CRISPR-based strategies for studying regulatory elements and chromatin structure in mammalian gene control

Abstract

The development of high-throughput methods has enabled the genome-wide identification of putative regulatory elements in a wide variety of mammalian cells at an unprecedented resolution. Extensive genomic studies have revealed the important role of regulatory elements and genetic variation therein in disease formation and risk. In most cases, there is only correlative evidence for the roles of these elements and non-coding changes within these elements in pathogenesis. With the advent of genome- and epigenome-editing tools based on the CRISPR technology, it is now possible to test the functional relevance of the regulatory elements and alterations on a genomic scale. Here, we review the various CRISPR-based strategies that have been developed to functionally validate the candidate regulatory elements in mammals as well as the non-coding genetic variants found to be associated with human disease. We also discuss how these synthetic biology tools have helped to elucidate the role of three-dimensional nuclear architecture and higher-order chromatin organization in shaping functional genome and controlling gene expression.

Introduction

Genome-wide epigenetic functional elements such as cis- and trans-acting DNA regulatory elements and chromatin state have been mapped for many different cell types in mammals (Rivera and Ren 2013). However, the potential functions, biological roles, or mechanistic principles of those regulatory elements are typically unknown. The causal relationships between chromatin features and transcription are complex and cell-type dependent. Therefore, the functional validation of the candidate regulatory elements is one of the most important endeavors in the biology of gene regulation. Furthermore, identification and functional validation of causal variants in the non-coding genome that affect the activity of regulatory elements is critical in our understanding of genetic predisposition to disease and in developing therapeutic targets. With the advent of genome-editing tools, it is now possible to identify and test the functional relevance of the putative regulatory elements and non-coding alterations therein on a genomic scale.

CRISPR (clustered regulatory interspaced short palindromic repeat)/Cas9-based RNA-guided DNA endonuclease technology is transforming molecular genetics research and has quickly become the preferred platform for genome and epigenome editing (Cong et al. 2013; Lau and Suh 2017). The CRISPR technology has revolutionized the gene-editing technique, due to its simplicity in target design, affordability, high efficiency, versatility, and multiplexing capability (Cong et al. 2013). The commonly used CRISPR system can be implemented in mammalian cells by co-expressing Cas9 nuclease along with chimeric guide RNA (gRNA), which is derived from a synthetic fusion of the CRISPR RNA array (crRNA) and trans-activating crRNA (tracrRNA) (Cong et al. 2013). Apart from the classical applications of CRISPR technology in functional gene knockout, genetic variants knock-in, and gene-expression modulation, CRISPR technology is increasingly being used to study the epigenetic regulation and chromatin biology in mammals (Lau and Suh 2017; Montalbano et al. 2017).

The background of and the recent developments in the CRISPR-based targeted genome and epigenome-editing toolboxes have been extensively reviewed elsewhere (Barrangou and Doudna 2016; Lau and Suh 2017; Lo and Qi 2017; Pulecio et al. 2017). Various applications of CRISPR technologies for mammalian genome interrogation and biomedical research have also been recently reviewed (Komor et al. 2017). Herein, this review aims to provide a comprehensive overview of the various CRISPR-based strategies that have been developed to uncover the functional roles of mammalian regulatory genomic elements and chromatin structure, as well as their associated non-coding genetic variants discovered in humans. We also discuss recent applications of the CRISPR imaging techniques to characterize the functional roles of three-dimensional nuclear architecture and higher-order chromatin organization in shaping functional genome and controlling gene expression. In addition, we discuss the potential use of CRISPR technology to manipulate nuclear architecture and gene expression for therapeutic benefit.

CRISPR-based strategies for studying regulatory elements and chromatin structure in mammalian gene control

There are multiple levels of chromatin regulation to ensure correct spatial and temporal gene activation for normal organismal development and cellular function. One of the potential epigenetic regulatory mechanisms underlying chromatin dynamics during gene transcription is through the action of cis- and trans-acting regulatory DNA elements. Some of the well-characterized regulatory genomic elements are insulators (Flavahan et al. 2016), promoters (Hwang et al. 2013), and enhancers (Guo et al. 2015; Korkmaz et al. 2016). These regulatory elements cooperatively and dynamically modulate chromatin states and gene transcriptions in the nuclei (Dao et al. 2017; Guo et al. 2015; Lupianez et al. 2015). Recent evidence suggests that specific chromosome structures, such as insulated neighborhoods and higher-order topologically associating domains (TADs), play important roles in gene control (Hnisz et al. 2016). CRISPR technology has been a robust tool to characterize the functional roles of these genomic regulatory elements in regulating local chromatin loop formation, chromatin contacts, and gene–enhancer interactions (Fulco et al. 2016; Korkmaz et al. 2016; Montalbano et al. 2017). Given the prominent role of three-dimensional nuclear architecture and chromatin topology in reshaping the epigenetic landscape, CRISPR imaging is an attractive approach for tracking long-range epigenetic regulatory architecture, higher-order chromatin organization and dynamics (Guo et al. 2015; Lupianez et al. 2015).

Enhancers are the principle regulatory components of the genome that enable cell-type and cell-state specificities of gene expression (Zhu et al. 2017). Enhancers are typically characterized by the presence of histone modifications such as H3K27Ac and H3K4me¹, co-occupancy of multiple transcription factors that recruit co-activators and RNA polymerase II to target genes, depletion of nucleosome and thus DNase hypersensitivity, expression of enhancer RNAs (eRNAs), and hypomethylation at CpG dinucleotides (Heidari et al. 2014; Zhu et al. 2017). An enhancer can be located proximal to the gene it regulates, or far from the gene in a linear genomic DNA sequence, though the distal enhancer is spatially close to the target promoter and gene through chromatin looping (Heidari et al. 2014). Recently, transcription at intragenic enhancers was shown to attenuate host gene expression through transcription interference (Cinghu et al. 2017), suggesting differential utilization of the same regulatory element for disparate functions. A cluster of constituent enhancers or simply an isolated strong enhancer was proposed as a super-enhancer. Super-enhancers differ from constituent enhancers in size, cell-type-specific transcription factor density and content, the degree of transcriptional activation, and sensitivity to perturbation (Whyte et al. 2013). Compared to enhancers, super-enhancers might play a more significant role in human cell identity by having a large cluster of transcriptional enhancers in close genomic proximity with unusually high levels of master regulators and mediator binding (Hnisz et al. 2013; Pott and Lieb 2015). Both enhancer and super-enhancer were recently shown to have an equivalent regulatory role in gene regulation (Moorthy et al. 2017), suggesting the ambiguity in differentiating enhancers from super-enhancers. Nevertheless, similar strategies of the CRISPR technology could be applied to manipulate both enhancers and super-enhancers. Recently, it has been shown that the majority of disease-associated variants detected by genome-wide association studies (GWASs) are enriched in enhancers (ENCODE Project Consortium 2012; Finucane et al. 2015). Thus, it is critically important to identify and functionally validate enhancer elements and causal variants therein that interfere with enhancer function, thereby conferring disease risk through dysregulation of gene expression. Biological insights underlying disease-associated sequence variants can then be utilized to improve clinical outcomes, including developing effective strategies for disease prevention and/or therapeutics.

In Fig. 1, we illustrate various CRISPR/Cas9-based strategies to functionally validate putative enhancers predicted from the ChIP-seq analysis for transcription factor binding and histone mark enrichment and the genome architecture mapping (Beagrie et al. 2017). These include enhancer deletion (Fig. 1a, b), mutagenesis of transcription factor-binding sites (Fig. 1c), enhancer replacement (Fig. 1d), or duplication (Fig. 1e) of enhancers in their native genomic context. For example, CRISPR/Cas9-mediated deletion of a superenhancer reduced the expression of cancer-related genes and impaired some oncogenic properties (Zhang et al. 2016). In contrast, duplication of super-enhancers led to the overexpression of a key oncogenic transcription factor, which then activated other cancer-related genes in squamous cell carcinomas (Zhang et al. 2017). In another example, deletion of super-enhancer constituents at the Nanog (Blinka et al. 2016) or Sox2 (Li et al. 2014) locus in mouse embryonic stem cells reduced corresponding target gene expression and impaired some other key pluripotency genes, resulting in cellular differentiation. A more recent study revealed that CRISPR/Cas9-mediated deletion of super-enhancer constituents impacted cell-specific microRNA biogenesis, processing, and networks, uncovering a novel role of superenhancers in RNA processing beyond transcription (Suzuki et al. 2017). Another CRISPR-based strategy, unbiased mutagenesis, enabled the identification of enhancer elements involved in regulating genes associated with cancer drug resistance (Sanjana et al. 2016). In this approach, a high-throughput CRISPR screen was performed using ~ 18,000 single guide RNAs (sgRNAs) targeting non-coding regions to induce mutations in the regulatory regions of genes that were involved in cancer drug resistance (Sanjana et al. 2016). The authors found that engineered mutations at these non-coding genomic loci altered gene regulation and cancer drug resistance due to changes in transcription factor occupancy and epigenetic environments (Sanjana et al. 2016).

Fig. 1.

CRISPR/Cas9-based genome-editing strategies for studying regulatory genomic elements and chromatin structure. a Deletion of an enhancer. When two CRISPR sgRNAs with Cas9 are used to cleave in cis, deletion of the intervening DNA sequence is a common outcome. Removal of the enhancer by cleavage activity of two Cas9 nucleases abolishes the physical interaction between the distal enhancer and proximal promoter of a gene, resulting in decreased gene expression. b Deletion of super-enhancer constituents. Partial deletion of the super-enhancer region by cleavage activity of two Cas9 nucleases compromise the super-enhancer activity, resulting in decreased gene expression. c Mutagenesis of transcription factor-binding sites within an enhancer. Upon CRISPR/Cas9-induced single double-strand break, the restoration of 2 DNA ends by non-homologous end joining repair gives rise to small indels in the target site and results in mutagenesis. CRISPR/Cas9-mediated mutagenesis of transcription factor-binding sites within the enhancer inhibits the enhancer–promoter interaction, resulting in decreased gene expression. d Mutagenesis of an enhancer. Replacement of a wild-type enhancer with the mutagenic enhancer via CRISPR/Cas9-mediated homologous recombination, resulting in de novo enhancer–promoter interactions and changes in gene-expression pattern. e Duplication of super-enhancer constituents. DNA duplication can be generated by the introduction of donor DNA, with two homology arms on each side flanking the desired insertion, to induce repair of the CRISPR/Cas9-mediated double-strand break by homology-directed repair. Addition of the constituent super-enhancer could promote the superenhancer activity. f Topological chromatin domain disruption. Disruption of a boundary between two topologically associating domains will rewire the long-range regulatory architecture and gene–enhancer interactions. g Deletion of an insulator. Removal of the insulator by cleavage activity of two Cas9 nucleases enables the physical interaction between the distal enhancer and proximal promoter of a gene, resulting in increased gene expression. h Inverse orientation of CTCF-binding sites. CRISPR/Cas9-mediated replacement of a CTCF-binding site with reverse-oriented sequences will reconfigure the topology of chromatin loops between the distal enhancer and target promoters and alter gene-expression patterns. CTCF CCCTC-binding factor, SNP single nucleotide polymorphism, TAD topologically associating domain

In addition to the CRISPR/Cas9-based genome-editing system, CRISPR/dCas9 interference approaches could be used to characterize the enhancer and super-enhancer functions. The nuclease-dead Cas9 (dCas9) was derived by introducing D10A and H840A mutations on the SpCas9 (Qi et al. 2013). Though dCas9 is lacking endonuclease activity and is unable to cleave the DNA, the dCas9 can specifically interfere with transcriptional elongation, RNA polymerase binding, or transcription factor binding upon binding to the target promoter and gene body (Qi et al. 2013). Fusing dCas9 to a transcriptional repression or activation domain enabled more efficient modulation of gene expression and distal regulatory elements (Maeder et al. 2013; Thakore et al. 2015). As illustrated in Fig. 2, these include inhibition of enhancer–promoter interactions (Fig. 2a), repression of gene expression using dCas9-KRAB transcriptional repressors (Fig. 2b), or activation of gene expression by dCas9-VP64 transcriptional activators (Fig. 2c). For instance, a high-throughput CRISPR interference-based approach enabled a systematic mapping of mammalian enhancer–promoter interactions (Fulco et al. 2016). Using dCas9-KRAB-based repression under a dox-inducible promoter, multiple distal enhancers with complex enhancer–promoter connections were shown to regulate GATA1 and MYC gene expressions and cellular proliferation in erythroleukemia cells (Fulco et al. 2016). In another study, dCas9-VP64-based activation approach together with genome-wide sequencing data (DNase-seq, RNA-seq, and ChIP-seq) revealed that developmental regulation of chromatin accessibility and activation at enhancer elements underlined neuronal differentiation and function in the developing cerebellum (Frank et al. 2015). By recruiting dCas9-VP64 activators to each of the two putative enhancers flanking Grin2c in cultured cerebellar granule neurons of the developing mouse, an upregulation of Grin2c expression was observed, confirming the enhancer activity of late-opening DNase I hypersensitive sites nearby Grin2c (Frank et al. 2015). The dCas9-VP64-based activation approach has also been recently used to identify stimulation-responsive enhancers in unstimulated Jurkat T cells (Simeonov et al. 2017). By recruiting a dCas9-VP64 activator to sites across large genomic regions surrounding two key autoimmunity risk loci, CD69 and IL2RA, multiple stimulus-responsive enhancers and a functional IL2RA enhancer that harbors an autoimmunity non-coding risk variant were characterized (Simeonov et al. 2017). Upon stimulated by dCas9-VP64 activators, the wild-type IL2RA enhancer strongly increased the levels of IL2RA transcript. When the minor variant of this SNP was introduced to the IL2RA enhancer, this stimulation-dependent enhancer function diminished (Simeonov et al. 2017). To support this notion, Cas9-based genome editing has also been used to generate mice with the human autoimmune-associated SNP knocked-in or with a 12 bp deletion at this IL2RA enhancer (Simeonov et al. 2017). Accordingly, the naive T cells isolated from both enhancer-edited mice had lower IL2RA surface expression compared to wild-type mice 24 h after activation with anti-CD3/CD28 antibodies (Simeonov et al. 2017). Taken together, both CRISPR/Cas9 and CRISPR/dCas9 are attractive molecular biology toolboxes to provide better insights into gene–enhancer connectivity in mammalian nuclei and to dissect the contribution of enhancers and super-enhancers to human disease. Table 1 lists recent studies highlighting the successes of CRISPR-based strategies for functional validation of enhancer and super-enhancer elements and their target genes.

Fig. 2.

CRISPR/dCas9-based epigenome-editing strategies for studying regulatory genomic elements and chromatin dynamics. a Block transcription factor-binding sites on an enhancer. The binding of dCas9s to the transcription factor-binding sites causes a steric block that interferes with the binding of transcription factors and chromatin loop formation between promoter and enhancer, resulting in decreased target gene expression. b Repress enhancer activity. dCas9-KRAB can repress target gene expression by preventing the transcriptional complex formation, inducing H3K9 trimethylation at the enhancer, and recruiting other histone repressive modifiers. c Promote enhancer activity. dCas9-VP64 can promote transcriptional activation by recruiting multiple components of the transcription pre-initiation complex, as well as histone acetyltransferases to promote acetylation of histone H3K27 at targeted enhancers. d CLOuD9-mediated chromatin loop reorganization. Addition of abscisic acid brings two complementary dSpCas9-PYL1 and dSaCas9-ABI1 into proximity by dimerization of PYL1 and ABI1 domains, and lead to chromatin structure remodeling. Removal of abscisic acid restores the endogenous chromatin conformation. e Chromatin opening. Two SpdCas9s bind the proximal target sites of FnCas9 to induce the chromatin opening and to enable the FnCas9 nuclease to access and cleave the otherwise inaccessible target sites. f Chromatin dynamics tracking. CRISPRainbow and multicolor CRISPR utilize different fluorescent-tagged dCas9-sgRNAs (red, green, and blue) to allow multiplexed labeling of chromatin loci for tracking chromatin dynamics in a spatial and temporal manner, visualizing endogenous genomic loci in live cells. Co-localization of red, green, and blue fluorescents indicates a close physical interaction between TADs from different chromosomes. g Chromatin loop visualization. In the case of studying cis/trans-acting regulatory elements within a TAD, CRISPRainbow and multicolor CRISPR allow tracking the dynamic interaction of enhancers or silencers with a specific promoter for transcriptional regulation. When the dCas9-blue-labeled enhancer is in close contact with the dCas9-red-labeled promoter, gene A is activated and pink fluorescent can be visualized. Similarly, when the dCas9-green-labeled silencer is in close contact with the dCas9-red-labeled promoter, gene B is activated and yellow fluorescent can be visualized. h Telomere dynamics tracking. The combined use of dCas9-eGFP imaging and DNA sequential FISH probes enables multiplexed visualization of telomere dynamics and subtelomeric regions. i Chromosome painting. dCas9-eGFP or dCas9-mCherry with hundreds of different sgRNAs can be used to paint an entire chromosome in order to visualize a desired chromosomal territory, spatial arrangements of sister chromatids, and to track the movement of a particular chromosome in dividing live cells. FISH fluorescence in situ hybridization, KRAB Krüppel-associated box, SNP single nucleotide polymorphism, TAD topologically associated domain, TSS transcription initiation site, VP64 multiple repeats of the herpes simplex VP16 transactivation domain. (Color figure online)

Table 1.

CRISPR-based strategies for studying mammalian regulatory genomic elements and chromatin structure

Regulatory element	Associated gene or locus	CRISPR strategy	Biological or phenotypic impact	Tissue origin (cell line)	Reference
Enhancer	BCL11A	In situ saturating mutagenesis	Fetal hemoglobin level reinduction cells to ameliorate the β-hemoglobin disorders	Primary human erythroid progenitors and mouse transgenesis	Canver et al. (2015)
	Vicinity to MYC and GATA1	dCas9-KRAB interference under a dox-inducible promoter	Reduced GATA1 and its downstream gene expression, reduced MYC expression and affected cellular proliferation	Human erythroleukemia (K562)	Fulco et al. (2016)
	Sox2	Deleted enhancer	Reduced Sox2 expression, loss of pluripotency and differentiation potential	Mouse embryonic (ESCs)	Zhou et al. (2014)
	Prc2	Deleted poised enhancers	Impaired the induction of major anterior neural regulators upon mESC differentiation	Mouse embryonic (ESCs)	Cruz-Molina et al. (2017)
	Mmp13	Deleted enhancer	Altered transcription factor occupancy and expression for Mmp13	Mouse osteoblast (UAMS-PB)	Meyer et al. (2015)
	CDKN1A, CCND1	Targeted Cas9 to p53- and ERα-bound enhancers	Reduced CDKN1A, CCND1, and their eRNA expressions	Human fibroblast (BJ), breast (MCF-7, T47D)	Korkmaz et al. (2016)
	surrounding the NF1, NF2, and CUL3	Unbiased mutagenesis	Altered transcription factor occupancy and long-range and local epigenetic environments, altered cancer drug resistance	Human melanoma (A375)	Sanjana et al. (2016)
	Cebpa	Deleted enhancers	Reduced Cebpa levels, defected in the granulocytic lineage, lose myeloid transcription program, blocked in progenitor differentiation, loss of hematopoietic stem cells maintenance	Mouse hematopoietic stem cells, myeloid-primed progenitor cells	Avellino et al. (2016)
	Cebpa	Replacement with mutated enhancer	Reduced Cebpa mRNA levels	Mouse myeloid (32Dcl3)	Cooper et al. (2015), Guo et al. (2016)
	TES/TESCO	Deleted enhancer	Reduced Sox9 expression levels in XY fetal gonads	Mouse embryos	Gonen et al. (2017)
	Stat5	Deleted enhancer	Reduced Stat5 and Stat5-associated genes level in mammary epithelium	Mouse fertilized eggs	Metser et al. (2016)
	ERV-9 L retrotransposon	Deleted LTR	Suppressed transcription of the globin genes	Human erythroid progenitor and erythroleukemia (K562)	Hu et al. (2017)
	POU5F1	Deleted regulatory elements	Temporary loss of POU5F1 transcription	Human embryonic (ESCs)	Diao et al. (2016)
	genome-wide	deleted promoter	A set of mammalian promoters involved in the cis regulation of the expression of distal genes	Mouse fibroblast (P5424 T, NIH-3T3)	Dao et al. (2017)
	Grin2c	dCas9-VP64-based activation	Confirmed enhancer activity of late-opening DNase I hypersensitive sites nearby Grin2c	Mouse cerebellar granule neurons	Frank et al. (2015)
	Hb9, Isl1	Deleted enhancer	Stage-specific decreased in Hb9 and Isl1 expression during motor neuron maturation	Mouse embryonic (ESCs)	Rhee et al. (2016)
	Enhancer L	Deleted enhancer	Ablation of HLA-G expression at the maternal–fetal interface	Human placenta (JEG3 and primary trophoblasts)	Ferreira et al. (2016)
	Pcdh clusters	Inversions and duplications	Altered the expressions of Pcdh α and its isoforms and members of the Pcdh β and γ clusters	Human cell lines (HEC-1-B and HEK293T), mouse embryos	Li et al. (2015)
	Ifitm	Truncated enhancer	Impaired IFN-induced resistance to influenza A virus infection	Mouse embryos	Li et al. (2017b)
	CD69 and IL2RA	Activated enhancer	Identified stimulus-responsive immune enhancers that harbors an autoimmunity risk variant	Human immune cells (Jurkat T)	Simeonov et al. (2017)
	Alpha-globin	Deleted the MCS-R2 alpha-globin enhancer	Reduced alpha-globin expression and corrected the pathologic globin chain imbalance associated with beta-thalassemia	Primary human hematopoietic stem cells	Mettananda et al. (2017)
	ATP2B4	Deleted enhancer	Had abnormally high intracellular calcium levels, affected red blood cell hydration and malaria susceptibility	Human erythroid precursor cell (HUDEP-2)	Lessard et al. (2017)
	Meis1	Deleted intragenic enhancer	De-repressed host gene expression by preventing attenuation of host gene transcription during productive elongation, impacted pluripotency and induced cellular differentiation	Mouse embryonic (ESCs)	Cinghu et al. (2017)
	ERBS	Multiplex enhancer interference	Revealed collaborative control of gene expression by estrogen receptor alpha-bound enhancers	Human estrogen-responsive cell Ishikawa (endometrial)	Carleton et al. (2017)
	MYC, GATA1, CD69	CRISPR interference and activation	Identified target genes of disease-associated non-proteincoding DNA elements	Human K562, My-La, GM12878, Jurkat cells	Mumbach et al. (2017)
	Zbtb16	Deleted Zbtb16 enhancer	Specific RUNX1-bound Zbtb16 enhancer elements critically modulated lymphoid lineage-dependent expressions of Plzf	Mouse zygotes	Mao et al. (2017)
	EPHB3	mutated FOX-binding motifs	Abolished FOXA1 enhancer occupancy and EPHB3 expression, reduced H3K27ac levels, and caused complete chromatin condensation at the EPHB3 promoter and enhancer	Human colorectal cancer cells (LS174T)	Jagle et al. (2017)
Super-enhancer	miR-290-295, miR-1, miR-148a	Deleted miRNA superenhancer constituents	Decreased mature miR-290-295 production and suppressed expression of Ago2-inducible gene Lin28a, suppressed induction of miR-1 and myogenic differentiation markers, suppressed miR-148a induction during Pro-B differentiation	Mouse embryonic (ESCs), myotubes (MyoD), progenitor (Pro-B)	Suzuki et al. (2017)
	Nanog	Deleted super-enhancer constituents	Different effects on Nanog and neighbor gene expression	Mouse embryonic (ESCs)	Blinka et al. (2016)
	Sox2	Deleted entire super-enhancer	Reduced expressions of Sox2 and a number of key pluripotency genes and induced a differentiation gene signature	Mouse embryonic (ESCs)	Li et al. (2014)
	Dppa5a, Ooep	Deleted super-enhancer constituents	Reduced target gene expressions	Mouse embryonic (ESCs)	Moorthy et al. (2017)
	Slc25a37	Deleted super-enhancer constituents	Each super-enhancer constituent has varying effects on Slc25a37 expression	Mouse erythroid (G1E/G1ER)	Huang et al. (2016)
	MYC-LASE	repressed or deleted constituent enhancer	Reduced MYC and its target gene expression, impaired anchorage-independent and clonogenic growth	Human lung (NCI-H2009)	Zhang et al. (2016)
	KLF5- KLF12	Duplication of super-enhancers	Activated KLF5 oncogenic transcription factor	Human squamous cell carcinomas	Zhang et al. (2017)
	Fgf2, Egr2, Tjpg1, and Fst	Deleted super-enhancers	Blocked AngII induction of its nearest gene	Mouse vascular smooth muscle cells	Das et al. (2017)
	Il2ra	Mutated STAT5-binding sites	Altered IL-2-induced Il2ra gene expression	Mouse	(Li et al. 2017a)
Insulator	Within the PCDH enhancer	Inverse orientation of CTCF-binding sites	Reconfigured the topology of chromatin loops between the distal enhancer and target promoters and altered gene-expression patterns	Human uterus or endometrium (HEC-1B)	Guo et al. (2015)
	Spanning WNT6/IHH/EPHA4/PAX3	Disrupted topological chromatin domains	Rewired long-range regulatory architecture, gene–enhancer interactions, and resulted in pathogenic phenotypes of human limb malformations	Mouse limb tissue and patient-derived fibroblasts (primary cell)	Lupianez et al. (2015)
	At a domain boundary of PDGFRA	Disrupted a CTCF motif	Loss of insulation between topological domains, lead to aberrant enhancer–oncogene interaction, induced oncogene expression and promoted gliomagenesis	Human IDH gliomaspheres (primary cell)	Flavahan et al. (2016)
	HA6	Disrupted a CTCF site	Altered the initial associations among enhancers, promoters and insulators, influenced retinoic acid-induced HOXA expression, increased the interaction between RE4 and HOXA5	Human embryonal (NT2/D1)	Ishihara et al. (2016)
	Tyr	Deleted insulator	Decreased Tyr expression and coat pigmentation	Mouse fertilized eggs	Seruggia et al. (2015)
	Csn1s1	Deleted border CTCF site located between enhancer and neighboring genes	Activated Sult1d1 but silenced Sult1e1 gene, de novo interactions between the Sult1d1 promoter and several enhancers in the casein locus	Mouse fertilized eggs	Lee et al. (2017a)
	Bcl11a	Mutagenesis CTCF site	loss of BRD2 occupancy at CTCF co-bound sites, enabled regulatory influence to spread from one gene to another following the disruption in CTCF/BRD2-mediated transcriptional boundary	Mouse erythroblast (G1E-ER4)	Hsu et al. (2017)
	Zfp608, Sox4	Induced chromatin insulation	dCas9-mediated activation was insufficient for creating TAD boundaries de novo	Mouse embryonic stem cell (E14tg2a)	Bonev et al. (2017)
Chromatin loop	Genome-wide	Manipulated CTCF motifs at loop anchors	Chromatin loop domains disappeared	Human haploid cell (HAP1)	Sanborn et al. (2015)
	Malt1, Fbn2, Sox2	Deleted core CTCF-binding sites, re-insertion of oppositely oriented CTCF sequences	Loss CTCF and cohesin recruitment, disrupted or destabilized chromatin loops, could not re-establish chromatin loops with opposite CTCF-binding polarity	Mouse embryonic (ESCs)	de Wit et al. (2015)
	Car2	Deleted enhancer	Compromised Car2 transcription, abolished the interaction and looping between LDB1-bound enhancer and CTCF-bound Car2 promoter	Erythroid (mouse MEL, human K562)	Lee et al. (2017b)
	CFTR	Deleted insulator or enhancer	Removal of an upstream CTCF-binding insulator altered the interaction profile, deletion of a pivotal enhancer element decreased CFTR expression	Human colorectal (Caco2)	Yang et al. (2016)
	NGB	Deleted distal regulatory elements	Reduced NGB expression	Human neuroblastoma (SH-SY5Y)	Tam et al. (2017)
	Beta-globin	Deleted GATA1 site on promoter	Eliminated LDB1-mediated enhancer looping and rescued by LDB1 and CRISPR/Cas9-mediated forced enhancer looping	Mouse erythroid (MEL)	Krivega and Dean (2017)
	5′HS5, beta-globin	dCas9 enChIP-Seq	Identified physical interactions between genomic regions upon erythroid differentiation	Human erythroleukemia (K562)	Fujita et al. (2017)
	Beta-globin, POU5F1	dCas9 CLOuD9-mediated chromatin loop reorganization	Reversibly manipulated chromatin contacts and induced corresponding alterations in gene expression	Human erythroleukemia (K562), kidney (HEK293T)	Morgan et al. (2017)
	Alpha-globin gene cluster	Mutagenesis to remove a conserved upstream CTCF–cohesin boundary	Enlarged the topologically associated sub-domains (sub-TADs), allowed alpha-globin enhancers to interact more distance promoters, upregulated expression of genes within the extended sub-TAD	Mouse in vivo	Hanssen et al. (2017)
	MLL4	deleted an upstream MLL4-binding site	Diminished FOXP3 induction, compromised Treg cell differentiation	Human regulatory T cells (Treg cells)	Placek et al. (2017)
	IrxA cluster	Deleted an intergenic CTCF site within the IrxA cluster	Altered gene expression in the developing heart, affected cardiomyocyte development and maturation programs	Mouse embryonic heart	Gomez-Velazquez et al. (2017)
	Thy1, Cd5, Runx3	Deleted CTCF-binding sites	Increased cell-to-cell variation of gene expression, compromised enhancer–promoter interactions	Mouse T cell (EL4)	(Ren et al. 2017)
Chromatin dynamics	Genome-wide	Single fluorescently labeled dCas9-HaloTag under an inducible TRE tight promoter	Visualized diffusion and chromatin binding of dCas9 for tracking dCas9 dynamics in living cells, slower movement of dCas9 within heterochromatin, off-target binding events of dCas9 were short-lived	Mouse fibroblast (NIH 3T3)	Knight et al. (2015)
	Genome-wide	Multicolor fluorescently labeled dCas9	Determined the intranuclear distance between loci on different chromosomes, assessment of the DNA compaction of a specific chromosome region	Human bone (U2OS), retina (RPE-1)	Ma et al. (2015)
	Chromosome 9	dCas9-eGFP with hundreds of different sgRNAs to paint an entire chromosome	Visualized a desired chromosomal territory, spatial arrangements of sister chromatids and tracked the movement of a particular chromosome in dividing cells	Human cervix (HeLa)	Zhou et al. (2017)
	Genome-wide	Dual-color chromosomal loci imaging with modified sgRNA	Labeled and tracked telomeres, centromeres and genomic loci	Human breast (MDA-MB-231)	Shao et al. (2016)
	Chromosome 12	Dual-color labeling with fluorescently tagged dCas9	Live imaging of satellite sequences and repeat-enriched individual loci, tracked dynamics of chromatin interactions	Mouse embryonic (ESCs), fibroblast (3T3)	Fu et al. (2016)
	genome-wide	sgRNAs integrated with MS2-binding motifs	Enabled robust fluorescent signal amplification and multicolor labeling of low-repeat chromosomal regions, tracked native chromatin loci and transcriptional activities throughout the cell cycle	Human osteosarcoma (U2OS), retina (RPE) and cervix (HeLa)	Qin et al. (2017)
	Chromatin regulation-associated genes	A combinatorial dCas9-KRAB	Mapped genetic interactions and revealed a functional map of chromatin regulation	Human kidney (HEK293)	Du et al. (2017)
	Genome-wide	Combined dCas9-eGFP and CasID	Identified chromatin factors associated with repetitive telomeric and satellite DNA, determined locus-specific chromatin composition	Mouse muscle (C2C12)	Schmidtmann et al. (2016)
	genome-wide	dCas9-induced chromatin opening	Induced chromatin accessibility at previously inaccessible genomic loci by altering local transcription factor binding	Mouse embryonic (ESCs)	Barkal et al. (2016)
Telomere dynamics	Genome-wide	eGFP-tagged dCas9 with an inducible promoter Tet-on 3G	Robust imaging of repetitive elements in telomeres and coding genes, visualization of non-repetitive genomic sequences, tracked telomere dynamics and mitosis events in living cells	Human retina (RPE), urinary bladder (UMUC3), cervix (HeLa)	Chen et al. (2013)
	TERT	Inserted FLAG-HaloTag at telomere, labeled telomerase with HaloTag-substrate	Telomerase trafficking and recruitment to telomeres via three-dimensional diffusion	Human cervix (HeLa)	Schmidt et al. (2016)
	TERT	Enriched cells that underwent homologous recombination	Efficient immunopurification and tracked TERT proteins subcellular localization, mutated TERT promoter decreased telomerase activity and shorten telomeres	Human kidney (HEK293T), cervix (HeLa), urothelial (SCaBER)	Xi et al. (2015)
	telomere	Combined CRISPR imaging and DNA sequential FISH	Multiplexed visualization of telomere dynamics, and identified 12 unique subtelomeric regions	Mouse embryonic (ESCs)	Takei et al. (2017)
	Telomere	CRISPR/Cas9 RNA-directed nickase system	Enabled high-throughput single-molecule telomere characterization	Human diploid fibroblasts	McCaffrey et al. (2017)
Unmarked regulatory elements	Genome-wide	Multiplexed editing regulatory assay (MERA)	Controlled Tdgf1 expression but do not have typical enhancer epigenetic or chromatin features	Mouse embryonic (ESCs)	Rajagopal et al. (2016)

AngII angiotensinogen; ATP2B4 ATPase plasma membrane Ca2 + transporting 4, BCL11A B-cell CLL/lymphoma 11A, BRD2 bromodomain containing 2, Car2 carbonic anhydrase 2, CasID promiscuous biotin ligase BirA, CCND1 cyclin D1, CDKN1A cyclin dependent kinase inhibitor 1A, CEBPA CCAAT/enhancer-binding protein alpha, CFTR cystic fibrosis transmembrane conductance regulator, Csn1s1 casein alpha s1, CTCF CCCTC-binding factor, CUL3 cullin 3, Dppa5a developmental pluripotency associated 5A, Egr2 early growth response 2, EPHA4 EPH receptor A4, EPHB3 EPH receptor B3, ERBS estrogen receptor alpha-binding site, Fbn2 fibrillin 2, Fgf2 fibroblast growth factor 2, FISH fluorescence in situ hybridization, FOXA1 forkhead box A1, Foxp3 forkhead box P3, Fst follistatin, GATA1 GATA-binding protein 1, Grin2c glutamate ionotropic receptor NMDA-type subunit 2C, Hb9 motor neuron and pancreas homeobox 1, HOXA homeobox A, IDH isocitrate dehydrogenase, IFITM interferon-induced transmembrane protein, IFN interferon, IHH indian hedgehog, IL2RA interleukin 2 receptor subunit alpha, Isl1 ISL1 transcription factor LIM/homeodomain, KLF5 Krüppel-like factor 5, KRAB Krüppel-associated box, LDB1 LIM domain-binding 1, Lin28a lin-28 homolog A, LTR long terminal repeat, Malt1 MALT1 paracaspase, Meis1 Meis homeobox 1, MLL4 lysine methyltransferase, Mmp13 matrix metallopeptidase 13, MYC MYC proto-oncogene, NF1 neurofibromin 1, NF2 neurofibromin 2, NGB neuroglobin, Ooep oocyte-expressed protein, PAX3 paired box 3, Pcdh protocadherin, PDGFRA platelet-derived growth factor receptor alpha, PLZF promyelocytic leukemia zinc finger, POU5F1 POU class 5 homeobox 1, Runx1 runt-related transcription factor 1, Runx3 runt-related transcription factor 3, SLC25A37 solute carrier family 25 member 37, Sox2 SRY (sex-determining region Y)-box 2, Stat5 signal transducer and activator of transcription 5, Sult1d1 sulfotransferase family 1D, Tdgf1 teratocarcinoma-derived growth factor 1, TERT telomerase reverse transcriptase, TES testis-specific enhancer of Sox9, TESCO TES core element, Thy1 thymus cell antigen 1 theta, Tyr tyrosinase, WNT6 Wnt family member 6, Zbtb16 zinc finger and BTB domain containing 16, Zfp608 zinc finger protein 608

Beyond the linear genomic sequences, three-dimensional nuclear architecture and higher-order chromatin organization are increasingly recognized for their critical role in shaping functional genome and controlling gene expression (Sexton and Cavalli 2015). Individual chromosomes tend to occupy small portions of the nucleus, called chromosome territories, minimizing interactions between chromosomes, while individual chromosomes are partitioned into megabase-sized topologically associating domains (TADs), regions with relatively high intradomain DNA interactions (Sexton and Cavalli 2015). The borders of TADs are frequently marked by the binding of the insulator protein CTCF (Ghirlando and Felsenfeld 2016). The highly organized hierarchical structures are built up from the stabilization of progressively higher-order chromatin conformations that are characterized by numerous local and long-range contacts between genes and functional regulatory elements (Sexton and Cavalli 2015). Recent evidence suggests that the interactions between transcription regulatory elements generally occur within chromosomal loop structures, referred to as insulated neighborhoods, formed by the interaction of two CTCF molecules bound to different sites and reinforced by a cohesin molecule (Hnisz et al. 2016; Rao et al. 2017) . Enhancer-bound proteins tend to interact with genes within this CTCF–CTCF loops, thus functioning to insulate enhancers and genes within the loop from enhancers and genes outside the loop. Each chromosome is estimated to contain thousands of chromatin loop structures, which form the mechanistic basis of higher-order chromosome structures, such as TADs (Ong and Corces 2014). Understanding the topological configurations of chromatin will reveal valuable insights into how distal regulatory elements such as enhancers affect promoters, and how insulators can abrogate these effects in control of gene expression. The CRISPR/Cas9 and CRISPR/dCas9 systems have proven to be useful tools to study chromatin structure and dynamics.

CRISPR/Cas9-mediated deletion of a TAD boundary or underlying CTCF site can lead to loss of physical insulation and a subsequent fusion of the two adjacent TADs into a single domain, leading to de novo enhancer–promoter interactions and misexpression (Fig. 1f). Similarly, deletion of distal regulatory elements or mutagenesis of contact regions of enhancer–promoter could be adopted. For example, removal or mutagenesis of core CTCF-binding sites on cis-regulatory elements disrupted or destabilized chromatin loop domains, respectively, due to the loss of CTCF and cohesin recruitment, leading to alterations in enhancer–promoter interactions and gene-expression profiles (de Wit et al. 2015; Sanborn et al. 2015). A recent in vivo study found that disruption of a conserved CTCF–cohesin boundary extended the sub-TAD of the mouse alpha-globin gene cluster to adjacent CTCF–cohesin-binding sites (Hanssen et al. 2017). This in turn allowed alpha-globin enhancers to interact with more distance promoters and upregulated the expression of additional genes within the extended sub-TAD. In addition, a very recent study showed that multiple CTCF-binding sites reduced gene-expression noise by stabilizing promoter–enhancer interactions and decreasing the fluctuation of the promoter activity (Ren et al. 2017). Likewise, a number of recent studies have successfully utilized CRISPR technology to identify and characterize human disease-associated insulator elements (Table 1), which include insulator deletion (Fig. 1g), disruption or inverse orientation of CTCF-binding sites (Fig. 1h). For example, CRISPR-mediated disruption of the CTCF-binding motif was able to promote human gliomagenesis by inducing a loss of insulation between topological domains and allowing aberrant gene–enhancer interactions that trigger glioma oncogene expressions (Flavahan et al. 2016). In another study, CRISPR was used to introduce structural variations associated with human limb malformations in mice to recapitulate the pathogenic phenotypes observed in human patients (Lupianez et al. 2015). In both cases of human and mouse, these structural variations affected CTCF-associated boundary elements by disrupting topological chromatin domains and leading to de novo gene–enhancer interactions and causing gene misexpression (Lupianez et al. 2015). Another excellent study has shown that CRISPR-mediated inversion of CTCF-binding sites within the protocadherin (Pcdh) enhancer reconfigured the topology of chromatin loops between the distal enhancer and target promoters, which in turn led to changes in gene-expression patterns (Guo et al. 2015). These three studies clearly demonstrated the important roles of insulator elements in orchestrating proper long-range regulatory interactions, chromatin looping, and gene regulation.

Recently, a robust CRISPR-dCas9 (CLOuD9) method has been developed to manipulate nuclear architecture and gene expression through targeted and reversible chromatin loop reorganization (Morgan et al. 2017). As illustrated in Fig. 2d, two complementary CLOuD9 constructs consisting of dSpCas9-PYL1 and dSaCas9-ABI1 formed a chromatin loop through dimerization of PYL1 and ABI1 domains upon addition of abscisic acid. Removal of abscisic acid could restore the native endogenous chromatin conformation. This CLOuD9 technique was successfully used to induce new chromatin looping formation between the human β-globin promoter and the HS2 region of the locus control region, without destroying the endogenous chromatin contact of these two regions (Morgan et al. 2017). This CLOuD9 approach was also successfully used to promote chromatin contacts between the Oct4 promoter and a distal enhancer, which in turn lead to dramatic increase in Oct4 gene expression (Morgan et al. 2017). The applications of CLOuD9 approach in these two loci clearly show that chromatin looping reorganization alone is sufficient to alter target gene expression (Morgan et al. 2017). The transcriptional misregulation due to disruptions in the spatial organization of genome architecture and aberrant regulatory interactions have been implicated in many human diseases (Kaiser and Semple 2017; Lupianez et al. 2016). Therefore, the CLOuD9 approach described here could be used for targeted epigenetic-oriented therapy by restoring normal epigenetic patterning, chromatin structure, and nuclear organization, thus correcting gene misexpression. For instance, in human IDH mutant gliomas, a mutation on CTCF-binding sites induced a loss of insulation between topological domains and led to aberrant oncogene activation by permitting a constitutive enhancer to aberrantly interact with a prominent glioma oncogene (Flavahan et al. 2016). Without the need to genetically modify the mutated DNA sequence on CTCF-binding sites, the CLOuD9 approach could be used to downregulate oncogene expression and inhibit gliomagenesis by inducing de novo chromatin loop formation and restoring the insulation between topological domains.

Similar to CLOuD9, a more recent study has developed another programmable DNA looping method using engineered bivalent dCas9-Zip complexes (Hao et al. 2017). In this new approach, two orthogonal dCas9 moieties consisting of dSpCas9-Zip and dStCas9-Zip heterodimerized and formed a DNA loop upon the binding of the dCas9-sgRNA complexes to two DNA target sites (Hao et al. 2017). Zip, a synthetic leucine zipper tag was used to destabilize homodimer interactions and maximize inter-helical interactions between dSpCas9-Zip and dStCas9-Zip (Hao et al. 2017). These bivalent dCas9-Zip complexes have been successfully used to improve enhancer–promoter contacts and to activate the endogenous gene expressions in bacteria by rewiring chromosomal DNA to bring distal enhancer site closer to the promoter (Hao et al. 2017). The DNA looping efficiency was further improved by creating multiple loops using multiple guide RNAs (Hao et al. 2017).

CRISPR-mediated gene editing efficiency is known to be affected by chromatin accessibility, with higher efficiency in euchromatin as compared to heterochromatin (Chen et al. 2016; Jensen et al. 2017). CLOuD9 and proximal CRISPR targeting have provided promising strategies to reconfigure local chromatin structure and to induce de novo chromatin loop formation even at heterochromatin. For example, the CLOuD9 has the ability to artificially induce chromatin loop formation in heterochromatin (Morgan et al. 2017), despite the challenge in assessing the target sites at the heterochromatin, as demonstrated with high-specificity Cas9 variants, including SpCas9-HF1 and eSpCas9 (1.1) (Chen et al. 2017b, 2016; Jensen et al. 2017). Moreover, a recently developed proximal CRISPR targeting method enabled the FnCas9 nuclease to access and cleave the otherwise inaccessible target sites (Chen et al. 2017a). In this approach, as illustrated in Fig. 2e, the catalytically dead SpCas9 (SpdCas9) was used to bind the proximal target sites to induce the reconfiguring and opening of the local chromatin structure, which in turn improved the FnCas9 accessibility and functionality (Chen et al. 2017a). The similar strategy described here has also been successfully used along with other Cas nuclease variants such as CjCas9, NcCas9, and FnCpf1 to improve gene editing efficiency and accessibility on difficult-to-cleave genomic sites (Chen et al. 2017a).

In addition, various CRISPR/dCas9-based strategies have been used to track physical interactions between genomic regions, spatial chromatin mobility, and localization in the nucleus. For example, CRISPRainbow (Ma et al. 2016), multicolor CRISPR (Ma et al. 2015), and single fluorescently labeled dCas9-HaloTag (Knight et al. 2015) were used to determine genomic DNA co-localization, dynamics, and interlocus interactions in living cells in space and time. As illustrated in Fig. 2f, using the pairs of differently fluorescent-tagged dCas9-sgRNAs (red, green, and blue), the intranuclear distance between loci on different chromosomes and the DNA compaction of specific chromatin regions within a chromosome could be assessed. The CRISPR-based multiple color imaging also allowed tracking of chromatin loop formation following enhancer–promoter interactions within a TAD (Fig. 2g). The combined use of dCas9-eGFP imaging and DNA sequential FISH probes enabled multiplexed visualization and tracking of the dynamics of telomerase recruitment to telomere and subtelomeric regions (Schmidt et al. 2016; Takei et al. 2017) (Fig. 2h). Compared to tracking satellite sequences and repeat-enriched individual loci such as telomeric regions, it is more challenging to use fluorescent-tagged dCas9s for labeling and visualizing non-repetitive regions within a chromosome. However, the recently developed dCas9-based chromosome painting technique provides a solution to this issue (Zhou et al. 2017).

Chromosome painting could be achieved using fluorescent-tagged dCas9s with a large number of different sgRNAs to bind and paint an entire chromosome (Fig. 2i). An array of sgRNAs tiling along the target locus and the strong fluorescent signals generated following chromosome condensation of dCas9-binding chromatin regions allowed visualizing a desired chromosomal territory, spatial arrangements of sister chromatids, and tracking the movement of a particular chromosome in dividing cells (Chen et al. 2013; Zhou et al. 2017). To achieve a higher photostability and long-term multicolor labeling of chromosomal loci in live cells, a recently developed dual-color imaging system using a modified sgRNA consisted of RNA aptamer insertions that bind fluorescent protein-tagged effectors was adopted (Shao et al. 2016). To amplify the fluorescent signals generated on low- and non-repetitive chromosome loci upon dCas9 binding, a re-engineered sgRNA containing up to 16 MS2 motifs that bind to the fluorescently tagged bacteriophage coat protein MCPs was used (Qin et al. 2017). A dual-color live imaging system consisting of two chimeric transcripts of viral RNA stem-loop motifs (MS2 and PP7) has also been developed to recruit two different fluorescently tagged viral RNA-binding proteins (MCP and PCP) to the sgRNA target sites (Fu et al. 2016). For example, sgRNA-MS2 and sgRNA-PP7 chimeric transcripts have been successfully used to recruit multiple molecules of MCP-eGFP and PCP-mCherry for each sgRNA scaffold to co-labeling two individual loci or satellite regions (Fu et al. 2016). Sensitivity in fluorescence imaging of dCas9-sgRNA-binding chromatin regions was further improved using peptide-repeat-based amplification systems such as dCas9-SunTag (Tanenbaum et al. 2014) and its modified version (Morita et al. 2016) to recruit multiple copies of a fluorescently labeled antibody-fusion protein. The combined use of dCas9-SunTag and all-in-one multiplex sgRNA expressing system generated strongest fluorescent signals for robust imaging of a single non-repetitive genomic locus (Shao et al. 2017).

A CRISPR affinity purification in situ of regulatory elements (CAPTURE) approach is another innovative way to unbiasedly identify locus-specific chromatin-regulating protein complexes, long-range DNA interactions, disease-associated cis-regulatory elements, and regulatory composition-based hierarchical structure (Liu et al. 2017b). The CAPTURE method enabled high resolution, high-throughput, and selective capture of the locus-specific chromatin interactions in their native context (Liu et al. 2017b). The CAPTURE system comprised three components: a biotinylated dCas9, a biotin ligase BirA, and sequence-specific sgRNAs (Liu et al. 2017b). Upon in vivo biotinylation of dCas9 by the biotin ligase BirA at the DNA target site, high-affinity streptavidin purification is used to bind biotinylated dCas9 for isolating the genomic locus-associated macromolecules (Liu et al. 2017b). The purified protein, RNA, and DNA complexes can then be analyzed by mass spectrometry-based proteomics and high-throughput sequencing (Liu et al. 2017b).

CRISPR-based strategies for studying non-coding genetic variants associated with chromatin structure and regulatory genomic elements

The vast majority of genetic variants associated with human disease are enriched in non-coding genomic regulatory regions that can impact transcription factor bindings, chromatin folding, and states (Deplancke et al. 2016; Zhang and Lupski 2015). These results suggest that disease-associated variants impose risk by altering functional DNA elements that regulate gene expression. Since human genetic studies such as GWAS detect only statistical associations, not functional signals, identification of truly causal variants and elucidating how they lead to dysregulation of target gene expression remain a significant challenge in the post-GWAS era. CRISPR technology that allows efficient and scalable targeted genome editing has been indispensable for functional validation studies of the candidate non-coding causal variants in the endogenous genome. Table 2 lists recent studies highlighting the successes of CRISPR-based strategies for functional validation of disease-associated non-coding variants and their impact on target gene expression. For example, CRISPR/Cas9-mediated deletion of enhancer regions harboring disease risk variants has been successfully applied to dissect their functional roles on target gene expression (Bailey et al. 2016; Jin et al. 2016; Yao et al. 2014; Ye et al. 2016). This includes the CRISPR/Cas9-mediated deletion of a TCF7L2 intronic region that harbors a well-known causal variant for type 2 diabetes, leading to abrogation of chromatin contacts between the TCF7L2 intron and ACSL5 promoter that resides within the same TAD in human colon tissue (Xia et al. 2016). Moreover, the combined use of genome-wide epigenetic information and CRISPR technology has enabled the identification and validation of a candidate causal variant in a distal enhancer element associated with Parkinson’s disease that altered the binding of a brain-specific transcription factor and consequent target gene expression (Soldner et al. 2016). Deletion of an aorta-specific enhancer region containing a non-coding variant, rs9349379, has elucidated a distal regulatory role of this vascular disease risk variant on the altered expression of the endothelin 1 (EDN1) gene (Gupta et al. 2017). By employing a combined approach of expression quantitative trait loci (eQTL), circular chromatin conformation capture (4C), and genome editing, it was found that the distal IRX3 and IRX5 genes were more plausible target genes of the obesity-associated variants in the FTO locus (Claussnitzer et al. 2015; Smemo et al. 2014). These two homeobox transcription factors are located 0.5 and 1 million bp away, respectively, from the GWAS signal. It has been demonstrated that a functional enhancer variant in the FTO GWAS locus (located in FTO intron 1) disrupted binding of a transcriptional repressor (ARID5B) in a mesenchymal preadipocyte-specific enhancer, resulting in upregulation of both IRX3 and IRX5. Subsequently, the IRX3 and IRX5 overexpression shifted cell fate of adipocyte precursor towards white adipocyte, repressed mitochondrial thermogenesis, and increased lipid storage. CRISPR/Cas9-mediated correction of the causal variant (rs1421085) in the ARID5B repressor-binding site restored its binding and the normal expression levels of IRX3 and IRX5. This in turn activated browning expression programs and restored mitochondrial thermogenesis in primary adipocytes derived from an obesity patient, demonstrating that CRISPR technology can be successfully used for therapeutic benefits (Claussnitzer et al. 2015).

Table 2.

CRISPR-based strategies for studying non-coding genetic variants associated with chromatin structure and regulatory genomic elements

Regulatory element	Non-coding SNP	Associated gene or locus	CRISPR strategy	Associated human disease	Mechanistic insight	Cell type	Reference
Enhancer	rs356168	SNCA	Replacement	Parkinson’s disease	Altered SNCA expression by affecting binding of the brain-specific transcription factors EMX2 and NKX6-1	Human stem cell-derived neural precursor cells and neurons	Soldner et al. (2016)
	rs2595104	PITX2c	Deletion	Atrial fibrillation	Altered PITX2c expression via interaction with TFAP2a	Human stem cell-derived cardiomyocytes	Ye et al. (2016)
	rs2742624	UPK3A	Deletion	Prostate cancer	Diminished UPK3A level	Human prostate carcinoma LNCaP	Jin et al. (2016)
	rs6983267	MYC	Deletion	Colon cancer	Downregulated MYC level	Human colorectal cancer HCT116	Yao et al. (2014)
	rs78744187	CEBPA	Mutagenesis	Hematopoiesis	Reduced CEBPA expression, impaired basophil differentiation and maturation	CD34 + human hematopoietic stem and progenitor cells	Guo et al. (2017)
	rs9383590	ESR1	Deletion	Breast cancer	Decreased ESR1 expression	Human breast cancer T-47D	Bailey et al. (2016)
	rs780094	GCKR	dCas9-VPR-mediated transcriptional activation	Type 2 diabetes	Induced GCKR expression by increasing H3K27Ac histone mark	Human liver cancer HepG2	Lopez Rodriguez et al. (2017)
	haplotype	BCL11A	Deletion	Beta-thalassemia	Reduced BCL11A expression	Human erythroid K562	Maroofi et al. (2017)
	rs9349379	PHACTR1	Deletion	Vascular disease	Resulted in higher expression of the endothelin-1 (EDN1) gene, suggesting a distal regulatory function of this SNP	Human stem cell-derived endothelial cells and vascular smooth muscle cells	Gupta et al. (2017)
	rs61839660	CaRE4	dCas9-VP64 activation	Autoimmunity	Diminished stimulation-dependent IL2RA enhancer function	Human immune Jurkat T cells	Simeonov et al. (2017)
	rs2279590	CLU	deletion	Pseudoexfoliation and Alzheimer’s	Affected CLU, EPHX2, and PTK2B gene expressions	Human HEK293 cells	Padhy et al. (2017)
	rs11055880 rs12142375	ATF7IP PDE4B	dCas9-KRAB repression	Breast cancer and leukemia	Reduced expression of ATF7IP and PDE4B genes	Human kidney (HEK293T)	Liu et al. (2017a)
Repressor	rs1421085	FTO	Correction	Obesity	Restored IRX3 and IRX5 repression, activated browning expression programs, and restored thermogenesis	Primary adipocytes from a patient	Claussnitzer et al. (2015)
	rs10486567 rs67152137 rs7808935	HOXA cluster	Deletion	Prostate cancer	Upregulated the expression of HOXA13 and HOTTIP	Human prostate epithelium cell (RWPE-1)	Luo et al. (2017)
Chromatin loop	rs7903146	TCF7L2	Deletion	Type 2 diabetes	Reduced ACSL5 expression, abolished chromatin contacts with the ACSL5 promoter	Human colorectal cancer HCT116	Xia et al. (2016)
Chromatin dynamics	rs1198588	MIR137	Mutagenesis	Schizophrenia	Altered MIR137 expression, dendrite arborization, and synapse maturation	hiPSC-derived excitatory neurons	Forrest et al. (2017)

ACSL5 acyl-CoA synthetase long chain family member 5, ATF7IP activating transcription factor 7 interacting protein, CEBPA CCAAT/enhancer-binding protein alpha, CLU clusterin, CaRE4 CRISPRa-responsive element 4, EMX2 empty spiracles homeobox 2, EPHX2 epoxide hydrolase-2, ESR1 estrogen receptor 1, FTO alpha-ketoglutarate-dependent dioxygenase ,HOTTIP HOXA distal transcript antisense RNA, HOXA13 homeobox A13 ,IL2RA interleukin 2 receptor subunit alpha, IRX3 iroquois homeobox 3, MYC MYC proto-oncogene, NKX6-1 NK6 homeobox 1, PDE4B phosphodiesterase 4B, PHACTR1 phosphatase and actin regulator 1, PITX2c paired-like homeodomain 2C, PTK2B protein tyrosine kinase 2 beta, SNCA synuclein alpha, TCF7L2 transcription factor 7 like 2, TFAP2a transcription factor AP-2 alpha, UPK3A uroplakin-3A

Conclusions and perspective

The CRISPR is certainly a powerful cutting-edge biotechnology platform that can be used to study epigenetic regulation and chromatin biology in mammals. Various strategies based on CRISPR/Cas9 and CRISPR/dCas9 systems have been successfully applied to characterize the functional roles of regulatory genomic elements and chromatin structure that are involved in determining gene transcriptional activities by reshaping the epigenetic landscape. Given the important roles of non-coding genetic variants in human disease and other complex traits, CRISPR-based approaches have been useful in resolving the challenges associated with the functional interpretation of these non-coding variants through direct interrogation and perturbation of these non-coding variants in their native genetic context. By manipulating nuclear architecture and gene expression via de novo chromatin loop formation and chromatin remodeling, it is now possible to restore normal epigenetic patterning and correcting transcriptional misregulation. Taken together, CRISPR technology provides a promising avenue for targeted epigenetic-oriented therapy for early intervention and for effective treatment.