Exercise and ageing impact the kynurenine/tryptophan pathway and acylcarnitine metabolite pools in skeletal muscle of older adults

Exercise‐induced perturbation of skeletal muscle metabolites is a probable mediator of long‐term health benefits in older adults. Although specific metabolites have been identified to be impacted by age, physical activity and exercise, the depth of coverage of the muscle metabolome is still limited. Here, we investigated resting and exercise‐induced metabolite distribution in muscle from well‐phenotyped older adults who were active or sedentary, and a group of active young adults. Percutaneous biopsies of the vastus lateralis were obtained before, immediately after and 3 h following a bout of endurance cycling. Metabolite profile in muscle biopsies was determined by tandem mass spectrometry. Mitochondrial energetics in permeabilized fibre bundles was assessed by high resolution respirometry and fibre type proportion was assessed by immunohistology. We found that metabolites of the kynurenine/tryptophan pathway were impacted by age and activity. Specifically, kynurenine was elevated in muscle from older adults, whereas downstream metabolites of kynurenine (kynurenic acid and NAD+) were elevated in muscle from active adults and associated with cardiorespiratory fitness and muscle oxidative capacity. Acylcarnitines, a potential marker of impaired metabolic health, were elevated in muscle from physically active participants. Surprisingly, despite baseline group difference, acute exercise‐induced alterations in whole‐body substrate utilization, as well as muscle acylcarnitines and ketone bodies, were remarkably similar between groups. Our data identified novel muscle metabolite signatures that associate with the healthy ageing phenotype provoked by physical activity and reveal that the metabolic responsiveness of muscle to acute endurance exercise is retained [NB]:AUTHOR: Please ensure that the appropriate material has been provide for Table S2, as well as for Figures S1 to S7, as also cited in the text with age regardless of activity levels.


Introduction
Human life expectancy has greatly increased over the last 50 years, resulting in a rapidly growing ageing population worldwide (Centers for Disease Control & Prevention, 2003). This demographic shift is an enormous healthcare challenge because old age is a major risk factor for many chronic diseases (Kennedy et al., 2014). There is increased urgency to develop therapeutic strategies that prolong the healthy period of life and delay the development of chronic disease during ageing (i.e. to improve healthspan).
Impaired cellular metabolism is a hallmark of ageing and may be associated with the heightened risk of chronic diseases (Cartee et al., 2016;López-Otín et al., 2013). Physical activity, including structured exercise training, is well known to provide many health benefits to older adults (Cartee et al., 2016). Skeletal muscle plays an important role in mediating the positive effect of physical activity via improved cellular metabolism. Specifically, older adults engaged in chronic endurance exercise training exhibit preserved muscle bioenergetics and cardiorespiratory fitness compared to age-matched sedentary counterparts, a phenotype that is comparable to individuals almost 30 years younger (Chambers et al., 2020;Gries et al., 2018;Heath et al., 1981;Trappe et al., 2013). At the level of muscle, the age-related loss in various mitochondrial parameters, including respiration, ATP production and markers of content such as citrate synthase activity and mtDNA copy number, are mitigated in endurance-trained older adults (Cartee et al., 2016;Coggan et al., 1992Coggan et al., , 1993Distefano et al., 2018;Gries et al., 2018;Short et al., 2003).
The molecular events by which exercise enhances a healthy skeletal muscle phenotype in older adults are not well understood. Acute alterations in muscle metabolites during the post-exercise recovery period probably propagate the molecular cascades needed for exercise-induced adaptations, including improved cardiorespiratory fitness and muscle-specific adaptations (Egan & Sharples, 2022;Egan & Zierath, 2013;Seaborne & Sharples, 2020). However, the metabolites that respond to acute endurance exercise in older muscle and the impact that physical activity level has on the response to acute exercise is unclear. In the present study, we used an unbiased metabolomics approach to identify metabolite signatures in skeletal muscle that are associated with ageand physical activity phenotypes at rest and in response to a bout of endurance exercise. Our general hypothesis was that the response to acute exercise at the whole-body and muscle metabolome level would be blunted in the older adult groups compared to young counterparts. Our study revealed that key metabolite signatures in skeletal muscle, including from the kynurenine/tryptophan pathway, acylcarnitines and NAD + , are linked to phenotype 0 J. Matthew Hinkley is a postdoctoral fellow at the AdventHealth Translational Research Institute in the laboratory of Dr Paul Coen. As a clinical translational scientist, he has combined basic and applied scientific approaches to identify mechanisms associated with age-related impairments in physical and metabolic function in humans. His current research has evaluated the molecular basis of physical dysfunction with sedentary ageing, and how acute and chronic endurance exercise alter the molecular profile of skeletal muscle in older adults. and clinical characteristics pertinent to healthy ageing, including cardiorespiratory fitness and mitochondrial energetics.

Ethical approval
All experimental procedures were approved by the AdventHealth Orlando (Orlando, FL, USA) Institutional Review Board (IRBnet# 554559-144) and were performed in accordance with the standards set forth by the Declaration of Helsinki, except for registration in a database. Participants provided their written informed consent before completing any data collection procedures.

Study design and participants
Forty-four men and women were recruited from the Orlando, Florida, area. All participants were in good general health, defined as not having chronic medical conditions and no contraindications for exercise, and were weight stable for the last 6 months, with normal resting blood pressure (<150 mmHg systolic, <90 mmHg diastolic). Participants were assigned to one of three groups based on age and physical activity/training status: young active/trained, older active/trained and older sedentary. Active participants were required to engage in endurance exercise training (e.g. running, cycling and/or swimming) at least 3 days week -1 for at least 5 years without an extensive layoff (>6 months), whereas sedentary individuals performed structured exercise ≤1 day week -1 . Additional inclusionary/exclusionary criteria included: (1) weight stable (±4.5 kg) over the past 6 months; (2) body mass index between 20 and 35 kg m -2 ; (3) normal blood pressure (<150 mmHg systolic, <90 mmHg diastolic); (4) not taking medications known to influence muscle metabolism; (5) no chronic medical conditions; and (6) no contraindications to exercise (high blood pressure, abnormal ECG, etc.). All testing procedures were completed over four study visits at the Translational Research Institute at AdventHealth Orlando. Visit 1 consisted of physical and skeletal muscle function assessments; visit 2 consisted of a maximal exercise test on J Physiol 601.11 a cycle ergometer to determine cardiorespiratory fitness; visit 3 consisted of imaging techniques to assess body composition; and visit 4 consisted of the acute exercise trial.

Body composition
Body composition was assessed by dual energy X-ray absorptiometry using a GE lunar iDXA whole-body scanner (GE Healthcare, Chicao, IL, USA). Body composition (total fat mass, total lean mass and total fat-free mass) was analysed with encore (Distefano et al., 2018).

Cardiorespiratory fitness
Maximum aerobic capacity (V O 2 peak ) was examined during a graded exercise protocol on a cycle ergometer as previously described (Coen et al., 2013). Heart rate, blood pressure and ECG were constantly monitored throughout the test. Tests were terminated when volitional exhaustion was reached or when criteria outlined by the American College of Sports Medicine guidelines was observed in participants (ACSM College of Sports Medicine et al., 2006).

Physical activity monitoring
Physical activity was assessed using a triaxial accelerometer (SenseWear® Pro Armband; BodyMedia Inc., Pittsburgh, PA, USA) between study visits 2 and 3. The monitor integrates motion sensor data from the tri-axial accelerometer to estimate the energy cost of free-living activity. The participant wore the monitor consistently for 5−7 days, except when showering or bathing. Days with a wear time of at least 85% were used for the analysis (Distefano et al., 2018).

Acute exercise trial
Participants were instructed to refrain from exercise 48 h prior to the acute exercise trial to minimize the effects of acute exercise on baseline metabolic profiles. Upon arrival at the laboratory following an overnight fast, whole-body indirect calorimetry was performed using an open circuit spirometry metabolic monitor system (Parvo Medics, Sandy, UT, USA) to estimate rates of post-absorptive fat and glucose-oxidation from respiratory gas exchange at baseline (resting metabolic rate). Following the procedure, participants consumed a small low glycaemic index meal (200 kcal, 15% protein, 35% fat and 50% carbohydrate) and, 15 min later, a muscle biospecimen was obtained using the muscle biopsy technique, as described below.
Each participant then performed an acute bout of endurance exercise as previously described (Distefano et al., 2018;Rubenstein et al., 2022). Briefly, participants performed a 5 min warm-up consisting of light cycling on an electronically braked cycle ergometer. After the warm-up, the participants cycled for 40 min at ∼70% of heart rate reserve, which was calculated as heart rate maximum during theV O 2 peak test minus supine resting heart rate. Heart rate, perceived exertion and blood pressure were measured every 5 min during the test. Indirect calorimetry was performed using an open-circuit spirometry metabolic monitoring system (Parvo Medics) and was performed from the seated rest period until 30-mins into the acute exercise bout (38 min total time examined).V O 2 and respiratory quotient (RQ), calorimetry determinants of exercise intensity, as well as energy expenditure were taken as an average throughout the exercise bout once steady-state was achieved (∼10-15 min into the 40 min bout). Gross exercise efficiency (GEE) was calculated as the ratio of work output to exercise energy expenditure during the submaximal bout as previously described (Distefano et al., 2018).
Immediately following the exercise trial, a muscle biopsy was performed on the leg opposite from the baseline biopsy procedure. Participants then rested in a supine position for 3 h, after which a third biopsy was taken from the same leg as the baseline biopsy procedure. The incision site was ∼5 cm proximal to the initial biopsy site. Additionally, whole-body indirect calorimetry was performed following each biopsy procedure to assess post-exercise alterations in whole-body substrate oxidation.

Blood analysis
Blood samples were collected using a catheter placed in the antecubital space prior to, immediately after and 3 h after the exercise bout for assessment of lactate (L-Lactate Assay Kit; Abcam, Cambridge, UK) and free fatty acids (NEFA-HR2 Assay Kit; FUJIFILM Wako Diagnostics, Neuss, Germany) using commercially available kits in accordance with the manufacturer's instructions and analysed on a Biotek Cytation3 Imaging Reader (Thermo Fisher Scientific, Waltham, MA, USA).

Percutaneous muscle biopsies
Percutaneous muscle biopsies were obtained with suction from the middle region of the vastus lateralis under local anaesthesia (2% buffered lidocaine) as previously described (Pruchnic et al., 2004). Following the biopsy, excess blood, visible fat and connective tissue were removed from the specimen. Approximately 50 mg of muscle tissue was immediately frozen in liquid nitrogen (−190°C) and stored at −80°C until metabolite analysis, whereas ∼10 mg of muscle tissue was used for respirometry experiments and ∼30 mg of muscle tissue was preserved in optimal cutting temperature compound and frozen in liquid nitrogen-cooled isopentane for histological analysis. Muscle biopsies were collected between November 2014 and August 2017, and all samples were stored in monitored and alarmed freezers at −80°C until analysis. Metabolomics analysis was run on all samples in March 2018.

Mitochondrial energetics
The mitochondrial respiratory capacity of permeabilized muscle fibres was assessed using high-resolution respirometry (Oxygraph-2k; Oroboros Instruments, Innsbruck, Austria) as previously described (Coen et al., 2013;Distefano et al., 2017Distefano et al., , 2018. Briefly, fibre bundles (∼1-3 mg) were gently teased apart, permeabilized for 20 min using saponin (50 μg mL -1 ) and washed in Buffer Z (pH 7.4 with KOH; 105 mm K-MES, 30 mm KCl, 10 mm KH 2 PO 4 , 5 mm MgCl 2 -6H 2 O, 5 mg mL -1 bovine serum albumin, 1 mm egtazic acid). Mitochondrial respiration was measured in Buffer Z with blebbistatin (25 μm) to prevent fibres from contracting. Assays were run in duplicate, at 37°C, in the range 400−200 nmol O 2 mL -1 . Complex I supported LEAK (LI) respiration was determined in the presence of malate (2 mm) and glutamate (5 mm). A saturated level of ADP (4 mm) was added to stimulate complex I supported oxidative phosphorylation (PI). Succinate (10 mm) was added to stimulate complex II respiration, eliciting maximal complex I and II supported oxidative phosphorylation (PI.II). Titration of carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (2 μm per injection) was then performed to determine maximal electron transfer system capacity (EI.II). Mitochondrial integrity was assessed by the addition of cytochrome c (10 μm), with values eliciting an increase in respiration >10% not being used for the analysis. Steady state oxygen flux was determined for each respiratory state using DatLab 5 (Oroboros Instruments) and normalized to fibre bundle wet weight. Respiratory measures have been previously published in a subset of participants from the young active, older active and older sedentary groups (Distefano et al., 2018).

Histological analysis for fibre type
Muscle biospecimens were used for histological analysis of fibre type as previously described (Rubenstein et al., 2022;Standley et al., 2020). A subset of participants (Young Active/Trained = 10 (six male/four female); Older Active/Trained = 12 (10 male/two female); Older Sedentary = 14 (eight male/four female) were used for this analysis as a result of insufficient tissue being obtained from the biopsy procedure from some participants. Mounted muscle samples were sectioned (10 μm) on a cryostat (Cryotome E; Thermo Shando, Pittsburgh, PA, USA) at −20°C and placed on glass slides. For muscle fibre type, sections were blocked in 10% goat serum and incubated overnight with primary antibodies directed toward type I [BA-F8 (IgG2b); Developmental Studies Hybridoma Bank (DSHB), Iowa City, IA, USA], type IIa [SC-71 (IgG1); DSHB] and type IIx (6H1 IgM; DSHB). The following day, sections were washed, incubated with fluorescence-labelled secondary antibodies [DyLight 405 (IgG2b), Alexa Fluor 488 (IgG1) and Alexa 555 (IgM); Thermo Fisher Scientific] and a coverslip with fluorescence mounting media was placed over the section (Prolong Gold; Thermo Fisher Scientific). Sections were imaged using an inverted fluorescence microscope (Nikon, Tokyo, Japan) and image analysis was performed using NIS Elements, version 4.20.01 (Nikon). Fibre type measures in a subset of the older active/endurance trained and older sedentary adults have been previously published (Rubenstein et al., 2022). A representative image containing all muscle fibre types, including hybrid (type IIa/IIx) muscle fibres, is provided in Figure S1 (Supplemental materials (Figure S1-S7) are available at URL https://figshare.com/s/611314d23cf7f190a44b; https: //doi.org/10.6084/m9.figshare.21983297).

Skeletal muscle metabolomics analysis
Metabolomic analyses were performed using non-targeted and targeted protocols as previously described (Hinkley et al., 2020;Tolstikov et al., 2014). Metabolite extraction from percutaneous muscle biopsies was achieved via cold homogenization in a solvent mixture of isopropanol:acetonitrile:water (3:3:2 v/v). Tissue samples were homogenized at 4°C using an Omni Bead Ruptor 24 (Omni International, Kennesaw, GA, USA) with muscle tissue samples combined with 400 μL of cold extraction solvent in 2 mL microtubes containing 1.4 mm ceramic beads (Omni International). Tissue slurries were transferred to 1.5 mL Eppendorf tubes for centrifugation (18,000 g for 10 min at 4°C) and the supernatant was collected and evaporated to dryness overnight using a SpeedVac Concentrator Savant DNA120 (Thermo Fisher Scientific). Samples were reconstituted in a ratio of 30 mg of tissue/400 μL of solvent (3:3:2 isopropanol:acetonitrile:water). Resuspended metabolite extracts were divided into three parts for separate analyses that included gas chromatography-time-of-flight-high-resolution mass spectrometry (GC-TOF-HRMS), reversed phase-liquid chromatography (RP-LC)-HRMS and hydrophilic J Physiol 601.11 interaction chromatography (HILIC)-LC-tandem mass spectrometry (MS/MS) instrumentation. For each of the analyses, a quality control was performed using metabolite standards mixture and pooled samples using the three separate instruments.
Samples for GC-TOF-HRMS analysis were initially dried using a SpeedVac Concentrator and pre-chilled samples were further dried using a Free-Zone lyophilizer (Labconco, Kansas City, MO, USA). Dried samples were then derivatized with methoxyamine hydrochloride in pyridine and N-methyl-N-trimethylsilyltrifluoroacetamide prior to analysis. Gas chromatography was performed using an Agilent 7890A gas chromatograph (Agilent, Palo Alto, CA, USA) that was interfaced to a high-resolution TOF Pegasus GC-HRT mass spectrometer (LECO, St Joseph, MI, USA). Automated injections were performed using a MPS2 programmable robotic multipurpose sampler (Gerstel, Muhlheim and de Ruhr, Germany). Acquisition parameters are available in Tolstikov et al. (2014). Data analysis was performed via ChromaTof (LECO).
Median data normalization (row scaling), logarithmic transformation and data scaling from GC-MS and LC-MS data were performed using MetaboAnalyst (https://www.metaboanalyst.ca) (median-normalized, log-transformed, autoscaled). A final metabolite table of raw peak areas (arbitrary units) was constructed by merging data acquired from all three methods. Peak lists from LC-MS/MS methods served as the basis for data merging, and duplicates from similar methodologies (HRMS) were removed.

Bioinformatic analysis
The metabolite profile was analysed with a user-defined bioinformatic procedure that included, raw data preprocessing, differentially expressed metabolite (DEM), agglomerative clustering and finally enriched correlation analysis. Briefly, raw the metabolite profile was log 2 -transformed, and quantile normalized. The normalized data were then analysed using Limma, an R/Bioconductor software package for metabolite discovery (Ritchie et al., 2015). Specifically, linear models were created for the study groups with sex as a covariate to identify metabolites differing significantly in expression at baseline (DEM-BS). Linear models also were created for paired samples, a special case of blocking with blocks of size two, for the analysis of the endurance exercise samples. The objective was to evaluate changes following an acute bout of endurance exercise (Post/Pre, 3 h Post/Pre, 3 h Post/Pre) (DEM-AT). Once established, the linear models were fitted using weighted least squares for each metabolite, moderation t statistics, moderated F statistic, log-fold changes (lgfc) and p values of differential expression computed by empirical Bayes moderation of the standard errors. Finally, the Benjamin-Hochberg method was used to adjust the p values.
Agglomerative clustering analysis was performed on a (g * k * N) lgfc matrix, a data frame with values of log 2 fold changes calculated from the DEM-AT analysis. The matrix represents exercise-driven metabolite dynamics detected in the muscle biopsy specimens. 'g' represents the number of study groups (young active, older active and older sedentary), 'k' is the number of timepoints where DEM analysis performed (Post/Pre, 3 h Post/Pre, 3 h Post/Pre) and 'N' is the number of DEMs identified. The clustering was performed in combination with Ward's minimum variance method and 'cutree' to minimize the total within-cluster variance and to establish optimal clusters Fig. S2A and B). Finally, the clusters were analysed using a principal component analysis (PCA), with factoextra, an R package to extract and visualize the output of multivariate data analyses Fig. S2 C and D) (Mundt, 2015). The objective is to assess the remodelling effects of the endurance exercise bout with respect to both the study groups and timepoint analysed (Post/Pre, 3 h Post/Pre and 3 h Post/Post). Once clusters of differentially expressed metabolites were established, correlation analyses were performed between these metabolites and clinical traits measuring muscle fibre proportion,V O 2 peak and mitochondrial respiration. The aim was to explore how changes in metabolites associated with the activity and ageing phenotypes.

Statistical analysis
Baseline group differences were examined using a one-way ANOVA with the group as the main factor, followed by Tukey's post hoc test. Chi-squared analysis was used to examine differences in the sex ratio (males/females) between the groups. For the acute exercise response, a repeated-measures ANOVA with Gross exercise efficiency (%) 19.7 ± 1.5 a 17.7 ± 2.4 a 13.6 ± 3.0 b <0.0001 Data are presented as the mean ± SD.V O 2 peak , maximal oxygen consumption during graded exercise test; mL kg -1 FFM min -1 , mL of oxygen consumed relative to fat-free mass per minute. Different superscript lowercase letters indicate significant differences between groups (P < 0.05). * Sex distribution across groups were determined by a chi-squared test.
group, timepoint and their interaction as factors in the models was used to determine the significant effect, followed by Tukey's post hoc test. Additionally, we performed repeated-measures analysis of covariance (ANCOVA, SAS Mixed procedure) with the group, timepoint and group × timepoint interaction as fixed effects, with sex and % heat rate reserve (HRR) as covariates and with the participant as the random effect. To adjust for the fact that we conducted multiple comparisons, we report false discovery rate (FDR)corrected P values (P < 0.05). Pearson partial correlation analyses were used to evaluate the relationship between individual metabolites and clinical phenotype measures at the same time as controlling for sex. Data are presented as the mean ± SD. All statistical tests were two-sided and analyses were performed in SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA) or Prism, version 6 (GraphPad Software Inc., San Diego, CA, USA). P < 0.05 was considered statistically significant.

Participant characteristics and muscle phenotyping
We performed metabolomics analysis and phenotyping of muscle biopsy specimens from three distinct study groups: young active (YA), older active (OA) and older sedentary (OS) (Fig. 1A). Participant characteristics are shown in Table 1. Per study design, physical activity levels and cardiorespiratory fitness were lower in the sedentary group compared to active individuals, whereas body mass was highest in the sedentary group. YA had higher cardiorespiratory fitness than OA and, at the level of the muscle, ex vivo mitochondrial respiration was highest in YA, with OS exhibiting the lowest levels of mitochondrial respiration (Table 2). YA and OA had a higher proportion of type I slow oxidative fibres, whereas OS had a higher proportion of type IIx fast glycolytic muscle fibres (Table 2). Together, these data indicate a higher oxidative phenotype in active/endurance trained adults that is evident at the whole-body (cardiorespiratory fitness) and muscle fibre (respiration) and is consistent with previous work (Coggan et al., 1990(Coggan et al., , 1992(Coggan et al., , 1993Distefano et al., 2018;Gries et al., 2018;Lanza et al., 2008).

Agglomerative clustering identifies phenotype-specific metabolite profiles in muscle
Metabolite profiling was performed with biopsy specimens obtained from the vastus lateralis of all participants at baseline prior to the acute bout of exercise (Pre) (Fig. 1A). We performed three separate analytical approaches (GC-TOF-HRMS, HILIC-LC-MS/MS and RP-LC-HRMS) to expand the total identifiable coverage of muscle metabolites and evaluated 431 metabolites in biopsy specimens from all participants (Tolstikov et al., 2014). We observed that 206 metabolites were differentially abundant amongst the groups in baseline samples. We performed agglomerative clustering to identify groups of metabolites that co-varied together within the dataset. Nine clusters with optimal levels of total within-cluster variance were identified that showed distinctive separation between clusters ( Fig. S2A and C). A full list of metabolites found in each cluster is provided in the Supporting information (Table S1). Strikingly, the abundance of the clustered metabolites varied in a phenotype-specific manner. The heatmap in Fig. 1B shows the row Z-score averages for each metabolite cluster, manually curated in a phenotype-specific manner.
Of the nine clusters, three appeared to be associated with the ageing phenotype, four clusters were associated with the active/trained phenotype and the remaining two clusters (clusters 3 and 7) contained metabolites that were Data are presented as the mean ± SD. LI, leak respiration through complex I of the electron transport chain; PI, maximal ADP-supported respiration through complex I of the electron transport chain; PI.II, maximal ADP-supported respiration through complexes I and II of the electron transport chain; EI.II, maximal electron transfer capacity of the mitochondria. Different superscript lowercase letters indicate significant differences between groups (P < 0.05).
highest in the OA group. In comparison with the robust differences amongst the phenotypical groups, only three metabolites (N-acetyl-l-lysine, capric acid and cytosine) were influenced by biological sex.

Chronic physical activity re-routes kynurenine metabolites in young and older muscle
We first evaluated metabolites from clusters 1, 2 and 4 that were associated with the ageing phenotype. Overall, 23 metabolites were elevated in YA in comparison to both OA and OS, whereas 59 metabolites were higher in OA/OS. Upon evaluation of individual metabolites found within these clusters, we observed several metabolites related to the kynurenine pathway. Although mostly known for its role in the brain, the effects of kynurenine metabolites on peripheral tissues have recently been highlighted (Cervenka et al., 2017). We observed elevated kynurenine (cluster 4) in OA compared to YA, whereas there was a trend (P = 0.07) for kynurenine to also be elevated in OS ( Fig. 2A). Muscle accumulation of kynurenine can arise from its uptake via the large neutral amino acid transporters (LATs) or catabolism of tryptophan (Martin et al., 2020). Tryptophan was found in cluster 6, an activity/trained associated cluster, and was elevated in OS muscle compared to YA and OA (Fig. 2B). The ratio of kynurenine to tryptophan, a proxy for activity of the major tryptophan degrading enzyme indoleamine-2,3-dioxygenase (IDO), was not different amongst the groups (ANOVA P = 0.0968) (Fig. 2C).
Taken together, we consider that these are the first human data to show age-and activity/training-related changes in kynurenine metabolite levels in skeletal muscle. Downstream metabolites of kynurenine, kynurenic acid and quinolinate (cluster 1) were both reduced in OS compared to YA ( Fig. 2D and E), whereas OA had intermediate levels. The degradation of tryptophan through the kynurenine pathway can result in the production of NAD + (Cervenka et al., 2017;Martin et al., 2020). Quinolinate, a metabolite produced from the initial hydroxylation of kynurenine, can be used to synthesize NAD + . Mechanisms that provoke an accumulation of NAD + levels are critical for older adults because multitissue reductions in NAD + is considered a hallmark of ageing across species (Monacelli et al., 2021). In line with the differences in quinolinate, muscle NAD + levels were reduced in OS compared to YA, whereas levels were preserved in OA (Fig. 2F). Collectively, these data support pre-clinical (Agudelo et al., 2014(Agudelo et al., , 2019 and clinical data (Agudelo et al., 2014;Allison et al., 2019;Janssens et al., 2022;Schlittler et al., 2016) suggesting that physical activity may reprogram muscle kynurenine and NAD + metabolism.
We next explored potential mechanisms that could explain the accumulation of metabolites downstream of kynurenine. The major enzyme that converts kynurenine to kynurenic acid is kynurenine aminotransferase (KAT). We observed differences in the ratio of kynurenic acid to kynurenine, a proxy of KAT activity, amongst groups, with a higher ratio in YA compared to both OA and OS (Fig. 2G). KAT activity was shown to be influenced isoleucine (I) were evaluated in muscle biospecimens obtained from young active (red circles, n = 15), older active (green circles, n = 13) and older sedentary adults (blue circles, n = 16). Metabolite levels are presented as raw peak areas per mg tissue weight (arbitrary units). Data are presented as the mean ± SD. * P < 0.05; * * * P < 0.001, one-way ANOVA with Tukey's multiple comparisons. Relationships between kynurenic acid levels and valine (J) and isoleucine (K) were examined in young active/endurance trained (red circles), older active/endurance trained (green circles) and older sedentary adults (blue circles by metabolites, particularly branched-chain amino acids (Badawy et al., 2014;Cervenka et al., 2017;Han et al., 2009). Numerous amino acids were found in cluster 2 (elevated in older muscle) and cluster 6 (elevated in older sedentary muscle). In particular, branched-chain amino acids valine and isoleucine were most abundant in muscle from OS, with lowest levels found in YA, whereas OA had intermediate levels ( Fig. 2H and I). In line with previous findings (Badawy et al., 2014;Cervenka et al., 2017;Han et al., 2009), higher levels of branched-chain amino acids were associated with lower kynurenic acid levels ( Fig. 2J and K).
To further evaluate whether alterations in kynurenine pathway metabolites were related to the active/trained phenotype, we correlated metabolite levels with whole-body and muscle phenotypic characteristics. Although kynurenine was not associated with any measures, kynurenic acid was positively associated with cardiorespiratory fitness (V O 2 peak ) and ex vivo mitochondrial energetics, whereas tryptophan was negatively associated (Fig. 2L). Conversion of kynurenine to kynurenic acid was shown to improve energy efficiency and exercise performance in rodents (Agudelo et al., 2019), with recent work showing a negative relationship between kynurenine levels and exercise efficiency in young and older humans (Janssens et al., 2022). In line with these reports, we observed that heightened kynurenic acid levels were associated with greater GEE (Fig. 2L). We next evaluated whether re-routing of tryptophan-kynurenine metabolites may be related to objectively assessed daily physical activity levels. We found that tryptophan levels were negatively associated with both average daily activity energy expenditure (EE) and steps day -1 , and was positively associated with daily sedentary time, whereas quinolinate was positively associated with average daily activity EE and steps day -1 (Fig. 2L). Kynurenic acid levels showed a trend similar to quinolinate, with a positive relationship with steps day -1 , and trended (P = 0.0536) towards a positive association with average daily activity EE (Fig. 2L). Furthermore, kynurenic acid levels were negatively associated with daily sedentary time (Fig. 2L). These exploratory data suggest that phenotypic characteristics of physical activity phenotype are associated with the re-routing of kynurenine pathway metabolites in human skeletal muscle. Future research using a larger cohort of participants, as well as cell and/or rodent models, is needed to confirm the involvement of kynurenine pathway metabolites on the oxidative profile of muscle.

Acylcarnitine metabolism is influenced by physical activity levels rather than ageing
Several metabolite species have been suggested to be involved in the ageing process, including an elevation in plasma acylcarnitines (Jarrell et al., 2020). Overall, 23 carnitine and acylcarnitine species were differentially abundant in muscle from the separate groups (Fig. 3A). However, only five of these metabolites were found in age-specific clusters. The putative alterations in acylcarnitine metabolism with ageing may stem from dysregulation of mitochondrial energetics in energy-consuming tissues such as muscle. As we and others have shown, the apparent age-related reductions in muscle mitochondrial energetics are probably driven to a large degree by secondary factors including a sedentary lifestyle and increased adiposity (Coggan et al., 1990;Distefano et al., 2017Distefano et al., , 2018Lanza et al., 2008;Zhang et al., 2021). We aimed to determine whether alterations in muscle acylcarnitine profiles in older adults are also linked to secondary factors (i.e. sedentary lifestyle) rather than ageing per se. To do this, we examined metabolites found in the activity/endurance training clusters (clusters 5, 6, 8 and 9) (Fig. 1B). We observed 100 metabolites in these clusters that were differentially abundant in muscle from YA and OA (48 higher and 52 lower) compared to OS. We found that several metabolites involved in acylcarnitine metabolism grouped in clusters that were highest in muscle from active adults, with one-quarter of metabolites found in clusters 5 and 9 being carnitine or acylcarnitine species (Fig. 3A; see also Supporting information, Table S1). Specifically, levels of muscle carnitine, as well as short-chain acylcarnitine species (including

. Acylcarnitine metabolism is elevated in active individuals and relate to clinical and biochemical characteristics of the oxidative phenotype
A, heatmap of carnitine and acylcarnitine species differentially expressed amongst the groups. Data are the group row Z-score average from transformed metabolomic data. Each column represents a group, and each row represents an individual metabolite. Baseline (pre-exercise) levels of carnitine (B), hydroxybutyrylcarnitine (C) and butyrylcarnitine (D) were evaluated in muscle biospecimens obtained from young active (red circles, n = 15), older active (green circles, n = 13) and older sedentary adults (blue circles, n = 16). Metabolite levels are presented as raw peak areas per mg tissue weight (arbitrary units). Data are presented as the mean ± SD. * P < 0.05; * * P < 0.01; * * * P < 0.001, one-way ANOVA with Tukey's multiple comparisons. Relationships between selected muscle metabolites carnitine (E, F) hydroxybutyrylcarnitine and butyrylcarnitine), were highest in the active groups, with no differences between the older and younger active groups (Fig. 3 B-D). In the circulation, acylcarnitine levels are reduced with exercise (Morville et al., 2020;Rodríguez-Gutiérrez et al., 2012;Xu et al., 2016;Zhang et al., 2017). Because increased circulating acylcarnitines are a potential marker of impaired metabolic health and are putative contributors to inflammation (Koves et al., 2008), our data suggest that a higher capacity to oxidize lipids (i.e. elevated mitochondrial energetics) may allow muscle to act as a metabolic sink in exercise trained individuals to reduce circulating acylcarnitines seen with ageing (Jarrell et al., 2020). We next aimed to determine how activity-related metabolites corresponded to clinical and biochemical characteristics associated with the oxidative phenotype. We performed Pearson correlation analyses on metabolites in clusters 5 and 9, which were elevated in active muscle, withV O 2 peak , ex vivo mitochondrial energetics, and muscle fibre type proportion. We then constructed a correlation network to map how the metabolites within these clusters interact with each of the phenotypic variables associated with exercise training. Our network analysis revealed links between muscle metabolite levels with cardiorespiratory fitness (V O 2 peak ) and ex vivo muscle mitochondrial respiration supported by complex I and II of the electron transport chain (PI.II) (Fig. S3). We observed that various carnitine-derived metabolites were positively associated withV O 2 peak and mitochondrial respiration (Fig. 3 E-H). Although the vast majority of acylcarnitines were associated witḣ V O 2 peak and ex vivo mitochondrial energetics, only a subset of metabolites within this class (butyrylcarnitine, S-2-hydroxyethyl-N-acetylcysteinylcarnitine and propionylcarnitine) were associated with the proportion of type I slow/oxidative muscle fibres (Fig. S3). This finding suggested that mitochondrial content, rather than fibre type per se, may be a primary driver for elevated acylcarnitine metabolism in endurance-trained muscle. Along with acylcarnitine metabolites, we also show that NAD + levels were positively related toV O 2 peak and mitochondrial respiration, proportion of type I muscle fibres (Fig. S3). These findings are remarkably similar to recent work by Janssens et al. (2022), who also saw a preservation in NAD + levels in older endurance trained adults, with metabolite levels positively associated with cardiorespiratory fitness and mitochondrial energetics. Together, these data suggest that several aspects of mitochondrial metabolism, including acylcarnitine and NAD + metabolism, may be a possible link between chronic physical activity and mitigation of the decline in cardiorespiratory fitness and mitochondrial energetics experienced by older adults.

Whole-body metabolic response to an acute bout of endurance exercise is similar amongst groups despite phenotypic differences
An acute bout of endurance exercise creates a dramatic metabolic demand on the body to meet the energetic requirements of contractile activity. Muscle accounts for ∼95% of energy requirements during moderate to vigorous exercise, with fuel provided by intramuscular stores and systemic sources (Goodpaster & Sparks, 2017;Romijn et al., 1993). It is suggested that muscle plasticity to resistance exercise is blunted in older adults (Breen & Phillips, 2011;Deane et al., 2019;Drummond et al., 2008;Fry et al., 2011;Haran et al., 2012); however, the metabolic response of muscle to an acute endurance exercise bout has not been fully explored in young and older adults, as well as phenotypically diverse older adults. To examine the metabolic response to exercise, all groups performed an acute bout of endurance exercise on a cycle ergometer (Fig. 4A). Despite distinct phenotypes, the groups displayed a similar response in whole-body substrate metabolism during and after the acute exercise bout. Specifically, participants exercised at a similar percent of their individualV O 2 peak (YA: 61.9 ± 6.3%; OA: 64.9 ± 6.0%; OS: 66.1 ± 12.5%; ANOVA = 0.4143) with a similar respiratory quotient (RQ; YA: 0.90 ± 0.04; OA: 0.88 ± 0.04; OS: 0.89 ± 0.06; ANOVA = 0.5518); however, the average percent heart rate reserve (HRR; YA: 74.3 ± 10.8%; OA: 66.9 ± 9.2%; OS: 62.5 ± 11.4%; ANOVA = 0.0126) during the exercise bout were higher in YA and lowest in OS. During the post-exercise recovery period, the three groups had a similar reduction in RQ (Fig. 4B) and increase in fat oxidation (Fig. 4C). This observation is in contrast to a study by Sial et al. (1996) and may be a result of differences in exercise intensity employed. The exercise intensity was ∼5% higher in younger adults compared to older counterparts in the study by Sial et al. (1996). Our findings closely resemble those reported by Tankersley et al. (1991), who showed that the respiratory quotient is similar between young and older adults when both groups exercise at a similar relative exercise intensity (%V O 2 peak ). The exercise-induced response of plasma lactate and free-fatty acids (FFA) was similar between    were obtained to examine the temporal changes in skeletal muscle metabolites in response to exercise. Along with muscle biopsies, indirect calorimetry was used to measure respiratory quotient [RQ; carbon dioxide produced (V CO 2 )/oxygen consumption (V O 2 )] (B) and whole-body fat oxidation (C) at baseline (Pre), immediately after (Post) and 3 h following (3 h Post) an acute bout of endurance exercise in young active (YA; red circles and lines, n = 13), older active (OA; green circles and line, n = 13) and older sedentary (OS; blue circles and lines, n = 14) adults.
Additionally, blood samples were obtained at the same timepoints to measure plasma lactate (D) and plasma free fatty acids (E). Data are presented as the mean ± SD. A repeated-measures ANOVA was used to determine significance with group, timepoint and their interactions as factors in the model. F, a principal component analysis (PCA) plot was generated for the top 50 metabolites differentially expressed with exercise in muscle biospecimens amongst the three timepoint comparisons (Post vs. Pre, black ellipsis; 3 h Post vs. Pre, red ellipsis; and 3 h Post vs. Post, green ellipsis). A large symbol found within ellipses (black square, red triangle and green circle) indicates the average amongst the entire dataset, whereas smaller symbols enclosed with a black symbol indicate group averages. G-I, heatmaps of skeletal muscle metabolite clusters that were altered by the acute exercise bout. Data are presented as row-normalized log-transformed data and clusters are ranked from greatest increase to greatest reduction in metabolites at each timepoint compared to the young active/trained group. Each column represents the average for each group ( groups, with lactate increasing immediately post-exercise and FFA increasing 3 h following exercise ( Fig. 4D and E). Although most studies report an immediate rise in plasma FFA immediately following exercise in the fasted state (Aird et al., 2018), we observed a delayed response.
It should be noted that participants were each given a low-glycaemic snack prior to exercise. Because feeding can alter various aspects of exercise metabolism, including reduced circulating FFA (Aird et al., 2018), we cannot exclude an impact of circulating FFA in the present study.

Metabolomic profiling reveals similar exercise-induced response in muscle of phenotypically different young and older adults
We next examined the metabolomic response to an acute bout of endurance exercise in muscle biopsies from the three groups. Biopsy samples were obtained from the vastus lateralis of all participants before, immediately after and 3 h following an acute bout of endurance exercise (Fig. 4A). Two hundred eighty-eight metabolites were altered in muscle at the immediate post and 3 h post-exercise samples. A PCA revealed a general separation between the timepoint comparisons (Fig. 4F).
To explore the exercise-responsive patterns amongst the groups, we performed agglomerative clustering analysis. Twelve separate clusters with optimal levels of total within-cluster variance ( Fig. S1B and D) were derived. The metabolites found in these clusters is provided in the Supporting information (Table S2). The magnitude of change in the metabolites between the timepoint comparisons (Post/Pre, 3 h Post/Pre and 3 h Post/Post) are presented in the heatmaps in Fig. 4 G-I. Differential metabolite analysis, presented as Venn diagrams and heatmaps in the Supplemental Figs S4-S6), revealed that, although there were metabolites that showed individualand/or phenotype-specific (i.e. age or activity/trained phenotypes) group differences, the overall metabolite response was generally similar amongst the three groups regardless of age and/or activity levels.

Responsiveness of muscle acylcarnitine and ketone body metabolites to an acute bout of endurance exercise
We next aimed to identify specific metabolite pools that were altered in muscle following an acute bout of endurance exercise. Acylcarnitines are elevated in human muscle following acute exercise (Decombaz et al., 1992;Hiatt et al., 1989;Minkler et al., 1995). Consistent with these findings, our analysis revealed that many metabolites within the exercise-responsive clusters were involved in acylcarnitine metabolism. Forty-two carnitine and acylcarnitine species were differentially abundant in response to exercise, with nine of the 12 clusters (see Supporting information, Table S2) containing carnitine and acylcarnitine metabolites. The heatmaps in Fig. 5A and B illustrate the change in muscle acylcarnitines immediately following exercise (Post/Pre) or 3 h into recovery (3 h Post/Post). The increase in acylcarnitines was accompanied by a reduction in carnitine levels in muscle immediately post-exercise which remained lower 3 h into recovery (Fig. 5C, F, and I). Acylcarnitines varied widely with respect to any further change in their levels 3 h following the bout of exercise. Specifically, butyryl carnitine increased immediately post-exercise and returned to or below baseline levels at the 3-h post-exercise timepoint (Fig. 5D, G, and J), whereas hydroxybutyrylcarnitine remained elevated at the later stages of exercise recovery (Fig. 5E, H and K). The levels of both butyrylcarnitine and hydroxybutyrylcarnitine at rest and in response to exercise were positively associated with whole-body fat oxidation at rest and following acute exercise was observed ( Fig. S7A and B), suggesting short-chain acylcarnitines may partially fuel the increased reliance of fats during the post-exercise recovery period.   Although the body primarily relies on carbohydrates and fats to fuel contractile activity, recent work has shown ketone bodies play an important role during and after exercise (Evans et al., 2017). In the current dataset, we observed elevated ketone bodies (2-hydroxybutyric acid and 3-hydroxybutric acid) in muscle following the acute bout of exercise ( Fig. 6A and B). Although unchanged immediately post-exercise ( Fig. 6C and D), ketone body accumulation was prevalent 3 h following the bout of exercise ( Fig. 6E and F), with a similar response amongst all the groups. FFA released from adipose tissue are imported by the liver and converted to ketone bodies which, in turn, are released into circulation followed by accumulation within extra-hepatic tissues (i.e. muscle). Interestingly, the changes in muscle ketone bodies into recovery tracked with the response of plasma FFA (Fig. 4E). Our correlative analysis revealed muscle 2-hydroxybutyric acid (Fig. 6G) and 3-hydroxybutyric acid (Fig. 6H) were positively associated with plasma FFA at rest and in response to an acute bout of endurance exercise. These data suggest that exercise-induced lipolysis may fuel liver ketogenesis, leading to a release and eventual accumulation of ketone bodies within muscle.
Finally, due to the group differences in %HRR observed during the exercise bout, we determined whether interparticipant differences in exercise intensity may influence the muscle metabolic response to acute exercise. To evaluate this, we performed repeated-measures ANCOVA (SAS Mixed procedure) with group, timepoint and group × timepoint interaction as fixed effects, with sex and %HRR as covariates, and with the participant as the random effect. We found that variance in %HRR did not impact exercise-induced changes in acylcarnitines (hydroxybutyrylcarnitine and butyrylcarnitine) and ketone body (2-hydroxybutyric acid and 3-hydroxybutyric acid) metabolite levels amongst the groups (data not shown). Together, these data suggest that, along with age and activity levels, exercise intensity does not appear to influence changes in muscle metabolite pools seen with acute endurance exercise.

Discussion
Combining molecular tissue profiling with deep human phenotyping approaches can play a key role in identifying cellular mediators of health and disease in humans (FitzGerald et al., 2018). In the present study, we unveiled novel metabolite profiles that define phenotypic and exercise-induced characteristics in muscle of young active, older active and older sedentary adults. We show for the first time in human muscle the impact of age and activity on metabolites of the kynurenine/tryptophan pathway. Specifically, tryptophan, kynurenine, kynurenic acid and quinolinate pools were altered with age, and correlated with muscle mitochondrial respiration, exercise efficiency and daily physical activity. Additionally, NAD + , an end-product of the kynurenine pathway, as well as acylcarnitine metabolite levels, appear to relate to clinical and biochemical features of the physically active phenotype. Despite these baseline differences, the response of various metabolite pools to a bout of endurance exercise was similar amongst the groups, including an elevation in muscle acylcarnitine and ketone bodies that correlated with changes in whole-body fat oxidation and circulating free-fatty acids. These data suggest that age-and/or physical activity levels do not impact the muscle metabolite response to acute endurance exercise.
Recent work has shown that kynurenine may mediate chronic inflammation, leading to increased frailty in older adults (Westbrook et al., 2020). Exercise appears to alter the metabolic fate of kynurenine because trained muscle contains higher levels of KAT activity, catalysing the conversion of kynurenine to kynurenic acid (Agudelo et al., 2014;Allison et al., 2019;Schlittler et al., 2016). The impact of acute and chronic physical activity on kynurenine metabolism has recently been explored in Figure 6. Ketone body accumulation correspond to changes in circulating free-fatty acids following acute exercise in young and older adults Individual response of 2-hydroxybutyric acid (A) and 3-hydroxybutyric acid (B) in muscle biospecimens obtained from young active/endurance trained (red circles and lines, n = 13), older active/endurance trained (green circles and lines, n = 13) and older sedentary (blue circles and lines, n = 14) adults at baseline (Pre), immediately following (Post) and 3 h after (3 h Post) an acute bout of endurance exercise. Metabolite levels are presented as raw peak areas per mg tissue weight (arbitrary units). Data are presented as the mean ± SD. A repeated-measures ANOVA was used to determine significance with group, timepoint and their interactions as factors in the model. Absolute change in 2-hydroxybutyric acid (C-E) and 3-hydroxybutyric acid (D and F) at the immediate post-exercise timepoint in comparison to baseline (Post-Pre, C and D) and at 3 h following the bout of exercise in comparison to immediately post-exercise (3 h Post-Pre, E and F) in young active/endurance trained (red circles), older active/endurance trained (green circles) and older sedentary (blue circles) adults. Dashed line indicates no change (delta = 0). Data are presented as the mean ± SD. A one-way ANOVA was used to determine significance between groups. No significant differences were observed between groups. Relationships between plasma free fatty acids (Mean FFA) and 2-hydroxybutyric acid (G) or 3-hydroxybutyric acid (H) were examined across three timepoints [before exercise (PRE), circles; immediately post-exercise (Post), squares; 3 h following exercise (3 h Post), triangles] in young active/endurance trained (red symbols), older active/endurance trained (green symbols) and older sedentary adults (blue symbols). Pearson correlation coefficient (r) and P values are presented. A dotted line indicates the 95% confidence interval. [Colour figure can be viewed at wileyonlinelibrary.com] muscle metabolically healthy and unhealthy (i.e. type 2 diabetics) human participants (Janssens et al., 2022;Jonsson et al., 2022;Martin et al., 2020;Savikj et al., 2022). Our current study addresses an important knowledge gap in the ageing literature by evaluating the distribution of kynurenine metabolites in muscle of young and older adults. Although kynurenine levels were elevated with age, trained individuals had higher levels of kynurenic acid regardless of age. Our findings are in contrast to those of Janssens et al. (2022), who observed an elevation in resting muscle kynurenic acid with age, with metabolite levels negatively associated with molecular and physiological parameters of healthy ageing. A possible explanation for the differences between studies is that we observed relationships between muscle kynurenic acid levels with a broader range of physical activity parameters, includinġ V O 2 peak , daily physical activity, energy expenditure as a result of physical activity and daily sedentary time. Together, our data are in line with pre-clinical (Agudelo et al., 2014(Agudelo et al., , 2019 and clinical data (Agudelo et al., 2014;Allison et al., 2019;Janssens et al., 2022;Schlittler et al., 2016) suggesting that kynurenine metabolism is sensitive to chronic physical activity.
Accumulation of kynurenine pathway metabolites within skeletal muscle have recently been described (Martin et al., 2020). Kynurenine accumulation is derived from either conversion of tryptophan by the IDO enzyme or direct uptake of kynurenine through the LAT transporter (Cervenka et al., 2017;Martin et al., 2020). IDO levels are low in muscle (Martin et al., 2020) and a lack of age-related differences in the ratio of tryptophan to kynurenine further indicates the accumulation of kynurenine may have occurred because of direct uptake by the muscle. In comparison with kynurenine, the accumulation of kynurenic acid in muscle appears to be dictated by the activity of KAT enzymes (Martin et al., 2020). In the present study, we observed an increase in kynurenic acid in muscle of trained individuals regardless of age, despite an age-related decrease in the kynurenic acid to kynurenine ratio, a proxy to KAT activity. Along with kynurenic acid, we also observed an increase in quinolinate levels in muscle of young and older active individuals. Quinolinate can be processed to NAD + (Cervenka et al., 2017;Martin et al., 2020), which we also observed in active individuals. However, the contribution of kynurenine pathway metabolites to the NAD + pool in active muscle is unclear and requires future investigation.
The cross-tissue benefits of exercise have spurred the pursuit of signalling molecules stimulated by muscle contraction that mediate communication across tissues (Chow et al., 2022;Murphy et al., 2020). Various tissues, including muscle, can act as an endocrine organ and secrete signalling molecules in response to exercise that may affect phenotypic characteristics of different tissues (Murphy et al., 2020). Lactate was the first signalling molecule shown to be modified by contractile activity (Fletcher, 1907). Current MS approaches enable a more comprehensive search for signalling molecules provoked by exercise. Exercise has a protective effect on brain health, and kynurenine pathway metabolites may be involved in this protection. Specifically, circulating levels of kynurenine are elevated in individuals with depression (Myint et al., 2007), with chronic endurance exercise training increasing the conversion of the neurotoxic kynurenine to kynurenic acid, which cannot cross the blood-brain barrier and is associated with protection against stress-induced depression (Agudelo et al., 2014). The contribution of kynurenine pathway metabolites can occur from several organ systems, with kynurenine mainly derived from the liver, whereas the conversion of kynurenine to kynurenic acid with exercise training is proposed to occur in muscle (Agudelo et al., 2014;Cervenka et al., 2017;Martin et al., 2020). However, recent evidence suggests that changes in systemic kynurenine metabolites with different perturbations may stem from tissues systems beyond muscle because Jonsson et al. (2022) have recently shown that the systemic reductions in kynurenine pathway metabolites following combined ingestion of branch-chained amino acids and one-legged endurance exercise were not associated with metabolite changes in muscle. In the present study, we did not examine changes in circulating levels of kynurenine J Physiol 601.11 pathway metabolites; therefore, how the changes in muscle kynurenine pathway metabolites with age and exercise training contribute to circulating levels remains unclear. Furthermore, whether the baseline differences in kynurenine metabolism produce improvements in ageing brain health, such as cognitive function and depressive symptoms, requires further investigation.
Exercise is a proven strategy to preserve cardiorespiratory fitness and muscle quality in older adults (Cartee et al., 2016;Gries & Trappe, 2022). Increased metabolic rate following acute exercise levels aids in the recovery of substrate stores, as well as stimulating cellular adaptations that, when repeated over time, can lead to phenotypic changes (Egan & Zierath, 2013). Muscle of older adults is purportedly less responsive to resistance exercise than young counterparts, leading to a dampened adaptative response to exercise training (Haran et al., 2012). By contrast, endurance exercise training was shown to stimulate similar metabolic adaptations between young and older participants (Harber et al., 2012;, suggesting that exercise responsiveness is retained with age. However, the metabolic response to an acute bout of endurance exercise has not been fully explored in older adults. In the present study, we demonstrated that young and old participants, regardless of phenotype, showed a similar reliance on fat oxidation into recovery. This whole-body response to exercise was translated to the level of the muscle. Short-chain acylcarnitines, a byproduct of fat and amino acid oxidation, were elevated to a similar extent in young and phenotypically diverse older adults. Interestingly, the change in acylcarnitines, which were lower at baseline in the older sedentary group, led to similar metabolite levels amongst the three groups at the immediate post-exercise timepoint, suggesting that exercise may acutely alleviate the metabolic defects seen in muscle of older sedentary individuals. These data provide evidence that the metabolic responsiveness of muscle is retained with age in active/trained and sedentary individuals. The present study provides evidence of novel ageand activity-related metabolite signatures in young and older adults. However, there are some limitations that should be considered. The study design did not include a young sedentary group to permit determination of activity-related metabolite signatures in muscle in young adults. Furthermore, although we accounted for biological sex in our analysis, the low number of female participants limited the ability to understand the role of sex on muscle metabolite distribution with age and physical activity, as well as in response to an acute bout of exercise. This is important because recent rodent work has shown sex dimorphism in the molecular response to exercise in multiple tissues (Group et al., 2022). Collectively, additional research with a larger cohort of individuals across the age and activity continuum is needed to confirm our findings and further understand the impact of age, activity/exercise status and biological sex on muscle metabolite distribution.
In conclusion, we have identified the kynurenine metabolites, acylcarnitines and NAD + as unique metabolite signatures in skeletal muscle that are associated with key phenotypic characteristics of young and older adults, including cardiorespiratory fitness and muscle mitochondrial energetics. Despite differing baseline profiles, the impact of exercise on the metabolite pool was remarkably similar amongst the experimental groups. Our data support the engagement in endurance exercise as a therapeutic intervention to improve muscle metabolism and ultimately healthspan for older adults.