Abnormalities in the Effects of Insulin and Exercise on Glucose- and Lipid Metabolism in Obesity and Type 2 Diabetes

BACKGROUND Insulin resistance (IR) plays a major role for the increased risk of cardiovascular disease (CVD) in obesity and type 2 diabetes (T2D). Targeted treatment of IR in affected tissues is, however, almost non-existing, except for weight-loss and physical activity. The central hypothesis of this proposal is that abnormalities in glucose and lipid metabolism as well as defects in molecular processes regulated by insulin and exercise can explain the link between IR, mitochondrial dysfunction and lipid accumulation observed in skeletal muscle in obesity and T2D. Focusing on these metabolic and molecular abnormalities, the major goal of our project is to discover novel targets for the treatment of IR and mimicking exercise, which can be used in the prevention and treatment of T2D and CVD. T2D, obesity and other high risk conditions are characterized by IR in skeletal muscle, the major site of insulin-mediated glucose disposal (1-3). In T2D, failure of the pancreatic β-cells to compensate for this abnormality causes hyperglycemia, which further aggravates the risk of CVD. Despite extensive research, the exact metabolic and molecular mechanisms underlying IR are not fully understood. At the metabolic level, the investigators and others have demonstrated that IR in obesity and T2D is characterized by impaired insulin-stimulated glucose uptake and glucose storage, but also reduced insulin-mediated increase in glucose oxidation and attenuated suppression of lipid oxidation and circulating levels of free fatty acids (lipolysis) in response to insulin (2,3). At the molecular level these metabolic abnormalities are associated with reduced insulin signaling to glucose transport through IRS-1, PI3K, Akt, AS160 and RAC1 and to glycogen synthesis through impaired insulin action on glycogen synthase in skeletal muscle (1-10). Moreover, IR is consistently associated with accumulation of lipid content and different measures of mitochondrial dysfunction in skeletal muscle of patient with T2D, obesity and high-risk individuals (1-3,11,12). In addition, there is accumulating evidence of increased glycolysis (13-15) and resistance to the beneficial effect of exercise in T2D and obesity (16-18). The changes in substrate metabolism in IR conditions seem to be associated with a specific metabolomic signature that includes increased circulating levels of branched-chain amino acids and α-hydroxybutyrate, and reduced levels of glycine (19-23). Finally, impaired adipose tissue expandability in T2D, obesity and other high risk conditions may play a role in the pathogenesis of IR and T2D by leading to increased secretion of inflammatory factors and overflow of free fatty acids to cause ectopic lipid deposition in liver and muscle (24). Genetic variants found to be associated with T2D, obesity and body composition in large GWAS-studies suggest a role for several transcriptional factors and signaling pathways in adipose tissue expansion and inflammation (25,26). Effects of exercise: Physical activity plays a fundamental role in the prevention and treatment of T2D (27-29). Thus, meta-analysis of several long-term (\> 8 weeks) exercise training studies have shown that exercise can improve body composition, insulin sensitivity and reduce Hb1Ac by approximately 0.67 % in patients with T2D (30). The beneficial effects of a single bout of moderate-intensity one-legged exercise include an immediate increase in muscle glucose uptake (31). This is followed by a more prolonged (\<48 h) increase in insulin sensitivity to stimulate glucose uptake in both lean healthy and insulin-resistant obese and type 2 diabetic individuals (32-33). In addition, the cumulative effect of repeated exercise bouts (exercise training) increases insulin sensitivity and VO2max, and is further known to increase skeletal muscle mitochondrial content and function in both nondiabetic and diabetic subjects (34-37). Furthermore, there is evidence that exercise training improves pancreatic beta-cell function in individuals with T2D (38). Exercise mode: The most appropriate exercise mode for IR individuals has not yet been established. A major problem in this respect is to find an effective mode of exercise (type, intensity and duration) which will ensure continued engagement (patient adherence) as well as continued beneficial effects on health (29). Thus, for IR individuals it is important to identify a mode of exercise that can safely and effectively maximize energy expenditure. This could be non-weight-bearing exercise modes such as cycling and rowing. Most studies so far have reported effects of acute exercise or endurance exercise training on stationery bikes. However, it has been shown that energy expenditure and fat oxidation during rowing is higher than during cycling (39,40). The strong association between IR, lipid accumulation and mitochondrial dysfunction in muscle makes fat-burning exercise of special interest (29,40). The contribution of fat oxidation to energy expenditure is highest during low- and moderate intensity exercise (45-65 % of VO2max) (41-43). Thus, the combination of cycling and rowing at moderate intensities could be an attractive exercise mode due to its non-weight-bearing nature and by the involvement of both lower and upper body muscle groups (29). In our studies, we will for the first time evaluate the effect and tolerability of a high intensity interval training protocol combining cycling and rowing on ergometers on insulin sensitivity, body composition, cardiorespiratory fitness, pancreatic beta-cell function and metabolic and molecular regulators of glucose uptake and energy metabolism in obesity and T2D versus weight-matched controls. Exercise resistance: Exercise training is an essential aspect in T2D prevention (27,28). Acute exercise and endurance training improves insulin sensitivity, also in obesity and T2D (34). Exercise activates various signaling pathways, which results in the upregulation of transcriptional networks that regulate mitochondrial biogenesis, glucose and fatty acid metabolism, and muscle remodeling (44). This includes increased levels of muscle-specific microRNAs (miRNA), activation of nuclear receptors such as the NR4As subfamily, the PPAR family and other transcription factors (45-50). This happens together with exercise-induced stimulation of PGC-1α, which promotes oxidative phosphorylation and fatty acid oxidation by enhancing mitochondrial biogenesis (51). This seems to be mediated by activation of AMPK, p38 MAPK, CamK-II, CREB and other upstream regulators (51-53). Moreover, exercise increases activity and/or levels of proteins involved in both exercise and insulin-mediated signaling to glucose transport such as Akt, AMPK, TBC1D4 and RAC1 (8,10,34,35,54,55). Within the past decade a number of reports have indicated an impaired response to exercise in different IR conditions. Thus, in early-onset T2D and prediabetic conditions such as obesity and first-degree relatives of patients with T2D a lack of increase in VO2max and insulin sensitivity in response to either acute exercise or exercise training has been reported (16-18,56,57). At the molecular level these abnormalities were associated with an impaired increase in PGC-1α and other regulators of mitochondrial biogenesis and dynamics (16-18,56,57). This suggests the existence of 'exercise resistance' in T2D and other common IR conditions. However, the transcriptional and signaling networks underlying these effects of acute exercise on insulin sensitivity and VO2 max in human muscle, and the possible metabolic and molecular alterations in obesity and T2D remain to be established. Moreover, due to inappropriate study designs in most of these studies, it is unclear to what extent these findings are related to overweight/obesity per se. A better understanding of the transcriptional and signaling networks regulated by acute exercise and exercise training will be crucial to design lifestyle interventions in IR individuals to prevent T2D. Mitochondrial dynamics and mitophagy: Numerous studies have implicated a link between IR and mitochondrial dysfunction in T2D, obesity and other high-risk conditions (1,11,12). This appears to involve both impaired mitochondrial biogenesis, as well as reduced content, intrinsic impairment and morphological changes of muscle mitochondria (58-66). It is believed that mitochondrial dysfunction together with an increased lipid supply in IR conditions causes lipid accumulation, which further aggravates IR by causing negative modulation of insulin signaling (1,11). It has been debated whether IR is the cause or consequence of mitochondrial dysfunction. Recently, the investigators and others have demonstrated that inherited IR results in different markers of reduced mitochondrial content (67,68). This is supported by data showing that insulin increases mRNA levels and abundance of mitochondrial proteins as well as ATP fluxes in human skeletal muscle (69-71). Interestingly, recent data indicate that insulin regulates fission (DRP1, FIS1) and fusion (MFN1/2, OPA1) proteins governing mitochondrial dynamics as well as enzymes involved in the removal of damaged mitochondria (mitophagy) such as LC3B, Mul1, and BNIP3/3L and AMPK and mTOR-mediated phosphorylation of ULK1 (72-76). Recent studies have shown that the mitochondrial fission factor (MFF) is a direct substrate of AMPK in response to exercise, and demonstrated that activation of AMPK promotes mitochondrial fission by stimulation of MMF and DRP1 (77). Thus, defects in the insulin- and exercise-mediated regulation of mitochondrial dynamics and mitophagy could contribute to mitochondrial dysfunction in IR. To what extent such abnormalities are present in skeletal muscle and/or fat in patients with T2D and high-risk individuals with obesity remains to be investigated. Lipid accumulation: Accumulation of lipids in skeletal muscle is strongly associated with IR and the development of T2D (2,78). Fatty acids are stored as triglycerides in organelles known as lipid droplets (LD). Triglyceride-biosynthesis involves several isoforms of GPATs, AGPATs, lipins and DGATs (79), while intramuscular lipolysis, e.g. in response to exercise, seems to involve ATGL, CGI-58 and HSL (80). Multiple proteins are bound to the surface of LD to regulate membrane trafficking, lipid turnover, fusion and formation of LD as well as interaction with other organelles including mitochondria (78,81,82). This includes the SNARE-proteins, STX5, SNAP23, VAMP4 as well as LD coating proteins (perilipins) such as PLIN2, PLIN3 and PLIN5 (formerly known as ADRP, TIP47 and OXPAT). There is evidence that exercise increase the physical interaction between LD and mitochondria in human skeletal muscle (83), and that the interaction with mitochondria involves specific LD-associated proteins such as PLIN5, PLIN3 and SNAP23 (84-86). How the multiple proteins involved in triglyceride-biosynthesis, exercise-mediated lipolysis, LD formation and interaction with mitochondria are regulated by insulin and exercise in human skeletal muscle in patients with T2D and high risk individuals with obesity have not previously been investigated in detail. Studies of these mechanisms in human skeletal muscle in vivo are warranted to evaluate their possible relation to lipid accumulation, mitochondrial dysfunction and impaired insulin signaling in IR. Phosphorylation and lysine acetylation of enzymes in major metabolic pathways: In addition to mitochondrial dysfunction and lipid accumulation, the investigators and others have provided evidence of enhanced glycolysis in skeletal muscle in obesity and T2D (13-15). In several proteomic studies, the investigators have mapped the proteomes and phosphoproteomes of human muscle and isolated mitochondria (97-100), Together with reports from other groups, these studies have shown that enzymes involved in glycolysis, fatty acid metabolism, Krebs cycle and OxPhos are highly abundant, and to a high degree modified by both phosphorylation and lysine acetylation in muscle (89-92). In addition, the investigators have shown that insulin regulates phosphorylation of mitochondrial proteins in human skeletal muscle (93). This included components of the newly defined mitochondrial inner membrane organizing system (MINOS), which is important for cristae morphology and mitochondrial respiration. Moreover, lysine acetylation of mitochondrial proteins were recently shown to correlate with insulin sensitivity, whereas acetylation of ADP/ATP translocase 1 (ANT1) was reduced in response to exercise (94). Of particular interest, a recent study indicated that lysine acetylation in 50% of the cases interacted with phosphorylation at the same peptides suggesting cross-talk, and by this mechanism regulated the activity of kinases such as AMPK, AKT and PKA (92). A better understanding of this interaction is crucial to understand the posttranslationel regulation of enzymes in major metabolic and signaling pathways. Aberrant lysine acetylation may involve the cytosolic and mitochondrial deacetylases, Sirt1 and Sirt3, which are known to regulate e.g. AMPK, PGC-1alpha and multiple mitochondrial proteins as well as critical enzymes in insulin signaling and other molecular pathways (95). The possible regulation of lysine acetylation and phosphorylation and their interaction by insulin and exercise in individuals with and without IR has not been investigated before, and may provide substantial important information related to IR, enhanced glycolysis, impaired insulin signaling, mitochondrial dysfunction and lipid accumulation. Adipose tissue expandability: Recent GWAS meta-analyses have identified multiple genetic loci associated with measures of body fat distribution (waist, hip and WHR), independent of BMI (25,26) Many of these loci are located within and/or near genes implicated in adipocyte development or function but also adipogenesis, angiogenesis, transcriptional regulation and insulin resistance as processes influencing differences in distribution. Adipose tissue distribution appears intrinsic to the individual and is likely to depend on heritable factors such as genetic variants. Body composition analysis has determined the relative expansion of subcutaneous fat (SAT) versus ectopic fat (visceral, liver and muscle) with overfeeding. Such studies have contributed to the adipose tissue expandability hypothesis whereby SAT has a finite capacity to expand and once capacity is exceeded ectopic triglyceride deposition occurs (24). The potential for SAT expandability confers protection from/predisposes to the adverse metabolic responses to overfeeding. The concept of a personal fat threshold suggests a large interindividual variation in SAT capacity with ectopic depot expansion/metabolic decompensation once one's own threshold is exceeded. The potential molecular and signaling pathways involved in impaired adipose tissue expandability in individuals with T2D and prediabetes remain to be established. Moreover, it is not known whether exercise may modulate adipose tissue expandability through specific molecular mechanisms. Metabolomics: During the last decade, a number of metabolites have been proposed as potential metabolic biomarkers of insulin resistance and glucose intolerance (19-21). In particular, altered circulating levels of metabolites related to pathways affected by insulin, such as lipolysis, ketogenesis, proteolysis, and glucose metabolism have been suggested (21). Metabolic profiling studies of non-diabetic individuals have reported altered levels of phospholipids, triglycerides, hexoses, α-hydroxybutyrate, glycine, glutamine and branched-chain and aromatic amino acids as early biomarkers of insulin resistance, glucose intolerance and the development of type 2 diabetes (19-23). However, it remains to be determined what extent acute exercise and exercise training changes the metabolites towards a normal metabolic signature in obesity and T2D. OVERALL HYPOTHESIS AND AIMS As outlined above, the investigators hypothesize that IR and its association with mitochondrial dysfunction and lipid accumulation in human skeletal muscle involves abnormalities in the insulin- and exercise-mediated regulation of 1. insulin sensitivity and energy metabolism, 2. insulin signaling, 3. mitochondrial dynamics and mitophagy, 4. lipid droplet function and interaction with mitochondria, 5. acetylation and phosphorylation of enzymes in major metabolic and signaling pathways, as well as 6. transcriptional and signalling networks regulating mitochondrial biogenesis and substrate metabolism. Moreover, the investigators hypothesize that at least some of these abnormalities are related to impaired adipose tissue expandability due to impaired regulation of specific transcription factors or signaling pathways. Finally, these abnormalities will result in metabolomic signatures, which can be used to detect and prevent the development of T2D and CVD. In this proposal, the investigators plan a series of studies, in which we combine state-of-the-art metabolic characterization and a novel exercise-training intervention with detailed investigations of blood samples, skeletal muscle and fat biopsies with the aim to identify defects in the above-mentioned novel regulatory systems and examine their potential relation to known abnormalities in insulin signaling, glucose and lipid metabolism and mitochondrial function in obesity and T2D. Further mechanistic insight will be obtained by further characterization of such defects in cultured myotubes and adipocytes as well as mice models in close collaboration with our research partners. The investigators ultimately expect that this will help us to identify novel targets for treatment of IR and mimicking exercise, which are currently missing in the treatment and prevention of T2D and CVD. SPECIFIC HYPOTHESIS Compared to matched controls, the investigators hypothesize that individuals with obesity and T2D are characterized by abnormalities in the fasting, resting state and/or an impaired effect of insulin, acute exercise and/or high intensity interval training recruiting several muscle groups on: 1. insulin sensitivity, body composition, cardiorespiratory fittness and substrate metabolism 2. Insulin secretion adjusted for insulin sensitivity 3. distal components and modulaters of insulin signaling in muscle and markers of mitochondrial dynamics and mitophagy in muscle and fat 4. regulators of lipid droplet function and interaction with mitochondria in muscle and fat 5. transcriptional and signalling networks regulating muscle metabolism 6. phosphorylation and acetylation of metabolic and signaling enzymes in muscle and fat 7. regulators of adipose tissue expandability in fat 8. metabolomic signature in plasma, fat and muscle METHODS Metabolic characterization and tissue biopsies Anthropometrics and biochemical analysis: Particpants in all cohorts will be examined at least 1 weeks prior to baseline clamp studies. This will include assessment of gender, age, BMI, blood pressure and ECG. Overnight fasting levels of anti-GAD65-antibody, HbA1c, screening blood tests, plasma glucose, FFA, lipid profile, adiponectin, leptin, serum insulin and C-peptide. VO2max is determined by a graded maximal test on a cycle ergometer using indirect calorimetry (8,35). Whole body composition (lean body mass, total, and regional fat mass) will be determined by DXA scans using a Hologic Discovery device (Waltham, MA, US). Lean and overweight/obese controls will be examined by an 2-h standard (75g) oral glucose tolerance test to exclude glucose intolerance (IGT). Euglycemic-hyperinsulinemic clamp: Before (baseline) and after the exercise training program, the participants are examined after a 12-h overnight fast by an euglycemic-hyperinsulinemic clamp (insulin 40mU/min/m2 for 4-h) as described (4-7). The clamps are combined with a 60 min intravenous glucose tolerance test (IVGGT) using a bolus of glucose (0.3 g/kg body weight). Insulin secretion is evaluated by estimates of first and second phase insulin responses during the first 10 min and the last 50 min, respectively. Tissue biopsies are taken as described below. The studies are combined with glucose tracers and indirect calorimetry allowing estimates of glucose disposal rates, glucose and lipid oxidation, and non-oxidative glucose metabolism and energy expenditure. Body fat (%) is determined by the bioimpedance method. Skeletal muscle and fat biopsies: In the basal and insulin-stimulated states of each clamp, skeletal muscle biopsies from m. vastus lateralis and subcutaneous abdominal fat biopsies are taken using a modified Bergström needle with suction under local anesthesia (4-9). Half of each muscle biopsy and all of each fat biopsy are rapidly frozen in liquid nitrogen within 30 s and stored at -130°C for later analyses. One third of each muscle biopsy is immediately transported in an ice-cold isolation buffer to the lab for high-resolution respirometry (see below) (106). Small pieces of muscle (\~5 mg) are embedded in Tissue-Tek and frozen in liquid nitrogen for immunohistochemistal analysis. Small cubes of muscle (1mm3) are pre-fixed in buffered glutaraldehyde and postfixation in osmium tetroxide for electron microscopy (73). Exercise interventions Acute exercise: After the baseline clamp, but prior to initiation of the exercise training program, the effects of acute exercise will be studied. After resting in supine position for 30 min the first set of blood samples and muscle biopsies are taken (see above). Then the subjects perform an aerobic exercise bout for 30 min on a cycle ergometer (followed by 30 min on a rowing machine (30 min) at an intensity of 60% of VO2peak. Blood samples are taken every 15 min, and oxygen consumption is measured by indirect calorimetry. After the exercise bout, the second muscle biopsies are taken while the subjects are resting in supine position, Finally, a third set of muscle biopsies will be taken 4 hours into recovery (see above). Exercise training: In all participants, the effects of 8-weeks supervised high internsity interval training (HIIT) will also be investigated. The HIIT protocol consists of 3 weekly sessions combing rowing and cycling in training blocks of 5 x 1 min high-intensity intervals each interspersed by 1 min active or resting recovery. Between the training blocks, the participants have a 4-min break in which they shifted from rowing to cycling and vice versa. The number of blocks per session is gradually increased with an extra block added every second week, going from 2 blocks in week 1-2 to 5 blocks in week 7-8. VO2max tests and DEXA scannings (body composition) are performed after the training period, but 48 hours before the final post-training clamp. VO2 max is determined in addition after week 4 and 6 in order to adjust the training intensities. Before and after (48 h) the training program all study participants will be examined by a euglycemic-hyperinsulinemic clamp and tissue biopsies as described above. Studies of skeletal muscle and adipose tissue biopsies Mitochondrial respiration: Measurements of mitochondrial respiration in permeabilized muscle fibers are performed in duplicate using a high-resolution respirometer (Oroboros Instruments) as described (96). Routine respiration (state 2) in the absence of adenylates is assessed by addition of malate and glutamate for Complex I substrate supply, followed by ADP-stimulated state 3 respiration. Succinate is then added to assess convergent electron input to Complex I and II. The integrity of the outer mitochondrial membrane is established by the addition of cytochrome c. This method allow studies of mitochondria in situ using very small biopsy samples. Quantitative real-time PCR (qRT-PCR) and western blotting: Total RNA is extracted from muscle and fat biopsies (obtained during the studies described above) using the TRIzol protocol with an extra phenol-chloroform step as described (67,97). Quantity of RNA is determined with a spectrophotometer, and RNA of high quality is assessed using Agilent 2100 Bioanalyser and degradometer software. Total RNA from all samples are treated with DNAse I and reverse transcribed to single-stranded cDNA using TaqMan reverse transcription reagents and random hexamer primers (Applied Biosystems). TaqMan gene expression assays for all relevant genes and TaqMan Universal Master Mix (Applied Biosystems) are used to quantify gene expression changes using Applied Biosystems Prism 7700. House-keeping genes will be tested for changes in response to the interventions, and the three most stable of these genes will be used to normalize gene expression levels. We will perform microarray-based mRNA and miRNA profiling using Affymetrix arrays in collaboration with the Department of Clinical Genetics at Odense University Hospital. Protein abundance and phosphorylation of all enzymes of interest will be studied in human skeletal muscle and fat biopsies (obtained during the studies described above) by Western blotting procedures as described (4-10) using commercial available antibodies, or using antibodies from our international collaborators. As outlined in the background of the proposal, we will measure mRNA levels and protein content and/or phosphorylation of multiple genes/enzymes involved in insulin signaling, Wnt-signaling and other regulators of adipose tissue expandability, mitochondrial dynamics and mitophagy, LD function, and transcriptional and signaling networks regulated by exercise. TEM and immunogold-labelling: Morphology, volume and localization as well as physical interaction between LD and mitochondria in muscle biopsies from the study cohorts before and after the different interventions are determined by transmission electron microscopy (TEM) using the principles of unbiased stereology as described (83,98). Immunogold-labelling using antibodies against e.g. PLIN5, PLIN3 and SNAP23 will be used in TEM studies as described 85,99) to localize and quantify the interaction of these LD and other proteins with mitochondria in human skeletal muscle. Unbiased and targeted MS/MS-based quantitative proteomics: First, we will perform discovery-mode, quantitative MS/MS based proteomic studies of the acetylomes and phosphoproteomes in muscle and fat of individuals with obesity and T2D using novel quantitative techniques including isobaric labelling (iTRAQ) or Tandem Mass Gag (TMT). Later a list of confirmed lysine acetylation and phosphorylation sites identified these studies on muscle and fat enzymes involved in major metabolic pathways and selected signaling cascades will be created and used for targeted analysis. We will take advantage of a novel targeted quantitative proteomic approach called selected reaction monitoring (SRM) using a new high-sensitivity LC-MS/MS system as described (100). By this method hundreds of lysine acetylation and phosphorylation sites can be reliable quantified in whole muscle lysates or isolated mitochondria in a single experiment. Metabolomics: Metabolomics is the study of the metabolome, which represents all metabolites found in a biological sample. This complexity demands sophisticated separation and detection methods. By using comprehensive and quantitative analysis, it is possible to detect a wide range of metabolites including precursors, derivatives, and degradation products within a large dynamic range (101). Recent developments in analytical chemistry, especially liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS) and the availability of extensive software tools for analyzing big data allow the detection, identification and quantification of very large numbers of molecules in the micro- and nano-molar range in body fluids, such as tissues, blood and urine (102). Compared to transcriptomics and proteomics, the metabolome is much more chemically diverse, and thus, multiple approaches involving different extraction and analytical methods must be applied to obtain unbiased and complete information about the entire metabolome. In this project, we will use both LC-MS and GC-MS-based metabolomics. Genetic manipulation of myotubes, adipocytes and mice: Genes and proteins found to be dysregulated in obesity and T2D will be used to design mechanistic studies using genetic manipulation of C2C12 muscle cells and 3T3-L1 adipocytes. We have currently established C2C12 cells and 3T3-L1 adipocytes and facilities for knockdown (siRNA) and overexpression of genes using transfection. When it comes to muscle- and adipose tissue specific transfection and/or knock-out of genes in mice models, our collaboration with professor Jørgen Wojtaszewski at the August Krogh gives us access to muscle-specific knock-out of AMPK α1/α2, mTORC2, RAC1 and PGC1α , which can be used to test their impact or involvement in an observed abnormality in human muscle. Moreover, our collaboration with the group of professor Fredrik Karpe at the Oxford Centre for Diabetes, Endocrinology and Metabolism (OCDEM) at University of Oxford, UK gives us access to carry out genetic manipulation of immortalized cultured human adipocytes for studies of critical enzymes in Wnt-signaling.