Composite microRNA–genetic risk score model links to migraine and implicates its pathogenesis

Abstract The neurobiological mechanisms driving the ictal–interictal fluctuations and the chronification of migraine remain elusive. We aimed to construct a composite genetic–microRNA (miRNA) model that could reflect the dynamic perturbations of the disease course and inform the pathogenesis of migraine. We prospectively recruited four groups of participants, including interictal episodic migraine (i.e. headache free for >72 h, apart from prior and subsequent attacks), ictal episodic migraine (i.e. during moderate to severe migraine attacks), chronic migraine and controls in the discovery cohort. Next-generation sequencing was used for miRNA profiling. The candidate miRNAs were validated with quantitative PCR in an independent validation cohort. Biological pathways associated with the microRNA regulome and interaction networks were explored. In addition, all participants received genotyping with the Axiom Genome-Wide Array TWB chip. A composite model was established, combining disease-associated miRNAs and genetic risk scores indicative of genetic susceptibility, with the objective of differentiating migraine from controls using a binary outcome. From a total of 120 participants in the discovery cohort and 197 participants in the validation cohort, we identified disease-state miRNA signatures (including miR-183, miR-25 and miR-320) that were ubiquitously higher or lower in patients with migraine compared with controls. We have also validated four disease-activity miRNA signatures (miR-1307-5p, miR-6810-5p, let-7e and miR-140-3p) that were differentially expressed only during the ictal stage of episodic migraine. Functional analysis suggested that prolactin and oestrogen signalling pathways might play important roles in the pathogenesis. Moreover, the composite miRNA–genetic risk score model differentiated patients from controls, achieving a positive predictive value of >90%. To conclude, we developed a composite miRNA–genetic risk score model, which might serve as a predictive tool for identifying high-risk individuals. Our findings might help to illuminate potential pathogenic mechanisms underlying the dysfunctional allostasis of migraine and pave the way for future precision medicine.