Plasma Proteomic Signatures of Physical Activity Provide Insights into Biological Impacts and its Protective Role against Dementia.
Level 3 - non-randomized controlled study
Observational cohort analysis combined with Mendelian randomization and mediation modeling.
PubMed 41582318 · doi:10.1249/MSS.0000000000003948
What was done
Investigators evaluated associations between physical activity (PA) and sedentary behavior (SB)—measured via three self-reported and three device-based metrics—and circulating levels of 2,911 plasma proteins among UK Biobank participants (maximum n = 39,160). Functional enrichment analysis was performed on identified proteins. Bidirectional Mendelian randomization was conducted to test for causal relationships between PA/SB and protein levels. Mediation analyses were then performed to determine which circulating proteins mediate the relationship between physical activity and incident all-cause dementia.
What was found
Observational analyses identified 41 proteins consistently associated with all PA metrics and 1,027 proteins associated with at least one metric. Both observational and Mendelian randomization analyses converged on proteins increased by PA (including ITGAV, ITGAM, MXRA8, CLEC4A, CLEC4M, LPL, and ADGRG2), proteins decreased by PA (including LEP, INHBC, CLMP, PTGDS, ADM, OGN, and PI3), and proteins linked to high-intensity PA (CA14, CA6, CA4, KIT, and ANGPT2). Enriched biological pathways included cell-matrix adhesion, integrin-mediated signaling, and collagen binding. Mediation analyses identified 21 unique proteins (including GDF15, ITGAV, ITGAM, ITGA11, HPGDS, GFAP, ADM, AHNAK, and DPP4) mediating the link between PA and all-cause dementia. Numerical effect sizes and confidence intervals were not reported in the abstract.
Why it matters
This study identifies specific molecular pathways and circulating protein mediators—notably integrin signaling, synaptic plasticity, neurogenesis, and inflammatory markers—through which physical activity may reduce the risk of all-cause dementia.
Limits
Exact numerical effect estimates, hazard ratios, and confidence intervals are not reported in the abstract. The UK Biobank cohort consists predominantly of middle-aged to older adults of European ancestry, which may limit generalizability. Mendelian randomization relies on key instrumental variable assumptions (such as the absence of horizontal pleiotropy), and mediation models remain subject to unmeasured residual confounding.
Cited by
- supports Higher circulating levels of GDF15 are epidemiologically associated with reduced engagement in voluntary physical activity, such as walking for pleasure and climbing stairs.