CRISPR/Cas9 Flexes Its Muscles: In Vivo Somatic Gene Editing for Muscular Dystrophy

Duchenne muscular dystrophy (DMD) is a fatal disease of skeletal muscle and heart; patients typically die from cardiopulmonary failure with an average life expectancy of 25 years. This disease is attributed to mutations in the gene encoding dystrophin, a vital protein linking the cytoskeleton of the myofibers and cardiomyocytes to the extracellular matrix.^1,2 The lack of functional dystrophin leads to structural instability of the sarcolemma membrane, causing damage to the muscle tissues, which are eventually replaced by fat and fibrotic tissues. DMD has always been considered a prime target for gene therapy, because its genetic etiology is well understood and conventional treatments are not effective in halting disease progression. Gene addition strategies and antisense oligonucleotide (AON)-based exon skipping have been shown to ameliorate the disease in animal models but have not yet been translated into clinical efficacy. In recent issues of Molecular Therapy and Science, four independent teams demonstrated for the first time that somatic in vivo genome editing has the potential to correct this devastating genetic disease.^3,4,5,6

DMD is a recessive X-linked form of muscular dystrophy with a worldwide incidence rate of 1 in 5,000 males. An important challenge in developing an effective gene therapy for DMD is the large size of the dystrophin-coding sequence (11.1 kb). Consequently, it cannot be readily accommodated into most commonly used viral vectors for muscle-directed gene therapy, particularly adeno-associated virus (AAV) vectors, because of intrinsic packaging constraints. Because they do not have such a strict size limit, transposons have recently been shown to enable stable delivery of the full-length dystrophin coding sequence into dystrophic stem/progenitor cells ex vivo.⁷ Truncated forms of the dystrophin complementary DNA that fit into an AAV vector have also been used⁸ (recently reviewed by Kawecka et al.⁹). These truncated micro/mini-dystrophin proteins retain partial function and resemble the dystrophin variant in patients with Becker muscular dystrophy, a much milder form of the disease.

Alternatively, AON-mediated exon-skipping therapies have been designed to restore the reading frame of the dystrophin transcript by masking splice donor or acceptor sites,^10,11 so as to skip the defective exon that contains the disease-causing mutation. Consequently, this approach also results in a partially functional Becker-like dystrophin variant. However, AON-mediated exon-skipping strategies have not yielded the expected clinical outcome in pivotal trials, although increased dystrophin expression was observed.¹² Moreover, the lack of efficacy with AONs was compounded by poor transfection of cardiac tissue, variable efficiencies of tissue uptake, requirement for repeated AON injections to maintain effective skipping, and the potential for AON-associated toxicities.

Genome editing is an attractive strategy to overcome these limitations of AON-based exon skipping to achieve stable dystrophin expression by skipping the defective exon at the DNA rather than the RNA level. Each of the new studies exploits the CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 (CRISPR-associated protein 9) system, which is a versatile platform for precise modification of the genome. In the presence of guide RNAs (gRNAs), Cas9 is directed to specific sites in the genome adjacent to a protospacer adjacent motif, causing a double-strand DNA break (DSB).^13,14,15,16 This DSB causes variable “indel” mutations at the target site via nonhomologous end joining. Alternatively, the DSB significantly increases homology-directed repair if a DNA template is provided that has homology to the target site. However, homology-directed repair is believed to be inefficient in postmitotic cells, including myofibers in vivo.

Previous in vitro studies had shown that CRISPR/Cas9 could be used to achieve relatively efficient in vitro genome editing of a defective dystrophin gene in distinct muscle stem/progenitor cell populations and induced pluripotent stem cells^17,18,19,20 or by genetic modification of embryos.²¹ Building upon these previous in vitro results, the new studies provide proof of concept for in vivo gene editing of the dystrophin gene in dystrophic mdx mice using this CRISPR/Cas9 system.^3,4,5,6 This represents an important advance in the field of gene editing, as it demonstrates the potential of CRISPR/Cas9 for muscle-directed gene repair in vivo. Restoration of the dystrophin reading frame by CRISPR/Cas9 resulted in dystrophin expression in both skeletal muscle and heart. Interestingly, correction was not restricted to mature myofibers but also affected the endogenous myogenic precursor cell pool,⁶ which can replenish mature differentiated muscle tissue. This could be an advantage toward achieving a lasting therapeutic effect in DMD patients, especially given the characteristic rapid turnover of muscle cells. Expression of dystrophin after CRISPR/Cas9-meditated in vivo editing was associated with a partial functional correction of the dystrophic phenotype consistent with an amelioration of muscle strength.

Somatic in vivo editing required coexpression of Cas9 and gRNAs specifically designed to excise exon 23 of the defective dystrophin gene in mdx mice (Figure 1). Different gene delivery platforms were used to deliver the Cas9 and gRNA constructs into the target cells. Initial proof-of-concept studies relied on the use of in vivo electroporation or adenoviral vectors,³ whereas the subsequent studies were based on either AAV8 or AAV9,^3,4,5,6 representing a more clinically relevant gene transfer modality. Both AAV8 and AAV9 mediate efficient gene transfer in skeletal muscle and heart^22,23,24 after systemic administration, at least in mice, offering a major advantage over other vector systems.

Figure 1.

In vivo somatic muscle-directed editing with CRISPR/Cas9. The Cas9 and guide RNA (gRNA) expression cassettes were packaged into an adeno-associated virus (AAV) vector that exhibits improved myotropic and cardiotropic properties (AAV8 or AAV9). The Staphylococcus aureus or S. pyrogenes Cas9 was expressed from a cytomegalovirus (CMV) promoter (or a smaller derivative), whereas the gRNAs were expressed from the U6 pol III promoter. The gRNAs were specifically designed to target the defective exon 23 of the endogenous mouse dystrophin gene of the mdx mouse, which contains a nonsense mutation resulting in a defective dystrophin protein. The vectors were injected either systemically or locally into the muscle in adult or newborn mdx mice. NHEJ, nonhomologous end joining.

These preclinical studies represent an important step forward toward the future clinical use of CRISPR/Cas9, given that as many as 80% of patients suffering from DMD could ultimately benefit from having a faulty exon removed. The specific design of the CRISPR/Cas9 strategy described in these studies resulted in the synthesis of a truncated Becker-like dystrophin variant. Although Becker muscular dystrophy is a milder form of the disease, it can still cause significant disability. Consequently, these CRISPR/Cas9-based strategies do not fully reconstitute all of the essential functions of dystrophin.

Moreover, several key issues would need to be addressed before clinical trials can be envisaged. To achieve efficient editing in skeletal muscle after systemic AAV vector administration, mdx mice were injected with vector doses that are several orders of magnitude higher than what would be clinically acceptable (e.g., ~10¹⁵ vg/kg in neonatal mdx intraperitoneally).⁶ That problem is compounded by the requirement for coadministration of both Cas9 and gRNA vectors, whereas an “all-in-one” vector that contains both components appeared to be less efficient.⁶ Moreover, only a partial correction in muscle strength could be achieved, consistent with a relatively low percentage (<10%) of dystrophin expression, compared to normal wild-type mice. This could be attributed, at least in part, to the use of suboptimal promoters used to drive Cas9 expression and/or the gRNA. Nevertheless, future improvements in vector design and/or tropism may ultimately overcome these limitations. Although dystrophin was expressed in the heart after CRISPR/Cas9-based in vivo editing in the mdx mice, one possible caveat of these studies is that these levels may not suffice to achieve robust correction of the cardiac defects. This is important, because cardiopulmonary failure represents the major cause of death in DMD patients. Follow-up studies would need to ascertain the effectiveness of the cardiac-related correction and optimize it.

Ideally, the Cas9 protein and the gRNAs should be expressed only transiently so as to achieve in vivo gene editing. Long-term expression of the CRISPR/Cas9 components may not be desirable, as it increases the risk of off-target integrations in nontarget genes. It is reassuring that none of the in vivo studies in the mdx mice reported any significant off-target editing, at least in the computationally predicted putative “off-target” sites. However, these types of computational algorithms are not comprehensive, and genome-wide analysis of putative “off-target” effects induced by CRISPR/Cas9 will be needed to better assess this risk.^25,26 The use of truncated gRNA or the latest-generation high-fidelity Cas9 (ref. ²⁷) may further reduce the risk of “off-target” effects. Moreover, because Cas9 is a foreign protein derived from Staphylococcus pyrogenes or Staphylococcus aureus, it could potentially provoke an immune response that would ultimately result in the elimination of the gene-edited cells.²⁸ Nevertheless, it is reassuring that the effect of the gene editing appeared to be relatively sustained, at least in the mdx mouse models. Provided that these outstanding issues can be addressed and that the efficacy and safety of the in vivo genome editing can be confirmed in large (canine) animal models of muscular dystrophy, the strategy represents an important avenue worth exploring further.

It would be premature to speculate that these emerging in vivo gene editing approaches could ultimately replace the more conventional gene therapy approaches based on gene addition. The overall efficiency of somatic in vivo gene editing in the muscle using CRISPR/Cas9 still lags behind what can typically be achieved using more conventional strategies based on the systemic administration of AAV vectors expressing either micro-dystrophin or exon-skipping constructs.⁷ Moreover, the use of improved expression cassettes (Rincon et al.²⁹ and Sarcar S, Rincón MY, Evens H, Tipanee J, Keyaerts M, Loperfido M, et al. (2016), unpublished results), alternative myotropic or cardiotropic AAV serotypes, and optimized therapeutic transgenes may further increase the chances of success of these conventional gene therapy strategies. These advances in “vectorology” may ultimately also cross-fertilize and benefit the field of gene editing as well. The translation of gene editing from the bench to the bedside will ultimately depend on the progress in the field of gene therapy at large. The simultaneous exploration of these different therapeutic strategies in parallel is therefore warranted and offers the best hope for those patients and their families, who are blighted by these devastating muscle disorders.