Abstract
nderstanding the metabolic processes in energy metabolism, particularly during fasted exercise, is a growing area of research. Previous work has focused on measuring metabolites pre and post exercise. This can provide information about the nal state of energy metabolism in the participants, but it does not show how these processes vary during the exercise and any subsequent post-exercise period. To address this, the work described here took fasted participants and subjected them to an exercise and rest protocol under laboratory settings, which allowed for breath and blood sampling both pre, during and post exercise. Analysis of the data produced from both the physiological measurements and the untargeted
how these processes vary during the exercise and any subsequent post-exercise period. To address this, the work described here took fasted participants and subjected them to an exercise and rest protocol under laboratory settings, which allowed for breath and blood sampling both pre, during and post exercise. Analysis of the data produced from both the physiological measurements and the untargeted metabolomics measurements showed clear switching between glycolytic and ketolytic metabolism, with the liquid chromatography-mass spectrometry (LC-MS) data showing the separate stages of ketolytic metabolism, notably the transport, release and breakdown of long chain fatty acids. Several signals, putatively identi ed as short peptides, were observed to change in a pattern similar to that of the ketolytic metabolites. This work highlights the power of untargeted metabolomic methods as an investigative tool for exercise science, both to follow known processes in a more complete way and discover possible novel biomarkers. Keywords: energy metabolism; sports metabolism; nutritional metabolomics; biomarker research; liquid chromatography-mass spectrometry (LC-MS); fasted exercise Metabolites2020,10, 399; doi:10.3390 /metabo10100399 /journal/metabolites
Metabolites2020,10, 399 2 of 11 1. Introduction The current understanding of metabolic processes in uenced by exercise is based mainly on targeted analysis methods as described by Egan et al. [1]. Changes in over 200 metabolites were examined using liquid chromatography mass spectrometry (LC-MS) by Lewis et al. [2]. Signi cant changes were identi ed in several metabolites in 25 subjects whose blood was collected before and after the Boston Marathon. These included large increases in lipolysis and 3-hydroxybutyrate, the endpoint of ketolytic metabolism. Van Hall et al. have shown that during high intensity exercise glycolytic metabolism is raised, resulting in increases in lactate [3]. Targeted methods have been used to investigate other indicators of energy metabolism, such as the work by Gibala et al. showing increased ux in the tri-carboxylic acid (TCA) cycle in skeletal muscle during moderate exercise and after intense exercise [4]. While these methods have allowed for the analysis of markers of exercise and health, a large number of intermediate metabolites that are key to fully understanding the biochemistry present are not seen. Using metabolomics, fundamental biochemical pathways involved in exercise can be probed using a single analytical platform, whereas targeted methods may require a wide range of analytical tools. The current state of metabolomics in sport and exercise science has been described by Heaney et al. [5], highlighting the use of targeted methods and exciting prospects of using untargeted metabolomics to identify biomarkers of health and performance e ects of exercise intervention. Untargeted metabolomics studies have shown elevated levels of lipid metabolism during training in endurance athletes and changes in acyl-carnitines have been linked to fatty acid oxidation [68]. While most studies currently focus on metabolites with a known function, untargeted metabolomics has also been used to identify metabolites with currently no known function related to exercise. For example, Malkar et al. identi ed changes in -valerolactam (2-piperidone) in response to exercise [9]. Interest in fasted exercising and ketolytic metabolism has grown in both research and the general public. While a large body of research exists comparing fasted exercise against fed exercise, there is currently
used to identify metabolites with currently no known function related to exercise. For example, Malkar et al. identi ed changes in -valerolactam (2-piperidone) in response to exercise [9]. Interest in fasted exercising and ketolytic metabolism has grown in both research and the general public. While a large body of research exists comparing fasted exercise against fed exercise, there is currently a lack of research looking at the changes in the metabolome over time during fasted exercise and subsequent rest. Previous work investigated blood glucose levels in diabetic subjects during high-intensity interval training (HIIT) and moderate-intensity continuous training (MICT) after fasting [10]. From this work, it was revealed that there was no di erence shown between pre and post HIIT and MICT for the length of time spent in the states of hyperglycaemia, hypoglycaemia or in a target glucose range over 24 h/during the nocturnal period. While this information is useful, it gives no insight into the other known energy processes which may in turn help to investigate cardiovascular health. Monitoring glucose levels for diabetics pre and post exercise provides valuable insights into how exercise can be e ectively utilised to minimise the impact of the disease. However, this may not translate well to the general population where an individual's glucose level is relatively stable. This work seeks to utilise untargeted metabolomics methods to provide a much more comprehensive picture of the metabolism of fasted individuals both during and post-exercise. Metabolites associated with the key energy pathways, along with transporter metabolites, can be seen allowing for identi cation of switching between energy sources. The untargeted nature of the method also means that novel compounds, that are either contained in databases but not yet associated to energy metabolism, or are currently not found in databases, can be investigated. 2. Results 2.1. Physiological Testing To assess the e ort and recovery of the participants, heart rate (HR) was recorded at 15-min intervals and breath samples were collected every 15 min to calculate the percentage of VO2max. As can be seen from Figure ,average HR across the participants was similar in both the pre-exercise
found in databases, can be investigated. 2. Results 2.1. Physiological Testing To assess the e ort and recovery of the participants, heart rate (HR) was recorded at 15-min intervals and breath samples were collected every 15 min to calculate the percentage of VO2max. As can be seen from Figure ,average HR across the participants was similar in both the pre-exercise and rest phases. The same trend is observed for VO2max measurements (Figure). These data indicate a similar exertion by all participants across the protocol and a similar level of endurance training.
Metabolites2020,10, 399 3 of 11 Baseline HR shows larger variation. The slight increase in variation of both HR and percentage of VO2 max measurements seen at 120 min is likely due to time trial pacing di erences between competitive cyclists and athletes who use cycling for training.Metabolites 2020, 10, x FOR PEER REVIEW 3 of 11 training. Baseline HR shows larger variation. The slight increase in variation of both HR and percentage of VO 2 max measurements seen at 120 min is likely due to time trial pacing differences between competitive cyclists and athletes who use cycling for training. Figure 1. Average heart rate recorded at 15-minute intervals during collection of breath samples. Small variations across subjects were observed during the exercise phases of the protocol. Upon rest, the average heart rate returned to the baseline and remained consistent across this phase. Figure 2. Average percentage of VO 2 max based on calculation from respired air collection every 15 min. This data shows that effort was well conserved at approximately 70% of VO 2 max for 60 min. Variation is seen across the time trail phase, with a general trend of slight reduction in effort. Variation increases towards the end of this section exercise phase, likely due to pacing differences in participants. Percentage of VO 2 max is well conserved across the rest phase, indicating a similar level of exertion for all participants. Average respiratory exchange ratios (RERs) across all participants show an initial increase from the baseline in the exercise phase, then a downward trend with a sharp drop during rest. These values then remain below the baseline, showing a slight increase towards 240 min, Figure 3. Some variation in work rate was seen across participants but this was small and likely due to differing levels of glycogen in their muscles. RER values of 0.7 or less are indicative of ketolytic metabolism. Values of 0.7 or less were seen for three participants and all participants were observed to approach this limit apart from participant 003. 0 20 40 60 80 100 120 140 160 180 200 0 15 30
small and likely due to differing levels of glycogen in their muscles. RER values of 0.7 or less are indicative of ketolytic metabolism. Values of 0.7 or less were seen for three participants and all participants were observed to approach this limit apart from participant 003. 0 20 40 60 80 100 120 140 160 180 200 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 Heart Rate (BPM) Time (Min) 0 10 20 30 40 50 60 70 80 90 100 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 % of VO 2 max Time (Min) Figure 1. Average heart rate recorded at 15-min intervals during collection of breath samples. Small variations across subjects were observed during the exercise phases of the protocol. Upon rest, the average heart rate returned to the baseline and remained consistent across this phase.Metabolites 2020, 10, x FOR PEER REVIEW 3 of 11 training. Baseline HR shows larger variation. The slight increase in variation of both HR and percentage of VO 2 max measurements seen at 120 min is likely due to time trial pacing differences between competitive cyclists and athletes who use cycling for training. Figure 1. Average heart rate recorded at 15-minute intervals during collection of breath samples. Small variations across subjects were observed during the exercise phases of the protocol. Upon rest, the average heart rate returned to the baseline and remained consistent across this phase. Figure 2. Average percentage of VO 2 max based on calculation from respired air collection every 15 min. This data shows that effort was well conserved at approximately 70% of VO 2 max for 60 min. Variation is seen across the time trail phase, with a general trend of slight reduction in effort. Variation increases towards the end of this section exercise phase, likely due to pacing differences in participants. Percentage of VO 2 max is well conserved across the rest phase, indicating a similar level of exertion for all participants. Average respiratory exchange ratios (RERs) across all
seen across the time trail phase, with a general trend of slight reduction in effort. Variation increases towards the end of this section exercise phase, likely due to pacing differences in participants. Percentage of VO 2 max is well conserved across the rest phase, indicating a similar level of exertion for all participants. Average respiratory exchange ratios (RERs) across all participants show an initial increase from the baseline in the exercise phase, then a downward trend with a sharp drop during rest. These values then remain below the baseline, showing a slight increase towards 240 min, Figure 3. Some variation in work rate was seen across participants but this was small and likely due to differing levels of glycogen in their muscles. RER values of 0.7 or less are indicative of ketolytic metabolism. Values of 0.7 or less were seen for three participants and all participants were observed to approach this limit apart from participant 003. 0 20 40 60 80 100 120 140 160 180 200 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 Heart Rate (BPM) Time (Min) 0 10 20 30 40 50 60 70 80 90 100 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 % of VO 2 max Time (Min) Figure 2. Average percentage of VO 2max based on calculation from respired air collection every 15 min. This data shows that e ort was well conserved at approximately 70% of VO 2max for 60 min. Variation is seen across the time trail phase, with a general trend of slight reduction in e ort. Variation increases towards the end of this section exercise phase, likely due to pacing di erences in participants. Percentage of VO 2max is well conserved across the rest phase, indicating a similar level of exertion for all participants. Average respiratory exchange ratios (RERs) across all participants show an initial increase from the baseline in the exercise phase, then a downward trend with a sharp drop during rest. These values then remain below the
pacing di erences in participants. Percentage of VO 2max is well conserved across the rest phase, indicating a similar level of exertion for all participants. Average respiratory exchange ratios (RERs) across all participants show an initial increase from the baseline in the exercise phase, then a downward trend with a sharp drop during rest. These values then remain below the baseline, showing a slight increase towards 240 min, Figure. Some variation in work rate was seen across participants but this was small and likely due to di ering levels of glycogen in their muscles. RER values of 0.7 or less are indicative of ketolytic metabolism. Values of 0.7 or less were seen for three participants and all participants were observed to approach this limit apart from participant 003.
Metabolites2020,10, 399 4 of 11Metabolites 2020, 10, x FOR PEER REVIEW 4 of 11 Figure 3. Average respiratory exchange ratio (RER) for all participants. Data suggests mainly glycolytic energy metabolism during the exercise phase of the protocol, then, an increase in fatty acid energy metabolism during rest with a trend back towards glycolytic metabolism at the end of the protocol. 2.2. Metabolomic Analysis Metabolomic analysis identified 851 likely metabolite signals, with 632 annotated on the basis of mass and retention time prediction using IDEOM [11] and 51 metabolites identified against authentic standards [12] (see Supplementary files). The implementation of a generalised linear model (GLM) identified 108 metabolites as significantly changing over time based on false discovery rate of less than 0.05 (S1). Due to the lack of replicates for individual subjects at each time point, the p-values for participant and the interaction term of time:participant were not presented. A heatmap was plotted for the data on the basis of fold-change comparisons to T0 to identify trends in metabolites. This can be seen in Figure 4. Figure 4. Metabolites showing fold changes across glycolytic and ketolytic energy metabolism pathways. Fold change values coloured from blue to red as fold change increases. Figure 3. Average respiratory exchange ratio (RER) for all participants. Data suggests mainly glycolytic energy metabolism during the exercise phase of the protocol, then, an increase in fatty acid energy metabolism during rest with a trend back towards glycolytic metabolism at the end of the protocol. 2.2. Metabolomic Analysis Metabolomic analysis identi ed 851 likely metabolite signals, with 632 annotated on the basis of mass and retention time prediction using IDEOM [11] and 51 metabolites identi ed against authentic standards [12] (see Supplementary les). The implementation of a generalised linear model (GLM) identi ed 108 metabolites as signi cantly changing over time based on false discovery rate of less than 0.05 (S1). Due to the lack of replicates for individual subjects at each time point, thep-values for participant and the interaction term of time: participant were not presented. A heatmap was plotted for the data on the basis of fold-change
model (GLM) identi ed 108 metabolites as signi cantly changing over time based on false discovery rate of less than 0.05 (S1). Due to the lack of replicates for individual subjects at each time point, thep-values for participant and the interaction term of time: participant were not presented. A heatmap was plotted for the data on the basis of fold-change comparisons to T0 to identify trends in metabolites. This can be seen in Figure.Metabolites 2020, 10, x FOR PEER REVIEW 4 of 11 Figure 3. Average respiratory exchange ratio (RER) for all participants. Data suggests mainly glycolytic energy metabolism during the exercise phase of the protocol, then, an increase in fatty acid energy metabolism during rest with a trend back towards glycolytic metabolism at the end of the protocol. 2.2. Metabolomic Analysis Metabolomic analysis identified 851 likely metabolite signals, with 632 annotated on the basis of mass and retention time prediction using IDEOM [11] and 51 metabolites identified against authentic standards [12] (see Supplementary files). The implementation of a generalised linear model (GLM) identified 108 metabolites as significantly changing over time based on false discovery rate of less than 0.05 (S1). Due to the lack of replicates for individual subjects at each time point, the p-values for participant and the interaction term of time:participant were not presented. A heatmap was plotted for the data on the basis of fold-change comparisons to T0 to identify trends in metabolites. This can be seen in Figure 4. Figure 4. Metabolites showing fold changes across glycolytic and ketolytic energy metabolism pathways. Fold change values coloured from blue to red as fold change increases. Figure 4. Metabolites showing fold changes across glycolytic and ketolytic energy metabolism pathways. Fold change values coloured from blue to red as fold change increases.
Metabolites2020,10, 399 5 of 11 2.3. Constant Load Phase During the constant load phase, changes were observed across glycolytic metabolism. The pro le of lactate over time resembles the physiological lactate concentrations measured in plasma and was seen to change signi cantly over the course of the protocol (correctedp-value=1.6 10 2 ). Pyruvate, another metabolite of glycolysis, and malate, a key TCA intermediate (correctedp-value=1.8 10 2 and 2.0 10 3 respectively). These metabolite pro les are consistent with glycolytic metabolism during the initial exercise phase. At 60 min, several metabolite signals showed changes compared to the baseline based on fold change. These are Glycerol-3-phosphate (FC=2.60) (identi ed based on authentic standard retention time), hexanoylcarnitine (3.41) and butanoylcarnitine (2.51), all metabolites involved in fatty acid transport and oxidation. These metabolites showed maximum fold changes against the baseline at 120 min and suggest a switch from glycolytic metabolism to ketolytic metabolism (Figure).Metabolites 2020, 10, x FOR PEER REVIEW 5 of 11 2.3. Constant Load Phase During the constant load phase, changes were observed across glycolytic metabolism. The profile of lactate over time resembles the physiological lactate concentrations measured in plasma and was seen to change significantly over the course of the protocol (corrected p-value = 1.6 × 10 −2 ). Pyruvate, another metabolite of glycolysis, and malate, a key TCA intermediate (corrected p-value = 1.8 × 10 −2 and 2.0 × 10 −3 respectively). These metabolite profiles are consistent with glycolytic metabolism during the initial exercise phase. At 60 min, several metabolite signals showed changes compared to the baseline based on fold change. These are Glycerol-3-phosphate (FC = 2.60) (identified based on authentic standard retention time), hexanoylcarnitine (3.41) and butanoylcarnitine (2.51), all metabolites involved in fatty acid transport and oxidation. These metabolites showed maximum fold changes against the baseline at 120 min and suggest a switch from glycolytic metabolism to ketolytic metabolism (Figure 5). Figure 5. Metabolites putatively identified as peptides. Values coloured from blue to red as fold change increases. 2.4. Time Trial Phase Metabolomics samples were taken at the start and end of the time trial phase. Several key ketolytic
Description
This study investigates energy metabolism changes during fasted exercise using metabolomics.