Pre-analytical Characterization of CNS-Derived Extracellular Vesicles from Human Saliva: Effect of Room Temperature and Cellular Origin

Abstract

Blood-derived extracellular vesicles (EVs) from neurons and astrocytes carrying Alzheimer's disease (AD) biomarkers can predict AD progression from mild cognitive impairment (MCI); yet their potential in saliva remains largely unexplored. Saliva-derived extracellular vesicles (sEVs) are promising non-invasive biomarkers for AD and other age-related dementias (ADRD), but their use has been hindered by unstandardized protocols for saliva collection, storage, and isolation of EVs from central nervous system (CNS) cells. Our research objective was two-fold. Firstly, we aimed to optimize the enrichment of sEVs from specific CNS cells derived from the same patient. Secondly, we aimed to assess how cellular origin and storage temperature (room temperature, 4°C, -20°C) PAGE \* Arabic \* MERGEFORMAT 4 This is a provisional file, not the final typeset article impact the stability and quantification of AD-related biomarkers and inflammatory cytokines. Using saliva collected by the passive drool method from participants in the Nathan Shock Healthy Aging Study (average age 71.3 years, n = 15); we isolated EVs from neurons, astrocytes, microglia, and oligodendrocyte origins using ExoQuick-TC and magnetic bead immunocapture. Executive function and attention were measured in all study participants using the Cognition Battery of the NIH Toolbox. Quantification via high-sensitivity immunoassays (MSD, SIMOA) revealed that key AD biomarkers (Aβ40, Aβ42, and total tau) were significantly enriched in astrocyte-derived EVs. p-tau217 was largely undetected in all fractions (sEVs and EV-depleted saliva). In contrast, TDP-43 was most abundant in EV-depleted saliva, while inflammatory cytokines were detected broadly across all fractions. Although storage temperature did not consistently impact biomarker levels, storage at -20°C was found to be optimal for biomarker quantification. Critically, lower levels of inflammatory cytokines (IFN-γ, IL-10, and IL-6) detected in EV-depleted saliva were associated with enhanced working memory, suggesting saliva-derived analytes can reflect a neuro-inflammatory state linked to cognitive performance. This study provides a critical proof-of-concept validation, characterizing and comparing multiple CNS-derived sEV fractions within the same individuals for the first time. By establishing a robust, standardized methodology for salivary EV analysis, this work paves the way for future longitudinal studies utilizing saliva to diagnosis and predict the progression of Alzheimer's disease.