Adenoviral vectors encoding CRISPR/Cas9 multiplexes rescue dystrophin synthesis in unselected populations of DMD muscle cells

Abstract

Mutations disrupting the reading frame of the ~2.4 Mb dystrophin-encoding DMD gene cause a fatal X-linked muscle-wasting disorder called Duchenne muscular dystrophy (DMD). Genome editing based on paired RNA-guided nucleases (RGNs) from CRISPR/Cas9 systems has been proposed for permanently repairing faulty DMD loci. However, such multiplexing strategies require the development and testing of delivery systems capable of introducing the various gene editing tools into target cells. Here, we investigated the suitability of adenoviral vectors (AdVs) for multiplexed DMD editing by packaging in single vector particles expression units encoding the Streptococcus pyogenes Cas9 nuclease and sequence-specific gRNA pairs. These RGN components were customized to trigger short- and long-range intragenic DMD excisions encompassing reading frame-disrupting exons in patient-derived muscle progenitor cells. By allowing synchronous and stoichiometric expression of the various RGN components, we demonstrate that dual RGN-encoding AdVs can correct over 10% of target DMD alleles, readily leading to the detection of Becker-like dystrophin proteins in unselected muscle cell populations. Moreover, we report that AdV-based gene editing can be tailored for removing mutations located within the over 500-kb major DMD mutational hotspot. Hence, this single DMD editing strategy can in principle tackle a broad spectrum of mutations present in more than 60% of patients with DMD.

Figure 1. Design and generation of AdVs encoding RNA-guided nuclease components.

(a) Schematic representation of the genome structure of control and DMD-targeting AdVs. The genomes of dual RGN-encoding AdV.Cas9^IN52.IN53 and AdV.Cas9^IN43.IN54 are depicted in relation to those of single RGN-encoding AdV.Cas9^EX51 and AdV.Cas9^EX53. The AdV.gRNA^S1 and AdV.Cas9, expressing exclusively an irrelevant gRNA and Cas9, respectively, served as controls. The AdV.Cas9^EX51 and AdV.Cas9^EX53 particles code for Cas9:gEX51 and Cas9:gEX53 complexes, respectively. The target sites of Cas9:gEX51 and Cas9:gEX53 complexes are located in exon 51 and exon 53 of the DMD locus, respectively. AdV.Cas9^IN52.IN53 and AdV.Cas9^IN43.IN54 deliver a pair of gRNA cassettes together with a Cas9 transcriptional unit. The gRNAs gIN52 and gIN53 address the Cas9 nuclease to DMD sequences in intron 52 and intron 53, respectively; the gRNAs gIN43 and gIN54 target the same nuclease to sequences located in intron 43 and intron 54, respectively. All vectors deployed in this study were isogenic in the sense that they were all assembled on the basis of a second-generation (i.e. E1- and E2A- deleted) AdV backbone coding for chimeric fibers (F⁵⁰) consisting of basal domains from human adenovirus serotype-5 and apical shaft and knob motifs from human adenovirus serotype-50. Each gRNA expression unit is under the transcriptional control of RNA Pol-III promoter (i.e. U6) and terminator sequences, whereas the Cas9 ORF is under the transcriptional control of the human PGK-1 promoter (PGK) and the simian virus 40 polyadenylation signal (SV40). Ψ, human adenovirus serotype-5 packaging signal; Ad L-ITR and Ad R-ITR, “left” and “right” adenoviral inverted terminal repeat, respectively; RF, reading frame (b) Ratios of genome copies (GC) to infectious units (IU). The GC and IU concentrations in purified preparations of the indicated AdVs were determined by using fluorometric and TCID₅₀ assays, respectively.

Figure 2. “All-in-one” AdV transduction of paired RGNs for DMD repair via short-range intragenic deletions.

(a) DMD repair through NHEJ-mediated exon 53 excisions in DMD.Δ45-52 muscle cells. Genomic deletions spanning exon 45 through exon 52 disrupt the DMD reading frame and yield out-of-frame mRNA transcripts carrying a premature stop codon within exon 53 (red box). RGN multiplexes consisting of Cas9 bound to intron-specific gIN52 and gIN53 gRNAs are designed to trigger the excision of a 1,123-bp DNA segment encompassing exon 53. This strategy is expected to induce the synthesis of a Becker-like dystrophin as splicing of exon 44 to exon 54 generate in-frame mRNA transcripts. gEX53, gRNA targeting Cas9 to a sequence within exon 53; SD, splice donor; SA, splice acceptor. Orange hexagons, CH1 and CH2 actin-binding domain 1; black hexagons, hinges; green boxes, spectrin-like repeats; orange oval, WW domain; blue pentagons, EFH1 and EFH2 hand-regions containing cysteine-rich motifs which, amongst other proteins, bind to β-dystroglycan; green pentagon, ZZ zinc finger domain; grey oval, C-terminal domain. Vertical arrowhead, protein junction formed by the DMD editing procedure. (b) qPCR analysis for determining the frequencies of DMD exon 53 deletion. The qPCR assay was applied on genomic DNA from DMD.Δ45-52 myoblasts transduced with “all-in-one” vector AdV.Cas9^IN52.IN53. The DMD.Δ45-52 myoblasts were transduced with AdV.Cas9^IN52.IN53 at the indicated multiplicities of infection (MOI). qPCR assays were carried out with primer pairs flanking the RGN-induced intronic junction (solid arrowheads). A primer pair targeting intron 43 sequences (open arrowheads) served as an internal control for determining the amounts of input DNA. MOI, multiplicity of infection; Error bars, standard deviations of technical replicates; ND, not detected. (c) Dystrophin western blot analysis after AdV.Cas9^EX53 and AdV.Cas9^IN52.IN53 transductions. The DMD.Δ45-52 myoblasts were transduced with AdV.Cas9^EX53 and AdV.Cas9^IN52.IN53 at multiplicities of infection (MOI) of 25 and 50 IU/cell. Cultures with myotubes differentiated from mock-transduced wild-type myoblasts (WT) served as positive controls. Negative controls were obtained from cultures containing myotubes differentiated from DMD.Δ45-52 myoblasts transduced at an MOI of 12.5 IU/cell with AdV.gRNA^S1. After 4 days in differentiation medium, western blot analysis was performed on unselected AdV-transduced muscle cell populations. Tubulin provided for a loading control antigen. E, empty lane.

Figure 3. “All-in-one” AdV transduction of paired RGNs for DMD repair via long-range intragenic deletions.

(a) DMD repair through NHEJ-mediated long-range, multi-exon, deletion. DMD-causing mutations located in the major DMD mutational hotspot (spanning exon 45 through exon 55) can, in principle, be tackled by a single multiplexing strategy based on the excision of a DNA segment encompassing exon 44 through exon 54. This multi-exon deletion strategy can build on the coordinated action of a pair of RGNs, consisting of Cas9 bound to the intron-specific gIN43 or gIN54 gRNAs. These RGNs are designed to repair the DMD reading frame as the joining of exon 43 to exon 55 is expected to yield in-frame mRNA species coding for an internally truncated Becker-like dystrophin. gEX51, gRNA targeting Cas9 to a sequence within exon 51. Vertical arrowhead, protein junction formed by the DMD editing procedure. For the naming of the dystrophin structural domains see legend of Fig. 2a. (b) qPCR analysis for establishing the frequencies of DMD multi-exon deletions. The qPCR assay was applied on genomic DNA from DMD.Δ48-50 and DMD.Δ45-52 myoblasts transduced with “all-in-one” vector AdV.Cas9^IN43.IN54. The DMD.Δ48-50 and DMD.Δ45-52 myoblasts were transduced with AdV.Cas9^IN43.IN54 at the indicated multiplicities of infection (MOI). qPCR assays were carried out with primer pairs flanking the expected RGN-induced intronic junction (solid arrowheads). qPCR amplifications with primer pair targeting intron 43 sequences (open arrowheads) provided for an internal control to determining the amounts of input DNA. Error bars, standard deviations of technical replicates; ND, not detected. (c) Dystrophin western blot analysis after AdV.Cas9^EX51 and AdV.Cas9^IN43.IN54 transductions. The DMD.Δ45-52 and DMD.Δ48-50 myoblasts were transduced with AdV.Cas9^IN43.IN54; the DMD.Δ48-50 myoblasts were transduced with AdV.Cas9^EX51. The multiplicities of infection (MOI) used are indicated. Cultures with myotubes differentiated from mock-transduced wild-type myoblasts (WT) served as positive controls. Negative controls were obtained from cultures containing DMD.Δ45-52 and DMD.Δ48-50 myotubes whose progenitors had been transduced with AdV.gRNA^S1 at an MOI of 12.5 IU/cell. After exposing DMD.Δ45-52 and DMD.Δ48-50 myoblasts to differentiation medium for 4 and 5 days, respectively, western blot analysis was performed on unselected AdV-transduced muscle cell populations. Tubulin provided for a loading control antigen. E, empty lane.

Figure 4. Dystrophin and β-dystroglycan immmunofluorescence microscopy on DMD-edited myotubes.

Immunostainings for dystrophin and β-dystroglycan were carried out in myotubes differentiated from DMD.Δ48-50 and DMD.Δ45-52 myoblasts transduced with the AdV.Cas9^IN43.IN54. Each numeral refers to multiplicities of infection (MOI). The same immunostainings were also done in myotubes differentiated from DMD.Δ48-50 myoblasts transduced with AdV.Cas9^EX51 at the indicated MOI. Myotubes derived from DMD.Δ48-50 and DMD.Δ45-52 transduced with AdV.gRNA^S1 at an MOI 12.5 IU/cell served as negative controls. Myotubes isolated from a healthy donor (WT) and transduced with AdV.gRNA^S1 provided for positive controls.