Alzheimer's disease (AD) is the most prevalent form of dementia among elderly people worldwide. Cerebrospinal fluid (CSF) is the optimal fluid source for AD biomarkers, while serum biomarkers are much more achievable. To search for novel diagnostic AD biomarkers, we performed a quantitative proteomic analysis of CSF and serum samples from AD and normal cognitive controls (NC). CSF and serum proteomes were analyzed via data-independent acquisition quantitative mass spectrometry. Our bioinformatic analysis was based on Gene Ontology (GO) functional annotation analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment. In comparison to the controls, 8 proteins were more abundant in AD CSF, and 60 were less abundant in AD CSF, whereas 55 proteins were more and 10 were less abundant in the serum samples. ATPase-associated activity for CSF and mitochondrial functions for CSF and serum were the most enriched GO terms of the DEPs. KEGG enrichment analysis showed that the most significant pathways for the differentially expressed proteins were the N-glycan biosynthesis pathways. The area under the curve (AUC) values for CSF sodium-/potassium-transporting ATPase subunit beta-1 (AT1B1), serglycin (SRGN), and thioredoxin-dependent peroxide reductase, mitochondrial (PRDX3) were 0.867 (p = 0.004), 0.833 (p = 0.008), and 0.783 (p = 0.025), respectively. A panel of the above three CSF proteins accurately differentiated AD (AUC = 0.933, p = 0.001) from NC. The AUC values for serum probable phospholipid-transporting ATPase IM (AT8B4) and SRGN were moderate. The AUC of the CSF SRGN + serum SRGN was 0.842 (p = 0.007). These novel AD biomarker candidates are mainly associated with inflammation, ATPase activity, oxidative stress, and mitochondrial dysfunction. Further studies are needed to investigate the molecular mechanisms by which these potential biomarkers are involved in AD.
Keywords: Alzheimer’s disease; bioinformatics analysis; differentially expressed proteins; proteomes.