Saliva proteome, metabolome and microbiome signatures for detection of Alzheimer's disease

Background: As the burden of Alzheimer’s disease (AD) escalates with an ageing population, the demand for early and accessible diagnostic methods becomes increasingly urgent. Saliva, with its non-invasive and cost-effective nature, presents a promising alternative to cerebrospinal fluid and plasma for biomarker discovery.

Methods: In this study, we conducted a comprehensive multi-omics analysis of saliva samples (n = 20 mild cognitive impairment (MCI), n = 20 Alzheimer’s disease and age- and n = 40 gender-matched cognitively normal individuals), from the South Australian Neurodegenerative Disease (SAND) cohort, integrating proteomics, metabolomics, and microbiome data with plasma measurements, including pTau181.

Results: Among the most promising findings, the protein Stratifin emerged as a top candidate, showing a strong negative correlation with plasma pTau181 (r = −0.49, p < 0.001) and achieving an AUC of 0.95 in distinguishing AD and MCI combined from controls. In the metabolomics analysis, 3-chlorotyrosine and L-tyrosine exhibited high correlations with disease severity progression, with AUCs of 0.93 and 0.96, respectively. Pathway analysis revealed significant alterations in vitamin B12 metabolism, with Transcobalamin-1 levels decreasing in saliva as AD progressed despite an increase in serum vitamin B12 levels (p = 0.008). Microbiome analysis identified shifts in bacterial composition, with a microbiome cluster containing species such as Lautropia mirabilis showing a significant decrease in abundance in MCI and AD samples. The overall findings were reinforced by weighted correlation network analysis, which identified key hubs and enriched pathways associated with AD.

Conclusions: Collectively, these data highlight the potential of saliva as a powerful medium for early AD diagnosis, offering a practical solution for large-scale screening and monitoring.