CSF immune activation in Alzheimer’s disease driven by tau‐ and not amyloid pathology

Abstract Background Genetic studies suggest involvement of the immune system in Alzheimer’s disease (AD) development, and neuroinflammation has been suggested as a key driver of AD. We here employ cerebrospinal fluid (CSF) markers to study whether immune‐ and glial activation are primary features of early amyloid plaque deposition, or follow significant tau pathology as the disease progresses. Method We included n=267 non‐demented cases with repeated CSF samples (1‐6 years). The presence of amyloid pathology (A), tau‐tangle pathology (T) and neurodegeneration (N) using Aß 42/40 ratio, p‐tau181 and total‐tau were determined at baseline and follow‐ups. We then selected participants (n=226) that remained stable A‐/T‐/N‐ (n=134), A+/T‐/N‐ (n=33) and A+/T+/N+ (n=34) over time and cases that progressed from A‐/T‐/N‐ to A+/T‐/N‐ (n=14) and A+/T‐/N‐ to A+/T or N+ (n=11). Other A/T/N categories (e.g. A‐/T or N+) were excluded. Mixed Linear Models (MLMs) were used to measure longitudinal changes in CSF sTREM2, YKL‐40, clusterin, fractalkine, MCP‐1, IFN‐γ, IL‐6, IL‐18 and IL‐10 between groups with stable A‐/T‐/N‐ group as the reference. All models included age, sex and number of APOE ε4 alleles as covariates. Result Compared to stable A‐/T‐/N‐, cases progressing from A‐/T‐/N‐ to A+/T‐/N‐ had lower sTREM2 levels (p=.022) over time, and both fractalkine (p=.053), IFN‐γ (p=.055) and clusterin levels (p=.068) showed trend levels towards lower levels. In stable A+/T‐/N‐ cases, YKL‐40 (p=.007) and fractalkine (p=.043) were lower than stable A‐/T‐/N‐ over time. A+/T+/N+ cases had markedly higher sTREM2, fractalkine and clusterin levels (all p<.001) that remained high over time. In this group the YKL‐40 levels were also higher at baseline (p<.001) and increased slightly over time (p=.032). There were no between‐group differences at neither baseline or over time in the CSF levels of MCP‐1, IL‐6, IL‐18 and IL‐10. Conclusion Amyloid pathology alone was not associated with increased immune and glial activation in our cohort, and several of the CSF immune markers were lower in both those progressing to A+/T‐/N‐ and the stable A+/T‐/N‐. An increase in activation markers was only shown in cases with tau‐pathology at baseline, or in cases that developed tau pathology during the follow‐up period, suggesting immune activation in response to tau pathology.