Work

46. a Transcriptomic Gradient of Neuronal and Immune Gene Programs in the Subgenual Anterior Cingulate Cortex in Schizophrenia

Year: 2025

Type: article

Abstract: Background Classic approaches to identifying transcriptomic changes in post-mortem brain tissue from individuals with psychiatric disorders often show limited concordance with risk genes identified by genome- and exome-wide association studies (GWAS). These analyses are challenged by limited sample size, genetic complexity, polygenicity, regional and cell-type variability, and environmental confounds like medicinal and recreational drug use. Here, we characterize transcriptomic differences in the relatively under-studied subgenual anterior cingulate cortex (sgACC) in bipolar disorder (BD), major depression (MDD), and schizophrenia, using a novel multivariate regression technique that accounts for complex interactions among features, including environmental confounds. Methods Deep RNA sequencing was performed on bulk post-mortem sgACC from individuals with BD (N=35), MDD (N=51), schizophrenia (N=44), and controls (N=55). We performed dimensionality reduction on toxicology data from 17 substances using multiple correspondence analysis, then analyzed case-control gene expression using 1) traditional differential gene expression analysis (DGE) and 2) group regularized canonical correlation analysis (GRCCA) – a multivariate regression method that accounts for feature interdependence by leveraging gene co-expression modules. We compared results with established neuropsychiatric risk genes, gene ontology pathways, and cell type-specific expression profiles using gene set enrichment analyses. Results GRCCA revealed a significant association with schizophrenia (P_perm=0.001), but not with BD or MIDD. The resulting gene weight vector correlated with schizophrenia-control log2(fold change) from DGE (R=0.43; P < 0.05). Both methods identified downregulation of immune and microglial genes and upregulation of ion transport and excitatory neuron genes; however, GRCCA showed stronger biological enrichments. Notably, GRCCA results were enriched for schizophrenia GWAS-implicated genes, whereas DGE results were not. Sensitivity analyses indicated that multivariate RCCA (i.e., without the grouping vector) is sufficient to identify the schizophrenia association, but incorporating biologically informative gene module assignments refines the genetic association. GRCCA run at the transcript level identified isoform-specific expression patterns of schizophrenia risk genes, highlighting a key role for alternative splicing. Discussion Together, our findings identify a schizophrenia-specific sgACC gene expression pattern enriched for known schizophrenia risk genes and implicating neuronal and immune pathways. These results demonstrate the utility of multivariate approaches in integrating transcriptomic and genetic signals in psychiatric disorders.

Source: European Neuropsychopharmacology

DOI: 10.1016/j.euroneuro.2025.08.151