Cerebrospinal Fluid Metabolomics, Brain Morphometry, and Psychiatric Disorders: A Mendelian Randomization Study

Objective To identify the causal roles of cerebrospinal fluid (CSF) metabolomics and brain morphometry in the development of psychiatric disorders, we conducted a bidirectional Mendelian Randomization (MR) study. Methods We performed a two-sample MR analysis to estimate the causal effects of 338 CSF metabolites and 83 brain-wide volumes on the risk of 6 psychiatric disorders: Attention Deficit Hyperactivity Disorder (ADHD), alcohol dependence, bipolar disorder (BD), cannabis use disorder (CUD), major depressive disorder (MDD), and schizophrenia (SCZ). Imputation and strict quality control were applied to CSF metabolite and genotype data, resulting in a final dataset of 291 baseline visits from 689 unrelated individuals of European ancestry (338 CSF metabolites quantified). GWAS summary data for brain morphometry were derived from the UK Biobank (36,778 unrelated white European participants, 54% female, with neuroimaging data). Summary statistics for psychiatric disorders were obtained from the Psychiatric Genomics Consortium (PGC): ADHD (38,691 cases/186,843 controls), alcohol dependence (10,206 cases/28,480 controls), BD (41,917 cases/371,549 controls), CUD (14,080 cases/343,726 controls), MDD (17,339 cases/53,426 controls), SCZ (68,125 cases/90,984 controls). We further explored whether brain morphometry mediates the pathway from CSF metabolomics to psychiatric disorders. Results After inverse variance weighting (IVW) and sensitivity analyses, 3 CSF metabolites showed causal effects on psychiatric disorders: 1. CSF cysteine levels increased BD risk (odds ratio [OR] = 1.23, 95% confidence interval [CI] = 1.11–1.37, P = 1.22×10⁻⁴), validated by an independent GWAS (OR = 2.46, 95% CI = 1.17–5.18, P = 0.02); 2. CSF betaine levels increased MDD risk (OR = 1.40, 95% CI = 1.18–1.66, P = 1.25×10⁻⁴); 3. CSF N-acetylglutamine levels increased MDD risk (OR = 1.42, 95% CI = 1.21–1.68, P = 4.47×10⁻⁴). Reverse MR showed no reverse causality (genetically predicted BD/MDD did not affect respective metabolite levels, P > 0.05). Brain morphometry had no causal effect on BD, and did not mediate the cysteine-BD association. Conclusions CSF metabolites (cysteine, betaine, N-acetylglutamine) have causal associations with BD or MDD, providing potential targets for diagnosis and treatment of these psychiatric disorders.