Long-range chromatin interactions in pathogenic gene expression control

ABSTRACT

A large number of distal cis-regulatory elements (cREs) have been annotated in the human genome, which plays a central role in orchestrating spatiotemporal gene expression. Since many cREs regulate non-adjacent genes, long-range cRE-promoter interactions are an important factor in the functional characterization of the engaged cREs. In this regard, recent studies have demonstrated that identification of long-range target genes can decipher the effect of genetic mutations residing within cREs on abnormal gene expression. In addition, investigation of altered long-range cREs-promoter interactions induced by chromosomal rearrangements has revealed their critical roles in pathogenic gene expression. In this review, we briefly discuss how the analysis of 3D chromatin structure can help us understand the functional impact of cREs harboring disease-associated genetic variants and how chromosomal rearrangements disrupting topologically associating domains can lead to pathogenic gene expression.

Introduction

During mammalian development, a fertilized egg gives birth to various cell types and organs throughout the delicate regulation of spatiotemporal gene expression. Systematic investigation of chromatin architecture in various developmental stages and cell-types has shown that such regulation processes are tightly controlled by cis-regulatory elements (cREs) [1–3]. Although cREs refer to all cis-acting noncoding DNA elements that regulate target gene expression such as promoters, enhancers, and silencers, in this review, we mainly focus on enhancer-like cREs. Specific histone modifications such as mono-methylation of histone H3 at lysine 4 (H3K4me1) and co-activator binding patterns are hallmarks of cREs [4]. In addition, DNA methylation and chromatin accessibility are good predictors to identify candidates of regulatory DNA elements [5,6]. Active cREs can be further specified by the enriched acetylation of histone H3 lysine 27 (H3K27ac), bi-directional transcripts (also known as enhancer RNAs), and even tri-methylation of histone H3 lysine 4 (H3K4me3) that is a hallmark of active promoters [7–10]. These chromatin signatures have been utilized to annotate hundreds of thousands of cREs in the mammalian genome through incorporation with machine-learning based computational methods [11–15].

Despite a remarkable achievement in annotating cREs, most of them often regulate non-adjacent genes over large genomic distances, making it challenging to characterize their functional properties [16–19]. Such long-range regulation can take place because chromatin fibers are folded into a higher-order structure, forming DNA loops where regulatory proteins bound at cREs can be in physical closeness to the promoters of their target genes. Thus, target gene identification based on spatial chromatin contacts is crucial in elucidating the regulatory potential of cREs during both normal and pathogenic gene expression control. In this aspect, we will briefly review how the analysis of chromatin interactions is critical to define target gene promoters of cREs and to understanding disease pathogenesis.

Methods to identify target genes based on long-range chromatin interactions

The establishment of Chromosome Conformation Capture (3C) [20] has evolved in the form of high-throughput versions such as 4C, 5C, and Hi-C to detect genome-wide chromatin interactions [21–24]. Through a combination of proximity ligation with next-generation sequencing, these approaches have identified many chromatin contacts of cRE-promoter pairs and provided a comprehensive view of the regulatory interactome [25,26].

To concentrate on well-annotated promoter connecting long-range chromatin interactions, Capture-C and promoter capture Hi-C methods have been developed [17,27–29]. These methods provide high-resolution cRE-promoter interaction maps with a relatively low sequencing cost by capturing ligated DNA fragments containing targeted promoters from 3C or Hi-C libraries. The unprecedented high-resolution cRE-promoter interaction maps have revealed that cell-type specific chromatin interactions often connect cell-type specific promoter and cRE pairs [17,28]. These data strongly support the essential function of long-range chromatin interactions in cell-type specific cRE usage.

To further examine the chromatin interactions bound by a particular protein, Chromatin Interaction Analysis by Paired-End-Tag sequencing (ChIA-PET), HiChIP, and Proximity Ligation-Assisted ChIP-seq (PLAC-seq) have been developed through a combination of ChIP and proximity ligation [30–33]. In ChIA-PET, chromatin immunoprecipitation of a particular protein is performed, followed by proximity ligation to collect chromatin interactions mediated by the target protein. In contrast, HiChIP and PLAC-seq conduct proximity ligation in nuclei prior to immunoprecipitation. Currently, HiChIP and PLAC-seq are widely used owing to their high efficiency and low input requirements.

The rapid accumulation of the rich resource in chromatin interactions enables the development of several computational methods to systematically identify target genes of cREs [34–36]. For example, PSYCHIC annotates several hundred thousand putative cREs and identified their target genes through the analysis of Hi-C data [35]. The development of computational methods in combining various genomic and epigenomic features will be further beneficial for the target gene annotation compared to solely based on chromatin contacts.

Functional annotation of GWAS-SNPs based on long-range chromatin interactions

Many important genetic variants associated with complex human diseases and traits have been identified thanks to the power of genome-wide association studies [37]. Interestingly, more than 95% of GWAS-identified SNPs are in noncoding sequences, and more than three-quarters are associated with open chromatin regions, suggesting their strong association with regulatory elements [38]. Thus, distal target gene information is essential to identify genes they affect. For example, multiple GWAS studies to identify genetic risk factors in obesity had consistently demonstrated variants within the FTO introns as significantly associated variants with the risk of obesity [39–41]. However, there was no clear explanation of how these variants regulate FTO gene expression and potentially cause obesity. This conundrum has been resolved through the analyses of long-range chromatin interactions that have revealed that obesity-associated GWAS-SNPs create cRE activity, a potent preadipocyte super-enhancer, and physically interact with distal promoters of IRX3 and IRX5 (Figure 1) [42,43]. Further experiments found that increased expression of IRX3 and IRX5 genes resulted in changes of body mass index [42,43]. As illustrated in the above example, target gene identification of noncoding associated GWAS-SNPs based on long-range chromatin interactions has been widely applied as a novel strategy to understand gene regulation mechanisms for various complex human diseases.

Figure 1.

Target gene identification of cREs harboring obesity-associated GWAS-SNPs

Visualization of significant long-range chromatin interactions (arcs) centered on obesity-associated GWAS-SNPs (rs1421085 and rs1558902) located at FTO intron in H1-derived mesenchymal stem cell. From the top, normalized Hi-C contact map with annotation of TADs (blue triangles), RefSeq gene annotation, normalized Hi-C contacts with obesity-associated GWAS-SNPs located in intronic regions of FTO gene, identified significant chromatin interactions, multiple histone modification marks, and GWAS-SNPs are shown together. The long-range chromatin interactions with the promoters of IRX3 and IRX5 are highlighted by red. All data were drawn from 3DIV database [36].

Pathogenic long-range chromatin interactions in human diseases

Large-scale chromosomal rearrangements such as structural variants (SVs) can disrupt the organization of topologically associating domains (TADs), thereby inducing aberrant gene expression (Table 1). TADs are megabase-scale local chromatin interaction domains that restrict the interactions between cREs and promoters outside of TAD [44,45]. Thus, changes in the TAD structure can rewire long-range chromatin interactions between cRE and promoter, leading to introducing ectopic cRE-gene interactions called enhancer-hijacking or enhancer-adoption [46–48]. Several studies have shown that TAD fusions, a fused form of multiple TADs due to genomic rearrangements, can activate proto-oncogenes by enhancer-hijacking mechanism [49–55]. Moreover, translocations and inversions spanning TAD boundaries can be responsible for TAD shuffling, which may disrupt originally linked cRE-promoter relationships [56]. Therefore, clarifying target genes of cREs based on rewired chromatin interactions can deeply elucidate the pathological consequence of noncoding-associated genomic rearrangement.

Table 1.

Rewired cRE-promoter interactions in pathogenic gene expression control

Disease	Type of enhancer usage	Variant type	Cause	Result	Reference
Limb malformation	Enhancer-hijacking	Inversion, duplication, deletion	TAD fusion by variants at CTCF-associated boundary domain	Overexpression of WNT or 6IHH or PAX3	47
T cell acute lymphoblastic leukemia	Enhancer-hijacking	Deletion	TAD fusion by variants at CTCF-associated boundary domain	Activation of LMO2 proto-oncogene	50
Medulloblastoma	Enhancer-hijacking	Tandem duplication, deletion, inversion, complex rearrangement	Relocating oncogene proximal to super-enhancer	Activation of GFI1 and GFI1B proto-oncogenes, and overexpression of PRDM6	52, 51
Oncogene activation	Enhancer-hijacking	Tandem duplication	neo-TAD appearance	Overexpression of IGF2	55
Neuroblastoma	Enhancer-hijacking	Copy number alterations	Relocating oncogene proximal to super-enhancer	Overexpression of TERT	53, 54
Branchiooculofacial syndrome	Enhancer disconnection	Inversion	TAD shuffling	Monoallelic and reduced TFAP2A expression	56
Acute myeloid leukemia	Enhancer-hijacking	Inversion	TAD shuffling	Activation of EVI1 proto-oncogene	49

Conclusion

The analysis of chromatin interactions has dramatically advanced our knowledge of understanding disease pathogenesis by identifying target gene information of noncoding located regulatory elements. However, a range of issues remains to be solved. First, the resolution of the 3C based methods to identify chromatin interactions is often too low, generally from kilobases to tens of kilobases. Thus, it hinders their ability to pinpoint the precise location of interacting DNA fragments. Second, more efforts are needed to distinguish the type of cREs interacting with their target promoters. Although we have discussed enhancer-like cREs in this review, especially within a TAD, silencers also act in a cis manner to regulate their target genes similar to enhancers [57]. Recent studies experimentally screened potential silencers, which turns out that strongly silencer-associated DNA elements also pose active chromatin marks that can be misinterpreted as enhancers [57–59]. Thus, discriminating multiple types of regulatory elements is critical in the functional annotation of noncoding genetic variants. Lastly, the identification of target genes at single-cell resolution is ultimately required. Single-cell based Hi-C methods [60–63] and a combination of imaging of the 3D chromatin structure [64,65] will revolutionize our understanding of long-range chromatin interactions and their regulatory potential. Nevertheless, the identification of target genes of cREs based on “C” technologies is highly beneficial in unraveling the regulatory potential of noncoding-associated GWAS-SNPs and large-scale genomic rearrangements, bridging the gap between coding and noncoding sequences.

Funding Statement

This work was supported by the Ministry of Science and ICT through the National Research Foundation in Republic of Korea under Grant number NRF-2020R1A2C4001464;

Disclosure statement

No potential conflict of interest was reported by the authors.