Dissecting the effect of blood colony-stimulating factors and receptors on Alzheimer's disease: the role of myeloid traits and inflammation.

BACKGROUND: Research has suggested a potential link between colony-stimulating factors (CSFs) and Alzheimer's disease (AD), but the findings have been inconsistent, and the causal relationship remains uncertain.

METHODS: We performed a Mendelian Randomization (MR) analysis to explore the association between blood levels of CSFs and their receptors with AD and its biomarkers. The study utilized summary-level data on blood levels of granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor, macrophage colony-stimulating factor, and their receptors from deCODE. Data on AD were sourced from the International Genomics of Alzheimer's Project (IGAP), UK Biobank (UKB), and FinnGen, while cerebrospinal fluid p-Tau and β-amyloid levels were from the largest GWAS currently available. Additional summary-level data on amyloid PET imaging and AD progression from ADNI and the Knight-ADRC were also included.

RESULTS: Genetically predicted one standard deviation increase in blood G-CSF was associated with a higher risk of AD across three datasets (IGAP: OR = 1.35, 95 %CI = 1.15-1.57, P < 0.001; IGAP+UKB: OR = 1.27, 95 %CI = 1.12-1.43, P < 0.001; FinnGen: OR = 1.43, 95 %CI = 1.26-1.61, P < 0.001), but not AD progression. Elevated blood G-CSF levels were inversely related to β-amyloid levels (β = -0.14, 95 %CI = -0.19 to -0.09, P < 0.001) and positively associated with p-Tau levels (β = 0.08, 95 %CI = 0.03 to 0.14, P = 0.001) in cerebrospinal fluid. Furthermore, genetically predicted blood G-CSF levels were positively associated with amyloid PET imaging in the brain (β = 0.10, 95 % CI: 0.06 to 0.14, P < 0.001). However, no significant associations were found between blood levels of other CSFs, and their receptors and AD risk. In contrast, there was little evidence supporting the impact of AD on CSFs levels. The multivariable MR analysis showed that the association between G-CSF and AD, along with its biomarkers, disappeared after adjusting for C-reactive protein levels, but not for neutrophil count.

CONCLUSION: These findings indicate a harmful role of G-CSF in the development of AD, primarily driven by inflammatory responses rather than neutrophil counts. Therefore, interventions targeting AD through CSFs, especially G-CSF, should be cautiously approached.