55. HIGHLIGHTING HUMAN BRAIN CELL TYPES AND REGIONS FOR THE GENETIC RISK OF SCHIZOPHRENIA AND OTHER PSYCHIATRIC DISORDERS

• Published in: European neuropsychopharmacology Vol. 75; pp. S85 - S86
• Main Authors: Yao, Shuyang, Harder, Arvid, Darki, Fahimeh, Li, Ang, Zeng, Jian, Nikouei, Kasra, Lundström, Johan, Wray, Naomi, Lu, Yi, Sullivan, Patrick, Hjerling-Leffler, Jens
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-10-2023
• ISSN: 0924-977X; 1873-7862
• DOI: 10.1016/j.euroneuro.2023.08.161

Understanding the meaning of the genomic findings of psychiatric disorders holds the potential to clarify the mechanism of these heritable disorders. Connecting genomic findings with single-cell transcriptomics can point towards specific cell types, providing a meaningful way to interpret GWAS findings. With recently available cell type atlas of the adult human brain, we aim to leverage the interpretation from cell type level to brain regions and their connectivity.
We integrated GWAS summary statistics of psychiatric disorders with the newly published adult human brain single-nucleus transcriptomic atlas that covers 106 regions across the brain. The atlas has annotated 31 superclusters and 461 clusters from over 3.3 million single-nuclei. Specifically, we evaluated whether the most specifically expressed genes per cell type (supercluster and clusters separately) were enriched of SNP-based heritability for each tested psychiatric disorder.
We identified multiple neuronal superclusters that are significantly enriched of the SNP-based heritability of schizophrenia, bipolar disorder, and major depressive disorder. For schizophrenia, the most significantly enriched excitatory neuronal clusters are among the Amygdala Excitatory, Hippocampal CA1-3, and Upper- and Deep-layer Intratelencephalic superclusters, and the most significantly enriched inhibitory neuronal clusters are among the Eccentric Medium Spiny Neurons. Based on cell type composition per sampled region, we further integrated cell-type-level results to brain regions. In general, cerebral cortical regions manifested similar significance of enrichment, which can be explained by their similar compositions of cell types. In contrast, certain regions in the hippocampus and amygdala showed the highest significance ((P < 5e-4), suggesting the potential importance of these regions for schizophrenia etiology. Finally, we attempted to distinguish between 47 schizophrenia cases and 47 controls using functional connectivity (fMRI) of the regions selected based on the significance of SNP-based heritability enrichment. With 80% training vs 20% testing sets, we obtained a classification accuracy of 0.72, with the connection between amygdala and hippocampus being the top contributor to the prediction model.
We have identified cell type in the human brain that are important for psychiatric disorder etiologies, and further leveraged these findings to highlight brain regions, especially the amygdala and the hippocampus. Results using fMRI data further imply potentially different connectivity of the selected brain regions between schizophrenia cases and controls. The study illustrates the great value of single-cell transcriptomics in understanding the polygenic risk of psychiatric disorders, and suggests the potential link between such genetic risks and brain regions and their connectivity.