The role of non-coding RNAs in the pathogenesis of migraine: from molecular insights to potential biomarkers.

Non-coding RNAs (ncRNAs) have emerged as critical post-transcriptional regulators in migraine pathophysiology, yet their clinical translational potential remains incompletely defined. This narrative review synthesizes current evidence on microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs) in migraine, organizing findings around a "mechanism-biomarker-intervention" framework. Key findings include: (1) multiple miRNAs (e.g., miR-155, miR-34a-5p, miR-342-3p) are differentially expressed in peripheral blood mononuclear cells (PBMCs) or serum of migraine patients, with expression patterns correlating with attack phase, chronicity, and treatment response to CGRP-targeted therapies; (2) the lncRNA NEAT1 regulates photophobic behavior via the miR-196a-5p/Trpm3 axis in preclinical models, though evidence derives exclusively from male mice-a significant limitation given the 3:1 female-to-male prevalence ratio of migraine; (3) recent studies have begun to provide direct evidence for circRNA involvement in migraine, including differential circRNA expression profiles in patient PBMCs and functional characterization of hsa_circ_0006168 in promoting central sensitization; and (4) lncRNAs PVT1 and MEG3 show differential expression between migraine with and without aura, suggesting subtype-specific biomarker potential. However, the field faces substantial challenges: most studies employ small, cross-sectional designs; biological matrices are inconsistently reported; evidence is predominantly preclinical; and no ncRNA-based diagnostic or therapeutic approach has entered clinical trials. We critically appraise these limitations and outline priorities for future research, including multicenter cohort validation, standardized detection protocols, and integration of multi-omics data. While ncRNAs represent promising candidates for migraine biomarkers and therapeutic targets, their clinical translation requires rigorous validation beyond current proof-of-concept studies.