Lipid metabolism links nutrient-exercise timing to insulin sensitivity in men classified as overweight or obese

Context Pre-exercise nutrient availability alters acute metabolic responses to exercise, which could modulate training responsiveness. We hypothesised that in men with overweight/obesity, acute exercise before versus after nutrient ingestion would increase whole-body and intramuscular lipid utilization, translating into greater increases in oral glucose insulin sensitivity over 6-weeks of training. Design and Participants We showed in men with overweight/obesity (mean±SD for BMI: 30.2±3.5 kg×m-2 for acute, crossover study, 30.9±4.5 kg×m-2 for randomized, controlled, training study) a single exercise bout before versus after nutrient provision increased lipid utilisation at the whole-body level, but also in both type I (p<0.01) and type II muscle fibres (p=0.02). We then used a 6-week training intervention to show sustained, 2-fold increases in lipid utilisation with exercise before versus after nutrient provision (p<0.01). Main Outcome Measures Postprandial glycemia was not differentially affected by exercise training before vs after nutrient provision (p>0.05), yet plasma was reduced with exercise training before, but not after nutrient provision (p=0.03), resulting in increased oral glucose insulin sensitivity when training was performed before versus after nutrient provision (25±38 vs −21±32 mL×min-1×m-2; p=0.01) and this was associated with increased lipid utilisation during exercise (r=0.50, p=0.02). Regular exercise prior to nutrient provision augmented remodelling of skeletal muscle phospholipids and protein content of the glucose transport protein GLUT4 (p<0.05). Conclusions Experiments investigating exercise training and metabolic health should consider nutrient-exercise timing, and exercise performed before versus after nutrient intake (i.e., in the fasted state) may exert beneficial effects on lipid utilisation and reduce postprandial insulinemia. Précis Exercise in the fasted-versus fed-state increased intramuscular and whole-body lipid use, translating into increased muscle adaptation and insulin sensitivity when regularly performed over 6 weeks.

To this end, the aim of the present work was to assess the acute and chronic effects 144 of manipulating nutrient-exercise timing on lipid metabolism, skeletal muscle 145 adaptations, and oral glucose insulin sensitivity in men with overweight or obesity. 146 We hypothesized that nutrient-exercise interactions would affect the acute metabolic 147 responses to exercise, with increased whole-body and intramuscular lipid utilization 148 with exercise before versus after nutrient provision. We also hypothesized that these 149 acute responses to exercise before versus after nutrient provision would result in greater training-induced increases in oral glucose insulin sensitivity in men classified 151 as overweight or obese.

153
Ethical Approval 154 This project comprised two experiments. We first assessed the acute metabolic and 155 mRNA responses to manipulating nutrient-exercise timing (Acute Study), followed 156 by a 6-week randomized controlled trial to assess the longer-term (i.e. training) 157 adaptations in response to nutrient-exercise timing (Training Study). All participants 158 provided informed written consent prior to participation. Potential participants were 159 excluded if they had any condition, or were taking any medication, known to alter    Participants ate their evening meal before 2000 h the evening prior to any exercise 211 sessions. Participants in BR-EX were given a drink in an opaque bottle made from 1.3 g carbohydrate·kg body mass -1 maltodextrin (MyProtein, Northwich, UK) with 213 vanilla flavoring (20% carbohydrate solution) for consumption 2-h before exercise.

214
They were asked not to eat or drink anything else (except water ad libitum) in this 215 period and confirmed they had consumed the drink before exercising. After exercise, 216 they were provided a taste matched placebo (water and vanilla flavoring) to consume 217 2-h after exercise and were asked not to consume anything else during this period.

218
Participants in EX-BR were given the same drinks, but with the order of the drinks 219 reversed. Participants in CON were given the same drinks for three days per week 220 during the intervention, with the carbohydrate drink as breakfast (0800-0900 h) and 221 the placebo for consumption with their lunch (1100-1300 h). These participants were 222 asked not to consume anything else between the drinks. There were no other diet 223 controls in the intervention. Blinding of the groups was deemed successful, because 224 at exit interview, 25 participants (83%) revealed they could not detect a difference 225 between the carbohydrate and placebo drinks or could not identify which contained 226 carbohydrate. Five participants determined which drink had carbohydrate (CON n=1, 227 BR-EX n=2, EX-BR n=1), but this is within the proportion that could do so at random.

229
Pre-and post-intervention, an oral glucose tolerance test (OGTT), a vastus lateralis 230 muscle sample (fasting, rested state) and an exercise test (to assess V O2 peak and 231 the capacity for lipid utilization during exercise in the fasted-state) were undertaken.

232
Post-intervention tests were between 24 h to 48 h (for muscle sampling) and 48 h to 233 72 h (for OGTT) after the last exercise training session, to reduce any residual effects 234 of the last exercise bout performed on these measurements. The primary outcome 235 for the Training Study was the pre-to post-intervention change in the glycaemic 236 and insulinemic responses to the OGTT, which were also used to derive an index of 237 oral glucose insulin sensitivity (as described subsequently).  and frozen at -20°C, before longer-term storage at -80°C. A proportion of the sample 283 (4 mL) was allowed to clot in a plain vacutainer prior to centrifugation, for serum.

284
Samples were analyzed for plasma glucose, glycerol and NEFA using an ILAB 650    Aldrich; reference #EZHCP-20K) concentrations were measured using commercially 299 available ELISA kits (intra-assay CV for insulin: 3.86% and for C-peptide: 4.26%).    Samples were homogenized with a dounce homogenizer, before a 60 min incubation 323 (4°C with rotation) and 10 min centrifugation (4°C and 20,000 g). The protein content 324 of the resultant supernatant was measured using a bicinchoninic acid assay.

Muscle glycogen (Acute Study)
Muscle glycogen concentrations were measured using a method described 363 previously (36). Briefly, 10-15 mg of frozen tissue was powdered and transferred into 364 a glass tube pre-cooled on dry ice. Thereafter, the samples were hydrolyzed by 365 adding a 500 µl of 2M HCL and then incubated for 2 h at 95 ºC. After cooling to room 366 temperature, 500µl 2M NaOH was added. Samples were centrifuged and the    to provide >90% power with α set at 0.05. In the Training Study, a sample size 479 estimation was completed using data from a training study in healthy, lean men (26).

480
In that study, a change in the plasma glucose AUC for an oral glucose tolerance test 481 mmol·min·L -1 for a CON group. With α set at 0.05, 9 participants were required for a 483 >90% chance of detecting this effect. We therefore recruited 30 participants to 484 account for the possibility of an unequal allocation of participants across three 485 groups when using a stratified randomization schedule. Participants were allocated 486 to the CON (n=9), BR-EX (n=12) or EX-BR (n=9) groups using this schedule, which 487 was generated by JPW and included a factor for Physical Activity Level (PAL) and matched-pairs signed rank tests) were employed. In the Acute Study, differences 497 between groups were assessed with paired t-tests or a two-way repeated measures 498 ANOVA (for variables dependent on time). In the Training Study, one-way ANOVAs 499 were used to assess differences between groups at baseline and two-way mixed-500 design ANOVAs were used to assess differences between groups in response to the 501 intervention (group x time). If interaction effects were identified, independent t-tests 502 were used to locate variance, with Holm-Bonferroni step-wise adjustments made.

503
Correlations between variables were explored using Pearson r or Spearman R for

515
Exercise before nutrient ingestion increases whole-body and skeletal muscle 516 lipid utilization but does not differentially modulate muscle gene expression 517 In the Acute Study, exercising before versus after nutrient provision increased the 518 acute plasma glucose and serum insulin responses to food consumption (Figure 2A     In the Training Study rates of whole-body lipid utilization were around 2-fold higher 560 with exercise before versus after nutrient provision and this difference between the 561 conditions was sustained throughout the whole 6-week intervention ( Figure 4A). As   3 and 4). However, exercise training 580 before, but not after nutrient intake reduced postprandial insulinemia (time x group 581 interaction p=0.03; Figure 4C). Exercise training before versus after nutrient 582 provision also increased the OGIS index (time x group interaction p=0.03; Figure   583 4D with pre-and post-intervention data in Table 3). The plasma C-peptide-to-insulin 584 ratio was not differentially altered by nutrient-exercise timing (time x group 585 interaction, p=0.12; Figure 4E). The change in the OGIS index in response to  Figure 8B). However, there was a ~2-fold increase in skeletal muscle GLUT4 protein 630 levels with exercise training performed before (p=0.04), but not after nutrient 631 provision (p=0.58) versus a non-exercise control group (Figure 8A). There was also 632 an increase in the protein levels of the CHC22 clathrin isoform and its associated adaptor protein (GGA2) relative to the CHC17 clathrin isoform, with exercise before 634 versus after nutrient provision (both p<0.05; Figure 8B). When we examined the 635 CHC22 isoform alone [data not shown but available online (51)]we noted baseline 636 differences which may have confounded the interpretation of these fold-changes due 637 to regression to the mean. We thus present the CHC22/CHC17 ratio ( Figure 8B) to 638 reflect GLUT4-associated clathrin-mediated membrane traffic relative to total 639 clathrin-mediated membrane traffic.    In the training study, we then showed that the acute increases in whole-body lipid 692 utilization during a single bout of exercise performed before versus after nutrient intake were sustained throughout 6-weeks of exercise training. Moreover, only 694 exercise training performed before nutrient intake reduced postprandial insulinemia 695 and increased the oral glucose tolerance test-derived estimate of peripheral insulin 696 sensitivity (i.e. the OGIS index). As the plasma C-peptide-to-insulin ratio was not 697 differentially altered by nutrient-exercise timing, the reduction in postprandial 698 insulinemia with exercise performed before versus after nutrient ingestion is likely to 699 be due to a reduction in insulin secretion rather than an increase in hepatic insulin 700 extraction (55). It should also be noted that difference between the exercise groups 701 for the change in the OGIS index was also broadly equivalent to the difference 702 between individuals classified as having a healthy phenotype compared to 703 individuals with impaired glucose tolerance (56).    It should also be noted that the responses observed for OGIS were an interaction 792 between groups, and thus the response to exercise before nutrient provision is an 793 increase relative to the non-exercise control group and the exercise after nutrient 794 intake group. Accordingly, these data may be specific to high-carbohydrate provision      Data are means (SD) for men classified as overweight or obese. BMI = Body Mass Index; V O2peak = peak oxygen uptake; PPO = peak power output; CON = control; BR-EX = breakfastexercise; EX-BR = exercise-breakfast.