Peripheral inflammatory biomarkers are associated with cognitive function and dementia: Framingham Heart Study Offspring cohort

Abstract

Inflammatory protein biomarkers induced by immune responses have been associated with cognitive decline and the pathogenesis of Alzheimer's disease (AD). Here, we investigate associations between a panel of inflammatory biomarkers and cognitive function and incident dementia outcomes in the well-characterized Framingham Heart Study Offspring cohort. Participants aged ≥40 years and dementia-free at Exam 7 who had a stored plasma sample were selected for profiling using the OLINK proteomics inflammation panel. Cross-sectional associations of the biomarkers with cognitive domain scores (N = 708, 53% female, 22% apolipoprotein E (APOE) ε4 carriers, 15% APOE ε2 carriers, mean age 61) and incident all-cause and AD dementia during up to 20 years of follow-up were tested. APOE genotype-stratified analyses were performed to explore effect modification. Higher levels of 12 and 3 proteins were associated with worse executive function and language domain factor scores, respectively. Several proteins were associated with more than one cognitive domain, including IL10, LIF-R, TWEAK, CCL19, IL-17C, MCP-4, and TGF-alpha. Stratified analyses suggested differential effects between APOE ε2 and ε4 carriers: most ε4 carrier associations were with executive function and memory domains, whereas most ε2 associations were with the visuospatial domain. Higher levels of TNFB and CDCP1 were associated with higher risks of incident all-cause and AD dementia. Our study found that TWEAK concentration was associated both with cognitive function and risks for AD dementia. The association of these inflammatory biomarkers with cognitive function and incident dementia may contribute to the discovery of therapeutic interventions for the prevention and treatment of cognitive decline.

Keywords: Alzheimer's disease; cognitive aging; peripheral inflammation; protein biomarkers.

FIGURE 1

Forest plots of combined and stratified protein effect size for significant associations within four cognitive domains using Model 1 (FDR ≤ 0.1). (a) Proteins associated with executive function domain score; (b) Proteins associated with language domain score; (c) Proteins associated with memory domain score; (d) Proteins associated with visuospatial domain score. The covariates adjusted in Model 1 included: sex, age, education level, time in years between Exam 7 (blood sample) and cognitive testing date, a retest indicator, and APOE genotype. The pink square represented the combined sample (N = 708), the blue circle represented the ε2 Carriers (N = 87), the red triangle represented the ε4 Carriers (N = 133), and the green diamond represented the ε3ε3 subgroup (N = 468). APOE, apolipoprotein E; CI, confidence interval; FDR, false discovery rate.

FIGURE 2

Forest plots of combined and stratified protein effect size for significant associations within four cognitive domains using Model 2 (FDR ≤ 0.1). (a) Proteins associated with executive function domain score; (b) Proteins associated with language domain score; (c) Proteins associated with memory domain score; (d) Proteins associated with visuospatial domain score. The covariates adjusted in Model 2 included: sex, age, education level, time in years between Exam 7 (blood sample) and cognitive testing date, a retest indicator, APOE genotype, indicators for prevalent cardiovascular diseases (CVD) such as prevalent stroke, prevalent CVD, and prevalent atrial fibrillation (AF) at Exam 7, and the CVD risk factors such as systolic and diastolic blood pressures (mmHg), diabetes status, treatment for hypertension, body‐mass index (kg/m²), current smoking status, total cholesterol level (mg/dL), high‐density lipoprotein cholesterol levels (HDL, measured in mg/dL), and use of lipid‐lowering agents at Exam 7. The pink square represented the combined sample (N = 708), the blue circle represented the ε2 Carriers (N = 87), the red triangle represented the ε4 Carriers (N = 133), and the green diamond represented the ε3ε3 subgroup (N = 468). APOE, apolipoprotein E, CI, confidence interval; FDR, false discovery rate.

FIGURE 3

Forest plots of combined and stratified protein effect size for significant associations with dementia outcomes using Model 1 (FDR ≤ 0.1). The covariates adjusted in Model 1 included: sex, age, education level, and APOE genotype. The pink square represented the combined sample (N = 837), the red triangle represented the ε4 Carriers (N = 164), and the green diamond represented the ε3ε3 subgroup (N = 551). AD, Alzheimer's disease; APOE, apolipoprotein E; CI, confidence interval; FDR, false discovery rate; HR, hazard ratio.

FIGURE 4

Forest plots of combined and stratified protein effect size for significant associations with dementia outcomes using Model 2 (FDR ≤ 0.1). The covariates adjusted in Model 2 included: sex, age, education level, APOE genotype, indicators for prevalent cardiovascular diseases (CVD) such as prevalent CVD, and prevalent atrial fibrillation (AF) at Exam 7, and the CVD risk factors such as systolic and diastolic blood pressures (mmHg), diabetes status, treatment for hypertension, body‐mass index (kg/m²), current smoking status, total cholesterol level (mg/dL), high‐density lipoprotein cholesterol levels (HDL, measured in mg/dL), and use of lipid‐lowering agents at Exam 7. The pink square represented the combined sample (N = 837), the red triangle represented the ε4 Carriers (N = 164), and the green diamond represented the ε3ε3 subgroup (N = 551). AD, Alzheimer's disease; APOE, apolipoprotein E; CI, confidence interval; FDR, false discovery rate; HR, hazard ratio.