The Converging Effects of Different Categories of Antidepressants on the Brain: A Systematic Meta-Analysis of Public Transcriptional Profiling Data From the Hippocampus and Cortex

Abstract

Depression can be treated with traditional pharmaceuticals targeting monoaminergic function, nontraditional drug classes and neuromodulatory interventions. To identify mechanisms of action shared across clinically-effective antidepressant treatment categories, we performed two systematic meta-analyses of public transcriptional profiling data from adult laboratory rodents (rats, mice). The outcome variable was gene expression, measured by microarray or RNA-Seq from bulk-dissected tissue from two depression-related brain regions (hippocampus, cortex). Relevant datasets were identified in the Gemma database of curated, reprocessed transcriptional profiling data using predefined search terms and inclusion/exclusion criteria (hippocampus: June 24, 2024, cortex: July 10, 2024). Differential expression results were extracted for all genes, minimizing bias. For each gene, a random effects meta-analysis model was fit to antidepressant vs. control effect sizes (Log2 Fold Changes) from each study for each brain region, with follow-up analyses exploring sources of effect heterogeneity. For the hippocampus, 15 relevant studies were identified, containing 22 antidepressant vs. control group comparisons (collective n = 313 samples), with approximately half representing traditional versus nontraditional antidepressants. Of 16 439 analyzed genes, 58 were consistently differentially expressed (False Discovery Rate (FDR) < 0.05) following treatment. Antidepressant effects were enriched in the dentate gyrus and in gene sets related to stress regulation, brain growth and plasticity, vasculature and glia, and immune function. Comparisons with single nucleus RNA-Seq confirmed effects on specific hippocampal cell types, including potential rejuvenation of dentate granule neurons. For the cortex, 13 studies were identified, containing 16 antidepressant vs. control group comparisons (collective n = 233 samples). Of 15 583 analyzed genes, only one was consistently differentially expressed (FDR < 0.05: Atp6v1b2), but overall expression patterns moderately resembled the hippocampus. These genes and pathways showing consistent differential expression across treatment categories may be promising targets for novel therapies. Future work should explore relevance to human clinical populations and potential heterogeneity introduced by sex and subregion.