Plasma YKL40 as a biomarker for brain aging and injury in three community cohorts: Biomarkers (non‐neuroimaging): Alzheimer's disease incidence, risk factors and biomarkers

Background YKL40 is expressed in reactive astrocytes and microglia and is upregulated in numerous neurological conditions associated with heightened neuroinflammation. We explored the utility of circulating YKL40 as a biomarker of accelerated brain aging and subclinical brain injury. Method YKL40 was measured using the Mesoscale Discovery (MSD) multi‐array assay system from stored plasma in three community‐based cohorts: the Atherosclerosis Risk in the Communities Study (ARIC), the Coronary Artery Risk Development in Young Adults (CARDIA), and the Framingham Heart Study (FHS). MRI‐derived outcomes of total brain volume, hippocampal volume, and white matter hyperintensity volume (WMHV) were expressed as a percentage of total intracranial volume. We derived a general cognitive factor from multiple neuropsychological tests capturing multiple domains using cohort‐specific principal component analysis with a forced single factor solution. We examined the association between log‐transformed and standardized YKL40 levels and MRI and cognitive phenotypes separately in each cohort using linear regression. Cohort specific results were then pooled in random‐effects meta‐analysis. Analyses were cross‐sectional and adjusted for age, age squared, sex, education (for cognition), and the time interval between the blood draw for YKL40 and outcome assessment. A second model included additional adjustments for vascular risk factors and presence of the APOE ε4 allele. Result Across cohorts, there were 2782 dementia‐free participants (mean age, 64 years; 45% male; Mean(SD) YKL40 levels = 50689(7114)pg/mL; YKL40 Coefficient of Variation = 7.5%). In random‐effects meta‐analysis, higher YKL40 levels were associated with smaller total brain volume (β±SE, ‐0.24±0.05 per percentage change in brain volume, p<0.0001) and poorer cognitive function (β±SE, ‐0.04±0.02 per unit change in cognitive composite, p=0.03). The association between YKL40 levels and total brain volume remained significant following full multivariable adjustments (β±SE, ‐0.23±0.05 p<0.0001) whereas the association between YKL40 and cognition was attenuated (β±SE, ‐0.04±0.02, p=0.06). YKL40 levels were not associated with hippocampal volume or WMHV (P>0.05). Conclusion Higher plasma YKL40 levels were associated with MRI and cognitive markers of accelerated brain aging, suggesting involvement of astrocytic or microglial mediated neuroinflammation in subclinical brain atrophy and cognitive impairment.