Discussion
In the present study, we investigated the effects of lifelong, moderate-intensity exercise on blood metabolites through an NMR-based metabolomics approach. By comparing a lifelong exercise group, two long-term exercise groups, and a group that did not exercise, we found that moderate-intensity exercise has a strong effect on the blood metabolome up until midlife. From midlife to elderhood, the effect of aging becomes stronger than that of exercise.
This unexpected finding comes in agreement with a study conducted by Falegan et al. (2016) who found that aerobic capacity could mitigate some but not all age-related changes in the metabolic profile of rats. In a very recent study Deda et al. (2018) detected changes after short-term and lifelong exercise in blood metabolites of female rats using LC-MS analysis, however, a limitation of this study is that no pre-exercise comparison of the exercise and non-exercise groups is shown. Most exercise metabolomic reports show the detected changes in human blood metabolome due to either short- or long-term exercise of young (Pechlivanis et al. 2013), aged (Mukherjee et al. 2014), or all ages individuals (Lewis et al. 2010). A few studies have shown the lifelong effect of exercise on cognitive processes (Pietrelli et al. 2018), cardiac function (Gama et al. 2010, Rockstein et al. 1981), and the antioxidant defense in many tissues (Cui et al. 2009, Gündüz et al. 2004), however, these studies did not involve any metabolomic analyses.
Thus, the uniqueness of our study is that the effect of lifelong exercise on the entire blood metabolome was analyzed on all groups pre and post-exercise by 1H NMR metabolomics analysis. NMR is a powerful metabolomics tool, as it can detect analytes which are difficult to be found with other technologies (Pechlivanis et al., 2010). Hence, the investigation of the effects of lifelong exercise and aging on blood metabolites through an NMR-based metabolomics approach may provide an insight in understanding better some aspects of exercise biochemistry. The relevance of the study lies in the fact that lifelong exercise has not been fully examined in the entire blood metabolome, possibly due to methodological difficulties.
Exercise and aging result in changes in numerous metabolites involved in different metabolic pathways. The results of our study are summarised in Figure …. The changes in the metabolites will be discussed in the alphabetical order of the metabolic pathway that they are involved for convenience.
Amino acid metabolism
The literature states that both exercise and aging result in fluctuations of amino acids, however, exercise tends to increase the total content of amino acids (Takeshita et al., 2011). This statement is somehow supported by our findings on the total content of amino acids, in which only group C, ie. the group that did not exercise at all, showed a significant decrease in the total content of amino acids from both 3 to 12 and 3 to 21 months.
Correlations of metabolites involved in amino acid metabolism were found at 12 months both because of exercise and aging. Correlations were also found at 21 months in groups A, B, and C. Aging resulted in a decrease in five (asparagine, aspartate, glutamate, glycine, and serine) out of the seven glucogenic amino acids that we identified. On the other hand, alanine and histidine, which are the remaining two glucogenic amino acids that we identified, increased. Group C, the group that did not exercise at all, was the only group that showed a decrease in the total content of glucogenic amino acids. These results could imply an aging-induced decrease in the dependence on glucogenic amino acids. In normal aging cell metabolism is reduced. However, no studies were found to support this finding, on the contrary, studies have shown that intense exercise results in a decrease in glucogenic amino acids, which may correspond to enhanced gluconeogenesis (Takeshita et al., 2011). At this point, we could hypothesize, that aging had a similar effect to that of intense exercise on glucogenic amino acids. Interestingly though, the levels of aspartate were higher at 21 months in the groups that exercised at the second half and the levels of glutamate were higher at the same age in the groups that exercised at the first half. This could imply an effect of lifelong and long-term exercise on aspartate and a prolonged effect of long-term exercise on glutamate.
Furthermore, group B, the lifelong exercising group, showed a significant increase in the total content of ketogenic amino acids on the second half of its life. It is of interest though, that the lifelong exercise group was the only group that had significantly higher values of lysine at 21 months. Group C, however, showed a significant decrease in the same group of amino acids on the first half of its life. This finding could imply that aging decreased the content of ketogenic amino acids. This last observation is supported by Houtkooper et al. (2011) and Okuda et al. (1987) who concluded that there is an aging-related diminished ketogenic capacity. Oddly enough though, the two exclusively ketogenic amino acids, leucine, and lysine were found to increase due to aging at 21 months. However, Hourkooper et al. (2011) also noticed aging-induced increases in lysine.
Another perspective concerning the amino acid metabolism is that two branched-chain amino acids (BCAA; isoleucine, leucine) and an aromatic amino acid (tyrosine) had higher values at 21 months. Elevated levels of these amino acids in the circulation are known to be significantly associated with age-related disorders, such as insulin resistance, which could lead to type 2 diabetes, and cardiovascular dysfunction (Falegan et al. 2016). Contrary to our findings are the results of Chaleckis et al. (2016) article, which identifies age-related differences in human blood metabolites, and it states that leucine and isoleucine were less abundant in elderly individuals. This difference could have either resulted due to differences in human and rat blood metabolome, due to a limitation of the study that it was not longitudinal, so different subjects were used to compare young and aged individuals, or according to Falegan et al. (2016) it could show that our rats were suffering from age-related disorders. Glycine, whose levels also decreased due to aging, is also found to be reduced years before the expression of prediabetes or type 2 diabetes (Klein et al. 2016).
Ammonia recycling
Exercise and aging resulted in correlations of many metabolites involved in ammonia recycling at 12 months. Three of them showed an aging-related significant decrease (asparagine, glutamate, serine). Decreases in asparagine, glutamate, and serine were also found by Houtkooper et al. (2011), who published changes in biomarkers of aging. Correlations in the same metabolic pathway were also found at 21 months in groups A, B, and C.
Betaine metabolism
A moderate correlation between two metabolites (dimethylglycine and betaine) involved in betaine metabolism was found in the lifelong exercise group at 21 months. Betaine is a methyl derivative of glycine and is metabolized to dimethylglycine and sarcosine (Cholewa et al., 2014). It is examined mainly as a dietary supplement, as improvements in lactate metabolism and hydration have been reported. However, little can be discussed about this pathway in our study.
Bile acid metabolism
A moderate correlation between two metabolites (glycine and taurine) involved in bile acid metabolism was found in the exercising groups at 12 months. At 21 months only group B showed a strong correlation between two metabolites involved in this specific pathway. Bile acid metabolism, which is the main way of eliminating excessive cholesterol in liver, can be activated by exercise training through the excretion of fecal bile acids (Farahnak et al., 2017).
Carbohydrate and lipid metabolism
Age resulted in correlations of metabolites involved in carbohydrate metabolism at 12 months. Of those metabolites, glucose showed a significant increase due to age. At 21 months correlations were found in groups A, B, and C. The increased levels of mannose, glucose, and glycerol at 12 months of age could possibly suggest enhanced carbohydrate and lipid metabolism up until midlife. Mannose, however, significantly decreased from 12 to 21 months and the exercising groups had higher values at 12 months. Pyruvate was also found increased at 21 months, suggesting enhanced glucose metabolism. Our results, in conjunction with another research, imply that glucose may be preferred over fatty acids as a fuel in older individuals, and is presumably the result of age-related changes in skeletal muscle’s respiratory capacity (Mittendorfer & Klein, 2001). It is of interest though, why these energy sources were not affected by exercise.
Gluconeogenesis is known to be activated in conditions of low glucose concentrations in plasma and intense exercise. However, many are the studies that claim that gluconeogenesis and blood glucose produced through it increases with age (Feng et al., 2016, Hachinohe et al., 2013, Lin et al., 2001). Specifically, a shift of energy metabolism away from glycolysis and towards gluconeogenesis is noticed (Lin et al., 2001). Feng et al. (2016) explain that gluconeogenesis might be elevated due to an increase in alanine (in our study, alanine indeed increased due to aging) and glutamine caused by sarcopenia, a disease by which muscle is lost in aging humans and animals. It is of interest though why the organism acts as if it is glucose-deprived when glucose uptake from brain and muscles in aged organisms is decreased. According to Feng et al. (2016), the increased demand and supply of glucose, combined with a reduction in usable glucose could lead to its disrupted homeostasis in aged organisms.
Citric acid cycle
Correlations between metabolites involved in the citric acid cycle (CAC) were seen at 12 months due to aging. Succinate, which is a TCA cycle intermediate, was affected by both aging, which decreased its levels, and exercise, which increased its levels. Huffman et al. (2014) support these findings, as they state that exercise training increases succinate dose-dependently. The CAC is an essential metabolic network in all oxidative organisms that provides anabolic processes in order to generate energy. It is also a functional target in the aging process in mammals (Yarian et al., 2006). Studies have identified an intricate link between CAC and reactive oxygen species (ROS) homeostasis, as the inefficient transfer of electrons results in generation of toxic ROS, which may help explain various metabolic diseases and aging (Mailloux et al., 2007, Yarian et al., 2006). Even subtle changes in just one metabolite could affect CAC intermediates and consequently affect signal transduction mechanisms (Yarian et al., 2006).
Ketone body metabolism
The evidence of moderately correlated metabolites involved in ketone body metabolism, in the exercising groups, as well as the increase of isobutyrate, a derivative of 3-hydroxybutyrate, which circulates sources of energy, in the same groups at 12 months, could possibly suggest an exercise-related activation of the specific pathway through periods of critically reduced CHO availability. However, at 21 months correlations of metabolites involved in ketone body metabolism were only found in group C, the non-exercising group. Ketone bodies are free fatty acid derivatives, which are converted to acetyl-CoA, via mitochondrial b-oxidation, and are used as a glucose-sparing energy source (Newman & Verdin, 2014). The literature states that ketogenesis is a critically important adaptive response, which during an energy crisis provides a substrate for brain (Evans et al., 2017).
Glucose – Alanine cycle
The exercise-related correlations of metabolites involved in the glucose-alanine cycle show its contribution to the energy demands of the exercising groups at 12 months. The exercise seemed to play a role as well at the correlations found at 21 months in groups B and D at the same pathway. In this pathway, lactate is formed from glucose and alanine through their conversion into pyruvate which is further reduced to lactate (Adeva-Andany et al., 2014). However, since it is a time-consuming process, the specific cycle is of limited importance to any type of exercise effort.
Phosphocreatine (CP)
In our study, we found low levels of CP at 21 months which suggests decreased CP re-synthesis when aging. In the review of Dalbo et al. (2009), it is stated that due to sarcopenia, the intramuscular CP levels are 5% lower in older compared to younger individuals. Although we did not find any significant changes due to exercise, it is important to note that from 12 to 21 months the lifelong exercise group had the least reduction in CP values (1%), group A, the group that exercised at the first half had a 7% reduction, group D, the group that exercised at the second half had a 16% reduction and the group that did not exercise at all, had the most reduction (18%).
Pyruvate metabolism and acetate
Acetate and lactate were found to be moderately correlated at 12 months in the exercising groups. Both metabolites take part in reactions involved in pyruvate metabolism. For example, in order to permit pyruvate, formed by glycolysis, to be fully oxidized by the TCA cycle, a molecular switch that interconverts acetyl-CoA and acetate needs to occur (Wolfe, 2005). This switch supplies the cell with opportunities to recover NAD+ and generate ATP.
Acetate, which is an important source of acetyl-CoA, increased up until midlife, and then it started decreasing until it reached significantly lower volumes at 21 months than at any other age point. Exercise, at the first half, helped at maintaining higher volumes at 12 months. This finding is supported by the fact that acetate levels also rise when fatty-acid oxidation rises (Shimazu et al. 2010). Formate had a similar trend to acetate, as it decreased due to aging and it increased due to exercise. The two metabolites had also a significant positive correlation. However, no studies were found to support these results. Significant negative correlations were also found between acetate and pyruvate, and formate and pyruvate. This could imply an important contribution of these metabolites in central metabolic pathways. Shimazu et al. (2010) speculate an important and underappreciated role of acetate metabolism in aging, due to its metabolism regulation by sirtuins, whose role are also important in the regulation of aging and longevity. Thus, further research between the relationship of acetate, formate, pyruvate, and aging is suggested.
Redox metabolism
A decrease in NAD implies a declined redox metabolism in the elderly, this result is supported by Gomes et al. (2013), who also mentions that the impairment of oxidative phosphorylation function during aging may have as a result depletion of the nuclear NAD pool. It is important to study redox metabolism as reduction and oxidation reactions are involved in our everyday life and are principal sources of energy. Although no oxidative stress biomarkers were identified, an interesting finding was that total glutathione of the lifelong exercise group and the group that exercised at the first half had lower values at both 12 and 21 months. Unfortunately, this result cannot lead us to a safe conclusion as we do not have values for GSH and GSSG separately.
Transcriptional and or translational pathways
Correlations of metabolites involved in transcription and translation pathways were also noticed due to exercise and aging at 12 months, as well as all four groups at 21 months. Specifically, threonine and asparagine, which are involved in transcriptional and translational pathways were found deceased due to aging, and cytidine was found increased. Group B and D showed a significant increase in tyrosine and only group B showed a significant decrease in threonine. Merry & Ristow (2016) mention changes in transcriptional responses due to the exercise-induced mitochondrial stress. Also, an increase in the translation rate of proteins is also noticed by exercise. Robinson et al. (2017) examined the effect of three different exercise modalities (HIIT, resistance training, and combined training) on the translation of mitochondrial proteins and two of them (HIIT and combined training) improved aerobic capacity, associated with their enhanced translation. As far as aging is concerned, it interferes with the transcriptional machinery in multiple ways, increasing diversity and heterogeneities (Alfego et al., 2018). Moreover, age is associated with changes in transcription regulatory networks, which may affect genes related to immune and defense responses as well as synaptic and neural activities (Ianov et al., 2016) and impact the function of cells or tissues and give rise to aging phenotypes and diseases (Buth & Brunet, 2017). Although aged organisms are more sensitive to errors in protein translation, studies have not detected a significant increase in mistranslated proteins during aging and cellular senescence (Ke et al., 2018). Nevertheless, protein synthesis decreases significantly (from 20 to 75%) with increasing age (Gonskikh & Polacek, 2017). This last statement is supported by Anisimova et al. (2018), who also mentions that aging affects the rate and type of damage accumulation in various organs and tissues having as a result altered translation efficiencies and changes in protein synthesis.
Nonetheless, exercise seemed to play an important role in maintaining lower body weight. This finding is supported by the fact that groups A and C, which did not exercise at the second half, weighed significantly more than the remaining two exercising groups, and the lifelong exercise group was the only group which did not show a significant increase in weight from midlife to elderhood. Specifically, at 21 months the lifelong exercising group weighed 10% less than the lifelong sedentary group and 6% less than the midlife to elderhood sedentary group. Garvey et al. (2015) also found that 8 weeks of activity significantly decreased body weight gain compared to rats without running wheel access. Furthermore, taking into account the previous results with the results of food intake which showed that group A had a significant increase in all periods except for the period that it did not train, which is from 12 to 21 months, in which it showed a significant decrease; as well as the results of the lifelong exercise group which showed a significant increase in all periods and group D which only showed an increase from 3 to 21 months and not between the first period of its life as it did not train at that time, we could hypothesize that the energy expenditure of the exercising rats was greater due to exercise.
Conclusion
Both exercise and aging had an impact on metabolites involved in different metabolic pathways. The results were mixed, some were expected, but others unexpected. The main outcomes of the study were that when aging the organism acts as if it is glucose-deprived and exercise does no longer affect the energy sources. However, exercise seemed to have a protective role over CP deprivation which is highly correlated with sarcopenia. Moreover, due to the greater energy expenditure of the exercising rats, their body weight was maintained lower until the end of their life. On no account do we diminish the importance of exercise at points that it might seem to have unexpected or no effects at all, on the contrary, we encourage people to become or remain active at all ages. The intensity and type of exercise play a big role in the outcome and effects of it and this could be a question open to future research. Furthermore, a similar experiment with caloric restriction and sampling of acute exercise at more age points could present notable results. Other recommendations for future research may be the exploration of the connection and pathways of acetate, formate, and pyruvate when exercising and aging, whose results were rather interesting.