Epigenome–proteome integration in aged skeletal muscle after exercise training

Epigenomic alterations in skeletal muscle after acute exercise and chronic training are becoming more established. In particular, modifications to the DNA methylome (e.g., but not limited to; Pilotto et al., 2024; Seaborne et al., 2018) and how they affect corresponding alterations in transcript expression (Turner et al., 2019 reviewed in Sharples (2024)). However, in skeletal muscle it is not yet well established how alterations in the DNA methylome after training feed through to alterations at the protein level. Exciting and recent multi-omic meta-analysis studies in humans have begun to investigate whether any relationships exist between the methylome and proteome after exercise training (Jacques et al., 2023, 2024). However, across the field in humans there are currently fewer proteome samples to compare with the methylome (especially relative to many transcriptome samples). Furthermore, when using paired samples, although shifts in DNA methylation occur in relevant exercise pathways following training (reviewed by Egan & Sharples, 2023; Sharples & Turner, 2023), methylome analyses that includes hundreds of thousands or even millions of DNA sites (with many sites within each gene) may become a challenge when directly comparing statistical approaches in proteomic data sets. The Molecular Transducers of Physical Activity Consortium (MoTrPAC) have recently begun to explore the multi-omic adaptations after endurance exercise training in genetically identical rats (reviewed in Sharples, 2024); however, the integration of the paired methylome–proteome has not yet been conducted in aged muscle following an exercise training intervention. Kevin Murach and Yuan Wen's teams attempt to bridge an important gap in epigenome–proteome association in skeletal muscle after exercise training in aged mice (22–24 months old) (Chambers et al., 2025). Importantly, late-life exercise has previously been associated with epigenetic reprogramming in mouse skeletal muscle (Murach et al., 2022). Higher levels of aerobic exercise in older age (Voisin et al., 2023), after disease (Gorski et al., 2023) and following resistance training (Blocquiaux et al., 2022; Ruple et al., 2021) have also been associated with restoration of the epigenetic landscape in human skeletal muscle. Using their established mouse model of 8 weeks progressive weighted wheel running (PoWeR) the authors implemented exercise training in aged mice (22–24 months old) and integrated the methylome and proteome of the resulting gastrocnemius muscle compared with a control group of age-matched sedentary control animals (Chambers et al., 2025). The prominent effect of PoWeR training was evident at the methylome (with 4,999 differentially methylated CpG sites identified, ∼13% of measured CpGs) and the proteome level (1,032 proteins significantly altered, with 90% of these upregulated). DNA methylome responses after PoWeR training identified alterations that have been described in human skeletal muscle after exercise, such as differential methylation of the nuclear receptor 4A (NR4A) genes (Maasar et al., 2021) and homeobox (HOX) genes that have been demonstrated to be influenced by age and physical activity levels (Turner et al., 2020). The group were also able to integrate the methylome and proteome data using an algorithm called binding and expression target analysis (BETA). For use in methylome data, BETA takes the locations of the differential methylated CpGs and combines the magnitude of change, significance and their location/distance relative to gene regulatory regions to provide a ‘regulatory potential score’. Therefore, this tool identifies genes and proteins that are more likely to be regulated by the epigenetic changes that are detected. Subsequently, pathway enrichment can also be performed on the genes that are predicted by BETA. This is a useful tool within the field, because there are often many single CpGs altered amongst thousands of genes (and many CpGs within the same gene) identified in skeletal muscle following acute exercise and after chronic training. Typically, to account for this, differential methylated region (DMR) analysis is performed to identify genes with multiple CpGs altered in the same direction in small regions of the same gene. BETA extends this approach with the magnitude of change and distance to known regulatory regions. Therefore, for discovery or hypothesis-generating purposes, BETA analysis could also help to further support DMR analysis and provide crucial information on the genes likely to be altered at the expression or protein level with a higher degree of confidence. Then, by using paired transcriptomic or proteomic analysis of the same samples, it is then possible directly to associate the hypothesis-generating BETA prediction of the epigenetic regulation of the gene with the quantitative confirmation of the alteration at the gene and/or protein level. Previously, using BETA, the group had identified epigenetic regulation of mitochondrial function genes in mouse skeletal muscle after mechanical overload-induced hypertrophy that matched muscle respirometry analysis (Ismaeel et al., 2024), suggesting an important epigenetic component in the regulation of metabolic pathways after muscle adaptation. In the present study, following PoWeR training, BETA predicted the epigenetic regulation of 216 upregulated proteins. Pathway enrichment of these proteins identified epigenetic regulation of protein translation, with the majority of genes belonging to ribosomal (RPL6, RPL18, RPL29, RPL31, RPLP0, RPS10 and RPS14) and mitochondrial ribosomal processes (MRPL1, MRPL2, MRPL4, MRPL10, MRPL11, MRPL14, MRPL28, MRPL37, MRPL43, MRPL45, MRPL46, MRPL53, MRPS11, MRPS22 and MRPS25) in addition to the eukaryotic initiation factors (EIFs) involved in the initiation of protein translation in the ribosome (EIF2B3, EIF2B4, EIF2B5 and EIF3D). Furthermore, together with the above alterations, NADH proteins and those in the electron transport chain (ETC) complex 1 were epigenetically regulated genes that were also upregulated at the protein level (NDUFA2, NDUFA3, NDUFA8, NDUFB3, NDUFB9, NDUFB11, NDUFC2, NDUFS1, NDUFS2, NDUFS6, NDUFS7 and NDUFV3). Overall, highlighting metabolic processes associated with energy metabolism, mitochondrial protein synthesis and protein translation seem to be the most prominent epigenetically regulated processess following exercise training in aged muscle. BETA also predicted epigenetic regulation of LACTB, MIB1 and UBR4, with MIB1 and UBR4 both being E3-ubiquitin ligases. Indeed, UBR4 and UBR5 have been identified as important in muscle mass regulation in model organisms and humans (Hughes et al., 2021; Hunt et al., 2019; Seaborne et al., 2019). Computationally the authors also undertook Landscape In Silico deletion Analysis (LISA) using the BETA upregulated proteins as an input. LISA uses ChiP-sequencing data and computation to predict transcriptional regulators likely to be involved in the gene (or ultimately the protein) response. Using LISA, Proto-Oncogene C-Myc (MYC) was highlighted as a transcription factor using both the genes corresponding to upregulated proteins predicted by BETA and the global proteome response. MYC overexpression in skeletal muscle can alter DNA methylation (Jones et al., 2023); therefore, this study adds to accumulating evidence that MYC is likely to be an important regulator of the epigenetic landscape that might subsequently impact the protein-level response to exercise training in ageing skeletal muscle. With respect to the exercise training protocol in mice, a limitation of the study is that PoWeR is not a perfect analogue for exercise training, e.g. resistance or aerobic. PoWeR does, however, include combined higher loading parameters together with a continuous contraction regime of both progressive resistance and continuous wheel running, leading to adaptation characteristic of both resistance and endurance training (reviewed by Murach et al., 2020). Where PoWeR also presents a more relevant physiological exercise stimulus without being invasive in comparison to other mouse models such as synergist ablation. Therefore, the PoWeR model, alongside other recent models, such as implanted electrical stimulation (Spillover) (Schmoll et al., 2018), serves as an important rodent model to investigate molecular responses and adaptation to exercise, perhaps lending itself to combined loading/intensity exercise stimuli that would perhaps be more reflective of the majority of mixed sports that combine both a resistance and a continuous element (Egan & Sharples, 2023). For methylome analyses, the authors use a custom Horvath mammalian array of ∼36K CpG site coverage on the DNA. Although this is much smaller coverage than Illumina EPIC arrays (now version 2 EPICs arrays cover ∼1 million CpGs) and much lower in comparison to bisulphite sequencing technology, the Horvath arrays are designed to cover conserved CpGs across mammalian species and are important in predicting epigenetic ageing. Therefore, making interpretation and translation across species more likely, in addition for it being possible to study known epigenetic biomarkers of ageing. Consequently, it is also interesting that, in the present study, the 36K assay identifies some age-related CpGs after exercise training in old mice, and again suggests that the epigenetic landscape altered by age itself could be influenced positively by exercise training (Blocquiaux et al., 2022; Gorski et al., 2023; Murach et al., 2022; Ruple et al., 2021; Turner et al., 2020; Voisin et al., 2023). In the future, although the largest proportion of nuclei in human skeletal muscle tissue are myonuclei, hence whole-tissue DNA analysis is perhaps reflective of changes across the tissue niche, it seems that there might be a distinct epigenomic contribution of the other cell types resident within skeletal muscle tissue (Wen et al., 2021). Therefore, it would be an interesting avenue for the field to explore the methylome at the single nuclei or single muscle fibre resolution and also to investigate other epigenetic modifications, such as chromatin accessibly (via ATAC-seq), that might precede or interact with DNA methylation changes to alter transcriptional and translational potential after exercise training (reviewed in: Sharples, 2024). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. No conflicts of interests are declared. Sole author. The Research Council of Norway (314157).