BACKGROUND: Schizophrenia (SCZ) is increasingly recognized as a complex, systemic disorder involving interactions between the central nervous system and peripheral biological processes. Growing evidence implicates dysregulation of the gut-brain axis in SCZ pathophysiology; however, the genetic mechanisms linking schizophrenia susceptibility to gut-related biological pathways remain poorly understood, particularly with respect to causal gene prioritization across tissues.
METHODS: To address this gap, we conducted a cross-tissue transcriptome-wide association study (TWAS) using FUSION across five tissues relevant to brain-gut interactions, including the hippocampus, frontal cortex (BA9), transverse colon, sigmoid colon, and whole blood. Candidate genes were further prioritized using MAGMA gene-based analysis based on schizophrenia genome-wide association study (GWAS) summary statistics from the Psychiatric Genomics Consortium and the IEU Open GWAS project. Integrative analyses combining summary-data-based Mendelian randomization (SMR) and Bayesian colocalization (PP.H4 > 0.75) were applied to refine candidate gene selection. Expression of the top candidate gene was examined by RT-qPCR in peripheral blood samples from 19 patients with SCZ and 21 healthy controls. Finally, two-sample Mendelian randomization (MR) analyses were performed to explore potential associations between the prioritized gene and 205 gut microbial metabolic pathways.
RESULTS: Cross-tissue integration of TWAS, MAGMA, and SMR identified five convergent schizophrenia susceptibility genes: TVP23B, NSUN2, RPL12, FOXN2, and THAP5. Bayesian colocalization analysis highlighted FOXN2 as the most robust candidate (PP.H4 =0.995). RT-qPCR analysis demonstrated significantly lower FOXN2 expression in peripheral blood from patients with schizophrenia compared with healthy controls (mean ± SD: 0.72±0.64 vs.1.13±0.42; t = 2.34, P = 0.02). In two-sample MR analyses, genetically proxied FOXN2 expression was modestly but significantly associated with 10 gut microbial metabolic pathways after false discovery rate correction, suggesting potential links between schizophrenia risk genes and gut microbial metabolic processes.
LIMITATIONS: This study relies on cis-eQTL-based transcriptomic integration and assumes a single causal variant in colocalization analyses. The number of available MR instruments was limited, and the clinical validation sample size was modest, with residual confounding not fully excluded. In addition, the GWAS and eQTL datasets were predominantly derived from populations of European ancestry, which may limit generalizability. The observed associations between FOXN2 and gut microbial metabolic pathways should therefore be considered exploratory and require further functional validation.
CONCLUSION: Our findings suggest that FOXN2 may contribute to the genetic architecture of schizophrenia and may be linked to gut-brain axis-related metabolic pathways. These results provide a basis for future experimental studies to elucidate the biological mechanisms underlying this association.
