Postnatal genome editing partially restores dystrophin expression in a mouse model of muscular dystrophy

See allHide authors and affiliations

Science  22 Jan 2016:
Vol. 351, Issue 6271, pp. 400-403
DOI: 10.1126/science.aad5725

Editing can help build stronger muscles

Much of the controversy surrounding the gene-editing technology called CRISPR/Cas9 centers on the ethics of germline editing of human embryos to correct disease-causing mutations. For certain disorders such as muscular dystrophy, it may be possible to achieve therapeutic benefit by editing the faulty gene in somatic cells. In proof-of-concept studies, Long et al., Nelson et al., and Tabebordbar et al. used adeno-associated virus-9 to deliver the CRISPR/Cas9 gene-editing system to young mice with a mutation in the gene coding for dystrophin, a muscle protein deficient in patients with Duchenne muscular dystrophy. Gene editing partially restored dystrophin protein expression in skeletal and cardiac muscle and improved skeletal muscle function.

Science, this issue p. 400, p. 403, p. 407


CRISPR/Cas9-mediated genome editing holds clinical potential for treating genetic diseases, such as Duchenne muscular dystrophy (DMD), which is caused by mutations in the dystrophin gene. To correct DMD by skipping mutant dystrophin exons in postnatal muscle tissue in vivo, we used adeno-associated virus–9 (AAV9) to deliver gene-editing components to postnatal mdx mice, a model of DMD. Different modes of AAV9 delivery were systematically tested, including intraperitoneal at postnatal day 1 (P1), intramuscular at P12, and retro-orbital at P18. Each of these methods restored dystrophin protein expression in cardiac and skeletal muscle to varying degrees, and expression increased from 3 to 12 weeks after injection. Postnatal gene editing also enhanced skeletal muscle function, as measured by grip strength tests 4 weeks after injection. This method provides a potential means of correcting mutations responsible for DMD and other monogenic disorders after birth.

Duchenne muscular dystrophy (DMD) is a fatal muscle disease affecting 1 in 3500 to 5000 boys. Cardiomyopathy and heart failure are common, incurable, and lethal consequences of DMD. The disease is caused by mutations in the gene encoding dystrophin, a large intracellular protein that links the dystroglycan complex at the cell surface with the underlying cytoskeleton, thereby maintaining integrity of muscle cell membranes during contraction (1, 2). In the absence of dystrophin, muscles degenerate, causing weakness and myopathy (3). Many therapeutic approaches for DMD have failed, at least in part because of the size of the dystrophin protein and the necessity for lifelong restoration of dystrophin expression in the myriad skeletal muscles of the body as well as the heart.

The CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 (CRISPR-associated protein 9) system allows precise modification of the genome and represents a potential means of correcting disease-causing mutations (4, 5). In the presence of single guide RNAs (sgRNAs), Cas9 is directed to specific sites in the genome adjacent to a protospacer adjacent motif (PAM), causing a double-strand break (DSB). When provided with an additional DNA template, a precise genomic modification is generated by homology-directed repair (HDR), whereas in the absence of an exogenous template, variable indel mutations are created at the target site via nonhomologous end joining (NHEJ) (6). Previously, we used CRISPR/Cas9 to correct a single nonsense mutation in Dmd by HDR in the germ line of mdx mice, which allowed the restoration of dystrophin protein expression (7). However, germline genomic editing is not feasible in humans (8) and HDR does not occur in postmitotic adult tissues, such as heart and skeletal muscle (9), necessitating alternative strategies of gene correction in postnatal tissues. Here, we devised a method to correct Dmd mutations by CRISPR/Cas9-mediated NHEJ (termed “Myoediting”) in postnatal muscle tissues after delivery of gene-editing components by means of adeno-associated virus–9 (AAV9), which displays high tropism for muscle (10, 11).

The dystrophin protein contains several domains (fig. S1), including an actin-binding domain at the N terminus, a central rod domain with a series of spectrin-like and actin-binding repeats, and WW and cysteine-rich domains at the C terminus that mediate binding to dystroglycan, dystrobrevin, and syntrophin (12). The actin-binding and cysteine-rich domains are essential for function, but many regions of the protein are dispensable (3). It has been estimated that as many as 80% of DMD patients could benefit from exon-skipping strategies that bypass mutations in nonessential regions of the gene and partially restore dystrophin expression (13). This approach has been validated in vitro by CRISPR/Cas9-mediated correction of Dmd mutations in patients’ induced pluripotent stem cells (14) and immortalized myoblasts (15). Similarly, adenovirus-mediated gene editing was shown to restore dystrophin expression in specific muscles of mdx mice after intramuscular injection (16), but adenoviral delivery is not therapeutically favorable (17).

Shown in Fig. 1A is the strategy whereby CRISPR/Cas9-mediated NHEJ can create internal genomic deletions to bypass the premature termination codon in exon 23 responsible for the dystrophic phenotype of mdx mice, potentially allowing reconstitution of the Dmd open reading frame. In principle, this approach could be applied to many mutations within the gene, including large deletions, duplications, and pseudoexons. An advantage of this approach is that it does not require precise correction of the disease-causing mutation. Instead, imprecise deletions that prevent splicing of mutant exons are sufficient to restore dystrophin protein expression.

Fig. 1 Permanent exon skipping in postnatal mdx mice by AAV-mediated Myoediting.

(A) Strategy for bypassing exon 23 of the mdx locus by NHEJ. (B) AAV vectors for expression of Cas9 (AAV-Cas9, upper), guide RNAs, and GFP (AAV-sgRNA, lower). ITR, inverted terminal repeat; RSV, Rous sarcoma virus promoter; U6, human U6 promoter. (C) Different modes of AAV9 delivery. Black arrows indicate time points for tissue collection after injection. (D) Rescue of dystrophin expression in mdx mouse by IM injection of Myoediting components. GFP and dystrophin immunostaining from serial sections of mdx mouse TA muscle are shown 3 and 6 weeks after AAV-IM injection of AAV-Cas9 sgRNAs at P12 (three male mdx mice in each group). Asterisks indicate serial section myofiber alignment. Scale bar, 40 μm. (E) RT-PCR of RNA from Myoedited mdx mice indicates deletion of exon 23 (termed ΔEx23, lower band) and shows increase in intensity of ΔEx23 bands from 4 to 12 weeks after AAV-RO injection (four male mdx mice in each group). Asterisk indicates the RT-PCR products with small deletions; M denotes size marker lane; bp indicates the length of the marker bands. (F) Sequence of the RT-PCR products of ΔEx23 band confirmed that exon 22 spliced directly to exon 24, excluding exon 23.

To test whether Myoediting could be adapted to skip the Dmd mutation in exon 23 in mdx mice, we first evaluated a pool of sgRNAs that potentially target the 5′ and 3′ ends of exon 23 (supplementary materials, fig. S2, and table S1). We co-injected Cas9 mRNA with sgRNA-mdx (directed toward the mutant sequence in exon 23) and either sgRNA-R3 or sgRNA-L8 (targeting the 3′ and 5′ end of exon 23, respectively) into mdx zygotes without a HDR template (fig. S3A). Strikingly, ~80% of progeny mice lacked exon 23 (termed mdx-ΔEx23) (fig. S3, B and C, and table S2), representing an increase in the efficiency of mdx editing relative to HDR (7). Genomic polymerase chain reaction (PCR) products from the target sites of exon 23 and reverse transcription PCR (RT-PCR) products of mdx-ΔEx23 mice were cloned and sequenced, confirming the skipping of exon 23 (fig. S3, D to F). As a result of skipping exon 23, the open reading frame of Dmd was restored, allowing dystrophin protein expression (fig. S4A). In mdx-ΔEx23 mice, serum creatine kinase levels (a measure of muscle membrane permeability) and grip-strength tests both showed restoration of muscle function (fig. S4, B and C). Control mdx mice without treatment (–) and mdx mice with Myoediting (+) were tested for potential off-target effects of Myoediting with sgRNA-R3 (fig. S5). Ten potential genome-wide off-target sites (OT-01 to OT-10) were predicted by the CRISPR design tool (; see supplementary materials and table S4) (7). Only the target site Dmd R3 of Myoedited mdx mice showed cleavage bands in the T7 endonuclease I (T7E1) assay, and no off-target effects were detected in the top 10 potential off-target sites (fig. S5).

To apply Myoediting to postnatal muscle tissues, we used AAV9, which displays tropism to cardiac and skeletal muscle (10, 11), to deliver Cas9 and sgRNAs to muscles of mice. AAV-guide RNAs were generated by cloning sgRNA-mdx and sgRNA-R3 into AAV-sgRNA vector containing a human U6 promoter and green fluorescent protein (GFP) (Fig. 1B). We generated AAV-Cas9 using a unique AAV-Cas9 vector (miniCMV-Cas9-shortPolyA) (18, 19), which uses a “mini”-CMV promoter/enhancer sequence to drive expression of the humanized Streptococcus pyogenes Cas9 (SpCas9). Different modes of AAV9 delivery and variations in timing of expression were systematically compared to identify the optimal method for Dmd Myoediting in postnatal mdx mice: (i) intramuscular (IM) at P12, (ii) retro-orbital (RO) at P18, and (iii) intraperitoneal (IP) at P1 (see supplementary materials) shown in Fig. 1C.

After IM injection of P12 mice with AAVs, muscle tissues were analyzed by immunostaining for dystrophin expression 3 weeks later (Fig. 1D and fig. S6A). Native GFP identified AAV-mediated gene expression in myofibers. Skeletal muscle from the IM-AAV–injected mice showed a mosaic pattern of dystrophin-positive fibers (Fig. 1D). The percentage of dystrophin-positive myofibers was calculated as a fraction of total estimated fibers. In the mdx mouse shown in Fig. 1D, 7.7 ± 3.1% of myofibers in the tibialis anterior (TA) muscle expressed dystrophin 3 weeks after IM-AAV injection (all errors reported are SD). Rescue increased to an estimated 25.5 ± 2.9% of myofibers by 6 weeks after IM-AAV injection (three male mdx mice per group) (fig. S6A). Hematoxylin and eosin (H&E) staining of muscle showed that histopathologic hallmarks of muscular dystrophy, such as necrotic myofibers, were diminished in TA muscle at 6 weeks after AAV delivery. Inflammatory cell invasion and centralized myofiber nuclei were minimal, in marked contrast to uninjected control mdx TA (fig. S6B).

RO injection of AAV into the venous sinus of the mouse represents an alternative to tail vein injection for the systemic administration via blood circulation in young mice. Muscle tissues from mice after RO-AAV injection at P18 were examined by RT-PCR (Fig. 1E). RT-PCR of RNA from Myoedited mdx mice showed that deletion of exon 23 (ΔEx23) allowed splicing from exon 22 to 24 (lower band) and the intensity of ΔEx23 bands was increased from 4 to 12 weeks after RO-AAV injection. Sequencing of RT-PCR products of the ΔEx23 band confirmed that exon 22 spliced to exon 24 (Fig. 1F). Muscle tissues were analyzed by immunohistochemistry (Fig. 2) and H&E staining (fig. S7). At 4 weeks after RO-AAV injection in mdx mice, 2.5 ± 1.1% of myofibers were dystrophin-positive, whereas 1.1 ± 0.3% of cardiomyocytes were dystrophin-positive. Progressive improvement with age was also observed from 4 to 8 and 12 weeks after RO-AAV injection. Rescue increased to an estimated 6.1 ± 3.2% of myofibers in TA muscle, and 5.0 ± 2.1% of cardiomyocytes, by 8 weeks after RO-AAV injection. At 12 weeks after injection, 4.6 ± 3.2% of myofibers were dystrophin-positive in TA muscle and 9.6 ± 3.9% of cardiomyocytes were dystrophin-positive. Western blot analysis confirmed the restoration of dystrophin expression in both heart and skeletal muscle (fig. S8).

Fig. 2 Rescue of dystrophin expression in postnatal mdx mice by retro-orbital injection of AAV-Cas9 sgRNAs.

(A) Dystrophin immunostaining of TA muscle is shown for wild-type (WT), mdx, and AAV-RO–treated mdx mice at 4, 8, and 12 weeks after injection (AAV-RO at P18, four male mdx mice in each group). TA muscle of unedited mdx control mice exhibits myonecrosis, indicated by cytoplasm-filling autofluorescence (highlighted with white asterisks). (B) Dystrophin immunostaining of the heart is illustrated for WT, mdx, and AAV-RO–treated mdx mice at 4, 8, and 12 weeks after injection (AAV-RO at P18, four male mdx mice in each group). Arrowheads indicate dystrophin-positive cardiomyocytes 4 weeks after AAV-RO injection into mdx mouse heart. Scale bar, 40 μm.

After IP injection of AAV editing components (fig. S9), dystrophin expression was rescued in 1.4 ± 1.2% of TA myofibers and 1.1 ± 1.1% of cardiomyocytes in treated mdx mice after 4 weeks. Higher percent correction was observed in mdx-injected mice at 8 weeks after IP-AAV injection with 1.8 ± 1.2% of dystrophin-positive myofibers and 3.2 ± 2.4% dystrophin-positive cardiomyocytes. Grip strength testing (see supplementary materials) showed a significant increase in strength of mdx mice at 4 weeks after IP-AAV injection relative to uninjected mdx controls (Fig. 3).

Fig. 3 Forelimb grip strength of mdx, mdx-AAV-IP, and wild-type mice 4 weeks after injection.

mdx, mdx-AAV-IP, and WT mice were subjected to grip strength testing to measure muscle performance (grams of force), and the mdx-AAV-IP mice showed enhanced muscle performance relative to mdx mice at 4 weeks of age (mdx male control, 34.7 ± 1.8%; mdx-AAV-IP male mice, 48.4 ± 2.5%; WT male, 71.8 ± 1.9%; mdx female control, 29.7 ± 1.4%; mdx-AAV-IP female mice, 45.5 ± 1.4%; WT female, 75 ± 2.4%). Numbers of mice in each group are labeled in the bar, six trials for each mouse. Data are means ± SEM. Significant differences between conditions are indicated (***P < 0.0005).

Semiquantitative immunohistochemistry was performed to quantify dystrophin expression levels, normalized to laminin, in wild-type (WT) and AAV-injected mdx mice (see supplementary materials). Integrated density measurements of sarcolemmal staining in TA myofibers showed dystrophin protein levels that were 23.7 ± 11.6% of WT after AAV-IP injection, 27.7 ± 6.6% of WT after AAV-RO injection, and 53.2 ± 18.5% of WT after AAV-IM injection (figs. S10 and S12A). Integrated density of dystrophin in cardiomyocytes showed dystrophin protein levels that were 52.4 ± 14.3% of WT after AAV-IP injection, 71.1 ± 21.0% of WT after AAV-RO injection, and 69.7 ± 19.8% of WT after AAV-IM injection (figs. S11 and S12B).

IM, RO, and IP injection all provide transducing potential in organs and muscle groups remote from the injection site, presumably through intravasculature circulation of AAV. Dystrophin expression in mdx mice was restored in vascular smooth muscle cells by all three modes of AAV delivery, but most effectively by RO (fig. S13A). In contrast, no mode of AAV delivery was able to cross the blood-brain barrier to restore dystrophin expression in hippocampal CA1/CA2 regions of mdx mice (fig. S13, B and C). AAV transduction across the blood-brain barrier and subsequent restoration of brain dystrophin expression will likely require other methods (2023). We also harvested sperm from AAV-injected male mdx mice and tested gene editing by T7E1 assay. No cleavage bands were detected (fig. S13D); however, more sensitive methods such as deep sequencing might be required to evaluate the risk of unexpected germline editing. Additionally, AAV vectors with tissue-specific promoters should enhance the safety of systemic gene editing.

Our results show that AAV-mediated Myoediting can rescue the reading frame and expression of dystrophin in postnatal mdx mice. The efficiency of restoration of dystrophin-positive myofibers increases with time, likely reflecting persistent expression of gene-editing components. Exon skipping by NHEJ-mediated genomic editing allows for the permanent removal of the disease-causing mutation and was about 10 times as efficient as gene correction by HDR (7). Myoediting by NHEJ does not require precise genetic modification. Instead, any types of indels that disrupt either a splice donor or acceptor sequence in a mutant exon result in exon skipping. It is noteworthy that the consensus sequence for splice acceptors is NAG, corresponding to the PAM sequence for Cas9 from S. pyogenes (NGG or NAG), so any exon can potentially be skipped by this approach.

It has been estimated that even low-level expression of dystrophin (4 to 15%) can partially ameliorate cardiomyopathy (24) and protect against eccentric contraction-induced injury in skeletal muscle (25). The efficiency of restoration of dystrophin expression observed after delivery of Myoediting components to mdx mice by AAV is therefore within the range expected to provide therapeutic benefit.

Off-target effects are a safety concern in the eventual translation of gene-editing methods to humans. We did not observe off-target mutations at 10 potential off-target sites in the mouse genome nor any abnormalities in mice after AAV9 delivery of Myoediting components. However, off-target mutations may occur at sites beyond those predicted in silico; hence, a comprehensive and unbiased analysis, such as whole-genome sequencing, would be an essential component of future efforts to establish the safety of this approach (2628). Given that Myoediting offers the potential for durable and progressive therapeutic response in postmitotic adult tissue, we propose that this methodology may warrant investigation as a way to restore muscle function in DMD patients, alone or in combination with other therapies (3, 29).

Supplementary Materials

Materials and Methods

Figs. S1 to S13

Tables S1 to S4

References (3033)

References and Notes

Acknowledgments: We thank D. Grimm and F. Schmidt for AAV-hCas9 and the sgRNA Cloning Vector plasmids; C. Wang for AAV9 packaging; J. Schneider and P. Mammen for discussions; C. Rodriguez for technical help; Z. Wang for input; S. Rovinsky and E. Plautz (UT Southwestern Neuro-Models Facility) for grip strength testing; X. Li and R. Gordillo (UT Southwestern Mouse Metabolic Phenotyping Core Facility) for CK level measurement; and J. Cabrera for graphics. Supported by NIH grants HL-077439, HL-111665, HL-093039, DK-099653, U01-HL-100401, and U54 HD 087351; Fondation Leducq Networks of Excellence; and the Robert A. Welch Foundation (grant 1-0025, E.N.O.). The pSpCas9(BB)-2A-GFP (PX458) plasmid is available for purchase from Addgene under a material transfer agreement. The University of Texas Southwestern Medical Center and the authors (E.N.O., R.B.-D., J.M.S., C.L., J.R.M.) have filed a patent application (#14/823,563) related to the use of CRISPR/Cas9 technology to treat muscle disease.
View Abstract

Navigate This Article