Joint analysis of GWAS and blood and brain QTL summary statistics reveals potential therapeutic targets for neuropsychiatric disorders

Abstract

Neurological and psychiatric disorders (NPDs) impose a substantial global burden. Many genetic associations have been reported for a range of NPDs, but the specific genes and proteins underlying susceptibility and their tissue- and cell type-specific manifestations remain poorly understood, hindering the development of targeted therapies. We integrated large-scale genome-wide association study (GWAS) summary statistics for 11 common NPDs with tissue- and cell type-specific multi-omics data from blood and brain using a Bayesian-based multi-omics Mendelian randomisation (MR) framework. We further performed pathway enrichment analyses to delineate functional pathways underlying these candidate signatures and used Connectivity Map (CMap) and ExPheWAS to prioritise therapeutic compounds mapping to NPD‑specific gene–protein signatures while systematically screening for potential off‑target side effects. We identified 26 and 60 significant gene–protein associations (posterior probability ≥ 0.7) across nine NPDs in blood and brain, respectively. Among these, 11 associations were resolved to specific brain-based cell types. Notably, we identified three previously unreported brain-specific gene–protein signatures, which to the best of our knowledge, with no prior evidence from GWAS, gene-based association studies or MR findings for the respective disorders: TRMT61B (associated with major depressive disorder [MDD] and Parkinson’s disease [PD] in bulk brain), EEFSEC (associated with schizophrenia [SCZ] in bulk brain and excitatory neurons), and SLC25A24 (associated with SCZ in bulk brain and at single-cell resolution across astrocytes, excitatory neurons, oligodendrocytes and oligodendrocyte progenitor cells). The association with TRMT61B implicates mitoribosome remodelling and mitochondrial translation in susceptibility to MDD and PD, whereas findings for EEFSEC and SLC25A24 suggest roles for serotonergic signalling dysregulation and disrupted mitochondrial homeostasis/calcium-sensitive signalling, respectively, in the pathophysiology of SCZ. We additionally delineated novel cell type-specific signals for established NPD-related signatures, including TRAF3 (associated with Alzheimer’s disease and multiple sclerosis in excitatory neurons) and SCFD1 (associated with amyotrophic lateral sclerosis in excitatory and inhibitory neurons). Our analyses also identified bortezomib as a potential candidate for drug re-purposing in SCZ. By resolving gene–protein associations at tissue and cell‑type resolution, our study provides new insights into the biological basis of risk for multiple NPDs. These insights provide a framework for advancing mechanistic understanding and therapeutic development, including opportunities for drug re-purposing for NPDs.