Unfolding neural diversity: how dynamic three-dimensional genome architecture regulates brain function and disease

Abstract

The advent of single cell multi-omic technologies has ushered in a revolution in how we study the impact of three-dimensional genome organization on brain cellular composition and function. Transcriptomic and epigenomic studies reveal enormous cellular diversity that is present in mammalian nervous systems, raising the question, "how does this diversity arise and for what is its use?" Advances in the field of three-dimensional nuclear architecture have illuminated our understanding of how genome folding gives rise to dynamic gene expression programs important in healthy brain function and in disease. In this review we highlight recent work defining how neuronal identity, maturation, and plasticity are shaped by genome architecture. We discuss how newly identified genetic variations influence genome architecture and contribute to the evolution of species-unique neuronal and behavioral functional traits. We include examples for both humans and model organisms in which maladaptive genomic architecture is a causal agent in disease. Finally, we make conclusions and address future perspectives of dynamic three-dimensional genome (4D nucelome) research.

Fig. 1. Characterizing the linear and 3D genome. Schematic illustrations of the linear and 3D genome structures.

A A DNA strand is unwound revealing that only cytosine gets methylated and that MeCP2 acts as a “reader” by binding to methylated cytosines. A clutch of tightly spaced nucleosomes with red lollipops depicting heterochromatin and loosely spaced nucleosomes with blue lollipops depicting euchromatin. A nucleosome has an octameric core with 4 histone types that can be modified. Right next to this is a nucleosome that is in the process of being remodeled with a zoom in box depicting how chaperones/remodelers like nBAF can move things around to open up, close, or switch out nucleosomes. A CRE that is devoid of nucleosomes and occupied by TF(s), a break in the DNA to indicate distance, and a promoter devoid of nucleosomes and bound by a TF(s). B, C Newer high resolution chromatin contact maps show that a hierarchical model of TADs inside of compartments does not capture the characteristics of the 3D genomic structure -- there are compartments and sub-compartments as well as TADs and sub-TADs (conceptually re-created from Beagan et al., 2020) [58].

Fig. 2. Defining cell types based on chromatin structure. Schematic illustrations of the role of 4D nucleomics in defining cell types.

A, B A combination of 4DN-type techniques is comparable in resolution with transcriptomics to define cell types. A Genome browser tracks for 3 cell types depicting ATAC, H3K27ac, DNA methylation, and HiC loops. B A cell type gene expression matrix for 3 cell types and 3 genes highlighting how cell types are defined by 4DN characteristics. C–F Neural precursors and glia (and most cells in the body) predominantly show inactive (I, red) chromatin localized at the nuclear exterior while active (A, green) chromatin is located centrally. Neurons do not seem to follow this rule as strictly, with the extreme example of nocturnal photoreceptors being complexly inverted.

Fig. 3. Development of characteristics based on chromatin structure.

Schematic illustrations of the role of 4D nucleomics in the development of species characteristics. A Depiction of the three stages of development which contain blastocyst/ES cells, embryo/neural progenitor cells, and adult/cortical neurons and the trend that differentiation induces novel enhancer-promoter (EP) contacts as determined with Hi-C data (conceptually re-created from Bonev et al., 2017) [102]. B Cartoon showing that there exists EP contacts that are unique to humans (conceptually re-created from Luo et al., 2021) [109]. C A human specific E-P loop that stimulates expression of FGFR2. Disruption of this human specific enhancer results in decreased self-renewal of neural progenitor cells. D A human specific enhancer that stimulates expression of EPHA7 -- deletion of this enhancer results in aberrant dendrite development (conceptually re-created from de la Torre-Ubieta L et al., 2018) [107].

Fig. 4. Role of chromatin structure and gene regulation in regulating neural plasticity.

Schematic illustrations of the role of 4D nucleomics in neural plasticity. A, B Electroconvulsive stimulation (ECT) induces dentate gyrus (DG) granule neuron activity in hippocampus. B ECT-induced granule neuron activity induces transcription factor activity (CREB) leading to transcription of immediate early genes (IEGs) -- IRGs then induce secondary responses genes (SRGs) (conceptually re-created from Su et al., 2017) [110]. C, D The enhancer-promoter landscape of activity-induced genes can correlate with induction time. E, F Delay tactile startle conditioning used as a motor learning paradigm depends on granule cells of the cerebellum (conceptually re-created from Yamada et al., 2019) [118]. F–H Nose pokes learned to be associated with condition stimuli requires topological association domain (TAD) formation and other chromatin contacts dependent on the cohesin subunit Rad21.

Fig. 5. Evolution of characteristics based on chromatin structure.

Schematic illustrations of the role of 4D nucleomics in the evolution of species characteristics. A Cartoon depicting chromatin loop contacts and toplogically associated domains (TADs). B Cartoon graph showing that synteny breaks are enriched at TAD boundaries and selected against inside of TADs. C A cartoon showing TAD rearrangement in skates/sharks leads to increased prickle1 expression (conceptually re-created from Marletaz et al., 2023) [123]), and (D) a cartoon showing that TAD rearrangement in mice/moles results in masculinization (conceptually re-created from Real et al., 2020 [126]).

Fig. 6. Defining disease based on chromatin structure.

Schematic illustrations of the role of 4D nucleomics in defining diseases. A, B This depicts a GWAS identified variant associated with AD that is a Hi-C verified enhancer to the TSPAN14 gene, with the risk variant exhibiting less open chromatin at the enhancer and lower gene expression. This gene product assists with proper trafficking of the protease ADAM10 to the cell surface, where it cleaves the extracellular domain of TREM2, which is the second highest hit in late onset AD screens (conceptually re-created from Yang et al., 2023) [128].