Temporal dynamics of the multi-omic response to endurance exercise training

Regular exercise is associated with reduced all-cause mortality and lower risk of cardiometabolic, neurological, oncologic and other diseases. Although exercise affects nearly all organ systems, prior molecular studies have typically examined one or two ‘omes’ at a single time point, often in one sex and in a limited set of tissues (commonly skeletal muscle, heart, or blood), leaving a gap in organism-wide understanding. The Molecular Transducers of Physical Activity Consortium (MoTrPAC) was established to build a comprehensive molecular map of exercise responses across tissues in animal models and in selected human tissues. This study presents the first whole-organism, temporal, multi-omic map of endurance exercise training in male and female rats to elucidate tissue-specific and shared molecular adaptations, their timing, and sex differences, with relevance to human health.

Previous work leveraging transcriptomics, epigenomics, proteomics, and metabolomics has revealed exercise-induced molecular changes but is typically constrained to one or two omes, a single time point, one sex, and a narrow set of tissues (often skeletal muscle, heart, or blood). Meta-analyses have characterized time trajectories in the transcriptomic response, and studies have shown exercise-induced changes in glucose metabolism, mitochondrial protein acetylation, and coordinated reprogramming of epigenome and transcriptome in skeletal muscle. However, comprehensive multi-omic, multi-tissue, temporal studies spanning both sexes are scarce. This motivated MoTrPAC to generate an organism-wide, multi-omic atlas of exercise responses.

Six-month-old male and female Fischer 344 rats underwent progressive treadmill endurance training for 1, 2, 4, or 8 weeks. Sex-matched sedentary rats served as controls. For trained animals, tissues were collected 48 hours after the last exercise bout to capture sustained adaptations. Phenotyping included VO2max, body fat percentage, lean mass, and body weight. Whole blood, plasma, and 18 solid tissues were profiled using up to nine omic modalities: transcriptomics (RNA-seq), proteomics (global protein abundance), post-translational modifications (phosphoproteomics, acetylproteomics, ubiquitylproteomics), epigenomics (DNA methylation via RRBS; chromatin accessibility via ATAC-seq), metabolomics and lipidomics, and multiplexed immunoassays. Molecular assays were prioritized by tissue quantity and biological relevance, with the gastrocnemius, heart, liver, and subcutaneous white adipose tissue having the most diverse assay coverage. In total, 9,466 assays across 19 tissues and 25 platforms produced 681,256 non-epigenetic and 14,334,496 epigenetic measurements, corresponding to 213,689 unique non-epigenetic and 2,799,307 unique epigenetic features. Differential analyses estimated a training P value per feature, with false discovery rate (FDR) control to define training-regulated features. Timewise summary statistics were computed by sex and time point. Multi-tissue gene mapping aggregated differential features to genes for pathway enrichment. Transcription factor motif enrichment and PTM-signature enrichment analyses inferred TF and kinase activity changes. Empirical Bayes graphical clustering summarized dynamic responses across time and sex, with pathway enrichment applied to feature clusters. Mitochondrial changes were assessed using mitochondrial RNA-seq reads and GSEA with MitoCarta MitoPathways on proteome and acetylome data. Network connectivity analyses (e.g., BioGRID) integrated multi-omic overlaps in skeletal muscle to identify functional clusters and putative regulators. Disease ontology enrichment and cross-study comparisons evaluated translational relevance.

- Endurance training induced widespread, multi-omic molecular adaptations across the organism. At 5% FDR, 35,439 training-regulated features were identified across tissues and omes. While effect sizes were modest (56% of per-feature maximum fold changes between 0.67 and 1.5), changes were pervasive. - Phenotypic adaptations: VO2max increased by 18% (males) and 16% (females) at 8 weeks; male body fat decreased by 5% at 8 weeks with no change in lean mass, whereas female body fat remained stable in trained animals but increased by 4% in sedentary controls; female body weight increased across intervention groups, with no change in males. - Tissue- and ome-specificity: The hypothalamus, cortex, testes, and vena cava had the smallest proportion of training-regulated transcripts; blood, brown and white adipose tissues, adrenal gland, and colon showed extensive transcriptomic changes. Gastrocnemius, heart, and liver exhibited substantial protein and PTM regulation. Metabolomic alterations were broad across tissues. - Multi-tissue gene responses: Among six deeply profiled tissues (gastrocnemius, heart, liver, WAT-SC, lung, kidney), 11,407 differential features mapped to 7,115 genes; 67% were tissue-specific, with the most in WAT-SC. Lung–WAT shared 249 genes enriched for immune pathways; heart–gastrocnemius shared genes enriched for mitochondrial metabolism (including OPA1 and MFN1). - Ubiquitous responses: Twenty-two genes were training-regulated in all six tissues, enriched for heat shock response pathways. Proteomics showed marked up-regulation of HSPs, including HSPA1B and HSP90AA1. Kininogenases KNG1 and KNG2 were also broadly regulated, implicating the kallikrein–kinin system in exercise benefits. - Regulatory inferences: TF motif enrichment reflected tissue identities (for example, MEF2 family in heart and skeletal muscle; hematopoietic TFs GABPA, ETS1, KLF3, ZNF143 in blood). PTM-signature analysis indicated altered kinase activities across tissues (e.g., AKT1, mTOR, MAPK). In heart, predicted decreases in AGC kinases (e.g., AKT1) and increases in tyrosine kinases (SRC, mTOR) were observed; increased phosphorylation at SRC targets GJA1 Y265 and CDH2 Y820 suggests extracellular remodeling and gap junction modulation. - Temporal dynamics and muscle adaptation: Graphical clustering showed major paths converging to sex-consistent up-regulation by week 8 in gastrocnemius, vastus lateralis, and heart. Shared enriched pathways included mitochondrial metabolism, biogenesis and translation, and cellular response to heat stress. - Immune modulation: KEGG immune pathway enrichment at 8 weeks showed male-specific up-regulation in white and brown adipose tissues, limited enrichment in striated muscle and brain, and down-regulation in lung and small intestine (more robust in females). Transcript profiles in adipose correlated strongly with B, T, and NK cell markers, suggesting immune cell recruitment/proliferation; small intestine showed down-regulation of inflammation-related transcripts and those linked to IBD risk loci (e.g., MHC class II), consistent with improved gut homeostasis. - Mitochondrial and metabolic remodeling: Oxidative phosphorylation and mitochondrial pathways were broadly enriched; mitochondrial biogenesis increased in skeletal muscle, heart, and liver. The heart showed coordinated up-regulation of nearly all glycolysis–gluconeogenesis enzymes. Metabolomic class enrichments were most numerous in liver, then heart, lung, and hippocampus. Specific metabolites reflected functional adaptations (e.g., kidney increases in 1‑methylhistidine across early time points; cortisol elevations; liver increase in 1‑methylnicotinamide at 8 weeks). - Liver lipid and acetylation changes: Liver displayed extensive regulation in lipid metabolism across proteome, acetylome, and lipidome, including increased phosphatidylcholines and decreased triacylglycerols at 8 weeks, and increased abundance and acetylation of peroxisomal and mitochondrial metabolic proteins. These adaptations suggest mechanisms for protection against hepatic steatosis. - Sex differences: 58% of 8-week training-regulated features exhibited sex-differentiated responses. Opposite responses were notable in adrenal transcripts, lung phosphosites and chromatin accessibility, WAT transcripts, and liver acetylation. Adrenal gland showed pronounced, sex-divergent transcriptional remodeling, implicating steroidogenesis and mitochondrial functions. - Translational relevance: Rat skeletal muscle transcriptomic and proteomic changes significantly overlapped with human long-term training and HIIT datasets. Disease ontology enrichment linked down-regulated genes in WAT, kidney, and liver to type 2 diabetes, cardiovascular disease, obesity, and kidney disease, underscoring clinical relevance.

This study delivers an organism-wide, temporal, multi-omic map of endurance training adaptations, addressing longstanding gaps in scope (multi-ome), breadth (multi-tissue), sex inclusion, and temporal resolution. The findings demonstrate that endurance training orchestrates coordinated molecular changes across tissues, prominently involving mitochondrial biogenesis and function, heat shock responses, metabolic reprogramming, and tissue-specific immune modulation. Regulatory analyses implicate key TFs (e.g., MEF2 in muscle) and kinases (e.g., SRC, mTOR, AKT1) in mediating these adaptations, with evidence for extracellular matrix remodeling in the heart and immune cell recruitment in adipose tissue (male-specific). The gut and lung show reduced inflammatory signatures, especially in females and early time points, respectively, suggesting improved barrier function and structural remodeling. The robust overlaps with human exercise datasets and disease ontology enrichments indicate that these rat multi-omic adaptations are highly relevant to human cardiometabolic and inflammatory diseases, providing mechanistic links between exercise and improved health outcomes (e.g., NAFLD, IBD, cardiovascular disease, tissue injury recovery).

The MoTrPAC multi-omic atlas reveals the temporal and sex-specific molecular architecture of endurance training across 19 tissues, identifying pervasive and tissue-specific adaptations in mitochondrial, metabolic, stress-response, and immune pathways. Novel insights include widespread heat shock protein up-regulation, kinase activity reprogramming in the heart, male-specific immune activation in adipose tissues, decreased inflammatory signaling in lung and intestine, and extensive liver acetylome and lipid remodeling suggestive of protection against steatosis. These data provide a foundational resource for hypothesis generation and translational research linking exercise to disease prevention and therapy. Future work should expand single-cell and spatial resolution, incorporate additional omic platforms (e.g., microbiome), sample acute post-exercise windows, and extend to human multi-tissue profiling to refine mechanistic understanding. All data and tools are publicly available via the MoTrPAC Data Hub and associated software and visualization resources.

Samples were collected 48 hours after the last exercise bout, capturing sustained rather than acute responses. Assays were performed on bulk tissues, limiting cell-type resolution and potentially conflating cellular composition changes with cell-intrinsic regulation. Some tissues had limited omic coverage, and certain relevant platforms (e.g., microbiome profiling) were not utilized. As an observational, hypothesis-generating multi-omic resource, findings require experimental validation. Sex differences and temporal dynamics may vary with strain, age, or training protocol, which were not explored here.