A primary adaptation endurance athletes possess compared to the nonathletic population is mitochondria biogenesis and improved VO₂ max, which enables better oxygen uptake, OXPHOS and FAO in skeletal muscles (Rivera-Brown and Frontera, 2012). Endurance exercise is the most potent physiological inducers of mitochondrial biogenesis. Regular endurance training within 4–6 weeks in humans and mammals has been shown to increase mitochondrial content from 30 to 100% (Hood et al., 2011) and to increase the volume density up to 40% (Hood, 2001; Lundby and Jacobs, 2016). In the same line, 5 months of endurance exercise induced systemic mitochondrial biogenesis, prevented mitochondrial DNA depletion and mutations, increased mitochondrial oxidative capacity and respiratory chain assembly, restored mitochondrial morphology, and blunted pathological levels of apoptosis in multiple tissues of mitochondrial DNA mutator mice (Safdar et al., 2011).

Mitochondrial biogenesis occurs via fusion, which is the merging of mitochondria, or fission, which is the separation of damaged mitochondria (Busquets-Cortés et al., 2016). The greater number of mitochondria found in trained athletes' muscles enable better FAO, OXPHOS, and oxygen usage, which spares carbohydrate oxidation and thus glycogen breakdown and lowers lactate production (Hood et al., 2011; Rivera-Brown and Frontera, 2012). Conversely, a decrease in the number of mitochondria is related to impaired OXPHOS and FAO capacity (Wai and Langer, 2016). More than two in five marathon runners report experiencing a rapid onset of fatigue and the inability to maintain the speed and intensity of performance due to carbohydrate depletion (Rapoport, 2010) suggesting that poor OXPHOS and FAO capacity could be a major underlying cause of fatigue in athletes (Figure 2).

Given the energy requirements during endurance exercise, and the recently described complex and reciprocal relationship between the gut microbiota and whole body energy metabolism, it is not surprising that efforts to identify the mechanisms in which gut microbiota enhance mitochondrial FAO and OXPHOS in elite athletes are increasing (Pyne et al., 2015).

The gut microbiota contains more than 100 trillion microorganisms (Rajilić-Stojanović and de Vos, 2014), which comprise approximately 160 species and 9 million genes (Li et al., 2014) which are key to host energy metabolism. In the gut, anaerobic bacteria ferment and extract energy from otherwise indigestible polysaccharides such as fiber and resistant starch, and synthesize the byproduct SCFA such as N-butyrate, propionate and acetate (Flint et al., 2008). High fiber diets lead to 400–600 mmol of SCFA in the cecum per day, which accounts for approximately 10% of human caloric requirements (den Besten et al., 2013). In colonocytes, butyrate is transported to mitochondria where it undergoes FAO in aerobic conditions and becomes acetyl-CoA, which enters the Krebs cycle resulting in NADH which then enters into the electron transport chain producing ATP production and CO₂ (Donohoe et al., 2011). Furthermore, butyrate is oxidized in mitochondria by a series of five enzymes including the mitochondrial enzyme acetoacetyl CoA thiolase, and butyrate oxidation has been shown to be impaired in ulcerative colitis patients (Roediger et al., 1993; Ahmad et al., 2000; Santhanam et al., 2007) suggesting that butyrate plays an important role in TCA activity in colonocytes and therefore overall colon health. Additionally, propionate and acetate can be carried into the bloodstream to various organs where they are used as substrates for mitochondrial oxidation, lipid production, and gluconeogenesis, which is synthesized from propionate (Nicholson et al., 2012).

Though human studies are lacking, various animal studies using germ-free (GF) animals or specific pathogen-free (SPF) animals have shown that the SCFAs, such as N-butyrate and acetate, may affect mitochondrial energy metabolism through a vast range of transcription factors that directly or indirectly control mitochondrial functions. The types and amount of SCFAs produced by gut microorganisms depends on the composition of the gut microbiota, metabolic interactions between microbial species and the amount and type of the main dietary macro- and micronutrients ingested (den Besten et al., 2013). The more plant-derived polysaccharides, oligosaccharides, resistant starch and dietary fiber one eats, the more these bacteria can ferment these indigestible food sources into beneficial SCFA.

Studies in C57BL/6 mice have shown that a sodium butyrate injection of 5% wt/wt in addition to a high fat diet (58% calories from fat) prevented insulin resistance by stimulating thermogenesis and fatty acid oxidation in skeletal muscle and brown adipose tissue mitochondria in part due to increased PGC-1α gene expression and protein activity (Gao et al., 2009). Interestingly, butyrate-mediated PGC-1α induction led to a transformation of skeletal muscle fibers from type II (glycolytic) to type I (oxidative), which are rich in mitochondria and stimulate FAO for ATP production (Gao et al., 2009). Similarly, rats fed human milk compared to cow or donkey milk displayed higher mitochondria energy efficiency associated with changes in the microbiota's capacity to produce N-butyrate (Trinchese et al., 2015). Therefore, it is interesting to speculate that N-butyrate could play an important role in PGC-1α activation, which is a biomarker of mitochondrial functions during endurance as PGC-1α gene expression has been shown to dramatically increase in skeletal muscles in response to exercise (Pilegaard et al., 2003; Bo et al., 2010; Little et al., 2010; Safdar et al., 2011).

N-butyrate and acetate may also affect mitochondrial function and metabolism via AMPK activation in colonocytes (Canfora et al., 2015; den Besten et al., 2015). AMPK has been shown to be a key energy sensor in skeletal muscles capable of regulating mitochondrial OXPHOS (Lantier et al., 2014). den Besten et al. (2015) discovered that SCFA activated the UCP2-AMPK-acetyl-CoA carboxylase (ACC) pathway which down regulated PPARγ gene expression resulting in decreased lipogenesis and increased AMP: ATP ratio. The increased AMP: ATP ratio can also activate AMPK in liver, adipose (Richter and Ruderman, 2009) and muscle tissues (Hood et al., 2011), which stimulates glucose uptake, mitochondria FAO and OXPHOS and decreases protein and lipid synthesis. Similarly, Mollica et al. (2017) discovered that N-butyrate improved respiratory capacity and FAO through the activation of the AMPK-ACC pathway, which promoted mitochondrial fusion in liver and reduced insulin resistance and fat accumulation in obese mice. SCFA are also key mediators of mitochondria energy metabolism because they serve as a ligand for free fatty acid receptors 2 and 3 (FFAR2, FFAR3), also known as G-coupled receptor protein 43 (GRP43) and GRP41 respectively, that regulate glucose and fatty acid metabolism (reviewed extensively by den Besten et al., 2013; Kimura et al., 2014), as well as GPR109A that activates pathways associated with energy homeostasis, lipid storage and food ingestion (Nicholson et al., 2012).

The impact of microbiota on mitochondrial functions has been further supported by studies intending to manipulate of gut microbiota through the use of probiotics. Administration of the probiotic Lactobacillus rhamnosus CNCMI–4317 was associated with greater Fiaf expression, also known as ANGTPL4 (angiopoietin-like 4), through PPAR-α dependent pathways (Jacouton et al., 2015) that modified the OXPHOS capacity of mitochondria. In line with this, Bäckhed et al. (2007) demonstrated that GF mice expressed a lean phenotype because they were protected from the obesogenic effects of a high fat and sugar Western diet due to elevated Fiaf expression in the intestines, as well as increased AMPK and PGC-1α expression in skeletal muscles and the liver, which regulate mitochondrial OXPHOS in muscle cells (Lantier et al., 2014). Finally, certain intestinal bacteria such as Eubacterium hallii and Anaerostipes caccae have the capacity to transform the byproduct of anaerobic glycolysis lactate into SCFA during glucose depletion thus creating an alternative energy source for the host (Duncan et al., 2004; Scott et al., 2013) while bypassing OXPHOS (Rogatzki et al., 2015).

Besides SCFA, secondary bile acids produced by the gut microbiota also play an important role in regulating mitochondrial energy metabolism. Anaerobic bacteria of the genera Bacteroides, Eubacterium, and Clostridium degrade 5–10% of the primary bile acids forming secondary bile acids (Gérard, 2013). Secondary bile acids interact with mitochondria by modulating transcription factors related to lipid and carbohydrate metabolism, including farnesoid X receptor (FXR) and G-coupled membrane protein 5 (TGR5) (Nie et al., 2015). FXR is a target of NAD-dependent protein deacetylase SIRT1 (reviewed by Kuipers et al., 2014) and regulates the steroid response element binding protein-1c (SREBP-1c), carbohydrate response element binding protein (ChREBP), and PPAR-α, which stimulates fatty acid uptake and oxidation (Joyce and Gahn, 2016).

There is increasing evidence that secondary bile acid metabolism might also directly modify SIRT1 and Fiaf expression as well as mitochondrial biogenesis, inflammation and intestinal barrier function in different types of cells (Gao et al., 2009; Korecka et al., 2013; Alex et al., 2014; Caron et al., 2014; Kazgan et al., 2014). SIRT1's role in gut homeostasis is also beginning to be elucidated shedding new light on the role gut microbiota-mitochondria crosstalk plays in energy metabolism (Kazgan et al., 2014; Lo Sasso et al., 2014). Because 12 weeks of voluntary running wheel in wild type mice enhanced biliary bile acid secretion and increased fecal bile acid and neutral sterol outputs compared to sedentary controls (Meissner et al., 2011), it is temping to speculate about the role of microbiota plays in energy metabolism through the modulation of bile acid, SCFA and indirect induction of SIRT1, Fiaf and FXR genes (Figure 2).

Unlike the beneficial effects commensal bacteria have on energy metabolism, pathogens such as Salmonella and Escherichia coli (Leschelle et al., 2005) can produce negative effects for the host mitochondria energy metabolism by degrading sulfur amino acids to produce hydrogen sulfide (H₂S) in the large intestines. H₂S is an important mediator of many physiological and pathological processes. High amounts of H₂S can inhibit a key component of the mitochondrial respiratory chain by penetrating cell membranes and inhibiting COX activity and energy production (Blachier et al., 2007; Mottawea et al., 2016). Pathobionts can also produce NO, which may affect host mitochondrial activity and favor bacterial infection (Vermeiren et al., 2012). Besides pathobionts, high protein diets, which are common in many elite athletes, can result in high levels of H₂S and urea in the gut (reviewed by Windey et al., 2012). For example, and fecal H₂S concentrations can reach up to 3.4 mM upon eating a high protein diet in humans (Leschelle et al., 2005). The high concentrations of gut-derived H₂S treatment led to decreased COX expression in human colonic HT-29 cells and thus reduced electron transfer of complexes I and II in the respiratory chain shifting oxidative metabolism toward glycolysis, increased lactate production and decreased ATP production (Leschelle et al., 2005). Similarly, Beaumont et al. (2016) concluded that exposure of high levels of H₂S to HT-29 human cells showed not only reduced mitochondrial oxygen consumption but also an increase in the expression of inflammatory genes such as IL-6, which was increased following a high protein diet. Mottawea et al. (2016) recently demonstrated that a proliferation of pathobionts, many of which are known to be potent H₂S producers, down regulated mitochondrial proteins.

Additionally, H₂S can inhibit butyrate β-oxidation in the colon, which is believed to contribute to ulcerative colitis (Leschelle et al., 2005; Blachier et al., 2007). In line with this, Le Chatelier et al. (2013) studied microbial diversity in obese vs. healthy individuals. They discovered that those individuals who had high bacterial diversity had higher hydrogen and organic acid production (i.e., N-butyrate, propionate) whereas those with low bacterial richness had less N-butyrate- producing bacteria, increased H₂S forming potential and Campylobacter/Shigella abundance and an overall inflammation-associated microbiota.

Given the lack of studies in human endurance athletes, it's difficult to make precise dietary recommendations for endurance athletes in regards to how to optimize SCFA and bile acid metabolism for energy production. However, high animal protein consumption during resting days and training may negatively affect the gut microbiota of elite athletes (e.g., production of potentially toxic byproducts such as H₂S). It is clear then that the interaction between diet and exercise needs to be further studied in order to better assess the contributions of diet and microbial activities in mitochondria functions and athletic performance.

During and after exercise, ROS, and RONS production increases in response to energy needs and are primarily produced from electron leak in the mitochondrial electron transport chain (Radak et al., 2013). Complex I of the mitochondrial electron transport chain is one of the greatest generators of ROS, RONS, and free radicals such as NO and superoxide anion (O2-), which are molecules that are unstable because they possess an impaired electron that causes oxidation reactions with proteins, lipids and DNA in order to become stable molecules (Fisher-Wellman and Bloomer, 2009; Hood et al., 2011; Gomes et al., 2012; Radak et al., 2013).

Exercise-induced ROS production during regular physical activity can lead to beneficial adaptations to exercise such as vasoregulation, fibroblast proliferation (Gomes et al., 2012), muscle hypertrophy, increased mitochondrial biogenesis, induction of antioxidant enzymes (Radak et al., 2013) and regulation of immune responses to eliminate antigens (Fisher-Wellman and Bloomer, 2009; Radak et al., 2013). Ten days of endurance exercise can elevate antioxidant production and decrease inflammation resulting in less intestinal permeability (Holland et al., 2015). The generally accepted hypothesis is that mitochondrial ROS can induce NF-κβ and activator protein 1 (AP-1) which activates antioxidant enzymes such as manganese superoxide dismutase (Mn-SOD) (Radak et al., 2013), catalase and SOD (Blaser et al., 2016), as well as PGC-1α (Steinberg et al., 2006), thereby having a homeostatic effect (Blaser et al., 2016). In PGC-1α knockout mice, St-Pierre et al. (2006) observed a decrease of Mn-SOD, copper–zinc superoxide dismutase levels, suggesting that PGC-1α knockout mice are more sensitive to oxidative stress (St-Pierre et al., 2006).

However, many athletes suffer from stress and enter into a vicious cycle of over exerting themselves with strenuous training and competitions, which in chronically high levels of ROS and RONS which can deplete the non-enzymatic antioxidant system and damage cellular function. In over trained athletes, the excessive release of stress hormones induced by physical as well as increased body oxygen uptake, might lead to the generation of ROS and RONS in the tissues that undergo ischemia and hypoperfusion (Mach et al., 2017). Ischemia-induced intestinal hyperpermeability typically occurs in humans exercising at 70% maximal oxygen consumption when blood supply is reduced by at least 50% (Holland et al., 2015). Therefore, athletes have two major sources of ROS and RONS: from the electron transport chain in mitochondria and the intestines by both epithelial cells and transmigrating neutrophils in the gut lumen (Figure 3).

It appears that both poorly trained individuals and athletes who overtrain are at a higher risk of suffering from oxidative stress (Radak et al., 2008), which is characterized by a substantial increase in ROS damage of lipids, proteins, and DNA (Hood et al., 2011; Radak et al., 2013). DNA oxidative damage increases the ratio of mutations in mitochondrial and nucleus genome (Green et al., 2011). Importantly, mitochondrial DNA is more susceptible to oxidative damage and mutation accumulation due to its proximity to ROS produced in mitochondria (Lee and Wei, 2005; Crane et al., 2013). Additionally, recent studies have revealed that ROS production also reduces telomere length while modifying mitochondrial biogenesis (Sahin et al., 2011). Telomere dysfunction activates p53-mediated pathway to repress the expression of PPAR-γ, PGC-1α, and PGC-1 (Sahin and DePinho, 2012). In turn, the repression of both co-activators impairs functional mitochondrial biogenesis leading to higher levels of ROS damaging both telomere and mitochondrial DNA, and starts a negative feedback loop. Moreover, several studies reported lately that positive correlations have been observed between telomere length and mitochondrial DNA copy number variations in healthy adults (Kim et al., 2013) and between telomere length and aging and several diseases (Garatachea et al., 2015).

Palazzetti et al. (2003) analyzed oxidative stress levels in male triathletes vs. sedentary individuals before and after a 4-week period of overtraining through blood and urine samples. They discovered that overtraining could compromise the antioxidant defense mechanism due to increased lipid peroxidation and an up regulation of plasma glutathione peroxidase (GSH-Px), which failed to prevent oxidative damage, which is in line with other studies in athletes who overtrain (Palezzetti et al., 2003). Bloomer et al. (2005) also showed that just 30 min of acute anaerobic and aerobic exercise can also lead to redox imbalance as demonstrated by a significant increase in glutathione oxidation (decreased GSH and increased GSSG post-exercise), increased lipid peroxidation (protein carbonyls) and minor increase in DNA oxidation (8-OHdG). In mice, Li et al. (2015) reported that acute exercise (76% VO₂ max for 15, 60, or 90 min) induced mitochondrial oxidative stress. This was confirmed by increased ROS production, which induced the inflammasome-NOD-like receptor family and pyrin domain containing 3 (NLRP3). Similarly, we have recently demonstrated in horses, which are considered to serve as an optimal in vivo model for characterizing the response to endurance exercise, that the elevated respiration rates during endurance exercise had led to the generation of more ROS than the antioxidant systems can scavenge (Mach et al., 2016, 2017). Of note, many studies in athletes have varying results on oxidative markers due to the heterogeneity of methodology, blood, urine and DNA analysis, athlete training status, type of exercise, exercise conditions, long-term effects of overtraining and athlete diet. Therefore, it is difficult to make generalizations about oxidative status pre and post-exercise in endurance athletes; however, it appears that in general, well-trained athletes who train 20–30 h per week likely have lower RONS production and better antioxidant defense mechanisms and thus redox balance (reviewed by Fisher-Wellman and Bloomer, 2009).

In the last years there has been a proliferation of experimental works, conducted mainly in animals, aimed to explore how the microbiota modulates mitochondrial ROS production. As mentioned above, evidence indicates that gut microbiota communicate with mitochondria via SCFA, which have antioxidative properties. The absence of microbial colonization in mice was associated to a reduced levels of SCFA and serum levels of GPx and catalase after endurance swimming, which are critical antioxidants for reducing oxidative stress (Hsu et al., 2015). Different studies using mice models (Dobashi et al., 2011, 2013) have reported that the gut microbiota regulate SOD activity, yet more studies are needed to better understand the gut microbiota's role in controlling the intestinal redox balance in response to long-term intense exercise, inflammation and dysbiosis.

Other studies have shown that SCFA may inhibit telomere length shortening (Garatachea et al., 2015). O'Callaghan et al. (2012) discovered that higher colonic SCFA production in rats was associated with reduced malondialdehyde levels (a marker of oxidative stress), telomere shortening and DNA damage. Though the exact mechanisms of how SCFA reduced telomere damage in this experiment are unknown, other studies have shown that N-butyrate possess antioxidant properties that can reduce H₂O2- induced DNA damage possibly by altering chromatin structures such as telomere length, and induce glutathione antioxidant production (Rosignoli et al., 2001; Hamer et al., 2009). N-butyrate can reduce colonic H₂O₂ levels and thus diminish oxidative damage by increasing COX-2 activity (Martinez et al., 2010). In summary, SCFA can improve mitochondrial function by reducing ROS levels through various possible mechanisms which can damage DNA and shorten telomeres which are both associated with aging and mitochondrial dysfunction.

The gut microbiota can also metabolize the aromatic amino acid tryptophan, which apart from being the precursor to the neurotransmitter serotonin also plays a role in synthesizing the cofactors for redox reactions nicotinamide adenine dinucleotide (NAD) and NAD phosphate (NADP) (Ito et al., 2003; Richard et al., 2009). Other bacteria might use tryptophan to produce quinolinic acid, which is a neuroactive metabolite of the kynurenine pathway often implicated in the pathogenesis of a variety of human neurological diseases and lipid peroxidation (Kurnasov et al., 2003). A comparative genomic analysis proved that several bacteria contain bacterial enzymes such as 3-hydroxy-kynureninase, necessary to convert tryptophan to quinolinic acid (Kurnasov et al., 2003). Strephtomyces antibioticus, Cyanidium caldarium, Karlingia rosea, and Xanthomonas pruni use tryptophan to biosynthesize quinolinic acid, and Pseudomonas aureofaciens have enzymatic activities such as tryptophan dioxygenase and kynureninase (Kurnasov et al., 2003). Other pathobionts such as Bacteroides thetaiotaomicron, Proteus vulgaris and E. coli possess the enzyme bacterial tryptophase, which metabolizes tryptophan into metabolites such as indole-3-pyruvate, indole-3-lactate, and indole-3-acetate (Boulangé et al., 2016). Indole-3-pyruvate can be converted downstream into indole-3-acetate, which can activate horseradish peroxides causing free radical formation and lipid peroxidation (Kumavath et al., 2017). Lastly, some intestinal pathobionts have adapted regulatory responses to oxidative stress (Chiang and Schellhorn, 2012) that can lead to either supports the growth of pathogens or inhibits N-butyrate production (Rivera-Chávez et al., 2017). For instance, the increased colonic nitrate content favors the proliferation of Enterobacteriaceae pathogens such as E. coli and Salmonella spp. through nitrate respiration (Rivera-Chávez et al., 2017; Zeng et al., 2017). These results illustrate that non-commensal bacteria could induce redox imbalance and inflammatory pathways in the gut.

Ghosh et al. (2011) believe that ROS production is a defense mechanism that can elicit cytotoxicity against the pathogen and reduce the burden of infection on the host. In fact, Ghosh et al. (2011) showed that Citrobacter rodentium infection in C57BL/6 resistant resulted in more Bacteroides and elevated levels of oxidative stress (reduced GSH ratio, induction of iNOS, and reduced antioxidant MnSOD/SOD2 protein expressions). A review by Lobet et al. (2015) describes how ROS production have been involved in the clearance of different intracellular pathogens such as Listeria monocytogenes, Staphilococcus typhimurium, or Toxoplasma gondii. However, it has been shown that in order to overcome the mitochondrial effect on the immune response and cell survival, numerous bacterial species of microbiota tend to directly reduce mitochondrial ROS production (Lobet et al., 2015). For instance, Mycobacterium tuberculosis downregulates LPS-induced TLR signaling pathways that reduce mitochondrial ROS production (Saint-Georges-Chaumet and Edeas, 2016). Other microbial toxins can upregulate activity of the detoxification enzyme mitochondrial superoxide dismutase, which results in a lower ROS content and reduces host cell apoptosis, as observed in Ehrlichia chaffeensis (Liu et al., 2012).

As reviewed by Mach and Fuster-Botella (in press), intense exercise raises plasma cortisol levels inducing an influx of neutrophils from bone marrow and in efflux of other leukocyte subsets, which causes an acute-phase inflammatory response. In contrast to habitual light exercise and fitness (Clarke et al., 2014), strenuous exercise causes an increase in the number of pro-inflammatory cytokines, such as TNFα, IL-1, IL-6, IL-1 receptor antagonist, TNF receptors, as well as anti-inflammatory modulators like IL-10, IL-8, and macrophage inflammatory protein-1, indicating a dose-response effect between biological responses to exercise and host immunity (reviewed by Clark and Mach, 2016).

A key mechanism of inflammation is the activation of inflammasomes, which are large cytosolic protein complexes that activate caspase-1, which induce the inflammatory cytokines IL-1 and IL-18 and facilitates other inflammatory mediators that are crucial for the innate immune response. Mitochondria play a central role in the initiation of inflammasomes and other inflammatory pathways and as such integrate autophagy, cell death, and inflammation (Green et al., 2011). Although the exact mechanisms are currently unknown, noninfectious agents such as ROS-triggering OXPHOS inhibitors and pro-inflammatory signals seems to activate NLRP3, which results in a mitochondria-mediated inflammatory responses (Green et al., 2011; de Zoete and Flavell, 2013). The activation of NLRP3 also promotes mitophagy (mitochondrial autophagy), a cellular process that eliminates malfunctioning mitochondria (Shimada et al., 2012). As previously mentioned, Li et al. (2015) reported that mice that performed acute exercise (76% VO₂ max for 15, 60, or 90 min) experienced exercise-induced mitochondrial stress, which was confirmed by increased ROS production and NLRP3 expression.

Another way endurance training can cause changes in immune responses and mitochondrial function is by reducing the gastrointestinal blood flow, oxygen and nutrients while increasing tissue hyperthermia, permeability of the gastrointestinal epithelial wall and the destruction of gut mucous thickness (Marlicz and Loniewski, 2015), which stimulates an inflammatory immune response. These processes are associated with cell damage and lipopolysaccharides (LPS) translocation outside of the gastrointestinal tract, which consequently triggers immune and inflammatory responses often resulting in increased intestinal permeability (reviewed by Clark and Mach, 2016) but also mitochondria functions. For example, studies in animal models have shown that LPS injection (3 mg/kg) in adult male cats resulted in a partial uncoupling of mitochondrial OXPHOS and a 40% reduction of COX activity (Crouser et al., 2002), but also to higher ROS production as a result of a reduction in the expression of uncoupling protein 2 (UPC2; Saint-Georges-Chaumet et al., 2015). Lee and Hüttemann (2014) postulated that TLRs that bind to LPS stimulate TNFα and IL-6 in production, which activates tyrosine kinase leading to downstream COX phosphorylation and impaired ATP production in mitochondria.

On the other hand, evidence is emerging that SIRT1 activity can attenuate LPS- induced inflammatory cytokines and signaling while improving mitochondrial functions (Caruso et al., 2014; Lo Sasso et al., 2014; Wei et al., 2017). Cheng et al. (2015) revealed that exercise activated SIRT1 activity, which reduced NF-κB, TLR-4, IL-1β, IL- 6, IL-8, and TNFα inflammatory activity. A murine sepsis model in C57BL/6 mice injected with LPS stimulated TLR4 and activated SIRT1, which induced NF-κB p65 deacetylation and deactivation thus inhibiting pro-inflammatory gene expression (Liu et al., 2012). SIRT1 could play an important role in maintaining intestinal barrier function through various mechanisms such as enhancing crypt proliferation and suppressing villous apoptosis (Wang et al., 2012), stimulating intestinal stem cell expansion in the gut (Igarashi and Guarente, 2016), regulating tight junction expression of zonulin ocludin-1, occluding, and claudin-1 during hypoxia (Ma Y. et al., 2014). SIRT1 can also reduce stress-induced inflammation and can improve intestinal ischemia/reperfusion-induced ROS accumulation and apoptosis via miR-34a-5p activation, which induces SIRT1-mediated suppression of intestinal ROS accumulation (Wang et al., 2016). However, some studies have shown that LPS-induced inflammatory signaling can inhibit SIRT1 activity (Shen et al., 2009; Fernandes et al., 2012; Storka et al., 2013) suggesting that SIRT1 activity or inactivity may be tissue specific and dependent upon the dose of LPS.

Other studies show that inflammatory cytokines such as TNFα and IL-6 can also inhibit AMPK activation, which negatively affects glucose metabolism and reduces FAO in mitochondria (Steinberg et al., 2006; Lim et al., 2010; Viollet et al., 2010; Andreasen et al., 2011). IL-6 which is another cytokine associated with exercise-induced muscle injury (Richter and Ruderman, 2009; Lantier et al., 2014) and intestinal permeability (Grootjans et al., 2010) can also activate AMPK upon prolonged exercise in skeletal muscle (Febbraio and Pedersen, 2005; Ruderman et al., 2006), which can improve peripheral glucose uptake and whole body insulin sensitivity (Glund et al., 2007). Furthermore, IL-6 synthesis in skeletal muscles can increase up to 100-fold in marathon runners (Febbraio and Pedersen, 2002). Ruderman et al. (2006) tested the metabolic effects IL-6 had on AMPK activation during exercise in IL-6 knockout mice. They concluded that mature 9-month old mice showed signs of dyslipidemia, impaired glucose tolerance and obesity possibly due to the lack of IL-6-induced activation of AMPK. Kelly et al. (2009) analyzed IL-6- induced AMPK activation in rat skeletal muscle cells in vivo injected with 25 ng/g animal weight. They discovered that IL-6 activated AMPK by increasing cyclic adenosine monophosphate (cAMP) concentration and the AMP: ATP ratio leading to energy synthesis in muscles. Another study by Lantier et al. (2014) showed that AMPKα1α2 knockout mice had a significantly reduced exercise performance and fatigue resistance as well as increased IL-6 expression possibly due to skeletal muscle injury. This study clearly provides evidence that AMPK inactivity reduced mitochondrial OXPHOS that resulted in reduced energy production in response to exercise.

Subproducts of commensal microbiota may also regulate mitochondrial inflammatory responses through different mechanisms during endurance exercise, the principal one likely being modulating the intestinal barrier-ROS production and LPS translocation. For example, bacteria-derived indole-3-propionate has beneficial effects for the host because it maintains intestinal barrier function by up regulating tight junction proteins and downregulating TNFα in enterocytes, which prevents the translocation of LPS and other pathogens thus reducing inflammatory immune responses (Boulangé et al., 2016). SCFA also downregulate LPS- induced pro-inflammatory mediators via histone decetylation in macrophages in the lamina propria as well as pro-inflammatory mediators such as NO, IL-6, and IL-12 (Chang et al., 2014). Additionally, SCFA inhibit NF-κB activation in lamina propia macrophages that reduced inflammation that is associated with ulcerative colitis (Lührs et al., 2002) and can also regulate AMPK (Canfora et al., 2015) by activating the UCP2-AMPK-acetyl-CoA carboxylase (ACC) pathway (den Besten et al., 2015) as well as the GPR43 protein, which play a protective role in the inflammatory activation and signaling. For example, Peng et al. (2009) determined that human colonic epithelial cell line Caco-2 treated with 2 mmol/L of N-butyrate presented an upregulation of AMPK activity, together with an increased transepithelial resistance, a marker of proper intestinal barrier function. Yet AMPK's activation and role in tight junction protein regulation remain unknown, though it's likely via phosphorylation (Elamin et al., 2013). Vieira et al. (2015) discovered that when the metabolic sensor receptor GPR43 sensed acetate in neutrophils in vitro, IL-18 production increased, which may promote gut epithelial integrity in colitis models (Zaki et al., 2010; Macia et al., 2015). Similarly, Macia et al. (2015) reported that diets rich in fibers increased the colonic acetate in mice that suffered DSS-induced colitis, and consequently increased the GPR43 expression and the IL-18 secretion. On the other hand, a diet that increased ketone metabolite β-hydroxybutyrate levels inhibited NLRP3 activity during intense exercise (Shao et al., 2015; Youm et al., 2015).

However, the inflamed microenvironment in the gut might confer a favorable environment for the expansion of pathogenic Enterobacteriaceae (Zeng et al., 2017). Enterobacteriacea, including E. coli, Klebsiella spp., and Proteus spp. are among the most commonly overgrown pathobionts in inflammatory conditions (Zeng et al., 2017). Inflammation in the gut can cause an elevation in oxygen to the intestinal lumen as blood flow and hemoglobin increase and induce the growth of pathogens in the gut (Zeng et al., 2017). Other bacteria, such as Chlamydia pneumonia (Shimada et al., 2012) and Salmonella typhimurium can induce NLRP3 via Caspase-11 macrophages (de Zoete and Flavell, 2013) and cause mitochondrial dysfunction.