Multi-omics investigation of metabolic dysregulation in depression: integrating metabolomics, weighted gene co-expression network analysis, and mendelian randomization

Background The etiology of depressive disorder, the leading cause of global mental disability, is characterized by systemic metabolic dysregulation. However, the causal metabolites and their mechanistic networks remain elusive. Methods We combined untargeted LC/GC-MS metabolomics (N=98 Chinese elderly), weighted gene co-expression network analysis (WGCNA), and two-sample Mendelian randomization (MR) using GWAS data (59,333 depression cases with 434,831 controls) to identify depression-associated metabolites and pathways. Results LC/GC-MS analysis identified 1,458 metabolites, with 84 differentially expressed in depression (VIP>1.5, p<0.05). WGCNA revealed a turquoise module enriched in amino acid metabolism (MM>0.7, p<0.05), while MR analysis confirmed 35 causal metabolites, including cysteine-alanine ratio (β=0.18, p=0.003) and serine levels (β=−0.24, p=0.001). Multi-omics integration highlighted glycine/serine/threonine metabolism (Impact = 0.35) and one-carbon folate cycle as core dysregulated pathways. Alterations were characterized by serine deficiency and phosphoserine accumulation, potentially reflecting disturbances in DNA methylation processes. Furthermore, elevated cysteine levels indicated a compensatory response to oxidative stress, and disruptions in purine metabolism pointed to mitochondrial dysfunction, particularly impaired mitochondrial ATP synthesis. Conclusion This study establishes a hierarchical metabolic framework for depression, prioritizing single-carbon metabolism and oxidative stress as central therapeutic targets. The findings emphasize methylation dysregulation and mitochondrial dysfunction in elderly depression, offering novel biomarkers for precision intervention.