Integrated Mendelian Randomization and Single-Cell Transcriptomics Analysis Identifies Critical Blood Biomarkers and Potential Mechanisms in Epilepsy

Background: Epilepsy has a genetic predisposition, yet causal factors and the dynamics of the immune environment in epilepsy are not fully understood.

Methods: We analyzed peripheral blood samples from epilepsy patients, identifying key genes associated with epilepsy risk through Mendelian randomization, using eQTLGen and genome-wide association studies. The peripheral immune environment's composition in epilepsy was explored using CIBERSORT. An epilepsy mouse model was established to validated the expression of key genes at the transcriptomic and proteomic levels through single-cell analysis. Relevant pathways were verified. Finally, we developed a predictive model for antiepileptic drug response in epilepsy patients.

Results: We found that CDC25B, DNMT1, GZMA, MTX1, and SSH2 expression decreases epilepsy risk, whereas FGD3, RAF1, and SH3BP5L increase it. Epilepsy patients exhibited an altered peripheral immune profile, notably with increased activated mast cells and decreased CD4 memory activated T cells and γδ T cells. Eight genes were significantly related to this immune environment. In the animal model, FGD3, SSH2, and DNMT1 were upregulated at both mRNA and protein levels in the hippocampus. FGD3 and SSH2 are specifically elevated in microglia and are primarily associated with actin regulation. The trained predictive model was deployed on an online platform.

Conclusions: This study elucidates key genes linked to epilepsy, delineates the epilepsy immune landscape, and highlights the interaction between these domains, providing insights into potential epilepsy mechanisms and treatments.