Functional Genomics of Bipolar Disorder

National Institute Of Mental Health

Investigators

Linked publications & trials

Abstract

NCT00001174 Despite strong evidence of heritability and discovery of genetic markers for major mental illness, little is known about how genetic differences affect gene expression in the brain or in individual brain cells, such as neurons and astrocytes. We have studied expression of genes and gene transcripts in postmortem subgenual anterior cingulate cortex (sgACC), a key component of limbic circuits linked to mental illness. Deep sequencing was carried out in RNA obtained postmortem from 200 donors diagnosed with bipolar disorder, schizophrenia, major depression, or no psychiatric disorder. Case-control comparisons detected modest expression differences that were similar across disorders, although transcript-level differences were more pronounced. The 250 rare transcripts that were differentially expressed were enriched for genes involved in synapse formation, cell junctions, and heterotrimeric G-protein complexes. Relative abundances of alternatively spliced transcripts were associated with common genetic variants that accounted for disproportionate fractions of diagnosis-specific heritability. Inherited genetic risk factors shape the brain transcriptome and contribute to diagnostic differences between broad classes of mental illness. Recent postmortem transcriptomic studies of human brain have shown hundreds of genes differentially expressed in donors with psychiatric illnesses. However, it is unclear which of these gene expression changes are the result of exposure to psychotropic medications rather than the illness itself. We explored this question using antipsychotic drug exposure in schizophrenia (SCZ) as a proof of concept. We compared differential gene expression in the prefrontal cortex from donors with SCZ who tested positive or negative for antipsychotics at the time of death. APD exposure was associated with numerous changes in the brain transcriptome, especially among donors exposed to atypical APDs. Brain transcriptome data from macaques chronically treated with APDs showed that APDs affect the expression of many functionally relevant genes, some of which show expression changes in the same directions as those observed in SCZ. In contrast, major cell type shifts inferred in SCZ were primarily unaffected by APD use. These results suggest that APD use confounds gene expression changes in postmortem brain tissue. Disentangling these effects will help identify causal genes and improve our neurobiological understanding of psychiatric disorders. We seek to model the impact of disease-related genes in cells derived from induced pluripotent stem cell (iPSC) lines. We have so far successfully reprogrammed fibroblasts into iPSCs from 89 study participants. We are developing a large iPSC-based resource and associated work-flows that constitute a living catalog of psychiatric risk alleles. iPSC-derived cells are studied with high-resolution microscopic imaging, electrophysiology, and gene expression methods. These data could reveal differences between control and patient-derived cells and the impact of known and novel therapeutic agents. We are also exploring ways to measure the functional impact of genetic mutations at the cellular level and use genome editing tools to establish a causal role for specific genetic mutations. Our ongoing copy number variant (CNV) studies examine known high risk CNVs in iPSC-derived neural cells and post-mortem brain. We have carried out extensive morphological and transcriptomic characterization of neural cells carrying duplication or deletion CNVs on chromosome 16p11.2, previously associated with BD and other psychiatric disorders. Transcriptomic analyses indicated that most genes in the CNV region show expression changes in neurons that were consistent with copy number, but many other genes were also dysregulated in carriers. Genes with concordant expression changes in carriers were enriched for several pathways, including neuronal growth and proliferation, synapse development, and cell migration. Consistent with these transcriptomic results, in vitro comparisons between carriers and matched non-carriers revealed major differences in cellular differentiation and growth, reduced synaptic structures, and reduced inter-neuronal signaling as measured by a multi-electrode array. While non-carrier NPCs easily differentiated into mature astrocytes using standard protocols, NPCs carrying the 16p11.2 duplication did not. Astrocytes play a critical role in the development and maintenance of healthy neurons and synapses. Exogenous astrocytes rescued most of the neurodevelopmental and synaptic deficits in 16p11.2 duplication carriers and restored inter-neuronal signaling toward normal levels. Astrocytes derived from carriers of the 16p11.2 duplication caused reduced synaptic development and reduced inter-neuronal signaling when cultured with neurons derived from non-carriers. These results demonstrate that CNVs on 16p11.2 confer convergent effects on gene expression, leading to neurodevelopmental and synaptic deficits that may play a causal role in psychiatric disorders. Despite their association with risk for a range of psychiatric disorders, little is known about the impact of recurrent neuropsychiatric CNVs on the human brain transcriptome. How do apparently disparate CNVs lead to convergent patterns of psychopathology? To address this, we performed single-nucleus RNAseq on a unique collection of brain tissue from carriers of high-risk CNVs. By screening national brain banks, we collected tissue from brain cortex of 13 carriers of neuropsychiatric CNVs, matched with 24 non-carriers. Differential gene expression analyses were performed across 9 cell types. As expected, the expression of genes within CNV regions corresponded to copy number in those regions, but differentially expressed genes (DEGs) were also widespread throughout the genome. Tissue from individuals with deletions exhibited more DEGs than those with duplications. In deletion carriers, DEGs converged on cellular stress, energy metabolism, and synaptic functions. This first transcriptomic analysis of high-risk CNVs in human brain at single-cell resolution suggests that deletions exert powerful effects on genes that converge on stress and energy metabolism pathways. These results are broadly consistent with previous in vitro studies, suggesting that diverse CNVs may lead to convergent defects in the brain. In the past year, we have screened a variety of psychotropic medications to identify drugs that may rescue or ameliorate neuronal dysfunction in BD and depression. Lithium and valproic acid (VPA) are popular first-line treatments for bipolar disorder. Extensive research has been done on the molecular effects of these two mood stabilizers but the mechanisms that mediate mood regulation remain elusive. We used human iPSC-derived neuronal cells to determine the impact of these two mood stabilizers on gene expression and regulation. Neurons were grown up to 6 weeks, treated with lithium, VPA, or no drug, then harvested for gene expression studies using RNAseq. This revealed many DEGs, esp. in VPA treated samples. Gene network analysis suggested Notch signaling as a key network influenced by both mood stabilizers. This project illustrates the use of iPSC-derived neurons to investigate gene mechanisms perturbed by known therapeutics. This system can also be used to screen novel therapeutics, such as ketamine. If successful, these projects will help unpack the biology behind genetic risk for mental illness and shed new light on how risk alleles exert biological changes in the brain. The findings may ultimately identify new targets that lead to better methods of diagnosis and treatment for neuropsychiatric disorders.