85. SHARED AND SUBTYPE-SPECIFIC NEUROPSYCHIATRIC AND PROTEOMIC ARCHITECTURE OF PRIMARY HEADACHE DISORDERS IN GREATER THAN 1.6 MILLION INDIVIDUALS

Background Headache disorders, including migraine, tension-type headache (TTH), and cluster headache, affect billions worldwide and are among the leading causes of disability. These conditions are highly comorbid with psychiatric, substance use, and chronic pain disorders, yet the shared and subtype-specific biological mechanisms underlying this overlap remain poorly understood. Most prior genetic studies have focused on European populations and have not systematically integrated multi-omic approaches to define molecular mechanisms across headache subtypes. Methods Leveraging data from the Million Veteran Program, UK Biobank, FinnGen, All of Us, Taiwan Biobank, and the International Cluster Headache Consortium, we performed multi-ancestry genome-wide association meta-analyses of six headache phenotypes: migraine (168,113 cases, 1,386,830 controls), migraine with aura (MA; 33,763 cases, 977,961 controls), migraine without aura (MO; 29,808 cases, 811,639 controls), TTH (191,667 cases, 603,196 controls), cluster headache (10,634 cases, 966,071 controls), and broad headache (310,859 cases, 874,419 controls). Genetic correlations with psychiatric, substance use, and complex medical traits were estimated using LD score regression, while phenome-wide association studies (PheWAS) in the Penn Medicine Biobank evaluated clinical pleiotropy. Multi-omic integration included brain proteome-wide association studies (PWAS), colocalization analyses using dorsolateral prefrontal cortex pQTL reference panels, functional enrichment analyses, and cell-type enrichment analyses across brain and trigeminal ganglion datasets. Results We identified 120 migraine (35 novel), 6 MA (3 novel), 6 MO (1 novel), 3 TTH (all novel), 12 cluster headache (2 novel), and 60 broad headache (33 novel) loci, with 35 loci shared across phenotypes. Shared genetic architecture included strong correlations with chronic pain, PTSD, major depressive disorder, anxiety, ADHD, insomnia, and neuroticism, alongside modest overlap with substance use traits. PheWAS analyses demonstrated pleiotropy across psychiatric, cardiovascular, respiratory, metabolic, gastrointestinal, and somatic pain phenotypes. Subtype-specific patterns implicated migraine in neuropsychiatric and vascular pathways, MA in cerebrovascular and ischemic mechanisms, TTH in stress-related and somatic pain amplification traits, and cluster headache in sensory neuronal and autonomic pathways. PWAS and colocalization analyses prioritized candidate brain protein mediators, including UFL1, PGD, NAE1, and NECAB1, and implicated oxidoreductase and mitochondrial metabolic pathways. Cell-type enrichment analyses highlighted sensory neuronal populations in cluster headache, glial and immune contributions in migraine and TTH, and hypothalamic-midbrain-thalamic circuitry in broad headache. Discussion These findings define a shared pain–psychiatric genetic core across primary headache disorders while identifying distinct subtype-specific neurovascular, sensory-autonomic, inflammatory, and metabolic mechanisms, highlighting biologically informed targets for precision therapeutics.