Genetic analyses prioritize a candidate peripheral ferroptosis-related immune–protein network in migraine
Abstract
Migraine is a disabling neurological disorder with substantial clinical heterogeneity. Ferroptosis-annotated genes encompass functions related to iron and redox homeostasis, lipid peroxidation, and immune regulation, but the genetically supported relationships connecting these processes with peripheral immune-cell phenotypes and circulating proteins have not been systematically examined. We integrated summary-data-based Mendelian randomization, two-sample Mendelian randomization, colocalization analysis, indirect-effect analysis, PBMC single-cell RNA sequencing, and druggability annotation. Five migraine genome-wide association study datasets were analyzed; GCST90043743 included 166 cases and was treated as a sensitivity dataset. A set of 718 ferroptosis-related genes was evaluated using blood mQTL and eQTL resources and plasma pQTL datasets, while immune-cell MR included 731 immunophenotypes. Evidence was first assessed across molecular data types and migraine outcomes and was then integrated with immune-cell traits, circulating proteins, single-cell expression context, and druggability information. Thirteen genes met the highest cross-omics evidence criteria: ACADSB, ADAMTS13, CDH1, CHMP1A, DPP4, GCH1, GSTM1, ISCU, MAPK9, MAPK14, MUC1, TFRC, and TLR4. A further 25 genes showed recurrent but incomplete cross-omics evidence. Of 2,087 colocalization tests, 313 met the colocalization support criterion, and all 13 highest-evidence genes had at least one result meeting this criterion. Immune-cell MR identified recurrent CD27-related B-cell and CD33-related myeloid traits. Thirty-four matched results from the two indirect-effect analyses were retained; 22 shared CD27-related B-cell traits between GSTM1- or TFRC-related gene-side results and ENPP2-related protein-side results. Additional matched results involved MUC1, myeloid dendritic-cell or naive CD8bright T-cell traits, and HMOX1 or IDO1; ALOX15, CD38-related B-cell or activated Treg traits, and IL1B; and CISD2, a CD27-related B-cell trait, and IFNG. PBMC single-cell RNA sequencing provided peripheral cell-type expression context, and targeted pseudobulk analysis showed increased IFNG expression in CD8+ T cells after Bonferroni correction. Druggability annotation highlighted ENPP2, IDO1, IL1B, and IFNG for translational follow-up. This study prioritizes a candidate peripheral ferroptosis-related immune–protein network associated with migraine susceptibility. The most recurrent relationship involved GSTM1 or TFRC, CD27-related B-cell traits, and ENPP2, with additional myeloid, B-cell/Treg, and inflammatory components. These findings nominate genetically informed candidates for biomarker research, immune-metabolic stratification, and experimental target investigation in migraine.