CSF levels of proinflammatory and anti-inflammatory cytokines as predictors of cognitive decline across Alzheimer’s disease spectrum
Kellen Petersen1, Bhargav Nallapu2, Richard Lipton1, Ali Ezzati3
1Albert Einstein College of Medicine, 2Neurology, Albert Einstein College of Medicine, 3University of California, Irvine
Objective:
To identify distinctive CSF inflammatory biomarkers associated with clinically meaningful cognitive decline (CMCD) over a 1-year follow-up period.
Background:
Increasing evidence suggests a prominent role of neuroinflammation in AD pathogenesis. Identifying key inflammatory cytokines that complement AD-specific biomarkers like amyloid-β (Aβ) and tau in predicting cognitive decline may enhance our understanding of AD pathogenesis and identify individuals at a higher risk of disease progression.
Design/Methods:
Participants were 242 older adults from the Alzheimer’s Disease Neuroimaging Initiative (ADNI): 65 cognitively normal (CN), 111 with mild cognitive impairment (MCI), and 66 with dementia. All had CSF biomarkers of Aβ, tau, inflammatory cytokines including TNF-α, TNFR1, TNFR2, TGF-β1, TGF-β2, TGF-β3, IL-6, IL-7, IL-10, IL-21, and ICAM-1. The primary outcome of interest was CMCD, defined as an increase of ≥4 on the Alzheimer’s Disease Assessment Scale Cognitive Subscore 11 (ADAS-11, scores 0-70, higher scores indicate worse cognition). Predictor variables included demographics (age, sex, and education), genetic information (APOE4 status), and CSF biomarkers of Aβ, phosphorylated tau181 (p-tau181), total-tau, and inflammation. Random forest machine learning models were used in conjunction with out-of-bag estimation to calculate predictor variable importance scores. Predictive performance was compared between models with and without including inflammatory biomarkers in the feature set.
Results:
At 1-year follow up, 25.6% of participants showed CMCD. When comparing models, those with added inflammatory biomarkers showed improved CMCD prediction: whole sample (AUC=0.79 vs AUC=0.75, p=0.022), CN (AUC=0.92 vs AUC=0.90, p=0.021), and MCI (AUC=0.83 vs AUC=0.73, p<0.001).The most important inflammatory predictors of CMCD, based on variable importance scores, were TGF-β3 (0.522), TNFR2 (0.411), IL-7 (0.227), ICAM-1 (0.179), and TNF-α (0.147).
Conclusions:
CSF inflammatory biomarkers, particularly TGF-β3 and TNFR2, are predictors of cognitive decline over a 1-year follow-up period. These biomarkers demonstrated added value in predicting cognitive decline in addition to the AD-specific biomarkers of Aβ and p-tau.