Jacques, Macsue (2020) The epigenetic basis of variable response to exercise training. PhD thesis, Victoria University.

Abstract

Exercise training provides health benefits to the general population, but there is considerable variability in the individual response to similar training. Some people have limited improvements following exercise (“low responders”), while others seem to improve considerably (“high responders”). To date, most exercise studies that have claimed to identify “low” or “high” responders assumed that if the participants were to repeat the same exercise training, they would show a similar response. However, within- subject variability has not been tested, which might lead to inaccurate classification of exercise responses at the individual level and the waste of precious research resources. Exposing individuals to a repeated or longer training intervention can assist in identifying the magnitude of responses to exercise training with better accuracy. Recent evidence also suggests that the response to exercise training may be influenced by epigenetic signatures. Epigenetics is a reversible process that affects how genes are expressed in cells, and it carries the memory of past cellular and environmental events. To date, no study has tested whether individual response is influenced by epigenetic marks. Thus, the overarching aim of this thesis is to identify the physiological, molecular, and epigenetic marks of exercise responses. Twenty young, healthy men from the Gene SMART (Skeletal Muscle Adaptive Response to Training) study completed a repeated and a longer exercise training intervention to measure within-subject variability and to obtain individual progress curves (See Figure 3.1 for study design). Participants underwent a four-week control period followed by four weeks of High-Intensity Interval Training (HIIT), had a washout period of > 1 year, and underwent another four weeks of HIIT followed by an additional 8 weeks of HIIT. The HIIT program was adjusted to individual fitness levels that were re-assessed every four weeks during the intervention to ensure improvements. Participants’ peak power output (Wpeak), lactate threshold (LT), and maximal oxygen uptake (VO2max) were assessed in duplicates at each time point. We used five known statistical methods to investigate changes in fitness and mixed models to estimate individual response. Muscle biopsies were collected at each time point to measure mitochondrial markers (i.e. mitochondrial respiration, citrate synthase, cytochrome C oxidase, succinate dehydrogenase, mitochondrial copy number, fibre typing, and myosin heavy chains PCRs), as well as genome-wide DNA methylation profiles in skeletal muscle using the Illumina HumanMethylation EPIC array. In Chapter 3, we show that at the group level, all physiological measures increased in a dose-response manner following HIIT (p<0.05). We found no changes in mitochondrial function and content or fibre type distributions. Baseline citrate synthase (CS) was associated with HIIT-induced changes in cytochrome-c oxidase (COX) and vice-versa (p < 0.05). At the individual level, we successfully identified trainability in physiological measurements using the repeated testing approach but failed to do so using the repeated intervention approach. We did not identify consistent individual response at the molecular level (mitochondrial function and content and fibre type distribution) using either approach, as measurements were highly variable within participants. We then investigated the reliability of the mitochondrial respiration technique (Chapter 4) by measuring the Technical Error of measurement (TEM) and the coefficient of variation (CV) for each mitochondrial complex. While the correlation between the two chambers was good for all complexes (R > 0.7 p < 0.001), the TEM was large (7.9 to 27 pmol·s-1·mg-1), and the CV was > 15% for all complexes. We performed statistical simulations to determine the sample size that would be required to detect a range of effect sizes at 80% power. We found that duplicate measurements on 75 participants are required to detect a 6% change in mitochondrial respiration after an intervention. Finally, Chapter 5 and 6 focus on the DNA methylation measures at the group and individual level respectively. For the first time at the group level, we have investigated DNA methylation patterns that are associated with fitness by combining three measurements of performance into a comprehensive z-score (Chapter 5). We found 12,107 DMPs that were associated with baseline fitness (z-score) (FDR < 0.005), 18.2% of which were hypomethylated and 81.8% hypermethylated with higher fitness levels. We identified 1,268 DMRs for baseline fitness, 15.3% of which hypomethylated and 85.7% hypermethylated. Hyper-DMRs were robustly over-represented in genic enhancers and flanking active TSS, and highly depleted in strong and weak candidate enhancers. Hypo and Hyper-DMRs had a moderate association with bivalent enhancers and promoters. Both hyper and hypo-DMRs presented a moderate representation in regions actively repressed by PolyComb proteins. Finally, significant DMRs were enriched for 26 GO terms, and these pathways were related to muscle system processes, actin cytoskeleton organization and regulation of actin filament and cytoskeleton processes. Next, we investigated the effects of exercise on the methylome, and surprisingly we observed an inverse pattern of DNA methylation profile after exercise. In summary, we found 568 DMPs that significantly changed after the 4 weeks (FDR < 0.005), and out of those only 1.4% were hypermethylated and 98.6% were hypomethylated. We identified 17 DMRs associated with changes in DNA methylation in response of 4 weeks of HIIT, and 100% of DMRs were hypomethylated. Lastly, we intersected DMPs that were significant for both baseline fitness z-score and after 4 weeks of HIIT. Five DMPs were significant, and they appeared to have inverse patterns for baseline z-score (more hypermethylation) and 4 weeks of HIIT (more hypomethylation). When we transitioned to the individual level study in Chapter 6, neither of our approaches (i.e. repeated intervention and repeated testing) yielded many significant results: only one DMP was significant (cg11260483, p-value: 3.22000e-10, adj.p-value: 0.00022) after the repeated intervention, and no DMPs significant after the repeated testing approach. The challenging experimental design of this thesis provided high resolution, longitudinal physiological and molecular profiles in skeletal muscle following repeated exercise training and testing. It yielded novel insights into the phenomenon of trainability in humans; young, healthy men displayed individual responses to HIIT at the physiological level, but not at the molecular level. This thesis also issued methodological considerations for protocols aimed at measuring individual response (Chapter 3). In particular, the high within-subject variability we observed led us to conclude that many repeated testings on the same individual at regular intervals during the training program, along with a moderate-to-large sample size, were necessary to estimate inter-individual variability in response to training. The mitochondrial respiration technique showed high technical variability (Chapter 4), making the measurement unreliable in our study with only n = 20 men and only two duplicates per individual. The typical sample sizes used in exercise training studies (n < 20) are likely insufficient to capture exercise-induced changes in mitochondrial respiration at the group level, let alone the individual level. Lastly, we observed a clear DNA methylation profile association with fitness levels (Chapter 5). However, when an exercise intervention was applied, we noticed a change in DNA methylation patterns that were inverse to those observed at baseline for the fitter participants. Such observations left us wondering on potential reasons to why this occurs. Thus, future research should also integrate the methylome with transcriptome and proteome to elucidate the mechanisms underlying adaptations to exercise training.

Search Google Scholar

Repository staff login