress-mediated mPTP opening (Supplementary Fig. 7b), whereas HS resembled HCR rather than LCR. The higher autophagy/mitophagy exhibited by HCR vs. LCR at 6 months (Figs ​(Figs1d1d and 2a, c) might, in part, explain the difference in respiration in the presence of lipid, observed between intact cardiomyocytes (Fig. ​(Fig.1a),1a), and isolated mitochondria (Fig. ​(Fig.1b),1b), which in the latter would be substantially altered by the isolation procedure. Aspects of enhanced mitochondrial function in intact cardiomyocytes that could be lost by cell disruption include loss of interactions between mitochondria, responsible for their network behavior, or the deficit of requisite cytoplasmic factors and/or signaling^6,43.

Open in a separate window

Fig. 2

Confocal live fluorescence imaging and electron microscopy (EM) of autophagy/mitophagy in LCR and HCR cardiomyocytes during aging.

a Depicted are the levels of autophagy/mitophagy quantified as % cell area covered by autophagy vesicles in isolated cardiomyocytes from 6 (LCR: 4 rats/experiments, n = 54; HCR: 4 rats/experiments, n = 53), 17 (LCR: 4 rats/experiments, n = 66; HCR: 5 rats/experiments, n = 70), and 24 (LCR: 5 rats/experiments, n = 94; HCR: 7 rats/experiments, n = 112) months old rats, using high-resolution confocal microscopy of formed autophagic vesicles labeled by the fluorescent CYTO-ID for autophagy detection (green) according to manufacturer’s instructions (see “Confocal fluorescence imaging” and “Electron microscopy” in “Methods”) and mitochondrial staining with 100 nM TMRM (red). See related Supplementary Fig. 8a for additional two-way ANOVA with Tukey’s multiple comparison test (GraphPad Prism 8.0) and all data points. b Quantification of total autophagy/mitophagy figures (autophagosomes [early] + autophagolysosomes [late]) observed by EM and normalized with respect to the cell volume using stereological procedures (see “Confocal fluorescence imaging” and “Electron microscopy” in “Methods”) in samples of LCR and HCR cardiomyocytes from 6 (LCR: 2 rats/experiments, n = 40; HCR: 2 rats/experiments, n = 40), 17 (LCR: 2 rats/experiments, n = 48; HCR: 2 rats/experiments, n = 47), and 24 (LCR: 2 rats/experiments, n = 37; HCR: rats/experiments, n = 35) months old rats. c Quantification of autophagic flux (late/(early + late)) figures observed by EM in samples of LCR and HCR cardiomyocytes from 6 (LCR: 2 rats/experiments, n = 38; HCR: 2 rats/experiments, n = 35; NS, p = 0.14), 17 (LCR: 2 rats/experiments, n = 49; HCR: 2 rats/experiments, n = 41; NS, p = 0.34), and 24 (LCR: 2 rats/experiments, n = 38; HCR: 2 rats/experiments, n = 27; NS, p = 0.17) months old rats. For b, c, see Supplementary Fig. 9a, b for additional two-way ANOVA with Tukey’s multiple comparison test (GraphPad Prism 8.0) and all data points. d Representative examples of cardiomyocytes; clockwise, from top left: CYTO-ID-stained autophagic vesicles, TMRM, transmission light microscopy, merge, and detailed (zoomed) image of a cardiomyocyte: CYTO-ID-stained autophagic vesicles, TMRM, transmission light microscopy, merge. e Representative electron micrographs of autophagy/mitophagy figures in cardiomyocytes isolated from LCR and HCR hearts at the indicated ages. Bar size = 0.5 µm in all panels. “Early” figures of autophagosomes (depicted in A, B, E, F, I, J) were identified by its content (morphologically intact, mostly mitochondria) and a double limiting membrane separated by a narrow electron-lucent cleft. “Late” figures of autophagolysosomes (depicted in C, D, G, H, K, L), delimited by single membranes, represent an advanced stage of autophagosomes where a lysosome has merged, and the content comprises partially degraded, irregular, electron-dense or pale material depending upon time of formation. f Interpretative scheme of autophagy/mitophagy describing the stages of this process measured by EM quantification. In all cases, data are represented as mean ± SEM. The statistical significance is indicated by *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; NS not significant.

Cross-sectionally, strain and age explain 11% and ~8% of changes in autophagy, respectively (Fig. ​(Fig.2a2a and Supplementary 8a). Importantly, autophagy is significantly higher in HCR vs. LCR at all ages tested (Supplementary Fig. 8a). The results of live-cell imaging were consistent with those obtained with quantitative EM, where total ([early] autophagosomes + [late] autophagolysosomes) (Fig. ​(Fig.2e)2e) and relative (late/[early + late]) were determined as metrics of autophagy/mitophagy (Fig. ​(Fig.2b)2b) and autophagic flux (Fig. ​(Fig.2c),2c), respectively, in a subset of the same pool of cardiac myocytes. The gallery of images displayed in Fig. ​Fig.2e2e shows representative EM micrographs obtained from isolated cardiomyocytes depicting mitophagy figures, in the two first columns, at 6, 17, and 24 months of age in LCR: A, E, L and HCR: B, F, J, respectively, and of autophagy, in the third and fourth columns, at the indicated ages and strain. “Early” figures of autophagosomes are surrounded by a double limiting membrane and the content, which in A, B, E, F, I, J, correspond to morphologically intact mitochondria. “Late” figures of autophagolysosomes (illustrated in C, D, G, H, K, L), surrounded by a single membrane, represent an advanced stage where the autophagosome has merged with a lysosome and its content becomes partially degraded and irregular, electron dense, or pale, depending on the time of formation.

The results show that, at young age, the autophagic flux exhibited a nonsignificant higher trend in HCR vs. LCR (Fig. ​(Fig.2c),2c), in agreement with live-cell imaging (Fig. ​(Fig.2a),2a), although the total number of autophagy/mitophagy figures from both strains was similar (Fig. ​(Fig.2b);2b); the latter can be explained by the fact that the CytoID probe labels autophagolysosomes, whereas EM counts mitophagy figures, in addition to autophagosomes and autophagolysosomes (Fig. 2e, f). The trend of increased autophagic flux in HCR vs. LCR was maintained across ages, accompanied by a significantly 36% higher total autophagy/mitophagy figures per constant volume density (data not shown) at mid-age, and a nonsignificant trend at the oldest age (Fig. ​(Fig.2b2b).

We also investigated the potential role of mitochondrial biogenesis using EM micrographs to determine mitochondrial numbers, size (area), and fractional area occupied by mitochondria⁴⁴. Only at 17 months of age, HCR exhibited a significantly higher 22% number (Supplementary Fig. 9c) of smaller mitochondria (Supplementary Fig. 9e) that covered a similar cell area (Supplementary Fig. 9d) compared to LCR; however, these metrics were unchanged at other ages. As at 17 months Rres/VO₂max (Fig. ​(Fig.1a1a and Supplementary Fig. 2a) and autophagy/mitophagy (Fig. ​(Fig.2)2) were higher 41/45% and 42% (confocal)/33% (EM) in HCR vs. LCR, respectively, we conclude that, except at middle age, where both autophagy/mitophagy and mitochondrial biogenesis contribute to the higher respiration observed, turnover rather than biogenesis was the main driver of mitochondrial quality in HCR. TWA further showed that age and strain exert ~7% and ~1%, respectively, influence on the variance of mitochondrial number (Supplementary Fig. 9c).

To further explore the interplay of mitochondrial health and fitness, we determined the resistance to mitochondrial permeability transition, measured as the time (t, in seconds) to reach the ROS-induced ROS release threshold for mPTP opening (t_mPTP). Indeed, t_mPTP was significantly longer in cardiomyocytes from age-matched HCR vs. LCR (Fig. 3a, b). TWA showed that age had a major impact (~50%) with strain exerting a relatively weak (~6%), but highly significant influence (Fig. 3a, b and Supplementary Fig. 8b). The resistance to opening of the mPTP with oxidative stress was significantly higher in HCR vs. LCR (Supplementary Fig. 8b), a result in agreement with better autophagy/mitophagy (Figs ​(Figs22 and ​and3,3, and Supplementary Fig. 8a, b).

Open in a separate window

Fig. 3

Confocal live fluorescence imaging of mitochondrial fitness and lipofuscin in LCR and HCR cardiomyocytes during aging.

a mPTP-ROS threshold (t_mPTP) as index of mitochondrial fitness was quantified in isolated TMRM-loaded cardiomyocytes from 6 (LCR: 4 rats/experiments, n = 47; HCR: 4 rats/experiments, n = 46), 17- (LCR: 4 rats/experiments, n = 44; HCR: 5 rats/experiments, n = 44), and 24 (LCR: 5 rats/experiments, n = 79; HCR: 7 rats/experiments, n = 91) months old rats, using high-resolution confocal fluorescent microscopy. t_mPTP reflects the average time required for the photoproduced ROS to cause mPTP opening induction. b Representative example of t_mPTP determination in LCR vs. HCR cardiomyocytes. c Lipofuscin quantification as % cell area covered by progressive age-dependent accumulation of lipofuscin as revealed by autofluorescence of cardiomyocytes from 6 (LCR: 3 rats/experiments, n = 55; HCR: rats/experiments, n = 56), 17 (LCR: 2 rats/experiments, n = 33; HCR: 2 rats/experiments, n = 34), and 24 (LCR: 5 rats/experiments, n = 80; HCR: 5 rats/experiments, n = 97) months old rats. d Zoomed image showing, in detail, accumulated lipofuscin as visualized by autofluorescence (blue) and TMRM-loaded mitochondria (red), and e representative examples of cardiomyocytes, full-scale, at 6 and 24 months of age. In all cases, data are represented as mean ± SEM. The statistical significance is indicated by *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; NS not significant.

Accumulation of lipofuscin, an undegradable cellular “wear-and-tear” byproduct, increased dramatically with age in both strains but was significantly higher only in old (24 months) LCR vs. HCR (Fig. 3c, d and Supplementary Fig. 8c). In older animals, typically perinuclear lipofuscin granules exhibited a widespread cellular localization, whereas at 6 months of age lipofuscin accumulation in cardiac myocytes from both strains was comparable and negligible, appearing restricted to small perinuclear areas (Fig. ​(Fig.3e).3e). TWA showed that age has a major impact (~55%) on lipofuscin accumulation, with strain exerting a negligible (~1%) effect (Fig. ​(Fig.3c3c and Supplementary Fig. 8c). The presence of lipofuscin granules in cardiomyocytes was independently confirmed by EM (data not shown).

Consistent with our initial hypothesis, autophagy/mitophagy and turnover are enhanced in cardiomyocytes of HCR vs. LCR, even from a young age (Fig. ​(Fig.1d),1d), which underlies, at least in part, the significant differences in mitochondrial Rres, VO₂max and fitness observed between both strains. Although there is an age-related decline in mitochondrial Rres and VO₂max independent of rat strain, the relative enhancement of these metrics in HCR vs. LCR is sustained during aging. Furthermore, HCR’s Rres is positively enhanced by the influence of substrate (namely the combination of glucose with Palm) (Supplementary Fig. 3b) and by its interaction with age (Supplementary Fig. 3a, b). The ensemble of data indicates that, comparatively, mitochondrial turnover via autophagy/mitophagy (Fig. 2a–c and Supplementary Fig. 8a) rather than biogenesis (Supplementary Fig. 9c, d) appears to be the main driver of the mitochondrial health benefits in HCR vs. LCR (Supplementary Fig. 9c). A stronger influence of age vs. strain on the variance of t_mPTP (50% vs. 6%) and lipofuscin (55% vs. ~1%) indicates that, together, these metrics behave as biomarkers of mitochondria and heart aging.

Method details

All the experiments described below, except for metabolomics, were performed following a blinding protocol, which was utilized throughout the study, and for all measurements. The identity of the rats (HS, HCR, LCR) was kept by the technician handling the animals/cells and was not revealed until after all the data were processed and analyzed by the investigators involved in the study. Therefore, any of the investigators collecting and analyzing data knew the identity of the animals until the end, including microscopy experiments (see below).

Cardiomyocytes isolation and high-throughput respiratory measurements

Cardiomyocytes were isolated from 6-, 17-, or 24-month-old rats following standard procedures as described in ref. ⁶⁷. Briefly, once isolated in Krebs Henseleit buffer (KHB) containing 1 mM CaCl₂, the cardiomyocytes suspension was let to sediment twice, 5 min each at 1 × g, to separate viable from unviable (round “meat balls”) cells as the former sediment faster than the latter. The viable cardiomyocytes were recovered directly from the bottom of the tube and carefully pipette-transferred to Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 1 g/l glucose, 100 units/ml penicillin, 100 μg/ml streptomycin, and 5% fetal bovine serum (Hyclone, Logan, UT) (DMEM medium), counted with a hemocytometer, and plated at a density of 500–4000 viable cells/well on a laminin covered multi-well plate. The laminin covered plates were incubated overnight, the day before the experiment, at room temperature with 0.01 mg/ml laminin solution freshly diluted in phosphate-buffered saline pH 7.4. The excess laminin solution was discarded shortly before seeding the cardiomyocytes. After seeding, the cells were maintained in DMEM medium at 37 °C in a humidified incubator with 5% CO₂ and 95% O₂ for 3 h to allow cells to attach to the laminin covered surface. Finally, the DMEM medium and dead cells (that do not attach) were removed by inverting the plate, the wells washed with KHB, followed by addition of 120 μl of KHB containing 1 mM glucose to each well for assaying OCR (final 80–85% cell viability, comparable between all experimental groups). We note that the approximate number of total cells recovered after the two sedimentation cycles were similar as age increased, and the recovery from the 3500–4000 cells per well initially plated, at the end of the attachment period, were not different as judged by the cell count performed after the OCR measurements. To avoid selection bias, isolations that rendered fewer viable cells than our standard 80–85% cell viability, were discarded.

Importantly, throughout the whole study, both LCR and HCR cardiomyocytes freshly isolated the same day were run paired in a 96-well plate for Rres, VO₂max, and, always, processed by the same operators. This enabled us to lessen, e.g., technical, equipment, differences to interfere with the response of each strain’s cardiomyocytes. Importantly, the same protocol utilized for Rres, VO₂max also applied for all other metrics as well, such as autophagy/mitophagy (confocal and EM), t_mPTP, lipofuscin, which were obtained using a subset of freshly isolated cells from the same isolation batch (see “Confocal fluorescence imaging” and “Electron microscopy”).

The protocol for measuring respiration in isolated cardiomyocytes comprised high-throughput OCR quantification using a Seahorse XFe96 equipment, in the absence of substrate (baseline), followed by successive measurements performed on the same cells when exposed to subsequent additions of substrate (OCR_substr), oligomycin (OCR_oligo), FCCP (OCR_FCCP), and respiratory inhibitors (antimycin + rotenone). Rres was quantified by subtracting OCR_substr from OCR_FCCP in the presence of either substrate or their combination (see Fig. ​Fig.1a1a and Supplementary 1). VO₂max was assigned to OCR_FCCP (Supplementary Fig. 2).

The abovementioned protocol carried out in the Seahorse XFe96 equipment, included three baseline measurements of the OCR, followed by OCR_substr comprising addition of 20 μl substrates’ stock solutions containing either glucose (Gluc, 4 mM final concentration + 1 mM already present in KHB = 5 mM) or Palm (200 μM bound to fatty acid-free bovine serum albumin (BSA), 4 : 1 Palm : BSA)⁶⁸, or the combination Gluc + Palm at their respective concentrations⁶⁹. Next, OCR_oligo was performed by adding 10 μM oligomycin followed by OCR_FCCP with 1.23 μM FCCP and 10 μM each of antimycin A/rotenone. Preliminary experiments were performed to demonstrate the linearity of the OCR response as a function of cell number and the expected behavior in the presence of the electron transport chain inhibitors, oligomycin, FCCP and antimycin A/rotenone⁷⁰. At the end of the protocol, the cells were detached from the bottom of the plate by trypsin (0.05%) digestion for 10 min at 37 °C followed by addition of FBS and 4% formaldehyde to fix the cell for counting in a hemocytometer. The normalized OCR values are expressed in pmol or nmol O₂/min/10⁴ cells.

Mitochondrial isolation and high-throughput bioenergetic measurements

Procedures for the isolation and handling of mitochondria from rat hearts were performed as previously described²⁷. Mitochondrial isolation was carried out in small (2–4 g) fresh heart tissue from 6-month-old rats, which were killed, the heart quickly excised, and the rest of the procedure was performed on ice. Briefly, the tissue was homogenized in the presence of 0.8 mg bacterial proteinase (type XXIV, Sigma), which was added just before starting the homogenization procedure, and the homogenate processed for mitochondrial isolation as previously described.

High-throughput Seahorse XFe96 OCR of isolated rat heart mitochondria under state 3 was determined in KCl-based assay medium in the presence of 5/5 mM G/M and 0.5 mM ADP. The oligomycin-sensitive OCR was obtained by subtracting from state 3 OCR the state 4 OCR measured in the presence of 10 μM oligomycin; n = 60/3 experiments, P < 0.0001.

The mitochondrial OCR was evaluated using the equivalent of 5 μg of mitochondrial protein/well plated on XF96 PET plates coated with polyethyleneimine, as previously described^69,71. The assay was carried out under three different substrate conditions: 5 mM glucose, 20 μM PCoA, or both substrates together. Under β-oxidation fueled conditions in the presence of 20 μM PCoA, 0.5 mM malate and 0.5mM l-carnitine were also present, in a Respiration Assay Medium (200 μl final assay volume) containing (in mM): 137 KCl, 2 KH₂PO4, 0.5 EGTA, 2.5 MgCl₂, and 20 HEPES at pH 7.2 and 37 °C in the presence of 0.2% fatty acid-free BSA⁷¹. The OCR corresponding to states 4 and 3 respiration was determined before and after addition of 500 μM ADP, respectively. Respiratory Control Ratios (state 3/state 4) of 5 or higher were obtained with 5 mM/5 mM G/M. Then, 1 μM FCCP was added to uncouple respiration and, finally, 10 μM each of antimycin A and rotenone, to confirm the mitochondrial origin of the O₂ consumption observed. The experimental rates are expressed in pmol min⁻¹ mg⁻¹ mitochondrial protein. As positive controls, samples were run in the absence of β-oxidation with NADH-linked substrates (G/M, 5/5 mM).

Confocal fluorescence imaging

Isolated cardiomyocytes were resuspended in HEPES buffer: 137 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO₄, 1.2 mM NaH₂PO₄, 15 mM glucose, 20 mM HEPES, and 1.0 mM CaCl₂ (pH 7.3). Lipofuscin autofluorescence was excited using 633 nm He-Ne laser; for all other confocal microscopy experiments, cardiomyocytes were loaded with an appropriate dye. Cells were imaged with an LSM-510 inverted confocal microscope, using a Zeiss Plan-Apochromat ×63/1.4 numerical aperture oil immersion objective (Carl Zeiss, Inc., Jena, Germany) with the optical slice set to 1 µm. Images were processed by MetaMorph software (Molecular Devices, San Jose, CA).

Imaging was performed with 15–20 randomly selected CytoID-loaded cardiomyocytes per cell batch isolation and analyzed using MetaMorph image analysis software. The % area occupied by the autophagolysosomes was referred to the whole cell area.

Determination of mPTP-ROS threshold

Experiments were conducted as described previously⁷, using a method to quantify the ROS susceptibility for the induction of mPTP in individual mitochondria within cardiac myocytes⁸. Briefly, isolated cardiomyocytes were loaded with 100 nM tetramethylrhodamine methyl ester (TMRM; Invitrogen {"type":"entrez-nucleotide","attrs":{"text":"I34361","term_id":"1825152","term_text":"I34361"}}I34361) for at least 2 h at room temperature. Line scan images at 2 Hz were recorded from ~22 mitochondria arrayed along individual myofibrils with λ_exc at 543 nm and collecting λ_em at >560 nm. The confocal pinhole was set to obtain spatial resolutions of 0.4 µm in the horizontal plane and 1 µm in the axial dimension. Repetitive laser scanning of this row of mitochondria in a myocyte loaded with TMRM results in incremental, additive exposure of only the laser-exposed area to the photodynamic production of ROS and consequent mPTP induction, which can be clearly identified by the immediate dissipation of ΔΨ_m in individual mitochondria, visualized as a point in time where “columns” of the line scan image suddenly lose TMRM fluorescence becoming black (Fig. ​(Fig.3b).3b). The ROS threshold for mPTP induction (t_mPTP) was determined as the average time necessary to elicit the first ΔΨ_m depolarization in the exposed row of mitochondria (Fig. 3a, b).

Autophagy

Cardiomyocytes were loaded with the autophagy dye from CYTO-ID Autophagy detection kit (ENZO 51031-K200), according the manufacturer protocol (dilution 1 : 500 in HEPES buffer) and incubated for 30 min at room temperature. The dye was washed out with HEPES buffer and cells imaged using λ_exc 488 nm and λ_em > 505 nm by confocal fluorescence microscopy (LSM-510 inverted confocal microscope).

Electron microscopy

Cardiomyocytes pellets were processed for stereological analysis. Fixation for EM was performed using 2% glutaraldehyde–3% paraformaldehyde in 0.1 M sodium cacodylate buffer (0.1 M sucrose, 2 mM CaCl₂). Samples were post fixed in 1% osmium tetroxide for 1 h at 4 °C in the same buffer, dehydrated, and then embedded in Embed 812 resin (Electron Microscopy Sciences, Hatfield, PA) through a series of resin resin-propylene oxide gradients to pure resin. Blocks were formed in fresh resin contained in silicon molds, and the resin was polymerized for 48–72 h at 65 °C. Ultrathin (40–60 nm) sections were stained with uranyl acetate and lead citrate, and then examined on a Jeol JEM1400 electron microscope (Servicio Centralizado de Apoyo a la Investigación, SCAI, University of Córdoba, Spain) by a blinded investigator (see above section: “Method details”). Micrographs at ×15,000 magnification were obtained from randomly selected areas of cardiomyocytes cytoplasm for quantitative analysis of autophagy/mitophagy figures. We determined two stereological parameters: (a) volume density of autophagy figures (Vv; i.e., the volume fraction of cardiomyocyte cytoplasm occupied by autophagy figures) and (b) numerical density of these figures (Nv, i.e., the number of autophagy figures per µm³ cell; see Fig. ​Fig.2b).2b). Volume density was obtained following a point analysis using a simple square lattice test system⁷². The numerical density was obtained using the formula

where “Na” is the numerical profile density (number of autophagy figures/μm² of cell), and k and β the autophagic figures size distribution and shape coefficient, respectively. These coefficients were calculated using the results of planimetric measurements in cardiomyocytes and assimilating figures shape to spheroids⁷². Planimetric measurements were performed using ImageJ software (NIH).

Autophagosomes were identified as double membrane vesicles with identifiable cargo and comparable density to the surrounded cytosol. Autolysosomes were denoted as single membrane vesicles containing non-identifiable cargo of density lower than the surrounded cytosol and fragmented organelles. Areas of vesicles and cytosol were measured using manual tracking with ImageJ software.

For mitochondrial metrics, only micrographs depicting longitudinal sections of cardiomyocytes with visible sarcomeres were utilized, and from each of these five to seven pictures were taken from four to six cells/fibers. From each picture, mitochondria were counted and measured to determine the number and area, doing the average of these metrics in each picture, and finally the average for each cell. With this procedure, a total of ~1600 up to ~2000 mitochondria per experimental group were counted/measured.

Cross-sectional metabolomics in serum, heart, and liver tissue from LCR and HCR rats

Metabolomic analysis was performed by the West Coast Metabolomics Center at UC Davis (Davis, CA) in the heart, liver, and serum from nonfasted animals as previously described^73,74. Data were acquired using the method as described by⁷⁵, briefly summarized in ref. ⁷⁴. Relative metabolite levels represent the MS peak amplitude normalized with respect to the total metabolites returned

We are grateful to the following colleagues for their input of information to this work: Dr. Mervyn Maze (University of San Francisco, San Francisco, CA), Dr. John P. Thyfault (University of Kansas Medical Center, Kansas City, Kansas), Dr. Henry J. Thompson (Colorado State University, Fort Collins, CO), and Dr. Colleen M. Novak (Kent State University, Kent, OH). We thank Dr. Michel Bernier for critically reading the manuscript. This work was supported by the Intramural Research Program of the National Institutes of Health, National Institute on Aging. The LCR-HCR rat model system is funded by the Office of Research Infrastructure Programs grant P40OD021331 (to L.G.K. and S.L.B.) from the National Institutes of Health. The rat models for low and high exercise capacity are maintained as an international resource with support from the Department of Physiology and Pharmacology, The University of Toledo College of Medicine and Life Sciences, Toledo, OH. Contact L.G.K., Lauren.Koch2@UToledo.edu, or S.L.B., brittons@umich.edu, for information on the rat models.