Identification of crosstalk genes and immune characteristics between Alzheimer’s disease and atherosclerosis

Background Advancements in modern medicine have extended human lifespan, but they have also led to an increase in age-related diseases such as Alzheimer’s disease (AD) and atherosclerosis (AS). Growing research evidence indicates a close connection between these two conditions. Methods We downloaded four gene expression datasets related to AD and AS from the Gene Expression Omnibus (GEO) database (GSE33000, GSE100927, GSE44770, and GSE43292) and performed differential gene expression (DEGs) analysis using the R package “limma”. Through Weighted gene correlation network analysis (WGCNA), we selected the gene modules most relevant to the diseases and intersected them with the DEGs to identify crosstalk genes (CGs) between AD and AS. Subsequently, we conducted functional enrichment analysis of the CGs using DAVID. To screen for potential diagnostic genes, we applied the least absolute shrinkage and selection operator (LASSO) regression and constructed a logistic regression model for disease prediction. We established a protein-protein interaction (PPI) network using STRING ( https://cn.string-db.org/ ) and Cytoscape and analyzed immune cell infiltration using the CIBERSORT algorithm. Additionally, NetworkAnalyst ( http://www.networkanalyst.ca ) was utilized for gene regulation and interaction analysis, and consensus clustering was employed to determine disease subtypes. All statistical analyses and visualizations were performed using various R packages, with a significance level set at p<0.05. Results Through intersection analysis of disease-associated gene modules identified by DEGs and WGCNA, we identified a total of 31 CGs co-existing between AD and AS, with their biological functions primarily associated with immune pathways. LASSO analysis helped us identify three genes (C1QA, MT1M, and RAMP1) as optimal diagnostic CGs for AD and AS. Based on this, we constructed predictive models for both diseases, whose accuracy was validated by external databases. By establishing a PPI network and employing four topological algorithms, we identified four hub genes (C1QB, CSF1R, TYROBP, and FCER1G) within the CGs, closely related to immune cell infiltration. NetworkAnalyst further revealed the regulatory networks of these hub genes. Finally, defining C1 and C2 subtypes for AD and AS respectively based on the expression profiles of CGs, we found the C2 subtype exhibited immune overactivation. Conclusion This study utilized gene expression matrices and various algorithms to explore the potential links between AD and AS. The identification of CGs revealed interactions between these two diseases, with immune and inflammatory imbalances playing crucial roles in their onset and progression. We hope these findings will provide valuable insights for future research on AD and AS.