Comprehensive multi-omics integration identifies differentially active enhancers during human brain development with clinical relevance

Using a comprehensive atlas and predictive models to reveal the complexity and evolution of brain-active regulatory elements

Most genetic variants associated with psychiatric disorders are located in noncoding regions of the genome. To investigate their functional implications, we integrate epigenetic data from the PsychENCODE Consortium and other published sources to construct a comprehensive atlas of candidate brain cis-regulatory elements. Using deep learning, we model these elements' sequence syntax and predict how binding sites for lineage-specific transcription factors contribute to cell type-specific gene regulation in various types of glia and neurons. The elements' evolutionary history suggests that new regulatory information in the brain emerges primarily via smaller sequence mutations within conserved mammalian elements rather than entirely new human- or primate-specific sequences. However, primate-specific candidate elements, particularly those active during fetal brain development and in excitatory neurons and astrocytes, are implicated in the heritability of brain-related human traits. Additionally, we introduce PsychSCREEN, a web-based platform offering interactive visualization of PsychENCODE-generated genetic and epigenetic data from diverse brain cell types in individuals with psychiatric disorders and healthy controls.

Lipotype acquisition during neural development is not recapitulated in stem cell-derived neurons

During development, different tissues acquire distinct lipotypes that are coupled to tissue function and homeostasis. In the brain, where complex membrane trafficking systems are required for neural function, specific glycerophospholipids, sphingolipids, and cholesterol are highly abundant, and defective lipid metabolism is associated with abnormal neural development and neurodegenerative disease. Notably, the production of specific lipotypes requires appropriate programming of the underlying lipid metabolic machinery during development, but when and how this occurs is unclear. To address this, we used high-resolution MSALL lipidomics to generate an extensive time-resolved resource of mouse brain development covering early embryonic and postnatal stages. This revealed a distinct bifurcation in the establishment of the neural lipotype, whereby the canonical lipid biomarkers 22:6-glycerophospholipids and 18:0-sphingolipids begin to be produced in utero, whereas cholesterol attains its characteristic high levels after birth. Using the resource as a reference, we next examined to which extent this can be recapitulated by commonly used protocols for in vitro neuronal differentiation of stem cells. Here, we found that the programming of the lipid metabolic machinery is incomplete and that stem cell-derived cells can only partially acquire a neural lipotype when the cell culture media is supplemented with brain-specific lipid precursors. Altogether, our work provides an extensive lipidomic resource for early mouse brain development and highlights a potential caveat when using stem cell-derived neuronal progenitors for mechanistic studies of lipid biochemistry, membrane biology and biophysics, which nonetheless can be mitigated by further optimizing in vitro differentiation protocols.

Genetic effects of sequence-conserved enhancer-like elements on human complex traits

BACKGROUND

The vast majority of findings from human genome-wide association studies (GWAS) map to non-coding sequences, complicating their mechanistic interpretations and clinical translations. Non-coding sequences that are evolutionarily conserved and biochemically active could offer clues to the mechanisms underpinning GWAS discoveries. However, genetic effects of such sequences have not been systematically examined across a wide range of human tissues and traits, hampering progress to fully understand regulatory causes of human complex traits.

RESULTS

Here we develop a simple yet effective strategy to identify functional elements exhibiting high levels of human-mouse sequence conservation and enhancer-like biochemical activity, which scales well to 313 epigenomic datasets across 106 human tissues and cell types. Combined with 468 GWAS of European (EUR) and East Asian (EAS) ancestries, these elements show tissue-specific enrichments of heritability and causal variants for many traits, which are significantly stronger than enrichments based on enhancers without sequence conservation. These elements also help prioritize candidate genes that are functionally relevant to body mass index (BMI) and schizophrenia but were not reported in previous GWAS with large sample sizes.

CONCLUSIONS

Our findings provide a comprehensive assessment of how sequence-conserved enhancer-like elements affect complex traits in diverse tissues and demonstrate a generalizable strategy of integrating evolutionary and biochemical data to elucidate human disease genetics.

Solving the unsolved genetic epilepsies: Current and future perspectives

Many patients with epilepsy undergo exome or genome sequencing as part of a diagnostic workup; however, many remain genetically unsolved. There are various factors that account for negative results in exome/genome sequencing for patients with epilepsy: (1) the underlying cause is not genetic; (2) there is a complex polygenic explanation; (3) the illness is monogenic but the causative gene remains to be linked to a human disorder; (4) family segregation with reduced penetrance; (5) somatic mosaicism or the complexity of, for example, a structural rearrangement; or (6) limited knowledge or diagnostic tools that hinder the proper classification of a variant, resulting in its designation as a variant of unknown significance. The objective of this review is to outline some of the diagnostic options that lie beyond the exome/genome, and that might become clinically relevant within the foreseeable future. These options include: (1) re-analysis of older exome/genome data as knowledge increases or symptoms change; (2) looking for somatic mosaicism or long-read sequencing to detect low-complexity repeat variants or specific structural variants missed by traditional exome/genome sequencing; (3) exploration of the non-coding genome including disruption of topologically associated domains, long range non-coding RNA, or other regulatory elements; and finally (4) transcriptomics, DNA methylation signatures, and metabolomics as complementary diagnostic methods that may be used in the assessment of variants of unknown significance. Some of these tools are currently not integrated into standard diagnostic workup. However, it is reasonable to expect that they will become increasingly available and improve current diagnostic capabilities, thereby enabling precision diagnosis in patients who are currently undiagnosed.

MYT1L is required for suppressing earlier neuronal development programs in the adult mouse brain

In vitro studies indicate the neurodevelopmental disorder gene myelin transcription factor 1-like (MYT1L) suppresses non-neuronal lineage genes during fibroblast-to-neuron direct differentiation. However, MYT1L's molecular and cellular functions in the adult mammalian brain have not been fully characterized. Here, we found that MYT1L loss leads to up-regulated deep layer (DL) gene expression, corresponding to an increased ratio of DL/UL neurons in the adult mouse cortex. To define potential mechanisms, we conducted Cleavage Under Targets & Release Using Nuclease (CUT&RUN) to map MYT1L binding targets and epigenetic changes following MYT1L loss in mouse developing cortex and adult prefrontal cortex (PFC). We found MYT1L mainly binds to open chromatin, but with different transcription factor co-occupancies between promoters and enhancers. Likewise, multiomic data set integration revealed that, at promoters, MYT1L loss does not change chromatin accessibility but increases H3K4me3 and H3K27ac, activating both a subset of earlier neuronal development genes as well as Bcl11b, a key regulator for DL neuron development. Meanwhile, we discovered that MYT1L normally represses the activity of neurogenic enhancers associated with neuronal migration and neuronal projection development by closing chromatin structures and promoting removal of active histone marks. Further, we showed that MYT1L interacts with HDAC2 and transcriptional repressor SIN3B in vivo, providing potential mechanisms underlying repressive effects on histone acetylation and gene expression. Overall, our findings provide a comprehensive map of MYT1L binding in vivo and mechanistic insights into how MYT1L loss leads to aberrant activation of earlier neuronal development programs in the adult mouse brain.

Illuminating links between cis-regulators and trans-acting variants in the human prefrontal cortex

BACKGROUND

Neuropsychiatric disorders afflict a large portion of the global population and constitute a significant source of disability worldwide. Although Genome-wide Association Studies (GWAS) have identified many disorder-associated variants, the underlying regulatory mechanisms linking them to disorders remain elusive, especially those involving distant genomic elements. Expression quantitative trait loci (eQTLs) constitute a powerful means of providing this missing link. However, most eQTL studies in human brains have focused exclusively on cis-eQTLs, which link variants to nearby genes (i.e., those within 1 Mb of a variant). A complete understanding of disease etiology requires a clearer understanding of trans-regulatory mechanisms, which, in turn, entails a detailed analysis of the relationships between variants and expression changes in distant genes.

METHODS

By leveraging large datasets from the PsychENCODE consortium, we conducted a genome-wide survey of trans-eQTLs in the human dorsolateral prefrontal cortex. We also performed colocalization and mediation analyses to identify mediators in trans-regulation and use trans-eQTLs to link GWAS loci to schizophrenia risk genes.

RESULTS

We identified ~80,000 candidate trans-eQTLs (at FDR<0.25) that influence the expression of ~10K target genes (i.e., "trans-eGenes"). We found that many variants associated with these candidate trans-eQTLs overlap with known cis-eQTLs. Moreover, for >60% of these variants (by colocalization), the cis-eQTL's target gene acts as a mediator for the trans-eQTL SNP's effect on the trans-eGene, highlighting examples of cis-mediation as essential for trans-regulation. Furthermore, many of these colocalized variants fall into a discernable pattern wherein cis-eQTL's target is a transcription factor or RNA-binding protein, which, in turn, targets the gene associated with the candidate trans-eQTL. Finally, we show that trans-regulatory mechanisms provide valuable insights into psychiatric disorders: beyond what had been possible using only cis-eQTLs, we link an additional 23 GWAS loci and 90 risk genes (using colocalization between candidate trans-eQTLs and schizophrenia GWAS loci).

CONCLUSIONS

We demonstrate that the transcriptional architecture of the human brain is orchestrated by both cis- and trans-regulatory variants and found that trans-eQTLs provide insights into brain-disease biology.

Disruption of the gene regulatory programme in neurodevelopmental disorders

Cortical development consists of a series of synchronised events, including fate transition of cortical progenitors, neuronal migration, specification and connectivity. It is becoming clear that gene expression programs governing these events rely on the interplay between signalling molecules, transcription factors and epigenetic mechanisms. When genetic or environmental factors disrupt expression of genes involved in important brain development processes, neurodevelopmental disorders can occur. This review aims to highlight how recent advances in technologies have helped uncover and imitate the gene regulatory mechanisms commonly disrupted in neurodevelopmental disorders.