To understand how a cell regulates its genes, it is not enough to read the DNA sequence linearly. The human genome is tightly folded inside the nucleus, and this three-dimensional folding dictates which regulatory elements can interact with which genes. The fundamental units of this spatial organization are Topologically Associating Domains (TADs)—insulated genomic neighborhoods where DNA loops frequently interact with itself, but rarely with neighboring domains.
As reviewed by Wujuan Zhong et al. in Understanding the function of regulatory DNA interactions in the interpretation of non-coding GWAS variants, these domains act as structural rooms. Inside a room, a powerful regulatory element called an enhancer can easily turn on its target gene promoter. Crucially, the walls of these rooms—TAD boundaries—prevent that enhancer from accidentally activating genes in the next room over. When these structural boundaries are disrupted, the insulation fails, leading to a pathological phenomenon known as enhancer hijacking.
Mechanisms of Boundary Failure
How do mutations break these physical walls? Structural variations (SVs) and non-coding mutations disrupt TAD boundaries through several distinct spatial mechanisms:
- TAD Fusion via Deletions: When a deletion removes a boundary element, the wall between two rooms vanishes. Chong Li et al. demonstrated this in An integrative TAD catalog in lymphoblastoid cell lines discloses the functional impact of deletions and insertions in human genomes, showing that deletions systematically cause adjacent regulatory domains to fuse into a single neighborhood, allowing cross-domain interactions.
- Creation of Neo-TADs via Translocations: Balanced chromosomal translocations can slice a domain and stitch it to a completely foreign region of the genome. Wenmin Sun et al. discovered in Altered chromatin topologies caused by balanced chromosomal translocation lead to central iris hypoplasia that these rearrangements create entirely new interaction patterns—termed "neo-TADs"—bringing distant enhancers into direct proximity with inappropriate gene promoters.
- Breakdown of Architectural Machinery: Boundaries also collapse when the molecular machinery that maintains them is mutated. Henrik Hornshøj et al. found in Pan-cancer screen for mutations in non-coding elements with conservation and cancer specificity reveals correlations with expression and survival that high frequencies of non-coding mutations cluster at binding sites for the cohesin complex, the motor protein responsible for loop extrusion and boundary maintenance. When specific cohesin components are lost, the structural integrity of these boundaries degrades; for instance, Zhaowei Chu et al. showed in STAG2 regulates interferon signaling in melanoma via enhancer loop reprogramming that the loss of the cohesin subunit STAG2 causes domains to expand outward, particularly at their boundaries, rewriting local enhancer-promoter loops.
Epigenetic Instability and Enhancer Hijacking in Cancer
In malignant cells, boundary disruption serves as a primary driver of oncogene activation. When a boundary fails, a gene that was safely sequestered can be exposed to a highly potent regulatory hub, such as a super-enhancer.
The authors of Noncoding mutations cause super-enhancer retargeting resulting in protein synthesis dysregulation during B cell lymphoma progression (Rebecca J. Leeman‐Neill et al.) demonstrated that somatic mutations frequently cluster near TAD boundaries, causing super-enhancers to "retarget" and misregulate protein synthesis machinery during lymphoma progression. Similarly, Yu Dong et al. established in Structural variation drives enhancer hijacking via 3D genome disruption in ccRCC that structural variants fuse previously isolated neighborhoods to activate oncogenes like SEMA5B in kidney cancer.
This structural collapse operates in tandem with chemical modifications to the DNA. Yiqun Zhang et al. showed in Global impact of somatic structural variation on the DNA methylome of human cancers that structural breakpoints breaching TAD boundaries systematically alter local DNA methylation patterns, driving recurrent gene silencing or activation.
Conversely, changes in methylation can themselves destabilize structure: Jennifer Derrien et al. discovered in The DNA methylation landscape of multiple myeloma shows extensive inter- and intrapatient heterogeneity that fuels transcriptomic variability that the large-scale erosion of DNA methylation increases the physical "floppiness" or structural flexibility of the 3D genome. This structural looseness allows genomic regions to switch more easily between active and inactive compartments, fueling transcriptomic chaos.
Phenotypic Consequences in Hereditary and Psychiatric Disorders
The impact of 3D boundary interference extends far beyond cancer, playing a critical role in hereditary and complex neuropsychiatric diseases.
In Mendelian disorders, disrupting non-coding spatial architecture can cause precise developmental defects. For example, Wenmin Sun et al. utilized Hi-C data analysis in Structural variations in a non-coding region at 1q32.1 are responsible for the NYS7 locus in two large families to show that large deletions in non-coding regions disrupt TAD organization, directly causing congenital motor nystagmus.
In complex psychiatric conditions, spatial misregulation is emerging as a major risk factor. Matthew Halvorsen et al. found in Increased burden of ultra-rare structural variants localizing to boundaries of topologically associated domains in schizophrenia that patients with schizophrenia carry an enriched burden of ultra-rare structural variants localized precisely at TAD boundaries, leading to the accidental activation or silencing of genes by distant regulatory elements.
These structural deletions can trigger cascading effects across the genome; Ralda Nehme et al. demonstrated in The 22q11.2 region regulates presynaptic gene-products linked to schizophrenia that large-scale deletions alter 3D folding to cause widespread trans-acting transcriptional dysregulation, specifically impairing presynaptic machinery and reducing neuronal activity.
Predicting and Stratifying Pathogenic Disruptions
Not every broken boundary results in disease. In a comprehensive pan-cancer analysis of over 2,600 genomes, Victoria T. Le et al. revealed in Disruption of chromatin folding domains by somatic genomic rearrangements in human cancer that while structural mutations frequently shatter and fuse genomic neighborhoods across all cancer types, the majority of these disruptions surprisingly do not lead to massive changes in gene expression.
Because the genome possesses some resilience, researchers must distinguish between benign structural changes and truly pathogenic disruptions. To address this, Bijia Chen et al. developed a method in Stratifying TAD boundaries pinpoints focal genomic regions of regulation, damage, and repair to stratify boundaries by their hierarchical complexity. They discovered that "high-level" boundaries—those separating multiple distinct domains—act as critical regulatory hubs and are highly sensitive hotspots for DNA damage and repair.
To automate the identification of dangerous variants, Sushant Kumar et al. created SVFX: a machine learning framework to quantify the pathogenicity of structural variants. By integrating tissue-specific epigenomic data with structural features, this framework models how rearrangements alter the physical proximity between enhancers and target genes. By capturing this spatial decoupling, machine learning models can accurately predict when a non-coding structural variant will cause disease, providing a vital tool for interpreting the pathogenic potential of the disrupted 3D genome.
Key Foundational Work
Global impact of somatic structural variation on the DNA methylome of human cancers (2019) — contributed 3 concept elements to this direction.
Pan-cancer screen for mutations in non-coding elements with conservation and cancer specificity reveals correlations with expression and survival (2018) — contributed 1 concept elements to this direction.