Resistance exercise (RE) gained much attention to counteract sarcopenia and cachexia [23] and is considered by the WHO as an essential training mode to maintain health. [24] Yet, it is not fully understood how RE affects distinct metabolic pathways in humans. Recently, Morville and colleagues investigated differences in the time-resolved plasma metabolome responses between acute RE and EE. [25] The authors determined 833 metabolites related to lipid-, carbohydrate-, energy, nucleotide and amino-acid metabolism and detected that more metabolites were increased after EE than in RE and, conversely, more metabolites were decreased after RE when compared to EE. Their results attribute RE as a unique exercise mode exhibiting partly a distinct metabolomic response when compared to EE. However, the plasma and blood metabolome does not always overlap, why the specific analysis of the skeletal muscle metabolome itself is essential. [12]
Our analysis of the skeletal muscle metabolome in response to resistance exercise (RE) provides valuable novel information for understanding the relative impact of RE on muscle metabolism under different training states. Here we widened the spectrum of metabolites and metabolic pathways investigated in muscle tissue in response to acute RE, allowing us to complement previous findings in blood 26. For well-known RE-related phenomena such as protein degradation/damage and degradation of ATP, the effect directions of related metabolites in muscle were largely consistent with those previously observed in blood, mostly confirming these results in the relevant tissue. A notable exception was the primary bile acid chenodeoxycholate, with potential signaling function for muscle growth and atrophy. [26] We observed increased chenodeoxycholate levels in response to acute RE in muscle, while levels in blood were previously reported to be decreased. [25] We further detected changes in metabolites after repeated RE that, to date, have not been analysed in human skeletal muscle biopsies. Among these, metabolites such as beta-citrylglutamate and chenodeoxycholate may play a deeper role in the adaptive process of skeletal muscle towards exercise. [26,27]
In addition to samples before and after a bout of RE in the untrained state, we compared post-exercise muscle samples of the same subjects in the unadapted and adapted states, before and after a prolonged RE intervention during which skeletal muscle has hypertrophied. Our results reveal that next to muscle growth, chronic resistance training additionally changes clusters of various intramuscular metabolites that were not directly affected by a bout of exercise in the untrained state in our study.
Metabolites associated with increased protein turnover after unaccustomed RE
As reported for blood plasma [25] , the modified amino acid 1-carboxyethylvaline also increased in muscle tissue in our study after acute unaccustomed RE. Carboxyethyl-derivatives of amino acids are typically generated through oxidation of glycated proteins [21], and their abundance is in concordance with protein damage and beginning protein breakdown in muscle samples of our study. [13]
Concordant with changes in 1-carboxyethylvaline, 3-(4-hydroxyphenyl)-lactate, a metabolite of tyrosine metabolism [28], significantly increased in response to acute RE. It is known that tyrosine levels in plasma increase after exhausting exercise and elevated amino acid-levels are also associated with increased protein degradation during exercise. [29,30] Hence, the increase in 3-(4-hydroxyphenyl)-lactate may indicate augmented tyrosine metabolism in response to acute RE-induced protein degradation.
In the context of protein damage, also 3-methylhistidine, an established marker for protein breakdown [31], significantly increased in muscle after RE. Most of the 3-methylhistidine is derived from skeletal muscle actin and myosin. [32] Indeed, RE induced myofibrillar damage and z-disk disruption was observed in muscle samples of our subjects [13] which is usually associated with a resulting increase in protein breakdown. [33] Proteasomal degradation as well as autophagic pathways like chaperone-assisted selective autophagy regulate the degradation and disposal of damaged skeletal muscle proteins which we already determined in samples of this study.[13,14]
In summary, changes in several metabolites detected in our study indicate increased protein breakdown in the early phase after RE.
Metabolites associated with the antioxidative system after unaccustomed RE
We further determined that the level of CoA-glutathione (CoA-GSH) in muscle decreased in response to acute unaccustomed RE. CoA-GSH is a metabolite of the GSH antioxidative system, which - in its reduced form - is a strong antioxidant in vivo. [34] During muscular work GSH effectively scavenges hydroxyl ions which are increasingly generated [35], resulting in the GSH oxidation product glutathione disulfide (GSSG). GSG-reductase replenishes GSH levels by reducing GSSG under NADPH usage. This mechanism might explain why immediately after intense RE increased GSH levels were observed while GSSG was reduced in human blood in a previous study. [36] In rat skeletal muscle high intense treadmill exercise increased GSH and GSSG whilst the ratio was maintained. [37] Analogously, CoA-GSH-reductase catalyses the reaction of the disulfide CoA-GSH to GSH and CoA [38] which may also support the maintenance of GSH, thereby reducing CoA-GSH levels as observed in our study. Based on previous findings obtained from blood plasma, our results in muscle likely reflect the impact of RE on increased oxidative stress in skeletal muscle. [39]
Interestingly, CoA-GSH clustered with three medium-chain dicarboxylic fatty acid metabolites showing concordant decreases post RE. Dicarboxylic fatty acids are generated through ω-oxidation of fatty acids and their increased metabolisation is observed in the fasting state [19] and it is known that GSH preserves fatty acids from being oxidized. [40] Furthermore, the dicarboxylic fatty acid sebacate supports glucose uptake of muscle cells [41] which is important for skeletal muscle during RE and to restore substrates after exercise.
Metabolites that support a growth-related environment in skeletal muscle in response to RE in the trained and untrained state
In contrast to recent observations in blood plasma [42], we saw a significant increase of the unconjugated primary bile acid chenodeoxycholate (CDCA) 45 min after RE in untrained skeletal muscle. Bile acids, mainly known for their function as detergents for ingested lipids in the intestine, have also been identified as signaling molecules, regulating energy metabolism via feedback mechanisms with endocrine factors of the fibroblast growth factor (FGF) family. [43] Regarding muscle growth after RE, FGF19 and its interrelation with CDCA is of particular interest: (i) pharmacological administration of FGF19 has been shown to induce hypertrophy in skeletal muscle of mice [44]; (ii) CDCA has been shown to induce FGF19 mRNA expression in primary human hepatocytes. [45] Considering this evidence, our finding that intramuscular CDCA levels significantly increased in response to acute RE, is in line with an induction of the CDCA/FGF19 axis through RE as previously proposed. Our findings therefore oppose the assumption that a decrease of CDCA in blood plasma after RE can be transferred to muscle and indicate a downregulation of FGF19 signaling. [42] The decrease of circulating CDCA as observed by Moreville et al., might be explained by increased uptake of CDCA by muscle tissue and thus does not necessarily contradict an induction of the bile acid/FGF19 axis through RE as hypothesized by the authors. We show here that substantial differences can exist between the plasma- and skeletal muscle metabolome and that the extrapolation from blood to muscle and vice versa can be misleading.
More studies are required to address the impact of CDCA and related metabolites on muscle adaptation and how plasma levels will be timely regulated compared to muscle tissue. Nonetheless, our study shows for the first time that this metabolite is likely involved in the skeletal muscle environment under conditions of tissue damage and muscle growth.
Beta-citrylglutamate (BCG) significantly increased after chronic training, i.e., when comparing metabolites after a bout of exercise between the unadapted and the adapted state. Though known for a while, the actual physiological role of beta-citrylglutamate is still elusive. It was first identified in newborn rat brain and testis and has been linked to brain development. [46] The brain level of BCG is the highest when the proliferation rate of neurons in cerebral cortex is high and it decreases when neurons mature. [47] In primary cultures of neurons from newborn mouse brain, BCG, which was suggested to serve as an Fe-carrier for aconitase [46], enhanced cell viability by accelerating mitochondrial activity. Rats that are infertile due to germ cell depletion show low beta-citrylglutamate concentrations suggesting its involvement in the metabolic support of cell proliferation. [48] BCG is a physiological substrate of the enzyme beta-citrylglutamate synthase-B, encoded by the RIMKLB gene. [49] In humans it is mainly expressed in testes but also other tissues including skeletal muscle. [50] Genetic variants in this gene were found to be associated with BCG (X-12748) levels in blood. [51] Interestingly, a meta-analysis of human muscle biopsy studies showed that RIMKLB expression increased by 43% (log2 0.52) after a bout of RE. [52] Therefore, it may be hypothesized that increased levels of BCG could support satellite cell responses that accompany skeletal muscle growth conditions also in shorter time frames than applied in our intervention. [53] While we did not observe an increase of BCG (p<0.05) after acute exercise in our study, a closer inspection of the individual metabolite profiles revealed that in fact four of six participants did show an increase in BCG (Supplementary Figure S1).
From the first unaccustomed bout to the last bout after chronic RE training, BCG-levels increased in all subjects including those for whom no increases were detected in acute unaccustomed exercise. This could indicate that repeated stimulation of muscle increasingly involves the synthetic pathway of BCG. However, as no second baseline biopsy was taken after repeated training, we cannot definitely say whether resting levels were also higher after 5 weeks or the acute response after RE was augmented.
Metabolites reflecting acutely increased energy and nucleotide metabolism in response to RE in the trained and untrained state
In metabolically active tissues the levels of NAD+/NADH and nucleotides are tightly linked to each other and ensure high rates of ATP turnover. [54] Hence, it is comprehensible that we observed changes in metabolites related to nucleotide and NADH turnover also in skeletal muscle. Increased ATP breakdown due to EE but also RE, exerts a rise in adenine nucleotide metabolites. [55] Those metabolites are metabolized within the purine nucleotide cycle to hypoxanthine, xanthine and xanthosine [56] which were shown to be elevated up to 180 min after EE and RE in plasma. [25] Hence, our study reveals that muscle is the origin for changes in blood of those nucleotides. Generally, the concentrations of nucleotide degradation products increase most after intense muscle contractions. Acute RE increases xanthine oxidase levels from 45 min up to 96 hours in human skeletal muscle. [57] Xanthine oxidase converts hypoxanthine to xanthine. Therefore, the strong increases in these metabolites including xanthosine likely reflects the impact of acute RE on adenine-nucleotide metabolism.
Xanthine can be further metabolized to IMP and xanthosinemonophosphate (XMP). [58] Interestingly, IMP decreased after acute RE in the unaccustomed state (Figure 4). Assuming that IMP will become rapidly metabolized within 45 min after RE in muscle, this may explain why IMP levels have declined while xanthine is still elevated. Indeed, lower IMP levels were also observed in human skeletal muscle acutely after sprint training. [59]
Xanthosine was significantly increased after the first bout of acute RE. Comparing the levels after a single bout of RE in the trained versus the untrained state, xanthosine was lower after the 5-week training, which could indicate a reduced increase of levels in response to a bout of RE after training or a reduced level at rest after training.
The nicotinamide adenine dinucleotide (NAD) metabolite N1-methyl-2-pyridone-5-carboxamide (2PY), is a metabolite of NAD-degradation. [60] Catabolism of NAD increases significantly during exercise when oxidative metabolism increases. [61] NAD is involved in ATP production and its turnover can be measured by the urinary outputs of 2PY and 4PY. Increased levels therefore likely reflect an acutely increased NAD turnover induced by an unaccustomed bout of RE. Increased levels of 2PY were detected in urine of mice subjected to a high fat diet and an associated increase of oxidative metabolism suggesting that 2PY reflects oxidative capacity. [62] Endurance exercise significantly enhances oxidative capacity of mitochondria and the rate of NAD turnover. [61] So far, specific changes in 2PY metabolite levels in human skeletal muscle in response to RE have not been investigated. But it has been shown that prolonged RE increases NAD and NADH levels in human subjects. [63] Similar as for xanthosine, we observed a decrease of 2PY when comparing the post-exercise levels in the trained versus the untrained state. This response may be explained by a reduced activity of enzymes involved in the generation of 2PY, a generally reduced requirement for NAD turnover after training or also an increased speed of 2PY excretion which may decline muscle 2PY levels.
Metabolites reflecting changes in the skeletal muscle lipid profile after a period of resistance training
Increased fatty acid metabolism has been shown to occur during and after EE but also after RE in skeletal muscle. [64,65] However, it is not clear in which direction repeated RE adapts lipid-derived metabolites in skeletal muscle. We detected that a cluster of phospholipids, mostly consisting of metabolites of glycerophosphoethanolamine plasmalogens, were reduced after RE in the trained state (Figure 5). Plasmalogens are a unique class of glycerophospholipids and important components of membrane structures. Plasmalogens are enriched in kidney, lung, and skeletal muscle. [66,67] Although increased metabolism of plasmalogens in resistance exercising human skeletal muscle to date have not been specifically determined, RE-induced changes could be assumed in a tissue that is constantly stimulated for tissue growth and challenged to maintain proteostasis.
Metabolites reflecting the modulation of skeletal muscle energy metabolism after a period of resistance training
Chronic RE training reduced fructose 1-6-bisphosphate (FBP) and acetylphosphate levels compared to unaccustomed exercise in our study. While FBP is a glycolytic intermediate generated by phosphofructokinase (PFK), acetylophosphate resynthesizes ATP. Chronic RE can increase oxidative capacity of skeletal muscle while glycolytic capacity rather decreases. [68] A reduction in glycolytic capacity in combination with increased oxidative capacity may have reduced levels of FBP in our study by a reduction in PFK activity and increased oxidative metabolisation of glycolytic products. We can only speculate about those events since we have not measured the levels or activity of the related enzymes.
Metabolites reflecting changes in the profile of N-acetylated ketogenic amino acids after a period of resistance training
After chronic training, we observed decreased levels of N-acetylated ketogenic amino acids (leucine, phenylalanine) and valine and the ketone body 3-hydroxybutyrate (BHBA). Usually, RE significantly augments the uptake of leucine which then stimulates mTORC-1 activity and consequently the early RE-induced increase in myofibrillar and sarcoplasmic protein synthesis. [69] Because expression of amino acid transporters is augmented in response to acute RE [70], increases in metabolites associated with BCAA metabolism would have been expected in skeletal muscle. Instead, we detected a reduced abundance of their N-acetylated forms. Acetylation is a very common physiological mechanism which alters the function of proteins. Acetylated proteins are known to increase in skeletal muscle in response to exercise. [71] Intense exercise has been shown to change the acetylation of histones and mitochondrial proteins. [72,73] Importantly, N-acetylated amino acids e.g. N-acetylleucine have opposite roles than the non-acetylated forms and N-acetylleucine has been shown to block p70s6k activation and induce a cell cycle arrest in cells. [74] In this regard it may be beneficial for muscle adaptation when the levels of N-acetylated amino acids are reduced. However, it is still unclear what their fate in skeletal muscle is and whether those amino acids are increasingly metabolized or the general acetylation is blunted. Another pathway that is activated by RE and involved in the regulation of protein synthesis and muscle hypertrophy is the cancer-like reprogramming inducing the synthesis of serine and glycine via PHGDH. [8,75] RE stimulates the expression of pyruvate kinase muscle 2 (PKM2), which in C2C12 myotubes has been shown to induce hypertrophy. [76] The present results, do not show a significant increases in serine and glycine levels, however, this could be due to the fact that these metabolites are immediately incorporated in the synthesized contractile and cytoskeletal proteins.
Interestingly, BHBA (3-hydroxybutyrate) significantly decreased after chronic training concordantly with the N-acetylated amino acids in our study. BHBA is a ketone body originating from lipid or protein metabolism and is increased during starvation or endurance exercise. [77] It serves as a fuel source in skeletal muscle and some authors recognized a correlation between BHBA levels as well as skeletal muscle function and cognitive capacity. [78] Interestingly, endurance trained humans have a significant greater capacity to oxidize ketones during exercise and therefore show reduced levels after exercise. [79] Based on this observation, reduced levels in our study may reflect either a training induced increase in the capacity to oxidize BHBA at the end of the study or also a chronic reduction in baseline levels.
Limitations
While we see the use of human samples as a particular strength of our study, we must acknowledge that our sample size is small and limiting statistical power. Furthermore, 45 min post RE was the only time point where we collected muscle biopsies after acute resistance exercise. At this time point, protein synthesis and degradation as well as restoration of tissue homeostasis and integrity is still being regulated. [13,33] Indeed, intense exercise stimuli affect molecular events for several days after stimulation. It can be assumed that also several hours after stimulation the metabolome is still changing. Therefore, we have detected only a snapshot of a changed intramuscular metabolome and further metabolites that may contribute to increased muscle anabolism and hypertrophy might not have been detected at this time point. [80] A further limitation is that we did not collect a second baseline biopsy at the end of the training period. As a consequence, we are not able to clearly differentiate whether changes in response to chronic training occur (i) due to changes in resting levels between the untrained and trained state (while the acute exercise-induced change remains the same between states) or (ii) due to changes in the steepness of increase or decrease in response to the acute exercise stimulus between the untrained and trained states (while the resting levels remain the same between states). Finally, whilst the measurement of steady state metabolite concentrations is useful to identify changes in the network of metabolic pathways, it does not inform about actual metabolic fluxes.