Abstract

Hippocampal differential gene expression converges across animal models of mood disorder: Results from an interactive meta-analysis pipeline encompassing five animal models

Sannah Y, Khalil H, Hagenauer MH, Marrocco J, Aydin C, Wei Q, Meng F, Nasca C, Nestler E, Meaney M, Watson SJ, Akil H
50th Annual Meeting of the Society for Neuroscience, Virtual. 2021.

The Hope for Depression Research Foundation (HDRF) focuses on the neurobiology of mood disorders and brings together several research groups which have been conducting transcriptomic studies within different animal models used for the study of depression. Each animal model typically represents one dimension of a plethora of changes that might be occurring within the brains of depressed individuals, so integrating the results from across these studies is essential to gain a better understanding of mood disorders. To facilitate comparison, we created an interactive meta-analysis pipeline that allows the user to identify consistent gene expression signatures across custom-defined sets of animal models, conditions, transcriptional platforms, and brain regions. To illustrate utility, we tested this pipeline using thirteen microarray and RNA-seq data sets derived from either the hippocampus or dentate gyrus of five animal models: selectively-bred high responder (HR) and low responder (LR) rats, Flinders sensitive and resistant rats, mice with glucocorticoid receptor overexpression (early life and lifetime), chronic social defeat stress (CSDS) mice, and chronic corticosterone-treated mice (total n=146). To ensure comparability, the raw data from each study were first run through a standardized pipeline for annotation, quality control, and data preprocessing. We calculated the effect size (Hedge’s g) for each gene from each study from the moderated t-statistic obtained as the result of a differential expression analysis conducted within the limma or limma-voom pipeline. This effect size was then used within a random effects meta-analysis model. Interestingly, we found 27 candidate genes (FDR < 0.1) which showed similar differential expression in the hippocampus across these animal models of mood disorder despite the fact that a superficial comparison indicated minimal intersection between the top results from each individual study. These differentially expressed genes were highly skewed towards down-regulation (89%) and especially enriched within two particular hippocampal cell types: 63% enriched within astrocytes, 30% within ependymal cells. Gene Set Enrichment Analysis similarly indicated a particular enrichment within cell type related gene sets, including a down regulation within gene sets specific to the choroid plexus, ependyma, astrocytes, interneurons, polydendrocytes, and oligodendrocytes, and upregulation within gene sets specific to CA1 pyramidal neurons. These results illustrate the power of integrating results from numerous distinct models to understand the neurobiology of mood disorder.

In memory of Bruce McEwen.