Inflammatory protein epigenetic scores (EpiScores) and cognitive function in the longitudinal Swedish adoption/twin...

Protein epigenetic scores (EpiScores) are DNA methylation (DNAm)-based proxies for circulating protein levels and may provide insights into inflammation-cognition relationships. Although some EpiScores have been linked to cognitive decline, it remains unclear whether these proxies show similar associations in other cohorts, and whether their effects vary across distinct cognitive domains. We aimed to evaluate the associations between inflammatory proteins EpiScores and cognitive functions including specific domains in the Swedish Adoption/Twin Study of Aging (SATSA) cohort. We analyzed 1332 repeated measurements from 520 SATSA participants across up to six testing waves (1992-2014). Inflammatory protein EpiScores (N = 27) were generated using previously published CpG site-specific weights. Outcomes were a global cognitive function (GCF) score and four specific cognitive domains (verbal ability, spatial ability, memory, and perceptual speed). We applied marginal linear models within the generalized estimating equations (GEE) framework to estimate population-averaged longitudinal associations, controlling the false discovery rate (FDR) using the Benjamini-Hochberg (BH) procedure (q < 0.05 considered significant). Two EpiScores-C-X-C motif chemokine ligand 9 (CXCL9) and vascular cell adhesion molecule 1 (VCAM1)-showed significant positive associations with GCF. In domain-specific outcomes, VCAM1 demonstrated a positive association with spatial ability, whereas C-X-C motif chemokine ligand 11 (CXCL11) had a negative association with verbal ability. These associations were broadly consistent across multiple sensitivity analyses. Inflammation-related DNAm signatures of certain proteins were associated with cognitive performance in SATSA samples. Moreover, the associations differ by cognitive domain. Future longitudinal studies are warranted to clarify the causal implications of these EpiScore-cognition associations and to explore population differences in DNAm patterns.