Investigating the Effects of Genetically Regulated Gene Expression on Neurological, Cardiovascular, and Psychiatric Health

Abstract

It is a matter of both clinical and scientific import to improve our understanding of how genetic variation influences the human brain in the contexts of both health and disease. Genetically regulated gene expression (GReX) often mediates the physiological consequences of germline variants as is evidenced by numerous transcriptome wide association studies (TWAS). To capture the consequences of GReX on human brain architecture, we conducted TWAS on over 3,500 heritable neuroimaging derived phenotypes (NIDPs) from the UK Biobank. The resulting transcriptional atlas of the human brain details associations between GReX of over 7,000 genes and NIDPs representing neurological structure, connectivity, and functional coactivation. GReX changes in both the brain and select somatic tissues demonstrated widespread, highly significant consequences for neuroanatomy. We identified 7 genes previously associated with structural heart measures that also predicted neurological changes, supporting transcription mediated, organ-level pleiotropy. The transcriptomic signature of Schizophrenia tagged a set of NIDPs that was statistically enriched for cortical regions affected in individuals with the disease. The same pattern was reduced in Parkinson’s and absent in Alzheimer’s, indicating that transcriptomic prediction of disease-relevant brain features is possible but sensitive to pathophysiology. Lastly, we leveraged whole blood TWAS reference panels trained in admixed populations enriched for African and indigenous American genetic ancestry to expand the set of brain-related GReX associations. Applying the admixed models to the UKB neuroimaging data, we expanded the set of GReX-NIDP associations in NeuroimaGene by 145%. Analysis of these data identified modules of genes that influenced network-like clusters of white matter tracts and adjoining cortical regions, many of which also demonstrated functional coactivation with each other. Together the work detailed here extends the molecular phenotyping of the brain using data from multiple ancestral backgrounds, highlights mechanistic avenues across the body through which genetic variation may impact brain health, and provides a sweeping array of candidate molecular mechanisms that may underlie empirical brain changes observed in the context of neurological and psychiatric disease.