Estimation of blood-based biomarkers of glial activation related to neuroinflammation

Abstract

BackgroundNeuroinflammation is a well-known feature of Alzheimer’s disease (AD), and a blood-based test for estimating the levels of neuroinflammation would be expected. In this study, we examined and validated a model using blood-based biomarkers to predict the level of glial activation due to neuroinflammation, as estimated by 11C-DPA-713 positron emission tomography (PET) imaging. MethodsWe included 15 patients with AD and 10 cognitively normal (CN) subjects. Stepwise backward deletion multiple regression analysis was used to determine the predictors of the TSPO-binding potential (BPND) estimated by PET imaging. The independent variables were age, sex, diagnosis, apolipoprotein E4 positivity, body mass index and the serum concentration of blood-based biomarkers, including monocyte chemotactic protein 1 (MCP-1), fractalkine, chitinase 3-like protein-1 (CHI3L1), soluble triggering receptor expressed on myeloid cells 2 (sTREM2), and clusterin. ResultsSex, diagnosis, and serum concentrations of MCP1 and sTREM2 were determined as predictors of TSPO-BPND in the Braak1-3 area. The serum concentrations of MCP1 and sTREM2 correlated positively with TSPO-BPND. In a leave one out (LOO) cross-validation (CV) analysis, the model gave a LOO CV R2 of 0.424, which indicated that this model can account for approximately 42.4% of the variance of brain TSPO-BPND. ConclusionsWe found that the model including serum MCP-1 and sTREM2 concentration and covariates of sex and diagnosis was the best for predicting brain TSPO-BPND. The detection of neuroinflammation in AD patients by blood-based biomarkers should be a sensitive and useful tool for making an early diagnosis and monitoring disease progression and treatment effectiveness.