Stress in a dish: modeling the impact of common genetic variation on stress response in hipsc-derived neurons in ptsd

Genetic studies of PTSD are complicated by the complex interplay between environmental traumatic exposures and genetics. To understand how genomic risk loci functionally mediate disorder risk, it is important to explore the regulatory activity of such loci in disorder-associated cell types and contexts. Here, we utilize hiPSC-derived neuronal models to develop a controlled model of stress exposure. Our previous work has identified a PTSD diagnosis-dependent gene expression signature in hiPSC-derived neurons to hydrocortisone (HCort). We have also demonstrated that common genetic variation impacts stress response in the post-mortem brain. Here, we integrate both signatures to show how common genetic variants may mediate differential response to stress in a cell type- and tissue-specific manner. In 304 post-mortem brains, we modeled a linear eQTL association using an interaction term between trauma exposure and genotype across four brain regions. To validate our signatures in vitro, we generated GABAergic and glutamatergic neurons from combat-exposed veterans with PTSD (n=20) and without (n=20), and treated neurons with HCort. We performed interactive-eQTL associations on HCort-treated neurons. Identified eQTLs were functionally fine-mapped by cell type deconvolution, integration with CHIP-seq datasets, and gene set enrichment of disorder-related genes. We identified 915 trauma-interactive eGenes across 4 post mortem brain regions. Up to 61% of interactive-eGenes replicated in HCort-treated neurons. eGenes enrich in previously reported post-mortem PTSD signatures, as well as psychiatric GWAS catalog genes. eQTLs show varying strengths and directions of effect by cell type, showing cell type and brain tissue-specificity. Trauma-interactive eQTLs are more likely to lie in NF-kB (p=1.37E-04) and GR (p=1.46E-02) binding sites, implicating transcription factor binding as a mechanism of genotype-dependent stress response. We find that common genetic variation regulates transcriptional stress response in the post-mortem brain in a manner that is replicable in an hiPSC-derived neuronal model of stress. This regulation is cell type- and tissue-specific and offers a mechanistic understanding into variant-level regulation that may be further explored via integration with psychiatric disorder GWAS.