Gene Editing for Duchenne Muscular Dystrophy Using the CRISPR/Cas9 Technology: The Importance of Fine-tuning the Approach

To the editor:

Four articles have been published recently that describe the use of CRISPR/Cas9 technology to mutate the Duchenne muscular dystrophy (DMD) gene.^1,2,3,4 In each study the authors deleted exon 23 of the mdx mouse by inducing double-strand breaks in the introns flanking this exon. Exon 23 contains a multiple of three nucleotides and a nonsense mutation. Thus, removal of exon 23 maintains the normal reading frame and restores dystrophin (Dys) expression as the nonsense mutation is deleted. However, the resulting Dys protein does not possess proper structure because of disruption of the spectrin-like repeats (SLRs) within the central rod domain. The amino acid junctions of the α-helices do not correlate with the DNA sequence of the exons, and when exon 23 is deleted there is a fusion of helices A and B of repeat 6 with the helix A of repeat 7, thereby missing a helix C.

These studies represent proof of principle that CRISPR/Cas9 technology can be used to correct the DMD gene and to restore the expression of an internally truncated Dys protein. However, this exon-deletion strategy will in most cases not result in an adequate SLR structure, as has been the case for the exon-skipping strategy using antisense oligonucleotides (AONs), which failed during clinical trials.

The 2.4-Mb DMD gene contains 79 exons coding for a very complex Dys protein.⁵ The central part of the protein is made of a rod domain containing 24 SLRs, each comprising three α-helixes (A, B, and C) forming a coiled-coil structure.^6,7 Each α-helix has a heptad structure with hydrophobic amino acids located in positions 1 and 4. The limits of the coding sequences of these helices do not correspond precisely to the limits of the exons. Whereas the N-terminal region of Dys is required for interacting with actin, a portion of exons 42–45 is required to interact with nNOS, and the C-terminal region interacts with the dystroglycan complex and syntrophin.⁸

Seventy percent of DMD patients have a deletion of one or more exons within the DMD gene that leads to a premature stop codon and the absence of the Dys protein.⁹ Patients with Becker muscular dystrophy (BMD) also have a deletion of one or several exons in the DMD gene,⁷ but because the latter deletions do not cause a frameshift, an internally deleted Dys protein is translated. These BMD patients nevertheless have more or less severe symptoms, depending on the structure of the resulting Dys.⁶ As demonstrated by Le Rumeur's group, the expression of a Dys protein with an inadequate SLR leads to a severe BMD,⁶ especially when the interaction with nNOS is abrogated.⁵ In the mdx mouse model only SLR 6 is affected by exon 23 removal, whereas most of the mutations in DMD patients are located in the hot-spot region containing exons 45–55 corresponding to SLRs 16–22. nNOS is known to interact with SLRs 16 and 17, and its correct interaction with Dys seems to prevail in BMD patients with less severe phenotypes.

A potential treatment for DMD involves therapeutic exon skipping. AONs are used to remove the exon flanking the patient deletion during splicing of the messenger RNA, so as to restore the normal reading frame and thus allow the expression of an internally deleted Dys.¹⁰ This approach has two drawbacks: (i) AONs must be readministered throughout the patient's life, and (ii) as for BMD, this approach may lead to a Dys protein containing an abnormal SLR and therefore a dysfunctional protein. Despite more than 360 articles on this approach in mice and cultured cells, a phase III clinical trial was halted by GlaxoSmithKline, because it failed to significantly improve the patient phenotype.¹¹ The failure of this clinical approach may be due in part to the fact that exon skipping aims to restore the reading frame of the dystrophin messenger RNA without paying attention to the structure of the resulting SLRs of the Dys protein. Alternative hypotheses are insufficient AON dosing or efficiency, potential toxicity of the AONs, poor delivery to the heart, and difficulties in reversing preexisting muscle weakness. However, it should be noted that the commercialization of an AON treatment was recently approved by the US Food and Drug Administration following a phase I clinical trial on only 12 patients. This approval was quite controversial.¹²

We recently published a fine-tuned approach using the CRIPSR/Cas9 technology to correct the DMD gene.¹³ Instead of cutting within introns, we induced double-strand breaks in the exons that flank the patient deletion. Careful selection of the target sequences for the guide RNAs made it possible to delete the intron sequence and parts of these exons, resulting in the formation of a hybrid exon that not only restores the normal reading frame but also encodes a Dys protein with an adequate SLR containing a normal succession of helices A, B, and C—and thus a correct heptad structure—and producing a protein that functions adequately. Figure 1 illustrates such a hybrid SLR formed by cutting in exons 50 and 54 with the CRISPR/Cas9 technique.¹³

Figure 1.

A hybrid spectrin-like repeat formed by cutting in exons 50 and 54 with the CRISPR/Cas9 technique.

CRISPR/Cas9 technology may lead to the development of a permanent treatment for DMD. Future experiments developing CRISPR/Cas9 therapy for DMD will have to be done in an adequate mouse model such as the hDMD/mdx mouse,¹⁴ containing in the mdx background the human DMD gene with deletion of one or several exons as observed in DMD patients. It will be essential to demonstrate that, following systemic delivery of the Cas9 gene and two guide RNAs with an adeno-associated viral vector, there is correction of a sufficient percentage of DMD genes to rescue the dystrophic phenotype. Moreover, as indicated by VandenDriessche and Chuah,¹⁵ the Cas9 protein has to be expressed only transiently to avoid an immune response and accumulation of off-target mutations.