On RNA-programmable gene modulation as a versatile set of principles targeting muscular dystrophies

Abstract

The repurposing of RNA-programmable CRISPR systems from genome editing into epigenome editing tools is gaining pace, including in research and development efforts directed at tackling human disorders. This momentum stems from the increasing knowledge regarding the epigenetic factors and networks underlying cell physiology and disease etiology and from the growing realization that genome editing principles involving chromosomal breaks generated by programmable nucleases are prone to unpredictable genetic changes and outcomes. Hence, engineered CRISPR systems are serving as versatile DNA-targeting scaffolds for heterologous and synthetic effector domains that, via locally recruiting transcription factors and chromatin remodeling complexes, seek interfering with loss-of-function and gain-of-function processes underlying recessive and dominant disorders, respectively. Here, after providing an overview about epigenetic drugs and CRISPR-Cas-based activation and interference platforms, we cover the testing of these platforms in the context of molecular therapies for muscular dystrophies. Finally, we examine attributes, obstacles, and deployment opportunities for CRISPR-based epigenetic modulating technologies.

Graphical abstract

Gonçalves and colleagues, after providing an overview on CRISPR-Cas9-based gene modulating platforms and the (epi)genetic mechanisms through which they operate, cover these systems in the context of molecular therapies for muscular dystrophies. Next, the authors end by examining the attributes, obstacles, and deployment opportunities for RNA-programmable epigenetic editors.

Introduction

Background on gene expression modulators: From epigenetic drugs to targeted biologics

Besides spatiotemporally regulating cellular phenotypes during development and cellular differentiation, the mechanisms underlying specific epigenome states play a critical role in both physiological and pathological conditions in mature cells. Consequently, several molecular regulators of the cellular epigenetic landscape have emerged as therapeutic targets for a range of malignancies. Indeed, therapies involving the inhibition of DNA methyltransferases (DNMTs) and histone deacetylases (HDACs) have been approved as treatments for cancer patients¹ and clinical trials are also evaluating the potential of these therapeutic targets in other disorders,² including muscular dystrophies.^3,4

Muscular dystrophies are characterized by a progressive weakness and loss of muscle mass resulting from genetic or epigenetic changes disrupting the regular complement of gene products underlying healthy and structurally intact muscle tissues.⁵ The genomic changes causing muscular dystrophies are often multiple and diverse, which complicates the development of therapies, especially those that need to be personalized to specific patient genotypes. Moreover, despite the expanding knowledge surrounding congenital muscular dystrophies, and the resulting fostering of translational research targeting these disorders, the relatively few approaches that have entered clinical testing went on to yield only partial resolution of disease symptoms.⁵ These results support reinforcing and diversifying research lines directed at identifying new candidate therapeutic agents and principles, especially those that are applicable to all or most patient cohorts. In this context, pharmacological approaches have been shown to successfully induce changes in the epigenome with measurable clinical outcomes, thus highlighting the potential of targeting epigenetic regulators as mutation-independent therapeutic strategies. Notably, it is known that HDACs⁶ play pivotal roles in the transcriptional regulation of muscle genes, contributing to the development, differentiation, and maintenance of muscle tissues.⁷ Moreover, altered levels of HDACs have been for instance detected in models of Duchenne muscular dystrophy (DMD)^7,8 and Emery-Dreifuss muscular dystrophy,⁹ highlighting the connection between HDAC deregulation and the pathogenesis of at least some muscular dystrophies. Based on this cumulative evidence, various epigenetics modulating drugs (“epi-drugs”), primarily pan-HDAC inhibitors, have been investigated in preclinical models. The testing of epi-drugs has included the targeting of DMD as well as other muscular dystrophies with similar pathological features of sarcolemma fragility. The beneficial effects of these molecules uncovered in multiple mouse models has, in turn, laid the groundwork for clinical translation. For instance, trichostatin A and givinostat, two pan-HDAC inhibitors, were proven to improve dystrophic muscle traits in mouse models.^10,11 In addition, givinostat and vorinostat were tested in preclinical settings for DMD, limb-girdle muscular dystrophies (LGMDs), and myotonic dystrophy type 1 (DM1).^11,12,13 Although the mechanisms of action of epi-drugs have not been fully elucidated, they might at least in part involve the regulation of compensatory genes, tissue remodeling (e.g., reduction of adipogenesis and fibrotic tissue differentiation), and inflammation that individually or combined alleviate pathology symptoms or, ideally, dampen disease progression.^4,7

Sirtuins constitute yet other relevant set of targets. Members from this class of proteins present nicotinamide adenine dinucleotide (NAD⁺)-dependent protein deacetylase activities that regulate multiple cellular processes including the control of gene expression.¹⁴ In this context, the SIRT1 gene product, the NAD⁺-dependent deacetylase protein sirtuin-1, is known to improve skeletal muscle function by suppressing oxidative stress.¹⁵ Indeed, a single pilot study conducted in DMD, Becker muscular dystrophy, and Fukuyama congenital muscular dystrophy patients involving the administration of resveratrol, a SIRT1 activator, revealed an improvement in motor function (UMIN-CTR Unique ID: UMIN000014836).¹⁶ Yet, among the epi-drugs under testing, only givinostat has successfully progressed to a clinical trial phase III (NCT0281797), confirming its effects in decelerating disease progression.¹⁷

Given the significance of epigenetic mechanisms in controlling development, cellular differentiation, and normal cellular physiology, targeting diseases states with epigenetic modulators that are more specific than the currently available epi-drugs is postulated to offer significantly heightened clinical efficacy while reducing toxicity and side effects. In this context, notwithstanding their complexity and large size, (epi)genetic modulators or editors built on clustered regularly interspaced short palindromic repeat (CRISPR) systems offer a promising alternative to global epigenome-modifying pleiotropic drugs owing to their higher specificity and mechanism-of-action versatility. Generically, the modus operandi of CRISPR-based (epi)genetic editors entails the direct recruitment of transcription activating domains (TADs) and/or chromatin remodeling factors to predefined cis-acting regulatory sequences. Ideally, these maneuvers change the chromatin marks that confer bona fide epigenetic remodeling and ensuing durable gene modulation, i.e., stable transcriptional activation or inhibition upon transient (hit-and-run) DNA engagement of defined effector domain(s).^18,19

In conclusion, the findings on the use of epi-drugs have laid the groundwork for ongoing and upcoming research efforts aimed at developing next-generation (epi)genetic editors that seek therapeutic outcomes via a more precise targeting of compensatory and/or disease-modifying genes. We next explore the landscape of such CRISPR-based gene modulating agents and their testing on the tackling of muscular dystrophies covering, in the process, their properties and future opportunities for deployment.

Repurposing of CRISPR systems: From gene breaking to modulating

Engineered CRISPR-Cas9 nuclease complexes consisting of sequence-tailored single guide RNAs (gRNAs) and Cas9 endonucleases remain among the most powerful genome editing tools since their inception in 2012²⁰ (Figure 1). However, the precision of genome editing using programmable nucleases is often hindered due to off-target and, more pervasively, on-target unwanted effects, e.g., installation of mutagenic small insertions and deletions (indels) and unpredictable local and genome-wide structural variants (e.g., duplications, large deletions, and translocations).^21,22 Most of these potentially deleterious genotoxic effects are the result of the repair of double-stranded DNA breaks (DSBs) by error-prone DNA end-joining processes, e.g., non-homologous end joining (NHEJ) and microhomology-mediated end joining (MMEJ). In addition, chromosomal DSBs can trigger P53-dependent apoptosis as well as haploinsufficiency and cell fitness losses especially when located in coding sequences.²³ Considering that, so far, genome editing directed at muscular dystrophies has mostly relied on adeno-associated viral vectors (AAVs) encoding CRISPR-derived nucleases,^24,25 it is worth noting the detection of substantial amounts of AAV vector DNA inserted at nuclease target sites in various tissues, including skeletal muscle.^26,27

Figure 1.

Main CRISPR technologies: a genetic Swiss army knife

Double-strand DNA break (DSB)-dependent CRISPR technologies. Engineered CRISPR-Cas9 complexes consist of a single guide RNA (gRNA) and a Cas9 enzyme with two nuclease domains (i.e., RuvC and HNH), each of which is responsible for catalyzing phosphodiester cleavage of the DNA double-helix strands. After Cas9 binding to a protospacer adjacent motif (PAM), target DNA cleavage ensues if, next to the PAM, is located a sequence (protospacer) complementary to the 5′ end of the gRNA (spacer). Typically, the repair of the site-specific DSB by non-homologous and homologous recombination processes is exploited for achieving gene knockouts (KO) and gene knockins (KI), respectively. The latter strategy requires the presence of an exogenous (donor) DNA containing the genetic information of interest often flanked by DNA tracts sharing homology to the target region. Single-strand DNA break (SSB)-dependent CRISPR technologies. Engineered CRISPR-Cas9 sequence- and strand-specific nucleases (“nickases”), generated by disabling either the HNH or RuvC domains of Cas9 yield base editors and prime editors once fused to DNA deaminase and reverse transcriptase domains, respectively. Base editors install single base pair substitutions with the aid of a conventional gRNA; while prime editors install single base pair substitutions, insertions, and deletions that are encoded in an extended prime editing gRNA (pegRNA).

DNA strand break-independent (SB-free) CRISPR technologies. Engineered catalytically inert (“dead”) Cas9 proteins (dCas9), generated by disabling both nuclease domains, serve as gRNA-programmable scaffolds that are biotinylated or endowed with a fluorescent reporter for, respectively, unbiased capture of chromatin-regulating proteins associated with specific loci or for gene visualization and chromosome dynamics studies in living cells. In addition, dCas9 can also serve as a scaffold for locus-specific recruitment of heterologous transcription activating domains (TADs), transcription repressing domains (TRDs), and/or other effectors that modify or install epigenetic marks consisting of histone post-translational modifications (PTMs) (e.g., acetylation and deacetylation) or DNA methylation and demethylation. The respective CRISPR activation and CRISPR interference complexes favor and hinder, respectively, gene expression via ultimately controlling the engagement of the endogenous transcriptional machinery at target loci of interest.

Hence, currently, there is a growing trend on the investigation of genome editing principles that instead of programmable nucleases use sequence- and strand-specific nucleases (“nickases”), as such,^28,29 or coupled to effector domains, e.g., deaminases and reverse transcriptases that yield, respectively, base editors and prime editors³⁰ (Figure 1). These DSB-independent strategies present significant safety advancements over programmable nuclease-based approaches in that nicks, in contrast to DSBs, are neither NHEJ nor MMEJ substrates and, in addition, constitute poor P53-activating lesions.^31,32 Albeit at much-reduced levels, it has nonetheless been reported that nickase-based techniques can still carry the risk of potentially harmful genetic alterations, such as indels and chromosomal translocations.^23,33,34 Indeed, under certain circumstances, nicks can evolve into DSBs such as when an advancing replication fork hits them and collapses.³⁵ Moreover, baseline levels of nickase-derived indels seem to vary in a locus sequence-dependent fashion.³⁶ Thus, next to these DSB-independent genome editing developments, there is also a gathering of momentum on the investigation of epigenome editing strategies that make use of catalytically inactive (“dead”) nucleases.^18,19 In this case, engineered CRISPR-Cas9 complexes derived from the prototypic adaptive immune system from Streptococcus pyogenes, are particularly useful scaffolds on which to build powerful tools for modulating disease-modifying genes given their robustness in engaging eukaryotic chromatin and ease of target DNA customization (Figure 1). Typically, CRISPR-Cas9 epigenome editing complexes consist of a dead Cas9 protein (dCas9) endowed with heterologous or synthetic effector domains conjugated to a sequence-tailored gRNA whose scaffold is either unmodified or modified with RNA aptamers for the recruitment of additional effector domains. These complexes seek altering transcriptional or epigenetic states that beneficially impact disease progression via modulating the expression of (1) compensatory genes, (2) genes promoting muscle regeneration, and/or (3) genes involved in controlling pathologic traits in muscle cells (e.g., muscle tissue remodelling and inflamation). In particular, upregulating or downregulating the expression of these endogenous target gene(s) results from fusing the gRNA-customizable dCas9 moiety to specific transcription activating or repressing domains, respectively, or to epigenetic modifiers that locally alter histone marks or DNA methylation composition at predefined chromosomal sequences (Figure 2).

Figure 2.

Main epigenetic mechanisms controlling endogenous gene expression

DNA methylation, top panel. DNA methylation regulates gene expression and epigenetic memory acquisition and occurs when a methyl group is added to cytosines in CpG dinucleotides. This process is catalyzed by DNMTs. DNMT1 maintains methylation patterns before and upon cell division and DNMT3A/B initiates de novo DNA methylation. DNA methylation is associated with gene repression through the recruitment of methyl-CpG-binding protein domains that engage repressive histone-modifying and chromatin-remodeling complexes. In addition, DNA methylation can directly inhibit transcription by blocking the access of transcription factors to their cognate target sequences. DNA methylation can also influence nuclear organization and chromosomal territory formation, further affecting transcription regulation. Histone post-translational modifications (PTMs), central panel. Histone PTMs such as acetylation, methylation, phosphorylation, ubiquitination, SUMOylation, and ADP-ribosylation, together, control chromatin compaction in a dynamic or inheritable manner, i.e., transcriptionally active open euchromatin versus transcriptionally inactive closed heterochromatin. The enzymes responsible for PTMs, writers and erasers, are highly dynamic and can mediate different histone modifications simultaneously. Whether PTMs at the nucleosomes (chromatin units formed by a complex of DNA, non-histone, and histone core proteins, i.e., two copies of H2A, H2B, H3, and H4), facilitate or inhibit transcription depends on the particular set and location of the residue marks (“histone code”). PTM marks often locate in histone tails at positively charged amino acids that, normally, tightly bind to the negatively charged DNA phosphate backbone favoring chromatin compaction. However, once acetylated by HATs, these residues acquire a more neutral character (hyperacetylated status) favoring, in this case, chromatin relaxation. In contrast, HDACs by removing acetyl groups from these residues (hypoacetylation status), promote chromatin compaction. Of notice, interactions between histone modifications and DNA methylation are intricate, influencing the signals of each other. Chromatin remodeling, bottom panel. Dueling ATP-dependent chromatin remodeling complexes such as imitation switch (ISWI) and switch/sucrose non-fermentable (SWI/SNF), read and bind to specific epigenetic marks to dynamically regulate higher-order chromatin structures that control DNA access to transcription activators, repressors, and epigenetic factors. Chromatin remodeling proteins share a similar ATPase motor that, via translocating DNA along the nucleosome face, disrupts histone-DNA contacts. Outcomes include, nucleosome assembly, sliding, eviction, and exchange. Finally, besides DNA methylation, histone modification, and ATP-driven chromatin remodeling, there are other mechanisms that collectively shape the epigenetic landscape of a cell, e.g., long non-coding RNAs and integration of histone variants in nucleosomes (not shown).

CRISPR activation platforms

To achieve gene upregulation, previously established native TADs have been fused to dCas9 scaffolds. Initial investigations reported fusions of S. pyogenes dCas9 to the commonly used VP64 activator (i.e., four tandem repeats of the herpes simplex virus protein 16 TAD, VP16³⁷), representing the first-generation CRISPR-based gene activation construct, also known as CRISPRa.^38,39,40 Subsequent advancements involved increasing the number of VP64 units fused to dCas9 to form constructs such as dCas9-VP160, dCas9-VP192, and VP64-dCas9-BFP-VP64, incorporating 10, 12, and 8 units of VP64, respectively.^41,42,43 To further attempt at mimicking endogenous transcription activation mechanisms, typically involving the combinatorial recruitment of numerous transcription factors and other auxiliary co-activating proteins, researchers have also started integrating multiple activators to boost chromatin remodeling and ensuing target gene expression. Such second-generation synthetic activators include the supernova tagging (SunTag) system, the synergistic activation mediator (SAM) system, and the dCas9-VPR system. The SunTag platform utilizes 10 repeats of a short epitope tag covalently linked to dCas9. This configuration allows for the recruitment of 10 single-chain variable fragments (scFvs) fused to VP64 domains.^44,45,46 Diversely, the SAM system incorporates a gRNA engineered with two bacteriophage MS2 hairpin aptamer loops that bind to their cognate MS2 coat protein (MCP) linked to a fusion of two different TADs (e.g., p65-HSF1).^47,48,49,50 The dCas9-VPR system consists in turn of the fusion of three different TADs (i.e., VP64, p65, and Rta) covalently linked to the dCas9.^51,52,53 This tripartite VPR system underwent further refinements in the form of the dCas9-VPRmini and dCas9-VPH constructs. The former is characterized by truncated, yet functional, versions of the p65 and Rta domains,⁵⁴ the latter comprises four repeats of VP48, p65, and HSF1 fused to dCas9.⁵⁵ In a recent study, the dCas9-VPH construct was additionally coupled to an SS18 domain able to recruit the full endogenous SWI/SNF chromatin-remodeling complex, thereby aiding exposing transcription factor binding sites for enhancing gene activation mediated through the VPH tripartite domain.⁵⁶ Notably, the VPR domain has exhibited toxicity when constitutively expressed in vivo in both Drosophila and Dravet syndrome transgenic mice.^57,58 Conversely, mice expressing dCas9-VP64 did not show adverse effects, indicating that the toxicity of dCas9-VPR is at least in part due to the p65 or Rta components rather than to the dCas9 or VP64 elements.⁵⁸ Furthermore, beyond refining the existing heterologous TADs, ongoing efforts are also being directed toward engineering new epigenetic effectors with the aim of reducing the toxicity associated with viral motifs, such as those seemingly present in VPR. In this regard, the CRISPR-DREAM transactivation system is a relevant development owing to its robustness and cell tolerability.⁵⁹ This system combines dCas proteins, guide RNAs engineered with MS2 aptamer loops, and MCP-linked multipartite transactivation modules built from human mechanosensitive transcription factors (e.g., MRTF-A, STAT1, eNRF2).^60,61 This platform has achieved specific and robust targeted gene activation across various mammalian cell types resulting in superior or comparable target gene upregulation when compared with the SAM system. Finally, the compactness of CRISPR-DREAM components facilitates their incorporation in payload-restricted delivery systems such as AAVs.⁵⁹

Critically, technologies based solely on TAD recruitment tend to merely induce transient maintenance of maximum gene upregulation levels. This poses a significant limitation to their broad applicability, especially in candidate clinical settings where a hit-and-run modus operandi would be preferable to avoid the need for constitutive presence of the exogenous effector constructs in cells. Hence, the ability to deposit heritable epigenetic marks that stably maintain euchromatic and heterochromatic states represents a more suitable and versatile approach for bona fide epigenetic therapies involving gene upregulation and silencing, respectively.⁶² Toward implementing this principle, chromatin remodeling modifiers have been fused to dCas9 to install activating or repressing epigenetic marks at predefined loci. A notable gene activation mark is that of acetylation of lysine 27 in histone H3 tails (i.e., H3K27Ac). The fusion of the core catalytic domain of the adenovirus E1A-associated protein p300 histone acetyltransferase (HAT) (p300^core) to the dCas9 protein (dCas9-p300^core), results in potent gene activation involving H3K27Ac acquisition when gRNAs addressing the complex to gene promoters and enhancers are provided.⁶³ Another study demonstrated a correlation between gene activation and hyperphosphorylation of histone H3 serine 28 (H3S28Ph) when fusing the hyperactive histone kinase MSK1 to dCas9 (dCas9-MSK1).⁶⁴

Beyond TAD recruitment and histone acetylation or phosphorylation of specific histone residues, DNA demethylation represents an additional epigenetic regulatory mechanism associated with gene activation.⁶⁵ Hence, researchers have for instance developed CRISPR-based tools to upregulate gene expression via targeting the catalytic domain of the TET1 DNA demethylase to regulatory regions. Independent studies reported successful gene activation of BDNF, MyoD, and BRCA1 in in vitro models when coupling dCas9-TET1 constructs to gRNAs targeting specific regulatory sequences.^66,67 Similar experiments combined a dCas9 and an MS2 aptamer-modified gRNA for tethering the Tet1 catalytic domain via its fusion to dCas9 or to MCP, respectively.⁶⁸ Moreover, combinatorial use of effector domains acting at the transcription and/or epigenetic levels is being pursued to achieve additive or, ideally, synergistic effects. For instance, the previously described SunTag system, initially designed to recruit multiple copies of the TAD VP64, has been adapted to recruit multiple copies of the TET1 demethylase core domain to desired loci resulting in targeted DNA demethylation and ensuing upregulation of the genes of interest.^69,70 In another study, researchers observed that at highly methylated promoter regions of tumor suppressor genes, enhanced gene activation is achieved by combining dCas-VP64 or dCas-VPR with the SAM complex.⁷¹ Similarly, dCas9-VPR has been combined with the p300 core to activate otherwise silenced endogenous genes. This method proved to be more effective than the administration of each component individually.⁷² The CRISPRon system is yet another combinatorial gene activation system that, in this case, consists of a dCas9-TET1 DNA demethylase construct coupled to a gRNA modified with MS2 aptamers for the recruitment of multiple TADs via their fusion to an MCP domain.⁷³ Together, these tools and associated findings provide important proof-of-concepts for utilizing combinations of transcriptional activators and epigenetic modifiers to induce robust expression of tightly regulated and silenced genes across multiple cell types in vitro and in vivo.

Finally, strategies are also under development for achieving tunable and temporal control over target gene modulation. For instance, a dCas9-p300 construct was successfully integrated with an auxin-controllable degron.⁷⁴ In addition, to avoid the need for exogenous transcription regulatory factors and to simultaneously achieve temporal control over target gene expression, a system has been devised in which a chemical epigenetic modifier (CEM) molecule binds endogenous factors in a dose-dependent manner. The CEM molecule couples, for instance, the endogenous bromodomain-containing protein 4 (BRD4), a chromatin remodeler that interacts with acetylated histones, to dCas9 complexes endowed with an FK506-binding protein (FKBP) moiety.⁷⁵ This chemical dimerization permits therefore not only gene activation but also tunable regulation of target gene expression levels depending on the availability and concentration of the CEM inducer.

CRISPR interference platforms

To widen the applicability of gene modulation approaches, complementary CRISPR interference systems are equally undergoing intense research and development. Similar to CRISPR activation platforms, gene interference systems can be built by endowing dCas9 scaffolds with, in this case, gene repressing or silencing effector domains. Different approaches have been studied in this regard which vary by their incorporation of diverse types of transcription-repressing domains, DNMTs and HDACs. One of the most broadly applied systems consists of dCas9 proteins fused to Krüppel-associated box (KRAB) transcriptional repressor domains that act as scaffolds for, among others, the endogenous co-repressing factors KRAB-associated protein 1 (KAP-1) and heterochromatin protein 1 (HP-1). This in situ recruitment of endogenous factors results in target sequences transiting from transcriptionally active euchromatin into transcriptionally inactive heterochromatin. Several generations of KRAB-based CRISPR interference constructs have been developed. An initial system used a dCas9 protein coupled to the KRAB domain from the zinc finger protein 10 (ZNF10).⁷⁶ Subsequently, the gene silencing moiety was substituted by a more potent KRAB domain derived from the zinc finger imprinted 3 (ZIM3) protein yielding the construct Zim3-dCas9.⁷⁷ Further studies have assessed the effects of KRAB targeting in conjunction with other heterologous domains of interest. For instance, combinatorial systems involving the methyl-CpG-binding protein MeCP2 (i.e., dCas9-KRAB-MeCP2)⁷⁸ and the SIN3A-interacting domain of MAD1, SID4x (i.e., SID4X-dCas9-KRAB),⁷⁹ were reported to display enhanced transcriptional inhibition when compared with constructs harboring the KRAB domain alone.^78,79 Next to these efforts, CRISPR interference systems designed for changing histone marks have also been tested. For instance, HDAC moieties, specifically from HDAC3, when fused to dCas9 were shown to efficiently catalyze the deacetylation of H3K27 when appropriately positioned within the Mecp2 promoter region.⁸⁰

Beyond acetylation, the methylation state of histones also influences the accessibility of regulatory regions (e.g., promoters and enhancers) to gene-activating or -repressing factors (Figure 2). For instance, the histone demethylase LSD1 is known to be implicated in enhancer inhibition⁸¹ and has, in fact, been tested for target gene modulation upon fusion to DNA-binding domains consisting of transcription activator-like effector (TALE) repeats.⁸² Subsequently, researchers explored the fusion of LSD1 to dCas9 for targeting the Oct4 enhancer region and reported a decrease in gene expression together with a decrease of di-methylated histone 3 lysine 4 (H3K4me2).⁸³ Moreover, the levels of H3K27ac, a marker of active enhancers, was reduced at the Oct4 target region.⁸³ Interestingly, when the proximal promoter region of Oct4 was targeted instead, changes in neither gene expression nor epigenetic marks were observed, indicating the enhancer-specific activity of the dCas9-LSD1 complexes.⁸³

In addition, in the context of Alzheimer’s disease, an ectopic DNA methylation strategy has been examined in which the fusion of dCas9 to Dnmt3a was demonstrated to increase local DNA methylation and ensuing gene repression.⁸⁴ To enhance DNA methylation-dependent effects, a dCas9 protein was fused to SunTag repetitive peptide epitopes for recruiting multiple copies of a scFv-linked DNMT3A methyltransferase, resulting in targeted methylation with minimal impact on genome-wide DNA methylation levels.⁸⁵ Moreover, combining a KRAB domain with the DNMTs Dnmt3a and Dnmt3L in a dCas9 fusion construct, known as the CRISPRoff system,⁷³ was shown to successfully silence target genes in a durable manner via the acquisition of heritable epigenetic marks, i.e., marks that persist after cell division and differentiation in the absence of the effector complexes. Interestingly, exposing cells initially treated with CRISPRoff to the aforementioned CRISPRon components leads to the reversibility of target gene silencing and of the associated epigenetic marks.⁷³

CRISPR activation for muscular dystrophies

Duchenne muscular dystrophy

DMD (OMIM no. 310200) stands out as the muscular dystrophy most targeted up until now by gene modulation platforms with one of these approaches reaching in fact clinical testing.⁸⁶ This progressive and lethal muscle-wasting disorder is caused by various loss-of-function mutations in the large DMD gene (∼2.4 Mb), whose muscle-specific dystrophin Dp427m isoform normally sustains the integrity of the striated musculature⁸⁷ (Figure 3A). Despite the absence of an effective treatment for DMD to date, several therapeutic strategies are under development.⁸⁸ These include DMD reading frame restoration via antisense oligonucleotide (AON)-mediated exon skipping and defective DMD gene complementation via either AAV-mediated micro-dystrophin transfer or full-length utrophin A isoform upregulation, a protein with high structural and functional similarity to dystrophin.⁸⁹

Figure 3.

Representative therapeutic approaches explored for muscular dystrophies using CRISPR activation and CRISPR interference systems

(A) CRISPR activation approaches. Disease phenotypes caused by loss-of-function gene mutations or alterations can, in principle, be complemented via the upregulation of endogenous (wild-type) compensatory gene products. In the case of Duchenne muscular dystrophy (DMD), the isoform A of utrophin is a candidate product for compensating for the absence of the muscle-specific isoform of dystrophin (Dp427m) in striated muscle cells. Whether utrophin upregulation can also ameliorate the milder phenotype of Becker muscular dystrophy (BMD) underpinned by truncated partially functional dystrophins, warrants further investigation. The laminin α1 and caveolin-1 proteins are additional candidate targets for addressing merosin-deficient congenital muscular dystrophy type 1A (MDC1A) and caveolae-associated muscular dystrophy, a type of limb girdle muscular dystrophy (LGMD), caused by the functional loss of laminin α2 and muscle-specific caveolin-3, respectively. (B) CRISPR interference approaches. Disease phenotypes caused by pathogenic gain-of-function mutations or alterations can, in principle, be counteracted via the downregulation or, ideally, silencing of endogenous (mutant) disease-causing genes or alleles. For instance, CRISPR interference systems addressed to DUX4 and DMPK alleles have therapeutic potential for facioscapulohumeral muscular dystrophy (FSHD) and myotonic dystrophy type 1 (DM1), respectively. FSHD is caused by a contraction of D4Z4 repeats leading to, among other outcomes, local loss of CpG methylation and ensuing transcriptional activation of unscheduled DUX4 gene expression in somatic cells. DM1 is instead caused by an expansion of trinucleotide CTG repeats located within the 3′ UTR of DMPK alleles. See the main text for further details.

Utrophin, expressed from the UTRN gene (an autosomal DMD gene paralog), is restricted to the neuromuscular and myotendinous junctions in post-natal muscle due to transcriptional and post-transcriptional regulatory mechanisms, e.g., via the binding of miRNAs at the 3′ untranslated region (UTR) of the UTRN mRNA. Importantly, increasing endogenous UTRN gene expression is a particularly appealing DMD therapy route as utrophin can compensate for the lack of dystrophin in linking the cytoskeleton to the sarcolemma of muscle cells (Figure 3A). Indeed, increase of utrophin expression by 3- to 4-fold in transgenic dystrophin-defective mice suffices to prevent the development of muscular dystrophy traits.^90,91 Notably, utrophin upregulation can be applied to all patients and, in principle, should not trigger immunological reactions as it is a patient-own ubiquitously expressed protein.

The first attempts at achieving utrophin upregulation entailed fusing TADs to artificial zinc finger domains designed for binding UTRN promoter A sequences. This so-called Jazz-ZF system,^92,93 was later refined into the UtroUp system by fusing six zinc finger motifs to CJ7, a TAD stronger than the previously used VP16 moieties.⁹⁴ As aforementioned, with the advent of CRISPR technologies,^18,19 researchers readily generated fusions between dCas9 and TADs (e.g., VP160). These reagents were shown to efficiently increase the amounts of utrophin protein in muscle cells isolated from DMD patients and lead to the stabilization of β-dystroglycan, a key component of the dystrophin-associated glycoprotein complex normally disrupted in the absence of dystrophin.⁹⁵ The beneficial role of endogenous utrophin upregulation on dystrophic disease traits was further confirmed in vivo by using the dystrophin-defective mdx mouse model. In particular, constructs encoding an MPH module (i.e., MCP-p65-HSF1 fusion protein) and a Utrn-targeting gRNA engineered with two MS2 aptamers for MPH binding were packaged in AAV particles. Co-injection of these particles into skeletal muscles of mdx mice together with AAV particles delivering Cas9 or dCas9 constructs resulted in significant utrophin upregulation and concomitant amelioration of disease symptoms when compared with controls.⁹⁶ Moreover, application of this dual AAV system was also proven to be beneficial when a gRNA was directed to the regulatory cis-acting sequences of the Fst gene, whose product, follistatin, is associated with muscle mass buildup via myostatin inhibition, a transforming growth factor β (TGF-β) family member that acts as a negative regulator of muscle growth. Besides providing important proof-of-concepts for CRISPR activation in an in vivo setting, these experiments hint at the potential for using multiplexing gRNA delivery in order to achieve combinatorial modulation of compensatory and disease-modifying genes, in this case, through human UTRN and FST co-targeting, respectively.

Experiments targeting SIRT6, an NAD⁺-dependent sirtuin family member with HDAC activity, revealed that this protein represses UTRN expression via H3K56Ac deacetylation in muscle cells identifying, therefore, histone acetylation as an additional regulator of utrophin synthesis.⁹⁷ Consistent with these data, addressing dCas9-Sirt6 and dCas9-p300 constructs to a specific UTRN gene enhancer yielded significantly decreased and increased, respectively, synthesis of utrophin in muscle cells.⁹⁷ As a result, next to direct UTRN gene upregulation, UTRN upregulation combined with orthogonal SIRT6 inhibition might provide for a candidate epigenetic approach for tackling DMD provided that untoward secondary effects do not emerge, e.g., substantial transcriptome deregulation.

In addition to utrophin, laminin-111 has also been suggested as a potentially beneficial molecule for counteracting DMD disease progression owing to its ability to partially replace the structural function of dystrophin through, in this case, the linking of the extracellular matrix to the sarcolemma. Supporting this view, independent studies have shown that intramuscular injections of laminin-111 complexes into mdx mice stabilize the sarcolemma and protects muscles from contraction-induced damages.⁹⁸ As the expression of Lama1, encoding the laminin-111 α1 subunit, is restricted to the embryogenic phase, a dCas9 protein fused to either the VP160 TAD or to the p300 HAT was addressed with gRNAs to the Lama1 gene promoter resulting in increased target gene transcription and protein expression in cultured muscle cells.⁹⁹ Subsequent experiments in a DMD mouse model using an electroporation method to deliver into muscle tissue plasmids encoding the dCas9-VP160 construct, led to Lama1 upregulation, as detected by immunofluorescence microscopy and western blot analyses.⁹⁹ However, experiments assessing functional endpoints were not performed.

The DMD gene contains at least seven tissue-specific promoters and two polyadenylation sites that govern the expression timing and tissue localization of the different dystrophin isoforms. While the well-studied isoform of dystrophin Dp427m is expressed in muscle cells (Figure 3A), the similarly long cortical isoform Dp427c and Purkinje isoform Dp427p of dystrophin are predominantly expressed in neurons of the cortex and hippocampus and in cerebellar Purkinje cells, respectively.^100,101 The high similarity among these proteins and Dp427m regarding their structures and sizes provides a rationale for ectopic upregulation of one of these potentially compensatory isoforms in muscle cells as candidate DMD therapeutic interventions. Following in vitro and in vivo testing of a Staphylococcus aureus dCas9 (dSaCas9) fused to a VP64 TAD targeted to the Dp427c promoter region, an n = 1 patient study commenced using AAV9 vector delivery.⁸⁶ Unfortunately, after receiving high loads of the AAV9 vector (i.e., 1 × 10¹⁴ vector genomes per kilogram of body weight) the patient succumbed due to exacerbated innate immune reactions. Preexisting cardiopulmonary dysfunctions progressed to acute respiratory distress syndrome and cardiac arrest within a week of the AAV9 administration. The patient died 2 days after the emergence of these acute complications. Upon postmortem tissue analysis, with the exception of the liver, dSaCas9 expression was not detectable, neither at the transcript nor protein levels. Therefore, Dp427c levels in skeletal muscle were not examined. This serious adverse event highlights the current challenges regarding the implementation of AAV-based genetic therapies for DMD.¹⁰² These challenges might be ameliorated or overcome via the development and implementation of muscle-tropic AAV vector particles or other alternative delivery systems.¹⁰³

Merosin-deficient congenital muscular dystrophy type 1A

Merosin-deficient congenital muscular dystrophy type 1A (MDC1A) (OMIM no. 607855) is an autosomal recessive disease originating from mutations in the LAMA2 gene leading to the absence of the laminin α2 chain, a major component of the extracellular basement membrane (Figure 3A). Clinical symptoms present at an early onset and include severe muscle weakness, skeletal abnormalities, peripheral neuropathy, lack of independent ambulation, and respiratory insufficiency. Importantly, as DMD, MDC1A might also benefit from the upregulation of compensatory gene modifiers that counteract its abnormal phenotypes, in this case LAMA1 encoding laminin α1.¹⁰⁴ Indeed, experiments have shown that transgenic expression of Lama1 in muscles and in the peripheral nervous system of laminin α2 chain-deficient mice ameliorates muscle wasting and peripheral nerve defects.¹⁰⁵ These data directly demonstrate the relevance of laminin α1 chains as candidate compensatory modifiers of MDC1A progression (Figure 3A). Based on these observations, researchers have reported a mutation-independent strategy based on the upregulation of laminin α1 chain synthesis via exploiting a dCas9-VP64 activator endowed with a gRNA targeting the Lama1 gene promoter. Significantly, prevention of muscle fibrosis and reversal of disease progression was demonstrated when the CRISPR activation components were administered in pre-symptomatic and symptomatic mice, respectively.¹⁰⁶ A subsequent study tested the same strategy to achieve CRISPR activation-mediated LAMA2 complementation in human fibroblasts derived from MDC1A individuals. In particular, a dSaCas9 fused to two VP64 TADs enabled LAMA1 gene upregulation, which in turn resulted in the normalization of the overactive migration and aberrant transcriptional profile of fibroblasts derived from MDC1A patients.¹⁰⁷ As a result, these data confirm LAMA1 gene upregulation as a potential therapeutic strategy for MDC1A involving functional LAMA2 complementation.

CRISPR interference for muscular dystrophies

Facioscapulohumeral muscular dystrophy

The common form of facioscapulohumeral muscular dystrophy (FSHD), FSHD1 (OMIM no. 158900), is an autosomal dominant condition characterized by the abnormal reactivation of the DUX4 retrogene, whose expression is normally epigenetically silenced in somatic tissues (Figure 3B). Clinical symptoms whose severity is highly variable in different patients, present at a late onset and include progressive weakness of skeletal muscles, especially those in the upper body, e.g., face and scapulae. In FSHD1, a contraction of D4Z4 macro-satellite repeats associated with the DUX4 locus is responsible for reduced CpG methylation, repressive histone marks loss, and partial chromatin relaxation all contributing to transcriptional de-repression of DUX4 expression in skeletal muscle and a pathogenic gain of function¹⁰⁸ (Figure 3B). This DUX4 misexpression results in diverse downstream effects, including altered RNA and protein metabolism, apoptosis, and muscle atrophy. Different studies have assessed the effectiveness of different CRISPR interference systems in decreasing aberrant DUX4 expression (Figure 3B). A study using the dCas9-KRAB system targeting the promoter and exon 1 of DUX4¹⁰⁹ has established a first proof-of-concept for the potential of CRISPR interference strategies in downregulating pathogenic DUX4 expression in FSHD1. In a follow-up study, the dCas9-KRAB system was instead addressed to the exon 3 region of DUX4.¹¹⁰ These experiments demonstrated a significant downregulation of pathogenic DUX4 transcripts with a concomitant enrichment of repressive histone marks at the DUX4 promoter in target cells. However, despite their effectiveness in downregulating pathogenic DUX4 transcripts, these dCas9-KRAB constructs did not prove to adequately maintain the repressive marks over time. Additional research using smaller CRISPR interference platforms consisting of a dSaCas9 fused to one of four alternative epigenetic repressors (i.e., HP1α, HP1γ, MeCP2TRD, or the SUV39H1 SET domain) were shown to successfully reduce the expression of pathogenic DUX4 transcripts in in vivo models of the disease.¹¹¹ Notably, the number of detected off-target events at the transcriptome level was lower than that observed with the dCas9-KRAB system tested in parallel. Nevertheless, it must be noted that functional analyses assessing the improvement, or lack thereof, of dystrophic phenotypes were not performed. A complementary study focused instead on testing indirect regulation of pathogenic DUX4 gene expression via modulating epigenetic regulators involved in this abnormal expression. This study harnessed the dCas9-KRAB transcriptional repressor system to target the promoter or exon1 of previously identified regulators (i.e., BRD2, BAZ1A, KDM4C, and SMARCA5) resulting in the downregulation of DUX4 gene expression.¹¹²

Myotropic dystrophy type1

Myotropic dystrophy type 1 (DM1) (OMIM no. 160900) is an autosomal dominant disorder principally characterized by progressive muscle weakness and wasting, myotonia (i.e., impairment of relaxation of skeletal muscles upon voluntary contraction or electrical stimulation), cataract propensity and, often, cardiac conduction abnormalities. DM1 results from the amplification of CTG microsatellites in the 3′ UTR of the serine-threonine DM1 protein kinase gene DMPK¹¹³ (Figure 3B). Expansions of this unstable trinucleotide repeat is associated with local chromatin compaction and range in size from 5 to 38 copies in non-pathogenic alleles to 50–5,000 in pathogenic alleles. The expression of repeat-expanded DMPK alleles deregulates genes in the DMPK genomic region and leads to an accumulation of abnormal DMPK transcripts that sequester muscle-blind-like splicing factors that aggregate in the nucleus as discernible foci (Figure 3B). Cumulatively, these changes disrupt the transcriptome of DM1 cells creating pathologic gain-of-function phenotypes. A CRISPR interference strategy involving dCas9-KRAB-mediated DMPK downregulation achieved a reduction of pathogenic transcript aggregates in DM1 patient muscle cells (up to 80%) resulting in a noticeable correction of the abnormal transcriptome and spliceopathy parameters. In addition, whole-cell patch-clamp assays demonstrated a normalization of altered ionic currents in treated DM1 myotubes when compared with untreated controls, establishing a proof-of-concept for CRISPR interference strategies as potential therapeutic approaches for DM1.¹¹⁴

Interestingly, another study reported that endowing unmodified dCas9 proteins with gRNAs targeting expanded repeats, including those underlying DM1, selectively reduces pathogenic transcript buildup and foci accumulation with a concomitant rescue of the spliceopathy phenotype in DM1 cell culture models.¹¹⁵ Moreover, building on this finding, the authors have applied AAV6 vector particles for delivering and expressing dSaCas9 and gRNA constructs targeting DMPK CTG repeats in a well-established transgenic DM1 mouse model, i.e., HSA^LR mice.¹¹⁶ These in vivo experiments provided evidence for an amelioration of DM1 disease traits in a portion of the myofibers of transduced mice, including a reduction of RNA foci numbers and myotonia.¹¹⁵ The binding of dCas9 complexes to repeat tracts presumably forms a roadblock to RNA polymerase II elongation with this transcriptional interference being directly proportional to the repeat tract length and more pronounced when targeting the non-template DNA strand.¹¹⁵ These findings are reminiscent of recent data demonstrating that dCas9 binding within transcription units in mammalian cells causes RNA polymerase II pausing and premature transcription termination when the gRNA anneals to the non-template DNA strand.¹¹⁷

Generic epigenetic approaches for muscular dystrophies

Myostatin, a transforming TGF-β family member, has crucial roles in skeletal muscle development and homeostasis as, besides controlling myofiber numbers during embryogenesis, it regulates muscle mass postnatally.¹¹⁸ Indeed, myostatin signaling inhibition as a means of increasing muscle mass is being explored as a therapeutic approach for muscular dystrophies and for other disorders associated with muscle wasting. Myostatin inhibition using soluble factors, e.g., small-molecule drugs, neutralizing monoclonal antibodies, myostatin receptor decoys, and follistatin administration, a natural antagonist of myostatin, have all been pursued.¹¹⁹ Alternatively, studies on post-transcriptional small RNA-mediated interference and AON-mediated exon skipping have also been initiated. These approaches were proven to ameliorate dystrophic traits in some but not all muscular dystrophy mouse models and, significantly, clinical translation has not yet been realized.¹¹⁹ As an alternative to myostatin inhibition via soluble molecules and small RNAs, the dCas9-KRAB system has been probed to either upregulate Fst whose follistatin product favors as aforementioned muscle mass growth via myostatin inhibition⁹⁶ or, alternatively, downregulate Acvr2B encoding the myostatin receptor activin A receptor type 2B.¹²⁰ The direct downregulation of the myostatin gene itself, MSTN, deploying a dCas9-KRAB approach has also been pursued for enhancing skeletal myogenesis. Yet, instead of a muscular dystrophy, this approach was tested in in vitro and in vivo rat models of obesity-associated stress urinary incontinence.¹²¹

Looking ahead

Beyond the group of muscular dystrophies currently being targeted by RNA-programmable epigenetic editors, there is in principle a broader scope for testing these CRISPR-based epigenetic modulating agents as candidate therapeutic options for this class of conditions with unmet medical needs. For instance, in caveolae-associated muscular dystrophy, an autosomal dominant form of LGMD classified as LGMD1C, the muscle-specific caveolin-3 protein, is altered. In contrast, caveolin-1 is ubiquitously expressed except in striated muscle cells. Due to their high similarity, caveolin-1 could potentially act as a functional substitute for caveolin-3.¹²² Another potential target is provided by the HSP70 chaperon. Specifically, in cell models of oculopharyngeal muscular dystrophy, an autosomal dominant late-onset myopathy characterized by nuclear aggregates of a polyalanine-expanded poly(A)-binding protein (PABPN1), it has been shown that pharmacological induction or ectopic expression of HSP70 correlates with a reduction of mutant PABPN1 aggregation.^123,124

Finally, modulating multiple disease-modifying targets via CRISPR activation or interference effectors combined with gRNA multiplexing might offer improved protection against pathological processes. Alternatively, integrating CRISPR activation with CRISPR interference effectors is yet another combinatorial principle that has still to be explored as a means of unleashing complementary mechanisms counteracting pathological endpoints. For instance, in the context of FSHD, one can envisage combining CRISPR interference at pathogenic DUX4 alleles with CRISPR activation of the FSHD disease modifier SMCHD1, an epigenetic silencing regulator whose binding to D4Z4 repeats is known to naturally counteract the hallmark of FSHD, i.e., unscheduled DUX4 expression in skeletal muscle.¹²⁵ In the context of DMD instead, utrophin upregulation using CRISPR activation could be complemented with CRISPR interference directed at silencing miRNAs known to post-transcriptionally inhibit UTRN expression in skeletal muscle.^126,127

Finally, the increasing knowledge regarding the key role of non-coding RNAs (e.g., miRNAs, circular RNAs, and long non-coding RNAs) on the post-transcriptional and epigenetic regulation of normal and abnormal muscle physiology, vastly enlarges the number of candidate targets for CRISPR-based gene modulation interventions.¹²⁸

Space for improvement and outlook

Despite the broad therapeutic potential of CRISPR-based (epi)genome editing approaches, supported by the buildup of promising results obtained in preclinical settings, limitations and bottlenecks warranting further research and development can be readily identified. This is especially so in the case of muscular dystrophies where the need to deliver therapeutic agents to extensive portions of affected tissue is often further compounded by the various and complex pathological processes occurring at different stages of disease progression. Hence, although phenotypic complexity provides a rationale for testing combinatorial approaches where the versatility and multiplexing capabilities of CRISPR-based epigenetic editors might be particularly valuable, it will be crucial first and foremost to achieve uniform, ideally, systemic delivery of the attendant tools. To this end, developments on viral vectors and non-viral or synthetic vectors with myotropic character, i.e., capable of achieving robust muscle entry with minimal capture by non-target tissues, such as the liver, are at a premium. Rationale engineering and directed evolution methodologies are making notable strides in this regard via the assembly and isolation, respectively, of myotropic AAV capsids.¹⁰³ It will be equally important to expand muscle-targeting capabilities to other delivery vehicles especially those that, similarly to AAVs, lack viral genes but that instead of AAVs have high packaging or loading capacity, e.g., high-capacity adenovectors,^24,129 extracellular vesicles such as exosomes,¹³⁰ and certain classes of lipid nanoparticles (LNPs).¹³¹ The loading capacity is particularly relevant in that CRISPR-based epigenetic modulators are assemblies of multiple components that might be ultimately refractory to miniaturization. Furthermore, single high-capacity vehicles allowing for all-in-one delivery should also facilitate the stoichiometric assembly of the components underlying combinatorial (epi)genetic modulation.¹³² Next to expanding the loading capacity, it will be likewise crucial to test the safety, efficacy, and immunogenicity resulting from delivering different types of molecular cargoes, i.e., DNA or mRNA encoding the epigenetic modulators versus the delivery of these reagents directly as ribonucleoprotein complexes.

Potential immunological roadblocks include (1) preexisting neutralizing immunity to Cas proteins,¹³³ viral vector components,^134,135 or LNP formulations (e.g., polyethylene glycol),¹³⁶ (2) innate sensors of exogenous DNA or RNA structures (e.g., cGAS-STING pathway effectors and Toll-like receptors), and (3) adaptive immune responses to foreign macromolecules. Regarding the latter aspect, studies in canine models of DMD reported Cas9-specific cytotoxic T-lymphocyte responses following AAV vector administrations. The infiltration of immune cells clears the initial rescue of dystrophin expression, thus representing a critical barrier for AAV-CRISPR gene therapy in large mammals.¹³⁷ In addition to prokaryotic Cas proteins, most CRISPR-based epigenetic modulators contain moieties derived from viral proteins. Hence, besides developing hypoimmunogenic Cas proteins,¹³³ it should be valuable to “humanize” effector domains, namely, via substituting viral for human, yet potent, TADs.⁵⁹ To further mitigate immunological consequences epigenetic editors should, as aforementioned, ideally work in a hit-and-run fashion whereby, upon target site engagement, bona fide epigenetic-driven chromatin remodeling ensues, yielding stable gene activation or repression. In dystrophic muscle subjected to cell turnover, hit-and-run epigenetic remodeling can be achieved via mRNA or RNP delivery or, alternatively, DNA delivery that, ideally, incorporates tissue-specific promoters and/or drug-dependent control devices.^{138,139,140,141} Despite their higher complexity, the latter systems offer the possibility for fine-tuning gene modulator levels in a strict tissue-specific manner.

Finally, boding well for targeted epigenetic interventions, recent experiments involving the administration of LNPs loaded with epigenetic repressors achieved heritable and durable Pcsk9 silencing in mouse livers with a concomitant control of circulating cholesterol levels.¹⁴² These data, besides expanding on earlier in vitro experiments,^73,143 have provided a proof-of-principle for targeted epigenetic interference with a pathologic trait, in this case hypercholesterolaemia, in an in vivo setting.¹⁴² Instead, it remains to be conclusively established that a similar heritable and stable epigenetic approach can be achieved by using a CRISPR activation system in a therapeutic context.

Concluding remarks

CRISPR activation systems specifically addressed to compensatory and/or disease modifier targets using single or multiple gRNAs open the perspective for stand-alone or complementary epigenetic therapies. Besides their non-mutagenic character, the appeal of these systems includes the fact that, via exploiting preexisting endogenous genetic information in target cells, they bypass the need for the delivery of large transgenes, such as those encoding dystrophin (>11 kb), whose sizes are not compatible with the space available in commonly used delivery vehicles, e.g., AAV vectors (i.e., ∼4.8 kb). Moreover, therapeutic gene expression is directly placed under endogenous regulatory sequences. Next to candidate therapeutic approaches involving CRISPR activation of compensatory and/or disease modifier targets, CRISPR interference systems can instead be specifically addressed to pathogenic gain-of-function (toxic) genes or alleles by capitalizing on gRNA tailoring to preexisting allelic polymorphisms or repetitive tracts. In conclusion, notwithstanding the numerous identified and to-be-identified obstacles on the path to fruition, rapidly evolving CRISPR activation and CRISPR interference systems are emerging as a complementary, powerful, and versatile set of tools for tackling genetic disorders, including those that like muscular dystrophies are associated with varied and, often, complex pathogenic traits.

Acknowledgments

This project has received funding from the European Union’s Horizon Europe research and innovation program under the Marie Skłodowska-Curie Actions grant agreement no. 101072427 (GetRadi – Gene Therapy of Rare Diseases). Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the European Research Executive Agency can be held responsible for them. This work was also supported by the Duchenne Parent Project NL (22.012). Research in the author’s laboratory is further supported by the Prinses Beatrix Spierfonds, and the Dutch Research Council (NWO)—Open Technology Program. S.C.G.S. held a fellowship from the European Union’s ERASMUS+ program. The figures were assembled with the aid of Biorender.com.

Author contributions

All authors co-wrote, edited, and approved the manuscript. S.C. and S.C.G.S. assembled the figures.

Declaration of interests

The authors declare no competing interests.