https://doi.org/10.1007/s00392-025-02625-4
1Universitätsmedizin der Johannes Gutenberg-Universität Mainz Präventive Kardiologie und Medizinische Prävention Mainz, Deutschland; 2Institut für Medizinische Biometrie; Epidemiologie und Informatik Mainz, Deutschland; 3Augenklinik und Poliklinik Mainz, Deutschland; 4Klinik und Poliklinik für Psychosomatische Medizin und Psychotherapie Mainz, Deutschland; 5Institut für Klinische Chemie und Laboratoriumsmedizin Mainz, Deutschland; 6Klinik für Psychiatrie und Psychotherapie Mainz, Deutschland; 7Klinik für Nephrologie, Rheumatologie, Nierentransplantation Mainz, Deutschland; 8Centrum für Thrombose und Hämostase Mainz, Deutschland; 9Universitätsmedizin der Johannes Gutenberg-Universität Mainz Kardiologie 1, Zentrum für Kardiologie Mainz, Deutschland
Background
Physical activity (PA) plays a valuable role in the prevention of various chronic conditions, particularly cardiovascular disease (CVD), significantly through beneficial anti-inflammatory responses. With this study it was aimed to identify the proteomic signature of PA and relevant pathways through targeted proteomics. The implications on health and CVD were explored in order to gain insight into the molecular mechanisms of the beneficial impact of PA.
Methods
Data from the Gutenberg Health Study, a population-based, prospective cohort study in Germany, were analysed. Subjects were profiled for 92 plasma proteins of systemic inflammation employing proximity extension assay technology (Olink Proteomics, Uppsala, Sweden). PA was assessed using the standardized Short QUestionnaire to ASsess Health-enhancing PA (SQUASH). Elastic net regularized linear regression, adjusted for age and sex, was used to select proteins related to the total weekly PA. A weighted score derived from the selected proteins was used in Cox proportional hazard models to predict all-cause death. The score was investigated in relation to cardiovascular risk factors (CVRFs), comorbidities, and subclinical markers. Protein-protein interaction network and pathway enrichment analyses were performed.
Results
The analysis sample included N=6,040 subjects (mean age 56.2 ± 10.7 years; 45.9% women), in whom 42 proteins were related to weekly PA. Among those, FGF23, DNER and TGFA exhibited the strongest relation. Network analysis revealed that proteins related to growth factors play a key role and are associated with chemokines, interleukins and members of the tumour necrosis factor superfamily. The PA-related proteins were associated with various molecular pathways, including receptor ligand activity, cytokine and chemokine activities, and multiple leukocyte processes. Diabetes mellitus (beta [95% CI]: -0.30 [-0.37; -0.23], p<0.0001), atrial fibrillation (-0.41 [-0.55; -0.28], p<0.0001), chronic kidney disease (-0.30 [-0.37; -0.23], p<0.0001), coronary artery disease (-0.39 [-0.49; -0.20], p<0.0001), and history of myocardial infarction (-0.32 [-0.43; -0.21], p<0.0001) were strong predictors of the protein score. The score was strongly predictive of a lower mortality (hazard ratio [95% CI] per standard deviation: 0.75 [0.68; 0.84], p<0.0001) independent of age, sex, CVRFs, and comorbidities. Sex-specific differences in the association between PA and protein levels were revealed, with notable effects observed in 12 proteins (top 3 FGF23, DNER, TGFA) which were downregulated in women compared to men.
Conclusion
The protein signature of PA revealed a comprehensive overview of the beneficial effects of PA on inflammatory mechanisms in the general population. The findings involved several biological pathways of immune response modulation, emphasizing their relevance to chronic conditions. The results underscore the benefit from regular, daily PA for managing systemic inflammation. However, it should be acknowledged that causality may be bidirectional, with individuals with lower systemic inflammation also being more likely to engage in PA. Further investigation of the direction and mechanisms of these associations is recommended in future studies.