The search for human skeletal muscle memory : exercise effects on the transcriptome and epigenome

Abstract

Regular physical activity is an environmental stimulus that is highly associated to many health benefits, while physical inactivity is detrimental for health and physical function. Regular exercise training is used in the prevention and treatment of a large number of disease conditions, including obesity, type II diabetes, cardiovascular disease and cancer, and reduces the risk for premature death. Most tissues adapt to exercise training, not least skeletal muscle tissue, which is highly plastic. The local adaptation of muscle is important not only for muscle function but also the health effects of training that affect the whole body. The cellular adaptations in skeletal muscle are driven by extra- and intracellular signals arising from the exercise stimulus, for example changes in shear stress, oxygen tension, energy levels, pH and temperature. Ultimately, these cellular perturbations lead to gene expression and protein alterations that improve muscle function. Thus, it is important from a clinical, as well as basic science perspective to understand the regulation of skeletal muscle gene activity and how activity changes contribute to the many health benefits of a physically active lifestyle.

The understanding of training-induced changes in gene expression and the underlying mechanisms have progressed extensively over the past 20 years. Still, many key mechanisms remain to be investigated. The overall purpose of this thesis was to investigate the influence of epigenetic mechanisms, i.e. DNA methylation and post-translational modifications of histones, on endurance training adaptation. Epigenetic mechanisms are important for cellular memory. Thus, another objective was to investigate if there were any residual intrinsic memory effects of previous endurance training, and if that could induce different responses to a repeated training period after detraining.

The results in this thesis are based on skeletal muscle biopsies from the vastus lateralis, taken before and after three months, or six weeks, of endurance training, or at rest in elite athletes and sedentary individuals. In the first study, the baseline skeletal muscle transcriptome was investigated. Studies using repeated skeletal muscle sampling regularly assume that potential changes are due to the intervention and not inherent variability between samples. The results showed, using global RNA sequencing analysis, that tissue homogeneity was remarkably high within a muscle and in the corresponding muscle of the contralateral leg of an individual, while the transcriptome difference between male and female skeletal muscle was substantial. This study also found 23 000 isoforms expressed in skeletal muscle at baseline, together with almost 2500 previously unannotated, novel transcripts, out of which at least five were protein-coding.

The transcriptome changes induced by three months of one-legged knee extension training were very significant. Over 3000 isoforms were found to be differentially expressed, as well as 34 of the novel transcripts discovered at baseline. The one-legged training regime meant that the other leg was included as an intraindivdual control leg, which was exposed to the same other environmental factors such as diet, stress, sleep etc. We found that the training response of the trained leg was very specific, although significant but markedly smaller changes occurred also in the untrained leg. At the protein level, a specific investigation of HIF (hypoxia inducible factor) was performed. HIF is activated by acute exercise, but was hypothesized to be attenuated by long-term training due to its inhibitory effect on mitochondrial energy production. A comparison of skeletal muscle from elite athletes with normally active individuals, showed that the negative regulators of HIF were higher in the elite athletes, indicating a reduced HIF activity in that group. This was supported by similar findings in a six-week bicycle training study.

Three months of endurance training induced changes in DNA methylation at almost 5000 specific sites across the human skeletal muscle genome that were associated to functionally relevant transcriptional changes. Many of these changes occurred in regulatory enhancer regions and the differentially methylated sites were associated to transcription factor binding sites for myogenic regulatory factors (increases in methylation) and the ETS family (decreases in methylation). Six weeks of bicycle training showed a strong trend towards a global downregulation of trimethylation of histone H3, lysine 27, previously described as a dynamic and predominantly inhibitory modification. The specific genes potentially affected by this histone modification in response to training are currently being analyzed using chromatin immunoprecipitation followed by sequencing.

After the initial three months of one-legged endurance training, a subset of the subjects came back after nine months of detraining and performed a second three-month training period. This time, they trained both legs in the exact same way as one leg was trained in the first period. One leg had thus been previously well-trained, while the other was previously untrained. Potential residual effects were investigated by comparing biopsies obtained from both legs before starting the second training period. At the transcriptome level, there were no indications of remaining effects, although the exertion perceived in the first training session of period 2 was lower in the previously trained leg. Repeated training induced similar changes physiologically and at the global transcriptome level between the two legs. There were specific differences in the gene activity changes between the legs, but with the current approach, we found no overall significant differences in the response to a repeated training period.

Collectively, the results in this thesis show that endurance exercise training induced associated changes in the epigenome and transcriptome of human skeletal muscle. The data included an in-depth analysis of the human skeletal muscle transcriptome at baseline and how it changes in response to repeated endurance training periods, with no detectable muscle memory of previous training at the transcriptome level. The results contribute to a better understanding of the molecular pathways involved in physiological adaptation to endurance training and can potentially be used to describe how training prevents disease development and different dysfunctions.