Gene Editing Applications as Future Cardiovascular Therapies

Abstract

Cardiovascular disease is the leading cause of global morbidity and mortality, despite advances in pharmacological and surgical interventions. The emergence of CRISPR-Cas9 genome editing technology offers promising approaches for correcting genetic causes of hereditary cardiovascular disorders and modulating pathogenic signaling pathways implicated in various heart diseases. However, several challenges with respect to in vivo delivery of gene editing components, as well as important safety considerations, remain to be addressed in the path toward possible clinical application. We review current gene editing strategies, their potential therapeutic applications in the context of a variety of cardiovascular disorders, and their respective merits, limitations, and regulatory considerations. The rapid advances in this field combined with the many opportunities for deploying gene editing therapies for cardiovascular disorders augur well for the future of this transformative technology.

1. INTRODUCTION

Cardiovascular disease is the leading cause of morbidity and mortality worldwide, accounting for approximately one-third of all deaths (126, 127, 146). Despite the efficacy of conventional treatments, including drug and surgical interventions, there remains a dire need for new therapeutic strategies (139). The advent of CRISPR-Cas9 (clustered regularly interspaced short palindromic repeats–CRISPR-associated protein 9) gene editing technology offers a promising therapeutic approach by enabling precise modifications of DNA sequences of disease-causing genes. Gene editing has facilitated in vivo modeling of cardiovascular disorders in animal models and human cells, providing platforms for therapeutic development. This review considers the current state of CRISPR technology and its potential applications for cardiovascular medicine, while also addressing the important challenges associated with its clinical translation.

2. GENE EDITING STRATEGIES

CRISPR-Cas9, originally identified in the prokaryotic adaptive immune system, recognizes and cleaves foreign nucleic acids using Cas enzymes, which function as sequence-specific targeting proteins and nucleases (53). Over the past decade, CRISPR-Cas9 has opened unprecedented therapeutic opportunities, owing to its precision, programmability, and applicability in diverse cell types. Current CRISPR-Cas9 genome editing tools include nucleases, base editors, and prime editors, enabling not only genetic modification but also gene regulation and epigenetic modification via CRISPR interference (CRISPRi) and CRISPR activation (CRISPRa). Additionally, RNA editing using CRISPR-Cas13 is gaining interest as a strategy for modifying messenger RNA (mRNA) sequences.

2.1. Cas Nucleases

The CRISPR-Cas9 system comprises a Cas9 endonuclease and a single guide RNA (sgRNA). The sgRNA directs Cas9 to a specific genomic target through base-pairing with a complementary DNA sequence (33, 53). Cas9 recognizes DNA sequences by identifying a protospacer adjacent motif (PAM) adjacent to the sgRNA-complementary sequence. The PAM, positioned precisely near the target site, is essential for Cas9 binding and catalytic activity. Upon recognizing both the PAM and the target sequence, Cas9 creates double-strand breaks (DSBs) in the DNA near the PAM site (134).

The traditional CRISPR-Cas9 system, which induces DSBs within the protospacer region of the DNA target sequence, has been extensively used for gene knockout, insertion, or replacement (135). CRISPR-Cas9-induced DSBs activate diverse cellular repair mechanisms, including nonhomologous end joining (NHEJ) and homology-directed repair (HDR) (7). NHEJ, the predominant DSB repair pathway, occurs in both proliferating and quiescent cells and involves direct ligation of DSB ends by the endogenous DNA repair machinery (15, 161). Although more efficient than HDR, NHEJ is error prone, often causing small insertions or deletions (INDELs) that can introduce frameshifts or premature stop codons in mRNA transcripts, potentially resulting in gene inactivation. Conversely, HDR is a more precise mechanism that can replace DNA sequences when a specific template is provided (Figure 1a).

Figure 1.

CRISPR-based genome and transcriptome editing tools. (a) The CRISPR-Cas9 system consists of a Cas9 endonuclease and an sgRNA. The sgRNA directs Cas9 to specific genomic targets by base-pairing with complementary DNA sequences, inducing DSBs at specific DNA positions. DSBs are repaired by NHEJ or HDR in the presence of a donor template. (b) In base editing, nCas9 is fused with a deaminase to achieve single-nucleotide editing. A CBE induces editing from C-G to T-A, and an ABE induces A-T to G-C editing. A CGBE creates C-G to G-C edits. An AYBE induces editing from A-T to C-G or A-T to T-A. (c) Prime editing, which allows the introduction of various DNA sequences at the target site, is performed with a PE and a pegRNA. The PE consists of an nCas9 fused with engineered RT. The pegRNA is composed of an sgRNA that anneals to a target site, a scaffold for nCas9, an RTT designed for the intended edits, and a PBS that binds to the nontarget strand. (d) In RNA base editing, the REPAIR system is comprised of a fusion protein that integrates a dCas13 with an adenosine deaminase acting on the ADAR2_DD, which facilitates A-to-I editing. dCas13 specifically binds to single-stranded RNA and is guided by an sgRNA, which specifies the target A by inducing an A-C mismatch within the mRNA-sgRNA duplex. The RESCUE system is composed of a fusion protein that integrates a dCas13 with an eADAR2_DD, which facilitates C-to-U editing. dCas13 specifically binds to single-stranded RNA and is guided by an sgRNA, which specifies the target C by inducing a C-C or C-U mismatch within the mRNA-sgRNA duplex. Abbreviations: A, adenine; ABE, adenine base editor; ADAR2_DD, adenosine deaminase acting on RNA 2 deaminase domain; AYBE, adenine transversion base editor; C, cytosine; Cas, CRISPR-associated protein; CBE, cytidine base editor; CBGE, C-to-G base editor; CRISPR, clustered regularly interspaced short palindromic repeats; dCas13, catalytically dead Cas13; DSB, double-strand break; eADAR2_DD , engineered adenosine deaminase acting on the RNA 2 deaminase domain; G, guanine; HDR, homology-directed repair; INDEL, insertion or deletion; I, inosine; mRNA, messenger RNA; nCas9, Cas9 nickase; NHEJ, nonhomologous end joining; PAM, protospacer adjacent motif; PBS, primer binding site; PE, prime editor; pegRNA, prime editing guide RNA; REPAIR, RNA editing for programmable A-to-I replacement; RESCUE, RNA editing for specific C-to-U exchange; RT, reverse transcriptase; RTT, reverse transcriptase template; sgRNA, single guide RNA; T, thymine, U, uracil.

Recent studies have challenged the traditional understanding that HDR is limited to proliferating cells in the S or G2 phases of the cell cycle by demonstrating that adeno-associated virus (AAV)-mediated delivery of HDR components can promote precise targeted integration in postmitotic cardiomyocytes (57). Intracardiac delivery of sgRNA and repair templates using high-dose AAV facilitated HDR in adult murine cardiomyocytes (65). Another study corroborated this finding by demonstrating efficient in vivo HDR in neonatal and mature murine cardiomyocytes following subcutaneous AAV injection (166). Although it is still relatively inefficient, these studies demonstrate the potential for precise genetic correction via HDR in mature cardiac tissue, indicating future therapeutic possibilities in the postnatal heart.

CRISPR-Cas9 gene editing also has the potential to treat heterozygous disease-causing missense mutations by inactivating dominant-negative mutant alleles (28). To silence the mutant allele, the Cas nuclease, guided by the sgRNA, can be directed to induce DSBs at the target site, which is then repaired by the NHEJ pathway. The consequent introduction of INDELs, causing frameshift mutations and premature stop codons, can create truncated, nonfunctional proteins, thereby silencing the mutant allele. However, this method may not always produce nonfunctional proteins, and some mutant alleles may retain function. Moreover, this approach requires absolute precision so as not to inadvertently silence the wild-type allele.

2.2. Base Editing

Base editing converts one DNA base into another without inducing DSBs, thereby reducing the risk of INDELs and subsequent frameshift mutations associated with DSB repair mechanisms (7, 115). Base editors are fusion proteins that combine a deaminase enzyme with either a nickase Cas9 (nCas9) or a catalytically dead Cas9 (dCas9) (121). Several recent studies have highlighted the potential of base editors for the correction of point mutations associated with various cardiovascular diseases (18, 109, 118, 122). Two classes of base editors were initially developed for converting cytosine (C):guanine (G)–to–thymine (T):adenine (A) and A:T-to-G:C base pairs: C base editors (CBEs) and A base editors (ABEs), respectively (40, 66) (Figure 1b).

Base editors function through DNA binding, R-loop formation, and targeted deamination. The Cas9 component, guided by an sgRNA, binds to the target DNA, creating an R-loop that exposes approximately five nucleotides of single-stranded DNA (ssDNA) for deamination (115). In CBEs, cytidine deaminases convert Cs within the R-loop into uracils (Us), which are read as Ts during DNA replication, thus achieving precise C-to-T edits (67). ABEs use deoxyadenosine deaminases such as TadA to convert adenosines to inosines (Is), read as Gs, enabling specific A-to-G edits without additional repair components (40, 121). This process creates mismatched base pairs, prompting the DNA repair machinery to complete the transition mutations. Nevertheless, these editors cannot generate transversion mutations, such as A-to-T or C-to-A alterations, emphasizing the need to develop additional base editors to introduce broader nucleotide substitutions.

Recently developed C-to-G base editors (CGBEs) have expanded the range of possible base conversions, allowing for C-to-G base transformations in mammalian cells (64, 68, 165) (Figure 1b). CGBEs comprise three key components: nCas9, a cytidine deaminase, and a U N-glycosylase (UNG) that specifically excises the U base. Unlike CBEs that use a U-DNA glycosylase inhibitor (UGI) protein to suppress UNG and promote C-to-T conversions, CGBEs exclude the UGI to facilitate C-to-G transversions (66, 165). During CBE editing, C-to-G conversion occurs as a by-product of a C-to-T mutation, likely due to the DNA repair of apurinic/apyrimidinic (AP) sites caused by the removal of U bases by UNG. To reduce C-to-G conversion frequency, UGI protein was used to inhibit UNG (66). By contrast, the CGBE exploits the AP state to achieve the C-to-G conversion (136, 165). The efficiency of CGBEs varied by architecture and target site, achieving up to 72.1% editing efficiency with potential off-target rates lower than 0.4% in mammalian cells (136).

Tong et al. (142) developed the first A transversion base editor (AYBE, Y = C or T) for efficient A-to-C and A-to-T edits in mammalian cells (Figure 1b). The AYBE comprises an ABE and N-methylpurine DNA glycosylase (MPG). MPG excises hypoxanthine from I, creating an AP site that activates the base excision repair (BER) pathway. The BER machinery recognizes this site and initiates repair processes, leading to A-to-C or A-to-T conversion through the incorporation of C or T at the AP site, with up to 72% conversion efficiency. This innovation addressed the previous limitation of base editors in A-to-C or A-to-T changes, showing reduced bystander edits and effectiveness across various mammalian cell types (23, 142). T base editors (TBEs) have been developed as another base editing technology, allowing the conversions of T-to-C or T-to-G, thereby broadening the scope of possible nucleotide modifications (141).

Despite significant advancements in precision genome editing, concerns remain regarding potential off-target effects, such as bystander editing and promiscuity. Base editors typically function within a 3- to 10-nucleotide editing window, modifying multiple nucleotides and potentially causing unintended edits of neighboring bases (115, 121). In particular, early-generation base editors showed genome-wide off-target deamination, potentially leading to harmful modifications. For instance, CBEs have been observed to induce off-target C-to-T conversions at frequencies 20-fold higher than those of spontaneous deamination in mouse embryos, while also causing extensive transcriptome-wide RNA C deamination in human cells, affecting 38–58% of the expressed genes, with editing frequencies ranging from 0.07% to 100% (45, 170). To address this issue, optimizing the deaminase domain is a promising approach. Variants such as YE1-BE3 reduced bystander editing while preserving high on-target efficiency (62). Similarly, introducing the V106W mutation in the TadA-8e domain significantly reduced off-target effects from 1.9–6.7% to 0.32–1.3% without compromising on-target efficiency (123). Continued research focusing on both diversification and improved precision will be crucial for realizing the full potential of base editing in therapeutic applications.

2.3. Prime Editing

Prime editing, a precise and versatile genome editing technique, overcomes the limitations of base editing by enabling all possible nucleotide changes, including transitions and transversions. Prime editors (PEs) also demonstrate the capacity to generate targeted, precise INDELs (7, 8). This system consists of PEs and a prime editing guide RNA (pegRNA). PEs are fusion proteins containing an nCas9 with engineered reverse transcriptase (RT) (Figure 1c). The pegRNA is a multifunctional component of the prime editing system, consisting of a spacer sequence, a scaffold for nCas9, an RT template, and a primer binding site (PBS). The spacer sequence guides nCas9 to the target site, whereas the RT template (RTT) at the extended 3′ end contains the desired genetic modification. The PBS hybridizes to the nicked 3′ end of the target DNA, initiating reverse transcription. This design allows the pegRNA both to direct the editing machinery to the genomic target sequence and to provide the template for the RT domain, producing an edited DNA strand (24). Consequently, a single pegRNA plays the roles of both the sgRNA and the DNA template in HDR-mediated genome editing. Furthermore, prime editing exhibits remarkably low off-target effects due to three checkpoints of complementary base-pairing for productive editing (24).

Recent advancements in pegRNA design have significantly improved prime editing efficiency. Engineered pegRNAs (epegRNAs) with structured RNA motifs at their 3′ termini enhance stability and prevent degradation of essential components, thereby improving editing efficiency in various cell types without increasing off-target editing (107). Additionally, bioinformatic tools such as PRIDICT have been developed to predict pegRNA efficiencies based on high-throughput screening data, enabling the efficient selection of optimal designs for specific edits (93, 94). Local chromatin environments affect prime editing outcomes across different genomic locations, leading to the development of models such as ePRIDICT, which demonstrates the significant impact of chromatin on prime editing (78, 94).

Despite the versatility of PEs in executing diverse genetic modifications, their editing efficiency is lower than that of base editors, especially in vivo (29, 83, 107). Therefore, it is important to consider when devising strategies both the desired editing and the required editing efficiency for therapeutic efficacy. Notably, for certain pathological conditions, even relatively modest editing efficiencies may suffice for therapeutic benefits. In a mouse model of Duchenne muscular dystrophy (DMD), an editing efficiency of ~15% has shown potential therapeutic effects (110). This underscores the need to tailor the gene editing approach to each disease context, balancing the precision of the editing with achievable efficiency.

2.4. Gene Regulation Using CRISPR Interference and CRISPR Activation

CRISPR-Cas9 can also be utilized to modulate gene expression through CRISPRi, using a dCas9 to inhibit gene expression without causing DSBs in the DNA (117). The CRISPRi system, composed of dCas9 and sgRNA, binds to the promoter region or open reading frame of the target gene, inhibiting transcription by disrupting the binding of transcription factors or obstructing RNA polymerase activity (31). To enhance the gene silencing efficiency in mammalian cells, the Krüppel-associated box (KRAB) was fused to the C terminus of dCas9, resulting in significantly greater gene-silencing efficacy than with dCas9 alone (31). Dual AAV8s expressing dCas9-KRAB and sgRNA effectively silenced Pcsk9, a key cholesterol regulator, in adult mouse liver. This intervention significantly reduced serum Pcsk9 and cholesterol levels for up to 24 weeks posttreatment (140).

CRISPRa utilizes dCas9 fused to transcriptional activators to enhance gene expression without altering the DNA. This system directs the dCas9-activator complex to specific genomic loci, typically targeting regions upstream of the transcriptional start site of the desired gene. By recruiting the transcriptional machinery, CRISPRa effectively increases the transcriptional activity of endogenous target genes. Various fusion proteins, such as dCas9-VP64 and dCas9-VPR (a fusion protein of VP64, p65, and Rta), have been developed to optimize activation efficiency and significantly enhance gene expression levels (20, 87).

Despite their potential, research into the therapeutic applications of CRISPRi and CRISPRa has been less extensive than that of base editing and prime editing. This likely reflects the ability of more precise gene editing tools to introduce permanent changes to the genome, whereas CRISPRi and CRISPRa may be generally less effective and require long-term expression for persistent therapeutic applications.

2.5. RNA Editing

Despite remarkable advances in DNA editing technology, unintended off-target editing remains a concern, presenting significant risks for therapeutic applications (32). RNA editing with CRISPR-Cas13 offers safety advantages compared to DNA editing, as it is transient and avoids permanent heritable changes, thereby reducing the risk of lasting unintended effects (27). This transient nature of RNA modifications is particularly significant in conditions that require rapid adjustments to gene expression without long-term consequences (79). Furthermore, in contrast to Cas9, Cas13 does not require a PAM sequence at the target site, offering greater flexibility in target selection. RNA-editing systems based on dCas13 enable transcription-level editing without introducing DSBs in DNA, thereby enhancing their safety profile (2, 27). In RNA base editing, a fusion protein consisting of dCas13 and adenosine deaminases acting on the RNA 2 deaminase domain (ADAR2_DD) facilitates A-to-I or C-to-U conversion at the RNA level [named the REPAIR (RNA editing for programmable A-to-I replacement) and RESCUE (RNA editing for specific C-to-U exchange) systems, respectively] (2, 27) (Figure 1d).

Despite these advantages, RNA editing faces challenges, such as relatively low editing efficiency, which is influenced by the secondary structures of both the sgRNA and target RNA (63). Additionally, in genes that are expressed at very high levels, as for cardiac-specific contractile genes, even low genomic-level editing efficiencies can result in significant changes at the transcript level due to the high expression of these genes (71, 72, 109). This amplification effect is not observed in RNA editing, where modifications are made directly to the transcripts. Moreover, despite the transient nature of RNA editing, off-target effects remain a significant concern (133).

3. THE GENETICS OF HEART DISEASES

Cardiovascular diseases are classified into monogenic and polygenic disorders based on the nature of the underlying mutations. Monogenic diseases, such as DMD, familial hypertrophic cardiomyopathy (HCM), and dilated cardiomyopathy (DCM), result from pathogenic mutations in single genes, allowing genome editing to target these specific mutations in the affected tissues (21, 144). By contrast, polygenic diseases, including common conditions such as coronary artery disease and heart failure, arise from multiple genetic and environmental factors. Therapeutic strategies for these conditions focus on editing noncausal genes to introduce beneficial variants or protective modifications. This section presents examples of representative genetic cardiovascular diseases and their corresponding gene editing strategies in human cells and animal models.

3.1. Duchenne Muscular Dystrophy

DMD is a severe X-linked genetic disorder characterized by progressive muscle degeneration and cardiac involvement (113). It predominantly affects young males, with diagnosis typically occurring between ages 3 and 5 years. Nearly all patients exhibit cardiac involvement by age 18, leading to premature death (114). Currently, there is no cure, and treatments mainly focus on reducing symptoms and improving quality of life (114).

DMD is caused by mutations in the dystrophin (DMD) gene, which encodes dystrophin, a muscle-specific membrane protein essential for myofiber integrity. Dystrophin functions as a shock absorber during myofiber contraction, linking the actin cytoskeleton to the contractile apparatus (35). The DMD gene, spanning 2.2 Mb and encompassing 79 exons, is one of the largest human genes, contributing to its high mutational frequency and the diverse range of mutations observed in patients (14, 54). Mutations include deletions (68.8%), duplications (11.2%), point mutations (10.4%), and INDELs (9.6%) (14, 21).

Exon skipping is a promising strategy for restoring dystrophin expression, allowing the bypass of out-of-frame exons and translation of the essential C terminus of the protein. The US Food and Drug Administration (FDA)-approved oligonucleotide-mediated exon-skipping therapies target mutations in exon 45, 51, or 53 in DMD patients. These drugs modulate precursor mRNA (pre-mRNA) splicing to produce truncated but partially functional dystrophin proteins. However, they require weekly intravenous administration and result in low levels of truncated dystrophin expression without dramatic phenotypic improvements (56, 80, 147). Moreover, while these treatments may slow the decline of muscle function, they do not represent a curative solution, and long-term improvements in patients’ functional abilities have not yet been definitively demonstrated (56, 80, 147). Therefore, challenges remain in fully addressing the needs of DMD patients.

CRISPR-Cas9 gene editing technology offers a promising method for the correction of pathogenic mutations in DMD, potentially providing a permanent therapeutic strategy to ameliorate the multifaceted pathological manifestations of the disease (110). Therapeutic strategies for DMD have focused on editing the DMD gene in postnatal muscle cells, primarily using AAV9 to deliver CRISPR components for testing therapeutic effects in vivo (3, 5, 9, 13, 17, 60, 84, 100, 101, 106, 138, 163). Initial studies deployed AAV9 to deliver Cas9 and a pair of sgRNAs, targeting the 3′ and 5′ ends of exon 23, resulting in exon 23 skipping and restoration of the open reading frame and recovery of dystrophin expression in a DMD mouse model (84). Work from our lab further demonstrated the efficacy of CRISPR-Cas9 gene editing in a canine DMD model using AAV9, achieving up to 92% restoration of dystrophin levels in the heart (3). Additionally, systemic delivery of AAV9 with CRISPR-Cas9 components reframed exon 45 in a DMD mouse model lacking exon 44, resulting in lifelong dystrophin expression and enhanced muscle durability. Stable gene correction and minimal off-target effects at 18 months highlight its potential as a durable therapeutic strategy for DMD (60).

Recent advances have extended the repertoire of editing strategies for DMD to base editing. An optimized ABE was employed to disrupt the splice donor site of exon 50 in Dmd, using an AAV9 split-intein trans-splicing system for larger payload delivery, resulting in exon 51 skipping and functional dystrophin expression (22). Base editing of DMD using engineered CBEs and base editing in a humanized DMD mouse model have also been reported (75, 81). In parallel, a prime editing–based approach was used to reframe DMD exon 52 by precisely inserting two nucleotides within the exon sequence, which restored the reading frame and enabled functional dystrophin production in human induced pluripotent stem cell–derived cardiomyocytes (iPSC-CMs) (22). A recent study demonstrated the efficacy of gene editing in a pig DMD model. Systemic AAV9 delivery of split-intein Streptococcus pyogenes Cas9 (SpCas9) with 2 sgRNAs flanking exon 51 into 4-week-old pigs lacking exon 52 restored dystrophin expression and prevented arrhythmias (103). Additionally, a rhesus macaque model of DMD was generated with mutations in exons 4 and 46, which are within DMD mutation hotspots (25). These advancements in large animal models highlight the potential to better mimic human DMD pathology, offering more translatable insights for therapeutic development.

RNA base editing has also been proposed as a therapeutic strategy for DMD. A mini-dCas13X-mediated RNA editing system (mxABE) achieved up to 84% A-to-I editing in mice, restoring dystrophin expression to over 50% of normal levels across multiple muscle tissues (76). In the heart, dystrophin expression was notably restored compared to the diaphragm and tibialis anterior muscles, with levels at approximately 60% of normal 3 weeks after AAV9 systemic injection, decreasing to 20% to 40% after 6 weeks, and further declining to 2% to 4% after 6 months. The basis for this decline in dystrophin expression over time remains to be determined but represents a potential therapeutic concern. Interestingly, A-to-I mutations in RNA were maintained throughout this period. These findings underscore the complexity of maintaining therapeutic dystrophin expression following gene editing, even with high editing efficiencies, while also highlighting the potential of RNA editing as a viable therapeutic strategy (76).

3.2. Hypertrophic Cardiomyopathy

HCM, the most common inherited cardiac disorder, involves left ventricular hypertrophy and can lead to heart failure, arrhythmias, and sudden cardiac death (89). There is no curative treatment for HCM except for heart transplantation. Although cardiac myosin inhibitors can partially mitigate the disease phenotype, their use is limited to specific patient populations and may also decrease systolic function and potentially exacerbate heart failure (52, 90).

HCM exhibits remarkable genetic and clinical heterogeneity, primarily caused by mutations in sarcomere protein genes, making it an important target for gene editing. The myosin heavy chain 7 (MYH7) gene, which encodes the β-myosin heavy chain protein, has been implicated in 35% to 40% of genotype-positive HCM cases (85). In contrast to humans, in which MYH7 is the predominant cardiac myosin, Myh6 is the most abundant cardiac myosin isoform in mice. Several studies have shown that the HCM phenotypes in mice harboring the pathogenic mutation R404Q in Myh6 or R403Q in MYH7 were mitigated by AAV9 delivery of a base editor to correct the respective mutations (18, 86, 122). AAV9 delivery of a base editor corrected Myh6^R404Q in mouse embryos, demonstrating effective mitigation of the HCM phenotype (86). Similarly, Chai et al. (18) targeted MYH7^R403Q using iPSC-CMs from HCM patients and a humanized mouse model and found that ABE delivery by AAV9 significantly mitigated HCM-associated pathological manifestations in newborn mice. In another study, AAV9-packaged ABE was used to correct the R403Q mutation in two HCM mouse models, R403Q-129SvEv and R403Q-129SvEv/S4, representing insidious and rapid-onset phenotypes, respectively. AAV9-ABE injection at 10–13 days of age significantly reduced the disease phenotype in both models. By contrast, Cas9 nuclease-mediated silencing of the mutant allele resulted in deleterious consequences, highlighting the advantages of the ABE system for HCM treatment (122). Cardiac myosin-binding protein C 3 (MYBPC3) is another significant HCM-causing gene, and Wu et al. (157) demonstrated that AAV9-ABE administration in a Mybpc3^R946X mouse model successfully ameliorated the HCM phenotype, including cardiac hypertrophy and dysfunction, with an editing efficiency of 9.56% at the DNA level.

Recent studies demonstrated the efficacy of RNA editing in DMD and HCM. Yang et al. (160) developed a high-precision CRISPR-Cas13d system (hpCas13d) that selectively targeted mutant Myh6 RNA in HCM. Using AAV9 delivery to cardiomyocytes in HCM mouse models via subcutaneous injection of 1 × 10¹¹ vg at postnatal day 3, they achieved allele-specific suppression of the mutant Myh6 without affecting the normal gene, effectively preventing cardiac hypertrophy in vivo (160) (Figure 2).

Figure 2.

Schematic strategies for rescuing monogenic and common cardiovascular diseases using CRISPR systems. In monogenic diseases such as hypertrophic cardiomyopathy, dilated cardiomyopathy, and hereditary arrhythmias, disease-causing variants lead to aberrant protein expression or gene knockouts. These mutations are then targeted for correction using gene editing tools, with the ultimate goal of developing gene editing therapeutic strategies. In common diseases such as ischemia/reperfusion injury and pressure overload heart failure, CRISPR technology can be applied to induce intentionally modified protein production or a gene knockout to disrupt pathogenic signaling. This technology enables not only the knockout of specific genes but also the precise regulation of protein activity. Abbreviation: CRISPR, clustered regularly interspaced short palindromic repeats.

3.3. Dilated Cardiomyopathy

DCM is a progressive condition characterized by enlargement and impaired contractility of the ventricles, particularly the left ventricle, which is the primary cardiac pump. This leads to reduced cardiac output, potentially resulting in heart failure, arrhythmia, and sudden cardiac death. Like HCM, DCM has no curative treatment except for heart transplantation (96). DCM is associated with mutations in more than 40 genes that encode proteins critical for cardiac muscle function (59). Autosomal dominant mutations in RNA-binding motif protein 20 (RBM20) have been associated with a particularly aggressive DCM, accounting for 2% to 6% of cases of familial DCM (42, 59, 125).

Recent studies have demonstrated the potential of gene editing techniques in addressing DCM. Nishiyama et al. (109) successfully used base and prime editing of pathogenic RBM20 mutations causing DCM, focusing on several mutations: R634Q and R636S in human iPSC-CMs and R636Q in a mouse model of severe DCM (which is equivalent to the R634Q mutation in humans). Correction of R634Q using ABE restored appropriate splicing patterns and proper nuclear localization of the RBM20 protein. Additionally, prime editing effectively corrected the R636S mutation in vitro, expanding the repertoire of targetable mutations. In a mouse model of the homozygous R636Q mutation with a severe DCM phenotype, AAV9-mediated ABE delivery at postnatal day 5 restored cardiac contraction and alleviated the DCM phenotype, achieving 66% editing efficiency in the heart at the complementary DNA (cDNA) level (109). Likewise, Grosch et al. (43) restored 75% of RBM20 nuclear localization in cardiomyocytes using ABE, significantly enhancing the editing efficiency with AAVMYO, an AAV serotype with increased cardiac tropism. Additionally, Cas13b RNA-mediated therapy using AAV9 successfully rescued the phenotype in DCM mice with the pathogenic R141W/+ mutation in TNNT2. Although there was no specific mention of the duration of therapeutic effects or off-target effects, this approach enabled efficient and specific knockdown of mutant TNNT2 (R141W) transcripts in vivo (77) (Figure 2).

3.4. Hereditary Arrhythmias

Inherited arrhythmias, caused by mutations in ion channel genes, represent a significant subset of genetic cardiac disorders that is responsible for 18% of sudden cardiac deaths (47). Current therapeutic approaches primarily involve β-blockers and implantable cardioverter-defibrillators (ICDs), although these interventions demonstrate only partial effectiveness (95).

Recent studies highlight the potential of CRISPR-mediated technology as a targeted approach for treating hereditary arrhythmias in vivo. For instance, knock-in mice with the PRKAG2^H530R mutation, linked to familial Wolff-Parkinson-White (WPW) syndrome, underwent systemic delivery of AAV9-sgRNA and AAV9-Cas9 to disrupt the mutant gene and mitigate arrhythmic phenotypes (158). Additionally, CRISPR-Cas9 was utilized to correct the R176Q/+ mutation in the ryanodine receptor 2 (RyR2) using AAV9 delivery at postnatal day 10, successfully preventing catecholaminergic polymorphic ventricular tachycardia (CPVT) (112). In another study, AAV9 encoding CRISPR-Cas9 components specifically targeted the R14del allele in humanized phospholamban (PLN)-R14del adult mice, which exhibit DCM associated with malignant arrhythmias. This intervention significantly improved cardiac function and reduced susceptibility to arrhythmias (28). Furthermore, AAV9 delivery of an ABE corrected the T1307M mutation in sodium voltage-gated channel alpha subunit 5 (Scn5a) in a long-QT syndrome type 3 (LQT3) mouse model. This treatment mitigated QT prolongation and reduced susceptibility to arrhythmias, including torsades de pointes and ventricular tachycardia (118) (Figure 2).

3.5. Blocking Cardiovascular Disease by Editing Pathogenic Signaling Pathways

Correcting pathogenic mutations, which are often heterogeneous and occur at low frequency, limits their broad applicability. Additionally, individual genes often have multiple mutations, requiring different gene editing strategies (49). Considering the effort and cost involved in targeting a single gene mutation, developing separate gene editing strategies for individual gene mutations is impractical, highlighting the need for more comprehensive gene editing approaches that target multiple mutations simultaneously (129). Instead of correcting a specific mutation, disrupting key pathogenic signaling pathways activated by various mutations can broaden the applicability of a single strategy. Notably, this approach could be applied to common nongenetic disorders by modifying noncausal genes to introduce beneficial variants or protective alterations (70-72, 104).

Hypercholesterolemia, characterized by elevated low-density lipoprotein cholesterol (LDL-C) levels, is a major risk factor for atherosclerotic cardiovascular disease, including myocardial infarction. While genetic factors, such as familial hypercholesterolemia (FH), play a crucial role in some cases, lifestyle choices, such as diet and physical activity, also contribute substantially to elevated cholesterol levels (44). Proprotein convertase subtilisin/kexin type 9 (PCSK9), synthesized in hepatocytes, is crucial in cholesterol metabolism and is a key target for new hypercholesterolemia treatments (108). Infrequent gain-of-function mutations in PCSK9 have been established as causative factors for FH (1). Conversely, loss-of-function variants in PCSK9, occurring in 2–3% of specific ethnic populations, are associated with lower plasma LDL-C levels and confer significant protection against coronary heart disease without adverse effects (26). Recent innovations in CRISPR-based gene editing have enabled new therapeutic strategies for hypercholesterolemia by targeting PCSK9 (30, 104). VERVE-101, a novel CRISPR-based gene editing therapy, has shown promise in early clinical trials (55). This single intravenous infusion of lipid nanoparticles (LNPs) containing sgRNA and ABE aims to permanently inactivate PCSK9 in hepatocytes, potentially offering a long-term solution for patients with refractory hypercholesterolemia. This approach could provide long-lasting LDL-C reduction without ongoing medication, addressing both genetic and lifestyle-induced hypercholesterolemia.

Another strategy to intervene in the pathogenesis of heart disease is to disrupt major signaling pathways activated in pathological conditions. Chronic hyperactivation of Ca²⁺/calmodulin–dependent protein kinase IIδ (CaMKIIδ) contributes to various cardiac disorders, including ischemia/reperfusion (I/R) injury, heart failure, cardiac hypertrophy, and arrhythmias, by disrupting Ca²⁺ homeostasis and prompting inflammatory or apoptotic signaling and fibrosis (10, 82, 105). Thus, the modulation of CaMKIIδ activity represents a promising therapeutic approach for various heart diseases.

CaMKIIδ undergoes oxidation and phosphorylation upon activation (36, 98). To disrupt its pathological activation, a phosphoresistant Camk2d T287A mouse model was generated, which exhibited cardioprotection against pressure overload–induced heart failure. Human iPSC-CMs harboring the CAMK2D T287A mutation were also protected from β-adrenergic stress (70) (Figure 2). Similarly, a CAMK2D mutant that was resistant to oxidative activation showed cardioprotection against I/R injury. Editing of both oxidation-sensitive methionines (M281V and M282V) conferred maximal cardioprotection against I/R stimuli in human iPSC-CMs. AAV9-mediated delivery of Camk2d editing components to adult mice post-I/R injury enhanced cardiac function and reduced apoptosis and fibrosis (72). For clinical translation, a humanized CAMK2D knock-in mouse model was developed. Intracardiac administration of editing components via MyoAAV 2A, a vector with efficient myocardial transduction, post-I/R injury demonstrated cardioprotective effects (71, 137) (Figure 2). These examples of the blockade to pathological signaling via gene editing suggest further potential applications for this approach.

Some cardiac diseases, such as myocarditis and takotsubo cardiomyopathy, are transient and typically do not require long-term treatment after recovery. Consequently, the necessity of genetic correction of the mutation diminishes once the disease is cured. However, the long-term effects of these permanent genetic alterations on physiological states after recovery remain poorly understood. Therefore, the choices for gene editing therapy must be considered carefully. Further research is required to assess the impact of these interventions on cellular homeostasis and systemic function over extended time periods (Table 1).

Table 1.

Summary of CRISPR-mediated gene editing strategies to rescue cardiovascular disease phenotypes in vivo

Disease	Animal model(s)	Strategy	CRISPR tool(s)	Delivery method(s)	Reference(s)
Duchenne muscular dystrophy (DMD)	mdx mice	Exon deletion	Cas9	AAV9	84, 106, 138
	Dmd ΔEx50 mice	Exon reframing	Cas9	AAV9	5
	mdx4cv mice	HDR and exon deletion	Cas9	AAV6	13
	DMD ΔEx50 dogs	Exon reframing and exon skipping	Cas9	AAV9	3
	Dmd ΔEx44 mice	Exon reframing and exon skipping	Cas9	AAV9 (including scAAV)	9, 60, 101, 163
	Dmd ΔEx43, 45, and 52 mice	Exon reframing and exon skipping	Cas9	AAV9 (including scAAV)	100
	Dmd ΔEx44 mice	Exon skipping	ABE8e-nSpCas9-NG	AAV9	17
	DmdΔ Ex51 mice	Exon skipping	ABEmax-SpCas9	AAV9	22
	DMDΔEx54mdx mice	Exon skipping	aTdCBE	AAV9	75
	DMDΔmE5051,hE50KI/Y mice	Exon skipping	ABE8e	AAV9	81
	DMDΔEx52 pigs	Exon deletion	Cas9	AAV9	103
	DMDEx30mut mice	PTC suppression	mxABE	AAV9	76
Hypertrophic cardiomyopathy (HCM)	Myh6R404Q mice	Nucleotide correction	ABEmax-NG	AAV9 and microinjection	86
	MYH7R403Q mice	Nucleotide correction	ABEmax-VRQR	AAV9	18
	R403Q-129SvEv and R403Q-129SvEv/S4 mice	Nucleotide correction	ABE8e	AAV9	122
	Mybpc3R946X mice	Nucleotide correction	SpRY-ABE8e	AAV9	157
	Myh6R872H/+ and Myh6R872H/R404Q mice	Disruption of the mutant allele	hpCas13d	AAV9	160
Dilated cardiomyopathy (DCM)	Rbm20R636Q mice	Nucleotide correction	ABEmax-VRQR-SpCas9	AAV9	109
	Rbm20R636Q and Rbm20R635L mice	Nucleotide correction	ABEmax and ABE8e	AAVMYO	43
	TNNT2R141W mice	Disruption of the mutant allele	hpCas13d	AAV9	77
Hereditary arrythmias	PRKAG2H530R mice	Disruption of the mutant allele	Cas9	AAV9	158
	RyR2R176Q/+ mice	Disruption of the mutant allele	Cas9	AAV9	112
	PLN14del mice	Disruption of the mutant allele	Cas9	AAV9	28
	Scn5aT1307M mice	Nucleotide correction	ABEmax	AAV9	118
Hypercholesterolemia	Wild-type mice (target: Pcsk9)	Gene silencing	dCas9-KRAB	AAV8	140
	Wild-type mice (target: Pcsk9)	Gene inactivation	Cas9	Adenovirus	30
	Wild-type mice and monkeys (target: Pcsk9)	Gene inactivation	ABE8.8	LNP	104
	Wild-type mice (target: Pcsk9)	Gene inactivation	ABE8e	eVLP	11
	Wild-type mice (target: Pcsk9)	Gene inactivation	PE	eVLP	6
Pressure overload heart failure	Camk2d T287A knock-in mice	Ablation of phosphorylation site	ABE8e-NG	NA	70
I/R cardiac injury	Wild-type mice (target: Camk2d)	Ablation of oxidation sites	ABE8e-SpRY	AAV9	72
	Humanized CAMK2D knock-in mice	Ablation of oxidation sites	ABE8e-SpRY	MyoAAV 2A	71

Abbreviations: AAV, adeno-associated virus; ABE, adenine base editor; Cas, CRISPR-associated protein; CBE, cytosine base editor; CAMK2D, calcium/calmodulin-dependent protein kinase II delta; DCM, dilated cardiomyopathy; DMD, Duchenne muscular dystrophy; DMD, dystrophin gene; eVLP, engineered viral-like particle; Ex, exon; HCM, hypertrophic cardiomyopathy; HDR, homology-directed repair; hpCas13d, high-precision CRISPR-Cas13d system; I/R, ischemia/reperfusion; KRAB, Krüppel-associated box; LNP, lipid nanoparticle; mxABE, mini-dCas13X-mediated RNA editing system; Mybpc3, myosin-binding protein C3; Myh6, myosin heavy chain 6; MYH7, myosin heavy chain 7; NA, not applicable; Pcsk9, proprotein convertase subtilisin/kexin type 9; PE, prime editor; PLN, phospholamban; PRKAG2, protein kinase AMP-activated noncatalytic subunit gamma 2; PTC, premature termination codon; Rbm20, RNA-binding motif protein 20; RyR2, ryanodine receptor 2; scAAV, self-complementary AAV; Scn5a, sodium voltage-gated channel alpha subunit 5; TNNT2, troponin T2, cardiac type.

4. CRISPR GENE EDITING MODELS OF CARDIOVASCULAR DISEASE

The advent of high-throughput sequencing technologies, including whole-exome and wholegenome sequencing (WGS), has dramatically expanded our capacity to identify genetic variations associated with cardiovascular diseases. These methods have uncovered numerous variants, such as single-nucleotide polymorphisms, insertions, deletions, and structural variations, potentially linked to heart disorders. However, the vast amount of genetic data complicates distinguishing pathogenic mutations from benign variants. CRISPR-Cas9 technology allows for the investigation of specific variants and their roles in cardiovascular diseases. This section explores the application of CRISPR-Cas9 technology in correcting and studying cardiovascular diseases across two key experimental platforms: iPSCs and their derived cardiomyocytes and mouse models.

4.1. Human Induced Pluripotent Stem Cells and Induced Pluripotent Stem Cell–Derived Cardiomyocytes

The inherent postmitotic nature of mature cardiomyocytes poses significant challenges to analyzing human cardiac tissue. The development of human iPSCs has overcome this limitation and provided a renewable and scalable source of cardiac cells (130). iPSCs provide a powerful platform for investigating human cardiovascular diseases in vitro by enabling the replication of patient-specific genetics and phenotypes of various cell types, particularly in iPSC-CMs (128). CRISPR-Cas9 genome editing provides a useful tool for cardiac disease modeling and characterizing genetic variants. This technology enables two critical applications in cardiac disease modeling. One approach involves creating corrected iPSC lines that only differ from the original patient-derived mutant iPSC lines at the locus of interest, allowing precise studies of specific genetic alterations and their phenotypic consequences without confounding factors. Another approach introduces disease-causing mutations into wild-type iPSCs, producing mutant lines that mimic patient mutations with the same genetic background as wild-type iPSCs. This is valuable for studying rare genetic variants or for creating cellular models of diseases for which patient-derived iPSCs are not available. These applications of CRISPR-Cas9 in iPSC-based cardiac disease modeling have significantly advanced our understanding of various cardiovascular disorders, including channelopathies and cardiomyopathies (18, 109, 156).

Moreover, the combination of iPSC technology and CRISPR-Cas9 gene editing has opened new possibilities for personalized medicine. Patient-specific cellular models allow researchers to assess the efficacy and safety of potential therapies in a more relevant context, potentially resulting in more targeted and effective treatments.

4.2. Mouse Models

Since the first CRISPR-Cas9-modified mouse model was developed in 2013, this technology has enabled rapid and efficient generation of both knockout and knock-in mouse models (151, 159). Traditional embryonic stem (ES) cell–based methods require lengthy culture processes that can take over a year for a single gene modification, whereas CRISPR-Cas9 can shorten the timeline to a few months for direct editing in fertilized eggs. Furthermore, CRISPR-Cas9 allows simultaneous editing of multiple genes, which is challenging in ES cells because of the complexities of targeting and validating each modification, often causing mosaicism. This efficiency and versatility have enabled the creation of complex genetic models, expediting our understanding of cardiovascular genetics (131).

Advancements of CRISPR-Cas9 technology have enabled more refined methods for cardiacspecific gene editing in mice. Carroll et al. (16) developed transgenic mice expressing SpCas9 under the control of the cardiac-specific α-myosin heavy chain (Myh6) promoter. They achieved efficient gene editing by intraperitoneal injection of AAV9 encoding an sgRNA targeting the Myh6 gene, successfully suppressing Myh6 expression and resulting in cardiac failure. In a different approach, Rosa26-Cas9-GFP mice expressing a Cre-inducible Cas9-P2A-GFP fusion protein were generated. Injection of AAV9 encoding sgRNAs and cardiac troponin T (cTnT) promoter–driven Cre into neonatal mice enabled cardiomyocyte-specific CRISPR-Cas9 gene editing (48).

In addition, as mentioned in Sections 3.2 and 3.3, our lab successfully performed cardiacspecific base editing through systemic injection of dual AAV9s encoding sgRNAs and a split base editor under the control of the cTnT promoter. This approach effectively corrected the mutations of Rbm20 and MYH7 that cause DCM and HCM, respectively (18, 109).

Importantly, this technology has enabled the development of humanized mice by replacing specific genomic regions, which contain specific exons or mutations, with human genetic sequences. This approach allows for human genome editing tools to be used in vivo but cannot predict potential off-target or unintended human editing events. This capability is particularly important for targeting hotspot regions with disease-causing mutations, providing a more comprehensive assessment of the specificity and safety of the genome editing strategy (18, 76).

Interspecies genotype–phenotype differences between mice and humans are important considerations in translational research. For instance, RBM20 mutations cause severe DCM phenotypes in humans with heterozygous mutations, whereas mice typically require homozygous mutations to exhibit similar phenotypes (42). This discrepancy may be attributed to species-specific biological differences, such as redundant pathways or compensatory mechanisms in mice that mitigate the effects of single-mutant alleles. Indeed, mice display greater genetic variation than humans (99). These interspecies variations extend beyond genotype–phenotype relationships to include differences in gene expression patterns, metabolic pathways, and physiological parameters such as heart size and heart rate (124). Consequently, therapeutic interventions that demonstrate efficacy in mouse models may not directly translate to humans because of these biological differences. Moreover, mouse models often simplify disease mechanisms by employing single gene mutations, whereas human pathologies are induced by complex genetic and environmental interactions. This methodological limitation reduces the predictive value of mouse studies for human therapeutics. To bridge this interspecies gap and enhance the translational potential of preclinical studies, large animal models are becoming increasingly important. These models offer physiological and anatomical similarities to humans, potentially providing more accurate predictions of therapeutic outcomes in clinical settings.

5. CHALLENGES OF GENE EDITING

Genome editing technology is now beginning to be applied to patients in clinical settings, and clinical trials are underway (162). However, there are several challenges to achieving clinical application, such as suitable delivery strategies and potential off-target effects.

5.1. Delivery Challenges

The delivery of genome editing components remains a major obstacle in therapeutic applications. AAV is currently the most widely used viral vector for delivering therapeutic agents owing to its safety profile and high transduction efficiency in various tissues. However, AAV faces several limitations. The inherent packaging capacity constraint of AAV vectors, limited to approximately 4.7 kb, presents a significant challenge for delivering SpCas9 and its associated sgRNA (91). This limitation has necessitated the development of dual-vector systems, wherein the Cas9 and sgRNA components are packaged separately (4). Alternatively, a split-intein trans-splicing system has been used, allowing the Cas9 protein to be divided and delivered using two AAV vectors (18, 22, 72, 109). However, editing efficiency may be affected by the need for both components to enter the cell for complete reconstitution of the system. Moreover, these approaches often require higher viral doses, which can lead to potential cardiotoxicity and other adverse effects (34, 73, 154). In clinical trials for DMD, the doses of AAV used have typically ranged from 5 × 10¹³ to 2 × 10¹⁴ vg/kg, highlighting the substantial viral load required for therapeutic effect (97). To mitigate these limitations, the use of smaller Cas variants that fit within a single AAV (all-in-one vectors) is an option. This reduces virus dose and production costs but often results in less editing efficiency (83, 164). Self-complementary AAV (scAAV) allows for up to a 20-fold reduction in viral dose for effective delivery and is more resistant to degradation than single-stranded AAV DNA (163). However, the smaller packaging capacity of scAAV (< 2.4 kb) restricts its use for sgRNA delivery (155).

Capsid modification is a promising approach to enhance AAV transduction efficiency. Recent studies have explored various strategies to engineer AAV capsids for improved cardiac and muscle targeting. Screening of 183 AAV variants in adult mice identified AAVMYO, a mutant AAV9 capsid with improved efficiency and specificity for muscle tissues, including heart (153). Tabebordbar et al. (137) developed MyoAAV, a new class of AAV capsids, which shows superior efficiency and selectivity in transducing muscle tissues and heart. Importantly, improved transduction efficiency with AAVMYO and MyoAAV allows for lower viral doses while maintaining therapeutic efficacy, potentially mitigating dose-related adverse effects.

The host immune response remains a significant obstacle in gene therapy using AAV. Both innate and adaptive immune responses to AAV capsids and transgene products can lead to the elimination of AAVs, thereby reducing their therapeutic efficacy (120). Preexisting antibodies in humans, due to prior exposure to wild-type AAVs or previous treatments, can neutralize systemically delivered AAVs, preventing them from entering target cells and posing a major obstacle to their effective clinical application (120). Additionally, excessive levels of neutralizing antibodies can result in adverse immune responses, including an increase in proinflammatory cytokine secretion and complement activation, which can complicate treatment outcomes (132).

To address the host immune response in AAV-mediated gene therapy, several strategies have been developed. Plasmapheresis, which removes circulating antibodies, including neutralizing antibodies against AAV, can enhance AAV delivery to target cells (111). However, this method has several challenges, such as the necessity for multiple sessions, the easy replacement of the antibody pool (the antibody rebound effect), and rendering patients susceptible to infections (111, 120). Immunosuppressive agents, including corticosteroids, can control the immune responses to AAV-mediated gene therapy but also raise the risk of infection or viral reactivation (148). Modifying AAV capsid proteins to alter surface epitopes can also help evade immune detection, making engineered capsids less recognizable by preexisting antibodies and improving delivery and efficacy (12) (Figure 3a).

Figure 3.

Comparison of various delivery strategies for gene editing components in cardiomyocytes. (a) AAV-based delivery of gene editing components is the most widely practiced method. Advantages include tissue specificity, high effectiveness, and preclinical and clinical experience. Disadvantages include limited packaging capacity, preexisting immunizations, high manufacturing costs, a single dose limitation, and viral integration. (b) LNPs have greater packaging capacity and can deliver DNA, mRNA, protein, or RNP. Other advantages include transient expression of cargo, low immunogenicity, scalable and inexpensive production, and allowance of multiple doses, while disadvantages include lack of tropism, a tendency to aggregate, and low efficiency in some tissues. (c) VLPs have advantages over both viral and nonviral delivery. VLPs exhibit high packaging capacity, allowing the delivery of mRNA, protein, or RNPs. They also show a low risk of viral genome integration and minimal off-target editing. Disadvantages include lack of in vivo evidence, instability, high manufacturing costs, and complex manufacturing techniques. Abbreviations: AAV, adeno-associated virus; LNP, lipid nanoparticle; mRNA, messenger RNA; RNP, ribonucleoprotein; VLP, virus-like particle.

Gene editing strategies that create DSBs in the target gene raise significant safety concerns owing to AAV integration at the target site. To address the limitations of AAV vectors in delivering genome editing components, alternative delivery methods should be explored. LNPs offer several advantages over AAVs, such as large packaging capacity that allows the delivery of mRNA, proteins, or ribonucleoproteins (RNPs) (32, 143). Optimizing the size, shape, coating, and surface chemistry of LNPs can enhance their efficacy without any preexisting immunity. Additionally, LNPs are more cost efficient and suitable for large-scale production compared to AAVs. LNPs enable transient expression, which reduces off-target editing, and allow multiple dosing due to lower immunogenicity (61, 119). Importantly, the delivery of LNPs has succeeded in clinical trials (41). However, despite their potential advantages, LNPs have significantly lower delivery efficiency than AAVs (102, 119). While LNPs show effective delivery to the liver, their delivery to the heart is limited, in part due to the absence of unique cell surface markers on cardiomyocytes aiding the penetration of nanoparticles (32, 108, 152). Thus, further advancements in nanoparticle properties are needed for efficient tissue-specific delivery to challenging target organs, including the heart (Figure 3b).

Viral-like particles (VLPs) represent a promising option for delivering gene editing components. These self-assembling, noninfectious structures mimic the form and size of a virus particle but lack viral genomes, offering high packaging capacity for the delivery of mRNA, proteins, or RNPs, with minimal off-target effects and low risk of viral genome integration (119). Most VLP architectures for mRNA or protein delivery are derived from retroviruses owing to their spherical shape and lack of rigid structural symmetry (119). Recently, engineered virus-like particles (eVLPs) for efficient in vivo base editing across multiple tissues were developed. Specifically, an eVLP base editor targeting DNA methyltransferase 1 (Dnmt1) achieved 53–55% editing efficiency through cerebroventricular injection into newborn mice, successfully installing a silent mutation in Dnmt1. Targeting Pcsk9 achieved 63% editing in liver tissue and reduced serum Pcsk9 protein levels by 78% following retro-orbital injections into adult mice. In the rd12 mouse model of genetic blindness, subretinal injection of ABE8e-NG-eVLPs corrected 21% of the retinal pigment epithelium 65 (Rpe65) R44X mutation, partially rescuing visual function (11). Furthermore, an optimized VLP (PE-eVLP) enabled efficient delivery of a prime editor in vivo. The eVLP prime editor introduced a 4-bp substitution at the Dnmt1 locus following intracerebroventricular injections into newborn mice, resulting in 47% editing. It also achieved 7.2% correction of the R44X mutation in Rpe65 and partially rescued visual function in the rd12 mouse model (6).

VLPs exhibit key characteristics of both viral and nonviral delivery technologies, making them a promising strategy to enhance the therapeutic potential of genome editing. However, several challenges must be addressed to achieve their clinical use. One significant challenge is the stability of recombinant particles, which is crucial for maintaining their integrity and function during storage and delivery. Compared to AAVs, VLP delivery lacks substantial evidence, particularly of effective delivery to cardiac tissue. Additionally, the safety profile of VLPs in vivo requires further characterization to ensure their suitability for therapeutic applications. Furthermore, the manufacturing costs of VLPs are high, and the process is complex (50, 119) (Figure 3c).

In summary, several means of delivering gene editors to target organs and cells in vivo are emerging, but none is ideal, and all have their strengths and weaknesses. The continued optimization of nonviral delivery strategies represents an important area of future investigation.

5.2. Off-Target Effects

Safety concerns are a significant hurdle in translating genome editing therapies in humans for clinical use. In utilizing CRISPR-Cas9 technology for cardiovascular disease correction, addressing off-target and off-organ effects is crucial, particularly in the heart, where off-target mutagenesis can cause fatal arrhythmic events, even from unintended mutations in a small number of cardiomyocytes (108). Recent advancements have developed several methods for the genome-wide unbiased identification of off-target events, including GUIDE-seq, Extru-seq, DISCOVER-Seq+, and Tracking-seq (69, 145, 167, 169). WGS is the most comprehensive method for detecting off-target mutations in humans, allowing direct comparison of genomic sequences before and after gene editing. However, the high cost and sensitivity needed to detect low-frequency off-target events pose significant limitations (46). Moreover, in light of individual sequence variations in coding and especially noncoding regions of the genome, possible off-target gene editing cannot be comprehensively predicted across large numbers of patients based on studies of representative genomes.

Additionally, continuous Cas9 expression in cells can lead to unintended off-target modifications, particularly with life-long gene editing strategies. Precise spatiotemporal control of Cas9 activity is therefore desirable. Various strategies have been explored to enable conditional regulation of the system, including small molecule activation, small molecule inhibition, cellspecific promoters, bioresponsive delivery carriers, gene regulation, and chemical and physical strategies such as light, thermal, ultrasound, and/or magnetic activation of the CRISPR-Cas9 system (168). While these approaches offer potential solutions, each has its limitations (168). In addition to various strategies for regulating Cas9 activity, anti-CRISPR (Acr) proteins, which inhibit CRISPR-Cas systems, provide an effective means of controlling gene editing. These proteins can inhibit Cas9, reducing off-target modification, but have several limitations, such as their large size, potential toxicity, and immunogenicity in Acr proteins themselves (88).

The complexity of monogenic diseases, let alone polygenic diseases, combined with potential unintended genomic alterations, complicates the prediction of the consequences of permanent genetic modifications. This uncertainty is exacerbated by our limited understanding of gene–environment interactions over time, potentially leading to unexpected phenotypic consequences or pathologies. Although RNA editing theoretically enables tunability and reversibility without permanent off-target effects, substantial transcriptome-wide off-targets have been observed in many RNA editing strategies (133).

5.3. Potential Immune Responses to Gene Editing Components

The CRISPR system, originating from bacteria, raises concerns regarding the potential immunogenicity associated with therapeutics using this technology. While immunogenicity is not necessarily serious, it can impact both the safety and efficacy of the therapy (37). Several studies have shown that a significant proportion of the human population possesses preexisting adaptive immune responses to Cas9. For instance, antibodies against SpCas9 and Streptococcus aureus Cas9 (SaCas9) were shown to be detectable in 58% and 78% of human blood samples, respectively (19). CRISPR components can activate both humoral and cellular immune responses, even in individuals with no preexisting immunity. Hakim et al. (51) showed that intramuscular and intravenous delivery of AAV8 and AAV9 containing CRISPR components in canine models of DMD led to the production of Cas9-specific antibodies and cytotoxic T-lymphocyte (CTL) responses. To address these challenges, researchers have deployed various strategies, including immunosuppressive administration, adaptation of Cas9-specific regulatory T cells, and immunosilencing of Cas9 (37, 38, 149, 150). Immune reactions in clinical trials have profoundly impacted gene therapy research, highlighting the need for CRISPR-based therapies with lower immunogenicity (92).

6. LOOKING TO THE FUTURE OF CARDIOVASCULAR GENE EDITING

Cardiovascular device therapy has significantly advanced with innovative technologies, yet the clinical landscape for cardiovascular diseases has remained relatively stagnant in terms of pharmacological treatment. Despite numerous potential therapeutic targets identified in cardiovascular research over the years, the clinical reality is that traditional medications such as antiplatelet drugs, diuretics, β-blockers, and angiotensin-converting enzyme (ACE) inhibitors remain the cornerstone of pharmacological treatment.

Gene editing technologies offer new therapeutic possibilities for treating hereditary diseases, potentially transforming human cardiovascular disease therapies. The recent FDA approval of the first CRISPR drug for sickle cell anemia underscores the potential of this technology in treating genetic disorders, paving the way for cardiovascular applications (39, 74). However, the clinical translation of CRISPR technology presents several challenges. A primary concern is determining the optimal therapeutic window for gene correction in cardiomyocytes, because the efficacy of genome editing in cardiac diseases may depend on the stage of disease progression. For example, homozygous Rbm20^R636Q mice treated at postnatal day 5 exhibited improvements in cardiac function that were comparable to those seen in heterozygous mice, indicating that earlier intervention might lead to even more significant benefits (109). This emphasizes the importance of timely therapeutic strategies, as earlier correction of mutations may enhance the likelihood of restoring normal cardiac function. Moreover, as cardiac disease progresses, associated fibrosis, inflammatory infiltration, and cardiomyocyte death diminish the efficacy of gene delivery.

The percentage of cardiomyocytes that require correction for clinical benefits remains undetermined and likely varies with the target gene and specific mutations. Adding to this uncertainty, CRISPR-Cas9-mediated cardiac editing using AAV vectors presents additional complexities. This approach may result in a mosaic pattern of genetic modifications and incomplete gene knockout (58). Such mosaic gene corrections in cardiomyocytes might potentially lead to electrical heterogeneity in the heart, thereby increasing the risk of arrhythmias (116). Cells with different genetic compositions can exhibit varying electrophysiological properties, potentially disrupting the coordinated electrical activity of the heart. There is also potential for compensatory remodeling in edited cells responding to dysfunction in adjacent unedited myocardium.

The multinucleated nature of human cardiomyocytes presents another challenge. As a consequence, multiple cardiomyocyte nuclei may need to be edited for complete disease correction, reducing overall efficiency. Fundamentally, the heart’s continuous function throughout life emphasizes the critical importance of understanding the long-term effects and safety of genome editing in this organ. In addition to technical challenges, the clinical translation of CRISPR technology raises ethical concerns regarding genetic modifications beyond therapeutic purposes. The complexity of this technology and the uncertainty of adverse events present challenges in obtaining informed consent from patients. Furthermore, the implementation of CRISPR technology may lead to increased treatment costs, potentially limiting its accessibility.

We expect that scientific advancements will ultimately overcome these obstacles through collective efforts. By carefully overcoming these challenges and taking advantage of each new approach, we may soon witness an era of precision medicine in the treatment of both monogenic and polygenic cardiovascular disorders.