Insulin resistance plays a key role in the development of type 2 diabetes. It arises from a combination of genetic predisposition and environmental and lifestyle factors including lack of physical exercise and poor nutrition habits. The increased risk of type 2 diabetes is molecularly based on defects in insulin signaling, insulin secretion, and inflammation. The present review aims to give an overview on the molecular mechanisms underlying the uptake of glucose and related signaling pathways after acute and chronic exercise. Physical exercise, as crucial part in the prevention and treatment of diabetes, has marked acute and chronic effects on glucose disposal and related inflammatory signaling pathways. Exercise can stimulate molecular signaling pathways leading to glucose transport into the cell. Furthermore, physical exercise has the potential to modulate inflammatory processes by affecting specific inflammatory signaling pathways which can interfere with signaling pathways of the glucose uptake. The intensity of physical training appears to be the primary determinant of the degree of metabolic improvement modulating the molecular signaling pathways in a dose-response pattern, whereas training modality seems to have a secondary role.

1. Introduction

Insulin resistance plays a key role in the development of type 2 diabetes and is caused by genetic predisposition and environmental and lifestyle factors including physical inactivity and poor nutrition habits [1]. These risk factors also contribute to obesity, which is a major determinant of glucometabolic impairment and systemic subclinical inflammation [2]. Physical activity, as cornerstone in the prevention and treatment of diabetes, has marked acute and chronic effects on the regulation of glucose uptake and on inflammatory processes [3, 4]. The glucometabolic impairment in type 2 diabetes results from alterations of different signaling pathways modulating glucose uptake comprising insulin- and exercise-induced signaling pathways. However, during exercise, glucose uptake is normal or near normal [5], pointing to an insulin-independent activation of relevant signaling pathways mediating exercise-induced glucose uptake. An insulin-resistant state is also associated with changes in immunological and hormonal cross talk involving interleukin 6 (IL-6), tumor necrosis factor alpha (TNF-α), or adiponectin. These cytokines and adipokines are part of inflammatory processes and immune defense and can also affect molecular signaling pathways modulating glucose uptake. Behavioral interventions as well as unstructured physical activity have been shown to positively influence inflammatory processes, which was accompanied by improvements in glucose uptake [6, 7].

Physical exercise is distinguished primarily in resistance training and endurance training. Endurance training imposes a high-frequency (repetition), low-power output demand on muscular contraction, whereas resistance exercise imposes a low-frequency, high-resistance demand [8]. These two traditional modalities can also be performed as high-intensity training (HIT). This training form comprises alternating cycles of intensive and extensive phases involving endurance training, also known as high-intensity interval training (HIIT), and resistance training or the supramaximal exercise form of sprint interval training (SIT) [9, 10].

The overarching aim of this review is to summarize the mechanisms and molecular signaling pathways mediating glucose uptake as well as related changes in the release of immune mediators upon acute and chronic exercise exposure. Furthermore, we aim to assess the role of training intensity and training modality for the modulation of the aforementioned processes.

2. Search Strategy and Evaluation of Data

We searched PubMed/MEDLINE without language restriction from database inception until January 20, 2016, using the following search terms: “signaling OR pathway OR GLUT4 OR glucose OR inflammation OR inflammatory OR cytokine” AND “exercise OR training OR endurance exercise OR resistance exercise OR contraction”. Reference lists of review articles and all included articles identified by the search were also examined for other potentially eligible studies. The search was limited to human and animal studies. Duplicates were removed. Search results for relevant intervention studies are summarized in Table 1 and shown in detail in Tables 2, 3, 4, and 5.

The current literature does not provide a clear definition for acute or chronic effects of training [11]. The training effect is influenced by the time period between termination of the last bout of exercise and measurement as well as by training intensity [4]. Measurements of training effects within a time period of 0–72 h after exercise termination can show acute effects, even in a chronic training process, which makes it difficult to distinguish between acute and chronic training effects. In this review, we define the effect of chronic training as the sum of all training sessions, according to previous work [4].

3. Effect of Exercise on Molecular Signaling Cascades

3.1. Insulin Receptor Substrate 1 (IRS-1)/Phosphatidylinositol 3-Kinase (PI3-K) and Akt/Protein Kinase B (Akt/PKB) Pathways

In conditions of rest, insulin regulates glucose transport into the muscle due to activation of a protein signaling cascade. After binding of insulin to its receptor, the insulin receptor is autophosphorylated. Insulin receptor substrate 1 (IRS-1) binds to the phosphorylated tyrosine residues of the insulin receptor and is subsequently phosphorylated by the tyrosine kinase of the insulin receptor. Binding of IRS-1 to the p85 subunit of phosphatidylinositol 3-kinase (PI3-K) results in activation of a PI3-K-dependent pathway comprising phosphoinositide-dependent kinase (PDK) and atypical protein kinase C (αPKC) [110]. Key downstream molecules modulating translocation of glucose transporter type 4 (GLUT4) to the plasma membrane comprise, besides Akt/protein kinase B (Akt/PKB), Ras-related C3 botulinum toxin substrate 1 (Rac1), the TBC1 domain family member 1 (TBC1D1), or the Akt substrate of 160 kDa (AS160) [32, 110, 111] (Figure 1).

In type 2 diabetes patients, despite a normal amount of GLUT4 transporters [112], insulin fails, in general, to induce adequate insulin signaling as assessed by IRS-1 tyrosine phosphorylation, Akt/PKB activity, and translocation of GLUT4 to the cell membrane [113116].

Exercise activates the insulin-signaling pathways, facilitating GLUT4 expression and translocation to the cell membrane. The effects of acute and chronic exercise on glucose uptake and insulin signaling are shown in Table 1.

Acute continuous endurance exercise with 45–60 min of training at 65–75% of maximum oxygen consumption () leads to higher rates of tyrosine phosphorylation of insulin receptor and IRS-1/2 and to increased activity of PI3-K in muscle of untrained healthy as well as insulin-resistant individuals [1215]. In contrast, short and light resistance with 5 sets of 8 repetitions of isokinetic leg extension shows no effect in endurance trained athletes [16] (Table 2). Furthermore, acute muscle contraction activates molecules of the distal insulin signaling which are known to be involved in GLUT4 translocation such as Rac1, AS160, and TBC1D1 [16, 47, 48] which will be described in more detail below. Recent animal studies have shown that only very intense muscle contraction in situ via sciatic nerve stimulation of multiple muscle types with 2–5 V as well as one bout of intense swimming for 120 min or 60 min of running with a speed of 22 m/min and incline of 10% led to an acute increase in phosphorylation and activity of key molecules like the different AKT isoforms (AKT-1, AKT-2, and AKT-3) and AS160 [4951] (Table 3).

In contrast to these studies, some human as well as animal studies reported no effect of acute exercise on proximal insulin signaling like changes in insulin receptor amount, IRS-1 phosphorylation, or PI3-K activity [1621]. In the study of Wojaszewski et al. [21], one-legged cycling exercise for 60 min at intensity of 18–23% of was not sufficient to induce changes in proximal insulin signaling in young trained individuals. Furthermore, 60 min of cycling at 75%   did not lead to changes in proximal signaling in untrained and obese individuals [19, 20]. In line with this, some animal studies found that a running speed of 18 m/min for 45 min as well as electrical stimulation with 1–3 V were also not sufficient to induce insulin signaling in skeletal muscle [17, 18].

The reason for these discrepant results in human and animal studies might lie in the differences in the intensity of training conditions in acute exercise. Moderate endurance exercise seems to acutely increase proximal signaling in untrained individuals [12, 13, 15], whereas short and light resistance and endurance training in trained individuals shows no effect [16, 21] (Tables 2 and 3). In addition, the time point after exercise when the effect of exercise is studied appears to be highly important. A recent review from Frøsig and Richter identified a critical time point of 3 to 4 h after exercise for exercise-induced increase in glucose uptake indicating a time-dependent course in the activation of exercise induced molecular signaling [22], which may be the reason for the unaltered signaling in measurements 16 and 24 h after exercise termination [19, 20]. Though training intensity and the time point of investigation appear to be important for exercise-induced activation of insulin signaling, there is still a lack of knowledge about the underlying mechanisms of acute exercise and effects of different training factors, such as modality and intensity, on insulin signaling.

Chronic exercise can also lead to higher rates of tyrosine phosphorylation of key molecules in the insulin signaling cascade in muscle of healthy as well as insulin-resistant individuals [52, 95, 96]. A recent exercise study observed enhanced whole-body insulin action and increased Akt and AS160 phosphorylation after 10 weeks of chronic resistance training with exercises for upper and lower body and running endurance training in untrained individuals [52] indicating an independence of exercise modality. Compared to untrained controls, trained humans show increased insulin-stimulated PI3-kinase activation. The positive association between PI3-kinase activation and endurance capacity () indicates that regular exercise leads to greater insulin-stimulated IRS-1-associated PI3-kinase activation in human skeletal muscle [14]. This is in line with recent animal studies showing that intense chronic endurance training in mice with a running speed of 20–32 m/min on a treadmill increases total AS160 phosphorylation [53].

3.2. AMPK Signaling Pathway

AMPK is a metabolic master switch regulating several intracellular systems and consists of two catalytic alpha-isoforms: α2- and α1-AMPK. AMPK is activated by phosphorylation by kinases such as liver kinase B1 (LKB1) [117] and is regulated by cellular energy demand. Increasing adenosine monophosphate/adenosine triphosphate (AMP/ATP) and creatine/phosphocreatine (Cr/PCr) ratios, reflecting for instance the glucose deprivation state [118], are important stimuli for AMPK activity. In line with this, activation of AMPK is positively associated with an increased skeletal muscle glucose uptake [23].

In obese diabetic and nondiabetic humans, exercise-induced stimulation of the AMPK activity is attenuated but can be fully activated by exercise with higher intensities of training as compared to healthy lean controls [24, 25]. Acute cycling endurance exercise at a moderate intensity of 50–70% of increased AMPK activity and resulted in a 2.7-fold increase in mRNA expression of AMPKα1 and AMPKα2 [24, 25].

Activation of AMPK by acute cycling exercise led to an enhanced glucose uptake in human skeletal muscle [26]. AMPK phosphorylation and activity showed an intensity-dependent response pattern. More intense (80% of ) acute cycling endurance exercise with the same amount of energy expenditure (400 kcal) resulted in a higher activation of signal transduction compared to less intense (40%  ) endurance exercise [27]. High-intensity interval training consisting of repeated sessions of intense work like all-out sprints for 30 sec (SIT) induces, with a minimum of effort (<80 kJ total), an increased phosphorylation of AMPK. Though this kind of training appears to mimic resistance exercise because of the intense, short-term muscle work, phosphorylation and activity of downstream targets linked to hypertrophy like p70 ribosomal S6 kinase and 4E binding protein 1 were unchanged [28] (Table 2).

While some acute exercise studies in animals showed that AMPK-deficient mice and LKB-deficient mice had a normal contraction-induced glucose uptake, which was independent of the knockout of the catalytic alpha-isoforms of AMPK [97, 98], other studies found that pharmacological inhibition of AMPK and LKB activity blunted contraction-induced glucose disposal in animal models by electrical stimulation [99, 100]. LKB1 knock-out in muscle provoked a reduced activity of the AMPKα2 isoform, and transgenic mice expressing a kinase-dead, dominant negative form of the AMPKα2 showed also a reduced AMPK activity and blunted glucose uptake. The authors assumed that either the maximal force production was reduced in this muscle, raising the possibility that the defect in glucose transport was due to a secondary decrease in force production and not impaired AMPKα2 activity, or the kinase-dead, dominant negative form of the AMPKα2 had a negative influence on glucose uptake [99, 100].

Chronic endurance as well as resistance exercise also induces AMPK activation and leads, furthermore, to changes in gene expression favoring GLUT4 translocation. AMPK phosphorylation is more strongly increased after 10 weeks of cycling at 65%–90% of maximum performance () exercise than after 10 weeks of leg-focused resistance training with an intensity of a 4-5-repetition maximum (RM) [8, 29]. In animal studies, chronic treadmill running as well as resistance training in the form of ladder climbing with weights activated AMPK phosphorylation and up-regulated expression of AMPK in rat pancreatic islets and skeletal muscle [30, 31] indicating that AMPK upregulation is independent of exercise modality in different tissues (Tables 2 and 3).

In light of the animal studies showing that AMPK-deficient mice have a normal contraction-induced glucose uptake [97, 98] other molecular pathways, comprising the Ca2+/calmodulin signaling pathway, appear to modulate exercise-induced glucose uptake and will be described in the following subsections.

3.3. Ca2+/Calmodulin Signaling Pathway

Changes of the calcium concentration in skeletal muscle cells lead to activation of signaling cascades that influence cellular metabolism including glucose uptake [119]. In diabetes, the calcium- (Ca2+-) dependent signaling pathway and subsequently glucose uptake are impaired [120]. Genome-wide studies for DNA methylation have shown that first-degree relatives of patients with diabetes have already altered DNA methylation of genes encoding proteins involved in calcium-dependent signaling compared to healthy individuals without positive family history. However, DNA methylation decreased after 6 months of cycling and aerobic exercise [33].

As result of skeletal muscle contraction, cytosolic Ca2+ concentration and consequently the number of Ca2+/calmodulin complexes increase. Further important key players in the Ca2+/calmodulin signaling pathway are the Ca2+/calmodulin-dependent protein kinases (CaMKs). These components are critical for exercise-induced glucose uptake [32]. Downstream components of the Ca2+/calmodulin signaling pathway are members of the histone deacetylase (HDAC) family and proteins of the myocyte enhancer factor 2 (MEF2) family leading to an enhanced expression rate of GLUT4 [8].

Ca2+ release and phosphorylation of CaMKII after acute endurance cycling exercise depend on training intensity. A matched amount of work with different intensities of 40% and 80% of led to an increase in CaMKII phosphorylation by 84% immediately after high-intensity but not low-intensity cycling endurance exercise indicating that greater force outputs result in enhanced Ca2+/calmodulin signaling [27]. Furthermore, also the duration of endurance exercise affects the Ca2+/calmodulin signaling pathway activity, with higher activity after longer duration. A 90-min acute cycling endurance exercise resulted in a progressive increase of CaMKII activity during exercise peaking at 90 min of training [34]. In line with this, a recent study comparing acute HIT cycling with traditional continuous cycling exercise showed a marked increase of CaMKII activity by HIT despite the same amount of total work after 30 min of 70%   [35] (Table 2).

In accordance with the acute exercise studies, a recent animal study showed that chronic endurance training on a treadmill increased the phosphorylation of CAMKII in pancreatic islets of rats in a dose-response manner [31].

In experimental mouse studies, incubation with the Ca2+/calmodulin inhibitor KN-93 decreased skeletal muscle glucose transport [121] and inhibited electrical contraction-induced CaMKII phosphorylation [102]. In addition to the decrease of contraction-induced glucose uptake via electrical stimulation, inhibition of CaMKII resulted in an increase of AMPK activity in a recent mice study, pointing to overlapping mechanisms between these two key signaling pathways: the Ca2+/calmodulin signaling pathway and the AMPK-signaling pathway [103]. Another key player in the context of glucose uptake-related signaling is the protein kinase mammalian target of rapamycin (mTOR) that will be addressed in the following section.

3.4. Mammalian Target of Rapamycin/Serine Kinase 6 (mTOR/p70SK6) Pathway

MTOR is a serine/threonine protein kinase that integrates diverse environmental cues by translating them into appropriate cellular responses. Disrupting the mTOR signaling pathway causes a decrease in glucose uptake in multiple cell types such as brain, muscle, and adipose tissue [122124] and can lead to insulin resistance [125].

High-force stimuli like resistance training lead to muscle adaptation preparing skeletal muscle for more intensive stress. This muscle adaptation which appears to be dysregulated in an insulin-resistant and diabetic state is initiated by the activation of the mTOR/ pathway [36, 126, 127]. This protein complex activates signaling cascades including binding proteins (elF4E), initiation factors (4E-BP1), and elongation factors (eEF2) leading to protein synthesis and subsequently to cellular hypertrophy [128]. MTOR also stimulates focal adhesion kinases (FAK) and increases FAK-phosphotransferase activity in order to activate muscle protein synthesis [129]. This adaptation is related to the intensity of the muscle contraction, increasing with higher training load. Acute cycling exercise of 70% of as well as leg-specific strength exercises of 70% of 1-RM increased mTOR phosphorylation. In particular, resistance training leads to higher activation of mTOR signaling compared to traditional endurance exercise despite a huge difference in workload (660 versus 130 kcal) [29, 3739].

Protein synthesis is regulated, in particular, by contraction-induced activation of the multiprotein complex mTORC1. This protein complex functions as a sensor or control unit which regulates the translation of proteins by assessing the cellular environment for optimal conditions and initiating translation of mRNA. Besides physical activity, potent stimulators of the mTOR/S6K pathway are insulin, insulin-like growth factor (IGF-1), cytokines like IL-6, sufficient amino acid levels in skeletal muscle, and full-energy depots [130].

During acute endurance as well as resistance exercise, mTOR signaling is inhibited via AMPK phosphorylation and signaling to suppress high-energy demanding procedures such as protein synthesis [4042]. However, after exercise, muscle protein synthesis increases in parallel to the activation of Akt/PKB (protein kinase B), mTOR, S6K, and eEF2.

One bout of intense treadmill walking at 70% of for 45 min in untrained old men as well as 70% of of one leg exercise for 60 min in untrained healthy young men led to significant activation of the insulin signaling as well as of the mTOR/SK6 pathway [43, 44]. In line with this, recent exercise studies showed that acute cycling-based HIT or intense leg-specific strength training [40] activates the mTOR signaling pathway in human muscle [39, 45]. Exercise-induced activation of mTOR signaling in leg-specific endurance and resistance training appears to be time-dependent with a continuous increase after termination of physical activity [40, 43]. In line with the human studies, mTOR signaling was upregulated in acute exercise studies in animals comprising treadmill running and electrical stimulation, with a time-dependent answer after exercise termination [46] (Tables 2 and 3).

Chronic exercise studies also demonstrate that long-term leg-specific resistance training with 4-5-RM in sedentary individuals and high intensity cycling with 70–85% of in untrained controls can activate the mTOR signaling pathway in human muscle [29, 38]. These results underline that the activation of mTOR signaling may be independent of exercise type as well as training history. Besides mTOR, there are other important downstream targets modulating glucose uptake that will be addressed in the following section.

3.5. Ras-Related C3 Botulinum Toxin Substrate 1 (Rac1), TBC1 Domain Family Members 1 and 2 (TBC1D1/2), and Akt Substrate of 160 kDa (AS160)

The proteins AS160, TBC1D1/2, and Rac1 are involved in insulin- as well as contraction-induced glucose uptake [131, 132] and are, therefore, points of convergence of these two pathways. These downstream targets are altered in an insulin-resistant or diabetic state showing a reduced signaling activity [101, 133136].

Acute endurance exercise studies in untrained and trained humans showed an increase in phosphorylation of TDC1D1/4 and AS160 in skeletal muscle in the first 4 hours after cycling and specific one-leg endurance exercise, especially under long-term training conditions with a training duration of at least 60 min at 65%   [16, 5456]. In line with these human studies, animal studies found that contraction-induced glucose uptake by electrical stimulation was also modulated by an increase in phosphorylation of AS160 and TBC1D1 proteins [57] (Tables 2 and 3).

Rac1, a key downstream target in the regulation of glucose uptake, was shown to modulate exercise- and insulin-stimulated GLUT4 translocation in human muscle, with an intensity-dependent response pattern, as shown in murine muscle [58, 59]. Animals were exercised at their 50% and 70% maximum running speed over 30 min on a treadmill, and the higher intensity program resulted in an larger increase of Rac1 activation. Given that the total amount of work differed between both measurements, the results are hard to interpret. The larger improvement may result from the higher intensity or from the greater amount of exercise. A future study comprising an alternative training protocol with identical energy expenditure but different intensities would help to clarify the role of exercise intensity in this context. Furthermore, in RAC1-deficient mice, GLUT4 translocation as well as glucose uptake decreased after acute electrical stimulated muscle contraction and insulin infusion as a sign of an inhibited signaling capacity [59, 60].

Glucose uptake and insulin signaling are influenced by inflammatory processes and specific cytokines [2]. The following section aims at shedding some light on the impact of inflammatory signaling on exercise-stimulated glucose uptake and insulin signaling.

4. Inflammation-Associated Signaling Pathways and Key Players

4.1. IκB Kinase/Nuclear Factor Kappa B Pathway (IKK/NF-κB)

Different environmental influences, for example, certain pathogens, can activate molecular signaling cascades leading to an inflammatory response mediated by the IKK/NF-κB pathway. Recognizing receptors are, in particular, Toll-like receptors (TLRs). TLR4 plays a key role in the activation of the pro-inflammatory NF-κB pathway. TLRs interact with pathogen-associated molecules, resulting in an activation of downstream signaling proteins, for example, MyD88 [137], and subsequently an immune reaction via cytokine release, for example, of IL-6 and TNF-α from adipose tissue. The adapter protein MyD88 also activates other inflammation-associated signaling pathways like MAPK signaling as described below in more detail [138]. TLRs are expressed on macrophages, which can be subdivided into pro-inflammatory M1 and anti-inflammatory M2 macrophages. Exercise studies have shown that physical activity modulates TLR-dependent pathways [2]. As a result, acute as well as chronic exercise can lead to reduced TLR expression [61] and phenotypic switching from M1 to M2 macrophages in adipose tissue of obese mice [62].

Cytokines like IL-6 or agents comprising microbial components trigger signaling cascades that converge in the activation of IκB kinase (IKK) enzyme complex and subsequently in a translocation of the protein complex NF-κB into the nucleus. This results in transcription of target genes for inflammatory immune reaction including cytokines like IL-6, TNF-α, and IL-15 [139]. Chronic activation of the NF-κB pathway contributes to insulin resistance and muscle wasting. Especially in type 2 diabetes, human muscle is characterized by an increased activity of this pathway [63].

Human and animal exercise studies have shown that acute as well as chronic exercise can reduce the activation of the IKK/NF-κB pathway. This attenuation of the inflammatory signaling was independent of the exercise modality, age, and training status [6369] (Tables 4 and 5).

During acute physical activity with a sufficient load, muscle contraction induces a marked increase of IL-6 expression in skeletal muscle but also suppresses IL-6 production in adipose tissue [70]. Increasing energy demands due to prolonged or intense acute training like marathon running or cycling at 88% of [7173] as well as shrinking depots of muscle glycogen [104] accelerate the increase of IL-6 plasma levels. Interestingly, a recent work from Castellani et al. showed that exercise induces also a specific increase of IL-6 in adipose tissue which occurred more rapidly in adipose tissue from trained mice in comparison to untrained mice when exercised at the same relative running speed on a treadmill. The authors speculated that the increase of IL-6 would be needed for the provision of lipids to the muscle and liver [74]. In line with this, Macpherson et al. showed an increasing IL-6 and decreasing M1 macrophages content in inguinal adipose tissue and an improved insulin action after an acute bout of treadmill running exercise in obese mice [75]. In line with the results of the acute exercises studies, chronic exercise also led to decreased activity of the IKK/NF-κB pathway after 8 weeks of cycling exercise at 70% of and intense whole-body strength exercise with 50–80% of 1-RM [63, 66] (Tables 4 and 5). Accordingly, a decreased plasma IL-6 concentration at rest as well as in response to chronic exercise appears to characterize a normal training adaptation [71].

The transient rise in IL-6 also appears to be responsible for the production of anti-inflammatory mediators like IL-10 or IL-1 receptor antagonist (IL-1RA). In particular IL-1RA prevents inflammatory processes by blocking signal transduction of the proinflammatory IL-1 and creates also an anti-inflammatory balance to the proinflammatory cytokine IL-1β [76, 77, 140]. Furthermore, elevated levels of IL-6 from skeletal muscle stimulate an anti-inflammatory signaling cascade that inhibits the secretion of proinflammatory cytokines like TNF-α or IL-1β, suppress the secretion of the acute-phase reactant C-reactive protein (CRP) from the liver, a general and unspecific marker for systemic inflammation [76, 77], downregulate monocyte TLR expression at both mRNA and cell surface protein levels, and finally inhibit the IKK/NF-κB pathway [6466]. Besides the TLR family, there are other receptor proteins like NOD-like receptors initiating inflammatory processes and subsequently modulating glucose uptake-related signaling which will be discussed in the following section.

4.2. Inflammasome Pathway

The NOD-like receptor (NLR) family is of key importance in the innate immune system. NLRs are responsible for recognizing pathogen and danger-associated molecular patterns. In response to stress signals, NLRs activate the inflammasome pathway which forms a multi-protein complex [2]. Participating components of inflammasome complexes are NLRs, neutrophilic alkaline phosphatases (NALPs), apoptosis-associated speck-like protein (ASC) and caspase-1. After its formation, this oligomer converts proinflammatory cytokines into active forms such as IL-1β. Increasing IL-1-β levels have been hypothesized to play a role in the progression of type 2 diabetes and its complications because its activity stimulates inflammatory processes leading to cell damage and apoptosis, in particular in pancreatic β-cells. Furthermore, IL-1β inhibits proximal and distal insulin signaling and mediates interorgan cross talk between adipocytes and the liver, contributing to systemic inflammation [2, 141143].

A recent review reported that chronic endurance and resistance training in mice decrease NLR family pyrin domain containing 3 (NLRP3) mRNA levels accompanied by reduced IL-18 levels, reflecting diminished activity of the NLR/inflammasome pathway [2]. IL-18 expression decreases under chronic intense endurance exercise conditions with sports like rowing, running, or cycling with an intensity which is at 70% of in humans [78, 79]. Only chronic training conditions, but not acute exercise, appear to reduce IL-18 mRNA expression [79]. In line with this, a recently published animal study with chronic treadmill running as endurance exercise and isometric strength training as resistance training showed a decrease of IL-18 expression in adipose tissue and plasma levels [80] (Table 5).

So far, there are no human exercise studies which measured acute or chronic effects of physical activity on the upstream elements of the inflammasome pathway. Further mechanistic studies are, therefore, needed to better understand the role of the inflammasome in the anti-inflammatory response to exercise. In contrast, the role of the C-Jun N-terminal kinase (JNK)/mitogen-activated protein kinase (MAPK) pathway in the modulation of exercise-dependent effects on glucose uptake and inflammatory response has been investigated by several animal as well as human studies.

4.3. C-Jun N-Terminal Kinase (JNK)/Mitogen-Activated Protein Kinase (MAPK) Pathway

Lipid accumulation in adipocytes and endoplasmic reticulum (ER) stress as well as a NF-κB dependent cytokine releases activate the JNK/MAPK pathway [139, 144]. This activation results in the serine phosphorylation of IRS-1 and the phosphorylation of the c-Jun component of activator protein-1 (AP-1). The phosphorylation of serine residues in insulin receptor substrate-1 leads to an impairment in the ability of IRS-1 to activate downstream phosphatidylinositol 3-kinase-dependent pathways which may cause insulin resistance [145147]. AP-1 is a transcription factor that mediates the gene expression of many cytokines. Subsequently, the JNK pathway leads to an inflammatory reaction, especially to TNF-α and IL-6 release [139]. JNKs are divided into 3 isoforms and belong to the MAPK family. The MAPK family comprises extracellular regulated kinases (ERKs), JNKs and p38, and mediates cell growth, differentiation, hypertrophy, apoptosis, and inflammation [144]. Furthermore, oxidative stress following reactive oxygen species (ROS) production induces JNKs and p38 MAPK activation reflecting an important immune defense mechanism [148]. JNK activation by skeletal muscle contraction is also associated with an increase in muscle IL-6 mRNA expression in mice acutely after endurance exercise in form of treadmill running [81].

Exercise studies in human and animal models showed that the JNK/MAPK pathway is activated in a dose-response pattern. In particular very intense acute exercise like marathon running or cycling at 70% of and intense dynamic pull exercise as resistance training with an one-repetition maximum (1-RM) of 85% stimulate JNK signaling in skeletal muscle [8284], independently of training modality. JNK activation results, as a physiological mechanism, in DNA repair and muscle regeneration [149]. In contrast, a recent animal study has shown that acute long-term exercise by swimming for 180 min reduces JNK phosphorylation and improves insulin signaling and sensitivity in adipose tissue from obese rat [69]. In particular, chronic endurance exercise in form of swimming and treadmill running contributes to a reduction in JNK phosphorylation and improves insulin signaling and sensitivity in adipose and hepatic tissue from obese rats [6769, 85].

Inflammatory signaling pathways are associated with insulin resistance and impaired glucose uptake, whereas adiponectin is an important, though controversially discussed, counterpart being positively associated with insulin sensitivity. This adipokine will be discussed in the following section.

4.4. Adiponectin

Adiponectin, an adipokine which is primarily released by white adipose tissue (WAT), appears to be a key player in glucose metabolism at least in rodents, whereas its relevance in humans is somewhat less clear [6]. The secreted adiponectin binds to its receptors AdipoR1 and AdipoR2 and activates AMPK, p38 MAPK, and peroxisome proliferator-activated receptor α (PPAR-α) following adaptor protein 1 (APPL1) release in skeletal muscle and liver [150]. As a result, adiponectin positively affects metabolism by increasing fatty acid oxidation and glucose uptake in muscle. Furthermore, it plays a critical role in the cross talk between different insulin-sensitive tissues [151, 152]. Adiponectin levels are decreased in patients with diabetes and low adiponectin levels are associated with insulin resistance and obesity [153, 154]. Recent mouse studies showed that pharmacological adiponectin agonists improve insulin sensitivity and other health-related parameters [155].

Only a limited number of acute exercise intervention studies focused on changes of adiponectin levels. In one study, circulating adiponectin levels increased 30 min after endurance exercise in the recovery phase [5]. The currently available data indicate that adiponectin levels change in dependence of exercise intensity, showing an increasing level by enhanced training intensity of 76%   in trained rowing athletes [86], whereas moderate and long-lasting cycling at 50% of for 120 min did not acutely increase adiponectin levels in trained individuals immediately after exercise [87] (Table 4).

Conflicting results were also observed under chronic exercise conditions. More intense endurance exercise in form of cycling and brisk walking at 70% of resulted in increases of adiponectin levels [88, 89]. Overweight and age seem to reduce the response of adiponectin to exercise [90]. In line with this, Simpson and Singh reported in their review that adiponectin expression levels are increased under high-intensity exercise conditions [91], regardless of training modality in untrained young lean or obese individuals, after chronic whole-body strength training or jogging [92, 93]. In line with this, Cho et al. showed that 40 minutes of HIT exercise on treadmill prevent the downregulation of AdipoR1 which was caused by a high fat diet in sedentary control animals [94] indicating the importance of intense training for the potential role of adiponectin.

In contrast, untrained and trained adiponectin knockout mice (AdKO) significantly increased glucose tolerance and insulin sensitivity after 8 weeks of treadmill running suggesting the presence of an unknown compensatory mechanism [53].

A recent meta-analysis found that chronic exercise did not significantly increase adiponectin levels. However, in subgroup analyses, all modalities tended to increase adiponectin. The lack of statistical power due to small group sizes may have contributed to the overall null-finding [3]. In contrast, lifestyle interventions with unstructured exercise alone or in combination with weight-reducing diet can positively influence adiponectin plasma levels [156]. Weight loss is an important factor contributing to increases in plasma levels of adiponectin [157159]. In conclusion, the impact of exercise on adiponectin levels needs further clarification. With respect to chronic effects it is important to investigate to what extent exercise effects on adiponectin may be mediated by weight loss.

4.5. Exercise, Inflammation, and Insulin Signaling

Circulating serum or plasma levels of cytokines are strongly linked with the onset of type 2 diabetes [160162]. The stimulation of inflammatory signaling cascades can lead to interference with the insulin signaling pathway [2]. During exercise, acute effects on cytokine regulation comprise an upregulation of both (i) proinflammatory cytokines (e.g., TNF-α, IL-1β, and IL-6) and (ii) anti-inflammatory cytokines (IL-1RA, IL-10) [106].

Long-term effects of physical exercise are known to reduce markers of inflammation by decreasing adipocytokine production and cytokine release from skeletal muscle [163165]. The relationship between glucose uptake and adiponectin, IL-6, and TNF-α is shown in Figure 1.

The mechanistic impact of inflammation on insulin signaling has been studied for several cytokines. Currently available data suggest that TNF-α plays a direct role in the development of insulin resistance by decreasing glucose uptake into adipocytes via suppression of insulin receptor activity, AMPK activation, and downregulation of GLUT4 expression [165168]. Acute exercise did not change the expression pattern of TNF-α [169], whereas the increase of TNF-α during high intense physical activity like marathon running appears to be a response to muscle damage [104, 107, 108, 170]. Large cohort studies show that physical activity or chronic endurance exercise in form of walking reduces systemic subclinical inflammation [92] and the impact of exercise rises in a dose-response pattern regulated by frequency and intensity, but inflammation remains unchanged when exercise intensity was only moderate [66, 109, 171, 172]. A moderate community-based walking program with 3000 steps more per day did not change TNF-α plasma levels [109] and a chronic resistance training with only 2 units per week of only 3 sets of 3 exercises had also no impact on TNF-α protein content. In line with this, TNF-α plasma levels were reduced by high-intensity chronic resistance training, even though fat mass has not changed [173]. Also animal studies show that chronic exercise training, in particular endurance training like treadmill running, can reduce TNF-α levels [62].

IL-6 is another important protein in this context and is expressed by several tissues. As a myokine, muscle-derived IL-6 is acutely upregulated during exercise exposure [106] and mediates a physiological cross talk with WAT and liver in order to regulate glucose metabolism [160]. However, long-term effects of regular exercise show marked decreases of IL-6 levels [77]. The role of IL-6 is complex, as also evident by its diverse effects on molecular signaling. In adipose tissue, IL-6 mediates inflammatory processes and causes insulin resistance by downregulating GLUT4 and IRS-1 expression [139]. Furthermore, increasing IL-6 levels block PI3-K, another key player in insulin signaling, and induce TLR4 gene expression leading to enhanced inflammatory processes [174, 175]. In addition, IL-6 induces the downstream NF-κB signaling pathway which impairs insulin signaling and subsequently induces insulin resistance in insulin-dependent tissues of obese humans and animals [2]. In particular, IL-6 and liver interact in the context of exercise. A human exercise study with long-term cycling has shown that contraction-induced IL-6 release increased endogenous glucose production (EGP), thus underlining the importance of IL-6 for glucose homeostasis [105].

Experimental studies using mouse models yielded controversial findings. IL-6-deficient mice can develop a glucose-intolerant and insulin-resistant state indicating that balanced IL-6 levels have a positive effect on glucose uptake. Furthermore, mouse studies showed that circulating IL-6 levels increase glucose uptake and improve insulin sensitivity in skeletal muscle via AMPK activation [176].

Besides, the inflammasome pathway downregulates insulin signaling. The inflammasome pathway which is part of the innate immune system converts proinflammatory cytokines into active forms such as IL-1β or IL-18 which are decreased in their levels after chronic exercise [2]. Important proinflammatory chemokines which are influenced by exercise are interleukin 8 (IL-8) and monocyte chemoattractant protein-1 (MCP-1). Both cytokines slightly increase after acute exercise; however, their circulating levels decrease after chronic exercise in human as well as animal model exhibiting an improved inflammation status [75, 77, 139, 160], as shown in Table 1.

5. Summary

Exercise is an important cornerstone in the prevention and treatment of metabolic disorders. Acute and chronic exercise activates different molecular signaling pathways that can counteract defects in signaling and associated metabolic processes (Table 1). Exercise interventions have shown that physical activity can increase GLUT4 protein expression and translation by activation of different molecular signaling pathways irrespective of the exercise modality. AMPK and Ca2+/calmodulin signaling pathways show a dose-response pattern and increase their activity with increasing intensity despite equal work rate in kcal when compared to less intense exercise.

The key players mTOR, AS160, TBC1D1/4, and Rac1 can be activated by exercise. Human exercise studies have demonstrated that acute and chronic physical activity, regardless of training modality, leads to increases in their activity and finally to improved glucose uptake. The change in activity reflects a dose-response pattern. MTOR and AS160 also exhibit a continuous time-dependent increase.

Metabolic disorders are accompanied by activated inflammation-related signaling pathways which result in elevated cytokine release. Proinflammatory immune mediators, like IL-1β, IL-6, or TNF-α, are important factors in the development of insulin resistance. Their expression is modulated by physical activity. In particular, chronic endurance and resistance training and high training intensity improve glucose uptake which is associated in the long term with decreased secretion of proinflammatory cytokines and increased release of anti-inflammatory proteins such as adiponectin.

In summary, the current literature points to a higher efficiency of more intense exercise because of a dose-response relationship regulating metabolic improvements. However, more high-quality exercise interventions as well as mechanistic studies have to be performed to fully understand the molecular mechanisms contributing to metabolic improvements.

6. Open Questions

Despite the high number of studies on exercise interventions and underlying mechanisms that have been conducted, we are far from understanding the details mediating the effects of exercise on glucose uptake. Single key players in this field were identified over time and confirmed with mechanistic human and animal studies. So far, there is still a lack of knowledge about the underlying mechanisms of exercise-induced glucose uptake in regard to training factors, such as point of termination or intensity, especially in proximal insulin signaling. When interpreting the responses to training, it is important to know, in particular when dealing with the issue of glucose uptake and related signaling pathways, when relative to the last bout, and preferably the last two bouts, the samples were collected to distinguish between acute and chronic training effects. Furthermore, exercise exposure can be considered the combined responses to intensity, bout duration, and bout frequency, where the product is usually considered to be total amount like total energy expenditure, but only a small number of intervention studies controlled for total work.

In regard to the key players of molecular signaling, the interplay of interacting pathways, such as the Ca2+/calmodulin signaling pathway and the AMPK pathway, is still elusive. The inflammatory signaling pathways involving IKK/NF-κB and the inflammasomes have not been sufficiently characterized in the context of the influence of acute as well as chronic exercise. The controversial results of the adiponectin exercise studies highlight potential species differences between men and mice and merit more mechanistic studies.

Furthermore, there is an intense need to detect to what extent the effects of physical exercise are independent of or explained by weight loss or change in body composition. Some of these questions require larger sample sizes and higher statistical power to quantify effects but also standardized methods for molecular measurements and high-quality study plans considering potential confounders.

Conflict of Interests

The authors declare that they have no conflict of interests.