Machine Learning-Based Integration Unveils RNA Methylation Regulator-Related Immune-Derived Gene Signatures in Ruptured Intracranial Aneurysm

Background/Objectives: Immune-inflammatory activation is a central feature of aneurysmal subarachnoid hemorrhage (aSAH), yet the epitranscriptomic mechanisms underlying this response remain insufficiently understood. This study aimed to investigate RNA methylation-associated immune dysregulation in aSAH and to identify potential biomarkers and signaling pathways. Methods: Four Gene Expression Omnibus datasets were analyzed to characterize RNA methylation regulator-related immune alterations in aSAH. Single-sample gene set enrichment analysis (ssGSEA), weighted gene co-expression network analysis (WGCNA), and intersection with ImmPort immune genes were used to identify candidate genes. A total of 159 machine learning combinations were evaluated for model construction and external validation. Two-sample Mendelian randomization, single-cell RNA sequencing (scRNA-seq), and CellChat analyses were further performed. Peripheral blood samples from patients with aSAH (n = 12) and matched healthy controls (n = 12) were used for total m6A quantification and quantitative real-time PCR (qRT-PCR) validation, while Western blotting and immunofluorescence were used to validate the protein expression of LIFR, GP130, IGF2BP2, and RBM15B. Results: Eleven RNA methylation regulators were differentially expressed between aSAH and controls in GSE122897. The WGCNA module most strongly associated with RNA methylation regulator-related scores was enriched in immune response and myeloid activation pathways. Intersection analysis identified 25 candidate immune-inflammatory genes associated with RNA methylation regulator-related transcriptional patterns. Among 159 algorithms, an XGBoost-LASSO pipeline selected oncostatin M (OSM) as the key variable, and the resulting RNA methylation regulator-related immune-derived gene signature (RMRIGS) showed good discrimination between aSAH and controls across training and validation cohorts. Mendelian randomization supported a protective association of genetically predicted OSM expression with subarachnoid hemorrhage risk (IVW OR = 0.66, p = 0.014). Single-cell analysis showed that Osm was predominantly enriched in infiltrating Ccr2+ macrophages, whereas Lifr and Il6st were broadly expressed in activated microglial subpopulations, indicating the presence of an Osm − (Lifr + Il6st) communication axis after SAH. Clinically, total m6A levels were increased in peripheral blood samples from patients with aSAH, and OSM, together with several RNA methylation regulators, was upregulated and associated with m6A-related changes. In experimental models, the protein expression levels of LIFR, GP130, IGF2BP2, and RBM15B were all increased after SAH-related stimulation. Conclusions: RNA methylation programs may be involved in immune dysregulation in aSAH. The OSM-centered RMRIGS was associated with disease status and may provide insight into the interaction between peripheral immune activation and post-SAH neuroinflammation. The potential involvement of the OSM–LIFR/GP130 signaling axis and its association with RNA methylation regulator-related alterations warrant further investigation.