Alzheimer’s disease (AD) is the leading cause of dementia. Monoclonal antibodies targeting Aβ (amyloid-β) have produced disease-modifying effects in some cases and led to regulatory approvals. The authors examined whether the repertoire of naturally occurring human antibodies (NAbs) differs between AD and cognitively normal (NC) individuals, reasoning that NAbs constitute a first line of immune defense against diverse disease-related antigens.
The study used epitomic profiling (EP) to characterize the specificities of circulating human antibodies at the level of peptide epitope segments (ES). EP was applied to compare NAb ES repertoires between AD and NC populations. Further analytical steps included correlation and network methods to assess co-expression and relationships among differentially expressed ES (DEES).
The principal result reported is a striking reduction in the diversity of antibody-specific peptide epitope segments in AD compared to NC: a 26-fold decline in ES diversity in the AD population. This decline indicates that the breadth of disease-associated NAb specificities is substantially lower in individuals with AD according to the EP measurements described.
Many of the DEES identified through EP map to protein sequences known to accumulate in AD pathology. The listed proteins include Aβ (Ab), tau, alpha-synuclein (a‑synuclein), TDP‑43, and TMEM106b. These mappings suggest that the reduced ES diversity in AD disproportionately affects antibody specificities recognizing sequences in proteins implicated in neurodegenerative aggregation.
Analysis of Aβ-directed ES revealed group-specific patterns: antibody reactivities targeting the amino-terminal regions of Aβ (residues 1–6 and 18–21) were elevated in the AD cohort. By contrast, ES mapping to Aβ residues 5–10 and 35–41 were elevated in the NC group. These differences indicate a redistribution of Aβ-targeting natural antibody specificities between the two populations.
DEES that map to the 2N4R tau isoform were reported to be highly cross-correlated across the population. The authors interpret this pattern as consistent with either polyspecific antibody binding (single antibodies recognizing multiple related sequences) or coordinated regulation of a set of antibodies that bind related tau sequences.
To identify co-expressed groups of DEES, the authors used weighted gene correlation network analysis (WGCNA). WGCNA produced modules of co-expressed DEES; on average approximately 25 DEES clustered within each WGCNA group. A major WGCNA cluster associated with AD contained three groups of overlapping DEES that together assemble into longer linear epitopes. One of those groups included sequences mapping to the Aβ amino terminus.
The authors applied UMAP dimensionality reduction to embed DEES in both AD and NC populations. The UMAP embeddings were used to visualize DEES that bind to the same antibody or to co-expressed sets of antibodies, supporting the network-based groupings identified by WGCNA.
The authors propose that DEES elevated in AD could serve as candidate blood biomarkers for Alzheimer’s disease. Conversely, DEES clusters elevated in NC may be associated with protection against misfolded-protein amyloid seeding and propagation and with maintaining neuronal resilience. These interpretations are presented as possible biological consequences of the observed repertoire differences.
This work is reported as a preprint and has not undergone peer review. Funding declared for the study includes Cure Alzheimer’s Fund. Specific methodological details, cohort sizes, statistical measures, and raw data descriptions beyond the summary-level findings were not reported in the source text provided here.