Bioengineered AAV9 and Optimised Microdystrophin Vectors Augment Phenotypic Rescue in a Murine Model of Duchenne Muscular Dystrophy
Mohankumar B. Senthilkumar, Sanya Sharma, Navaneeth Srinivasan, Anila Varghese, Pratiksha Sarangi, Vijayata Singh, Narendra Kumar, Devyani Yenurkar, Sudip Mukherjee, Sonal Amit, Sameer Bhatia, Santosh K. Misra, Ratna Dua Puri, Jeffrey Chamberlain, Giridhara R. Jayandharan
TL;DR
Researchers improved gene therapy for Duchenne muscular dystrophy using engineered AAV9 vectors and optimized microdystrophin, showing long-term benefits in mice.
Contribution
The study introduces a combination of AAV9 capsid engineering, promoter optimization, and codon-optimized microdystrophin to enhance gene therapy outcomes for DMD.
Findings
AAV9 vectors with a CAG promoter improved dystrophin expression in muscle fibers of mdx mice.
Engineered AAV9 capsids (N57Q, K51Q) significantly improved grip strength after gene therapy.
Systemic administration of AAV9K51Q vectors restored dystrophin in skeletal and cardiac muscles up to 14 months post-treatment.
Abstract
Duchenne muscular dystrophy (DMD) is a severe neuromuscular disorder without an effective cure. Adeno‐associated virus (AAV) based gene therapy has improved dystrophin function, with sub‐optimal clinical outcomes. We reasoned that a combination of rational engineering of AAV9 capsids modified at the post‐translational modification sites, optimal promoter selection, and codon‐optimisation of the microdystrophin (μDys) can enhance the AAV9 vector functionality. Our initial promoter screening demonstrated improved dystrophin expression in muscle fibres with a ubiquitous CAG promoter (1.61‐fold in CAG vs. MHCK7, p < 0.0001) in mdx mice. We then evaluated two engineered AAV9 capsids (N57Q, K51Q) containing CAG‐μDys intramuscularly in vivo, which demonstrated a significant improvement in grip strength 18 weeks after gene therapy. Subsequent evaluation of a codon‐optimised microdystrophin…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7- —The Wellcome Trust DBT India Alliance10.13039/501100009053
- —Indian Institute of Technology Kanpur10.13039/501100001403
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMuscle Physiology and Disorders · Virus-based gene therapy research · CRISPR and Genetic Engineering
Introduction
1
Duchenne muscular dystrophy (DMD) is a severe neuromuscular disorder in humans (1:3500 male births) caused by mutations in the DMD gene and the subsequent loss of dystrophin protein. This leads to a compromised dystrophin‐glycoprotein complex (DGC), crucial in connecting the cytoskeleton to the muscle fibre cell membrane. The phenotypic manifestations are severe, ranging from muscle fibre necrosis and inflammation, which manifests as cardiomyopathy, cognitive and ambulatory issues, due to compromised muscle function [1]. Many DMD patients eventually succumb to cardiac or respiratory failure in the second decade of their life [2].
The current standard of care for DMD patients includes the administration of corticosteroids such as deflazacort [3] and prednisone [4], that has limited efficacy in halting the disease progression and prolonging ambulation in treated patients [5, 6]. Newer therapeutic approaches through antisense oligonucleotide (ASO)‐mediated exon skipping [7], stop codon readthrough [8] and gene replacement [9] are promising. Gene replacement with Adeno‐associated virus vectors (AAV) has multiple advantages such as broader tissue tropism, non‐pathogenicity, and providing stable therapeutic gene expression [10]. Multiple serotypes of recombinant AAV have been extensively used for genetic therapies. Of these, AAV9 serotype has a greater transduction in skeletal and cardiac muscle cells [11] and hence, is an ideal vector for targeting muscle diseases such as DMD. Due to the maximum transgene packaging limit of ~4.4 kb in AAV vectors, microdystrophin—a truncated product of the dystrophin gene—is widely used for gene therapy of DMD. In preclinical models, AAV‐microdystrophin has demonstrated promising transduction to striated muscles [12], reduced cardiac fibrosis [13], and improved cardiac performance [14]. In clinical trials, use of AAV‐microdystrophin has shown a modest functional improvement in DMD patients, even though an extremely higher dose of 10^13^ –10^14^ vector genomes (vgs)/kg of AAV is required [9]. However, these high doses have resulted in inflammatory myopathy in dog pups and liver toxicity in rhesus macaques, attributed to severe immune response [9]. Incidentally, a high‐dose DMD clinical trial reported the death of two patients [15, 16]. In another clinical trial, evidence for the complement activation was found in a patient after high‐dose AAV injection (2 × 10^14^ vgs/kg) [9]. Multiple lines of evidence suggest that immune responses to AAV vectors are dose‐dependent [17, 18]. To address this crucial issue of high vector doses and concomitant immune response, we reasoned that a precise selection of appropriate promoter, the transgene, and capsid combination may be required to enhance therapeutic outcomes [19].
In this study, we utilised three overlapping strategies to develop an optimised AAV9 vector for DMD gene therapy. This includes selection of a hybrid chicken β‐actin promoter (CBA) and cytomegalovirus (CMV) early enhancer (CAG) for widespread and robust expression of dystrophin [20]. Additionally, a consensus ribosome binding site sequence, known as the Kozak sequence, was incorporated after the CAG promoter upstream of the transgene sequence to ensure increased and robust expression of the therapeutic protein [21]. Taking advantage of the superior transduction capability of the AAV9 serotype in muscle tissues, capsid engineering was performed at post‐translational modification (PTM) sites to generate AAV9 mutant vectors. Furthermore, to maximise transgene expression and overcome limitations such as codon bias and the availability of cognate tRNAs affecting translational efficiency, codon‐optimisation of the microdystrophin (CoμDys) transgene was carried out for the optimal expression in human skeletal muscles [22]. We have assessed the proof‐of‐concept efficacy of these triple‐engineered vectors in multiple cell line models in vitro and by both local and systemic administration in a mouse model (mdx) of DMD.
Materials and Methods
2
Cell Lines, Reagents and Plasmids
2.1
Human cervical carcinoma cells (HeLa) (ATCC, Manassas, VA, United States), C2C12 murine myoblast cell line (National Center for Cell Science, Pune, India), H9C2 (a kind gift from Dr. Ashok Kumar, IIT Kanpur), and AAV‐293 cells (Stratagene, San Diego, CA, USA) were used.
The control plasmid pssAAV‐MHCK7‐μDys (a kind gift from Dr. Jeffrey Chamberlain, University of Washington) contains human microdystrophin (R4‐R23/Δ71‐78) under the control of muscle‐specific MHCK7 promoter. The donor plasmid (pLV‐hsa‐μDys/eGFP) was purchased from Addgene (Plasmid #26810). This plasmid was used to chemically synthesise pssAAV‐CAG‐Kozak‐μDys (GenScript, Piscataway, NJ, USA). Further, a codon‐optimised version of human microdystrophin (pssAAV‐CAG‐Kozak‐CoμDys) was generated (GenScript). The AAV9 capsid mutants were generated in AAV9WT rep/cap plasmid.
Designing of Microdystrophin Plasmid Vectors
2.2
The truncated version of human dystrophin, that is, microdystrophin from donor plasmid (pLV‐hsa‐μDys) was digested and ligated in the recipient plasmid by replacing eGFP sequence downstream of CAG‐Kozak sequence between the inverted terminal repeats of the AAV back bone plasmid to generate the pssAAV‐CAG‐Kozak‐μDys (CAG‐Koz‐μDys) plasmid. The plasmids were synthesised (GenScript) and the cloning was confirmed by Sanger sequencing and restriction digestion analysis. pssAAVMHCK7‐μDys (MHCK7‐μDys) was used as the control. Further, the μDys sequence in the pssAAV‐CAG‐Kozak‐μDys was codon‐optimised for human muscle tissue and codon‐optimised μDys (CoμDys) was cloned into donor plasmid by replacing μDys sequence to generate pssAAV‐CAG‐Kozak‐CoμDys (GenScript).
Generation of Mutant AAV9 Capsid Vectors
2.3
Protein sequence for AAV9‐VP1 (GenBank: AAS99264.1) was retrieved and used as the reference. An online bioinformatics programme called NeddyPreddy was employed to identify putative Neddylation sites [23]. The medium and high threshold levels were established using the tool's output. An elevated cutoff score of 0.92 out of 1 led to the selection of the AAV9K51 target. Additionally, the glycosylation mutant AAV9 was produced by utilising the HexNAcylation (Glycosylation) site N57 in the VP1 protein of AAV9, which had been previously identified in our study [24]. Using specific primers and the QuikChange II XL site‐directed mutagenesis kit (Agilent Technologies, Santa Clara, CA, USA), a site‐directed mutagenesis (SDM) of the AAV9WT rep/cap plasmid (pAAVR2/C9 WT) was carried out to replace the Neddylation site (K51) and glycosylation site (N57) in the AAV9 capsid with glutamine residue (Q). The final constructs were confirmed by Sanger sequencing.
Production of Recombinant AAV Vectors
2.4
For AAV vector production, polyethyleneimine (PEI; Polysciences, Warrington, PA, USA)‐based triple transfection of AAV‐293 cells with equimolar concentration of wild‐type or mutant rep/cap plasmid (pAAVR2/C9 WT; pAAVR2/C9 K51Q; pAAVR2/C9 N57Q), transgene containing plasmid, CAG‐Koz‐μDys or MHCK7‐μDys or CAG‐Koz‐CoμDys, along with Adenoviral helper plasmid was performed. After 72 h, the cells were collected, processed, and the virus was purified and concentrated as described earlier [25]. Using this protocol, we generated AAV9WT‐CAG‐Koz‐μDys (AAV9WT‐μDys); AAV9WT‐MHCK7‐μDys (AAV9WT‐MHCK7‐μDys), AAV9K51Q‐CAG‐Koz‐μDys (AAV9K51Q‐μDys), AAV9N57Q‐CAG‐Koz‐μDys (AAV9N57Q‐μDys), and AAV9K51Q‐CoμDys vectors. Virus quantification was done by quantitative (qPCR) with polyadenylation (PolyA) signal‐specific primers (CFX96, Bio‐Rad, Hercules, CA, USA) using AAV2‐RSS (ATCC) as a standard.
Transfection Studies
2.5
The CAG‐Koz‐μDys plasmid and MHCK7‐μDys plasmid (500 ng) were transfected in HeLa, C2C12, and H9C2 cell lines using PEI. After 48 h, the cells were collected, and RNA was isolated using TRIzol (Invitrogen, Waltham MA, USA) [25]. cDNA was prepared from 1 μg of RNA using Verso cDNA synthesis kit (Thermofisher, MA, USA). The levels of μDys were measured by qPCR using primers (Forward primer (FP): AACAAAGTGCCCTACTA; Reverse primer (RP): AGGTTGTGCTGGTCCA). 18srRNA (FP: TTGACGGAAGGGCACCACCAG; RP: GCACCACCACCCACGGAATCG) was used as the reference gene for normalisation of qPCR data.
Transduction Assays
2.6
The transduction efficiency of the packaged vectors, AAV9WT‐μDys, AAV9WT‐MHCK7‐μDys, AAV9K51Q‐μDys, AAV9N57Q‐μDys, and AAV9K51Q‐CoμDys was assessed in vitro. Briefly, HeLa, C2C12, and H9C2 cells were infected with the vectors at 1 × 10^5^ vgs MOI (multiplicity of infection). After 48 h, the RNA from cells was isolated, and the μDys levels were measured by qPCR.
Immunocytochemistry was performed to validate the increased translation efficiency of AAV9K51Q vector packaged with an optimised transgene. Briefly, C2C12 cells were seeded at a density of 10^5^ cells per well. Cells were then mock treated or infected with AAV9K51Q‐μDys/AAV9K51Q‐CoμDys at an MOI of 1 × 10^5^. After 24 h, the cells were fixed in 4% paraformaldehyde. Further, the cells were incubated with primary antibody for Dystrophin (1:150, sc‐33,697 Santa Cruz, Dallas, TX, USA) for 1 h. The cells were then stained with secondary antibody goat anti‐mouse Alexa‐fluor 568 (1:250, A‐11004, Thermofisher) for 1 h. Cells were then counterstained with DAPI (1:1000, Sigma Aldrich) followed by mounting the coverslip over the glass slide with Flurosave (Sigma Aldrich). The images were obtained in the ZOE‐Fluorescent cell imager (Bio‐Rad, Hercules, CA, USA) and these images (n ≥ 8/group) were further used for semi‐quantification analysis (ImageJ software).
AAV9 Based Gene Therapy in mdx Mice
2.7
DMD mice (B6Ros. Cg‐Dmd^mdx‐4cv^/J) (mdx) were purchased from Jackson Laboratory (Bar Harbour, ME, USA). The animal experiments were approved by the IIT‐Kanpur Institutional Animal Ethics Committee (IAEC). For vector administration, 6–10 weeks old mice were used. For the initial studies to identify the optimum μDys transgene cassettes, 3 groups of mice, namely, Mock (n = 8), AAV9WT‐MHCK7‐μDys (n = 8), and AAV9WT‐μDys (n = 8) were utilised. The experimental mice were administered with the respective vectors at a dose of 3.42 × 10^11^ vgs/leg with a fixed volume of 20 μL in their TA muscle (Figure S3). Animals in the mock group were administered a similar volume of 1× PBS. In the second batch of in vivo experiments to study the efficacy of bioengineered AAV9 vectors, four groups‐Mock (n = 8), AAV9WT‐μDys (n = 8), AAV9K51Q‐μDys (n = 7), and AAV9N57Q‐μDys (n = 8) of mice were injected with the vectors at a dose of 2 × 10^11^ vgs/leg. Further, to characterise the efficacy of CoμDys, animals were grouped into mock‐treated (n = 10) or administered with AAV9K51Q‐μDys (n = 10) and AAV9K51Q‐CoμDys (n = 9) vectors (2 × 10^11^ vgs/leg) in their TA muscle. In the final set of experiments to evaluate the efficacy of the optimised AAV9K51Q‐CoμDys vector, animals were grouped into mock (n = 8) and AAV9K51Q‐CoμDys (n = 8). The animals were administered with either PBS or AAV9K51Q‐CoμDys vector at 4 × 10^12^ vgs/mice via the tail vein.
Muscle Strength Analysis
2.8
The muscular strength of the hind and four limbs of mock‐treated and vector‐treated DMD mice was measured using a grip strength metre (Bioseb, GS‐4, Chaville, France) as described earlier [26]. For each mouse, five peak grip strength readings were recorded, and the average was computed. The muscle grip strength in vector‐treated mice was assessed at different time points, ranging from 8 to 55 weeks after vector administration.
in vivo Muscle Contractile Function
2.9
Maximal isometric twitch and tetanic force exerted by the dorsiflexors of the animals that received systemic gene transfer were assessed by stimulating their TA muscle, which is the prime muscle of ankle dorsiflexors, using the 1300A 3‐in‐1 Whole Animal System (Aurora Scientific, Aurora, ON, Canada) as previously described [27, 28]. The maximal twitch and tetanic forces were individually measured for both right and left limbs of the mouse. The average maximal twitch and tetanic force (mN) for the animal was calculated and normalised to the body weight (mN/g) of the animal.
Echocardiography
2.10
Echocardiographic assessment of animals that received systemic gene transfer was performed 60 weeks after gene therapy. Echocardiography was performed using Vevo 3100 Imaging System (FUJIFILM VisualSonics, Toronto, ON, Canada) as previously described [13, 29]. B‐mode and M‐mode images were captured at the papillary muscle level to evaluate different cardiac ventricular measurements and other cardiac performance parameters.
Immunogenicity Assessment
2.11
Briefly, 62 weeks after systemic vector administration, peripheral blood was collected, incubated with APC‐labelled anti‐CD19, PerCP‐labelled anti‐CD4, FITC‐labelled anti‐CD3, and PE‐labelled anti‐CD8 (BD Biosciences) antibodies for 30 min at room temperature. Flow cytometry was performed on a BD Accuri C6 Plus (BD Biosciences, San Jose, CA). The data are represented as percentage of CD3+, CD4+, CD8+, and CD19+ cells [30].
Spleen from experimental animals that received systemic gene transfer were collected 62 weeks after vector administration. After RBC lysis, spleenocytes (1 × 10^6^) were seeded on IFN‐γ–coated plates (MabTech, Nacka Strand, Sweden) and stimulated with an AAV9 capsid peptide (RGLVLPGYKYLGPGNGLDKG, JPT Peptide Technologies) [31]. Concanavalin A served as a positive control. After 36 h of incubation at 37°C with 5% CO₂, spots were developed as per the manufacturer's protocol. The spot‐forming units were counted using an ELISPOT reader (AID GmbH, Germany).
Immunohistochemistry
2.12
Animals were euthanized at different time points (17, 24, 33, and 62 weeks post vector administration) and TA, gastrocnemius(GAS), diaphragm (DIA), and heart muscles were excised [32]. Muscle tissue was sectioned (~8 μm thickness) and fixed with 4% paraformaldehyde. After blocking, the TA, GAS, and DIA sections were incubated with primary antibody to Dystrophin (1:50) and laminin (1:300, PA1‐16730, Thermofisher) for 12 h at 4°C, and the secondary antibody‐goat anti‐rabbit Cy 3 (1:300, AB_2338002, Jackson ImmunoResearch, PA, US) and goat anti‐mouse Alexa‐fluor 647 (1:600, A_21235, Thermofisher) for 2 h at 4°C. Counterstaining was performed using DAPI (1:1000 v/v), and mounting was performed using FluorSave. TA muscle sections were additionally stained with primary antibodies β‐Dystroglycan (1:50, B‐DG‐CE,Leica Biosystems) and Dystrobrevin (1:50, 610766, BD Bioscience, NJ, USA) for 24 h at 4°C and followed by secondary antibody goat anti‐mouse Alexa‐fluor 568 (1:250, Thermofisher) for 24 h at 4°C. The immunostaining of cardiac sections was performed as above, with a slight modification. The heart tissue sections were fixed with ice‐cold acetone for 5 min, followed by a brief permeabilization in 0.4% Triton X‐100 for 15 min and blocking for 2 h, followed by primary antibody incubation for 24 h at 4°C and secondary antibody goat anti‐mouse Alexa‐fluor 568 (1:250) for 24 h at 4°C. All images were obtained in a confocal microscope at 40× magnification (LSM780NLO, CarlZeiss GmbH, Wein, Austria and AXR Nikon microscope). To quantify the total fibres, muscle sections were stained with Alexa fluor 568‐tagged wheat germ agglutinin (WGA) (W56133, Thermofisher) or laminin‐ stained muscle sections. Images were acquired from three different animals, with three independent sections per animal, to determine total fibre count from WGA or laminin‐stained sections and quantify dystrophin‐positive fibres. The quantification was performed using the multipoint tool in Image J as previously described [32, 33].
Morphometric Assessment in Vector‐Treated Mice
2.13
Haematoxylin and Eosin (H&E) staining and Masson's trichrome (MT) staining of a TA muscle section (8 μm thickness) from animals that received local gene transfer were performed using standard protocols to assess muscle morphological features and fibrosis, respectively. Additionally, for systemic vector‐administered animals, H&E staining was performed on TA, GAS, and DIA muscle sections. MT staining was performed in TA, GAS, and cardiac muscle sections to assess fibrosis 62 weeks after gene transfer. Images were captured in an inverted microscope (Dmi8, Leica Microsystems, Germany). Centrally nucleated myofibers were counted manually using a multi‐point tool in ImageJ software (n ≥ 8 images/group and n ≥ 2 tissues/group). MT staining images were used for the measurement of fibrotic areas (n ≥ 9 images/group and n = 3 tissues/group) using the colour deconvolution plugin in Image J software. The original images were segmented into three clusters. The blue areas of the image represent the collagen deposition whose total area was measured by correcting the threshold manually such that the collagen deposition overlaid with the original histology image [32, 34].
Vector Genome Quantification
2.14
Total genomic DNA was isolated from TA muscles (n = 4 per group) of the mice that received intramuscular vector administration. Vector genome copy numbers were determined by a quantitative PCR using ITR‐specific primers (Forward primer: GGAACCCCTAGTGATGGAGTT, Reverse primer: CGGCCTCAGTGAGCGA). For the systemically administered mice, the biodistribution of the vectors in different tissues (TA, GAS, Quadriceps, Brain, Heart, Liver, Lungs, Kidney, n = 4/tissue/group) was evaluated by a qPCR using CoμDys‐specific primers (Forward primer: CCGAGGATAGATGGGTGCTG, Reverse primer: CCGGTGGTGTGGATCTTGTT) as described earlier [32].
Haematological Parameters
2.15
Plasma from experimental animals was collected 62 weeks after gene transfer. The level of creatine kinase (CK) was measured in a Fully Automated Clinical Chemistry Analyser (EM 200) as per the manufacturer's protocol.
Statistical Analysis
2.16
Data are presented as the mean ± standard deviation (SD). Quantitative PCR data was analysed by the Biorad CFX Manager 3.1. Statistical significance was evaluated using an unpaired Student t‐test or one‐way ANOVA, as appropriate (GraphPad Prism 8.0.2 software).
Results
3
Design and Evaluation of a Ubiquitous Promoter Containing Microdystrophin Vector
3.1
We constructed the plasmid vector by sub‐cloning the ΔR4‐23/ΔCT microdystrophin gene (μDys) [35] from the donor plasmid pLV‐Hsa‐μDys to the recipient plasmid pssAAV‐CAG‐eGFP (GenScript, Piscataway, NJ, US). The recipient plasmid contains a hybrid promoter/enhancer sequence, that is, the combination of a CMV early immediate enhancer sequence and CBA promoter known as CAG promoter which has proven ubiquitous expression in various cell lines and cell types [20, 36]. To increase the translation efficiency of the transgene, a consensus ribosome binding site sequence termed as Kozak (GCCGCCACC) [37] was inserted upstream to the transgene to yield the final construct designated as pssAAV‐CAG‐Kozak‐microdystrophin (CAG‐Kozak‐μDys) to ensure the ubiquitous and greater translational efficacy of dystrophin (Figure S1). We then evaluated the efficacy of these plasmids, in vitro in different muscle cell lines like murine myoblast (C2C12) and rat cardiomyocytes (H9C2) and non‐myogenic cell line like human cervical carcinoma (HeLa) cells. The plasmid pssAAV‐MHCK7‐μDys (MHCK7‐μDys) [38] was used as a control. Following transfection, the comparison of steady‐state mRNA levels of ΔR4‐23/ΔCT microdystrophin revealed that the CAG‐Kozak construct demonstrated a 12‐fold (p < 0.001) increased expression compared to the MHCK7 promoter in HeLa cells (Figure S2A). Similarly, a significantly higher microdystrophin expression (5‐fold; p < 0.001 and 7‐fold; p < 0.01) was observed with the CAG‐Kozak driven construct in C2C12 cells and H9C2 cells respectively, when compared to the muscle‐specific promoter (Figure S2B,C). Subsequently, both the constructs of microdystrophin were packaged in AAV9 wild type (AAV9WT) capsid due to its tropism towards muscle tissue. The vector titers obtained are given in Table S1. Further evaluation of AAV9WT‐CAG‐Koz‐μDys (AAV9WT‐μDys) and AAV9WT‐MHCK7‐μDys vectors in vitro, revealed a three‐fold difference between CAG‐Kozak and MHCK7 driven microdystrophin expression levels in both HeLa and C2C12 (Figure S2D,E). Similarly, in vitro analysis in H9C2 cells revealed ~221 fold (p < 0.05) increased expression of μDys in AAV9WT‐μDys infected cells when compared to cells infected with AAV9WT‐MHCK7‐μDys vectors (Figure S2F).
We next evaluated the efficiency of these vectors in a murine model of DMD (mdx, n = 8 per group). An intramuscular administration of AAV9WT‐μDys and AAV9WT‐MHCK7‐μDys vectors in the tibialis anterior (TA) was performed at a dose of 3.4 × 10^11^ vgs/leg at a fixed volume of 20 μL per leg (Figure S3). Mock animals were administered the same volume of vehicle (PBS). Thirty‐three weeks after vector administration, TA muscles were assessed for dystrophin glycoprotein complex (DGC) proteins (Figure S4A,B). The CAG‐Kozak microdystrophin delivered by the AAV9WT vector had a localised expression of dystrophin, dystroglycan and dystrobrevin proteins. A similar expression pattern was observed in the muscles administered with AAV9WT‐MHCK7‐μDys. Further quantification showed that AAV9WT‐μDys (with the ubiquitous CAG promoter) group had a significantly higher percentage of dystrophin‐positive myofibers than the mock‐treated (29.9% vs 0.98%, p < 0.0001) and AAV9WT‐MHCK7‐μDys group (29.9% vs 18.48%, p < 0.0001) (Figure S4C). These results revealed the efficiency of the CAG‐Kozak construct in driving sustained microdystrophin expression at a longer follow‐up of 33 weeks. Therefore, CAG‐Kozak‐microdystrophin construct was employed for the subsequent experiments.
Development and Validation of Engineered AAV9 Vectors in vitro and in vivo
3.2
Subsequently, we developed optimised AAV9 vectors by engineering the capsids at PTM site and packaged them with pssAAV‐CAG‐Kozak‐μDys transgene to generate AAV9WT‐CAG‐Kozak‐μDys (AAV9WT‐μDys), AAV9K51Q‐CAG‐Kozak‐μDys (AAV9K51Q‐μDys) and AAV9N57Q‐CAG‐Kozak‐μDys (AAV9N57Q‐μDys) (Table S1). The engineered AAV9 vectors were assessed for their transduction efficiency in HeLa and C2C12 cells. Gene expression analysis for μDys revealed that AAV9K51Q‐μDys had 1.33‐fold (p < 0.05) higher expression when compared to AAV9WT‐μDys, whereas AAV9N57Q‐μDys showed 2.94‐fold (p < 0.01) increased expression of μDys when compared to AAV9WT‐μDys in HeLa cells (Figure S5A). In C2C12 cells, a significant increase in levels of μDys was observed in both AAV9K51Q‐μDys (~6‐fold, p < 0.001) and AAV9N57Q‐μDys (~3‐fold, p < 0.05) transduced cells when compared to AAV9WT‐μDys (Figure S5B). This data confirmed that the engineered AAV9 vectors are functional and demonstrated a relatively higher expression of dystrophin.
We further evaluated the efficiency of engineered AAV9 vectors in vivo by intramuscular vector administration. Four groups of animals (n = 8 per group) were mock‐administered or administered with AAV9WT‐μDys, AAV9K51Q‐μDys, and AAV9N57Q‐μDys vectors at a dose of 2 × 10^11^ vgs/leg. The phenotypic rescue was studied 12 weeks later. A significant increase in hind limb grip strength was observed in the AAV9K51Q‐μDys vector‐treated group compared to the group treated with the AAV9WT‐μDys (Mean force 0.73 Newton (N) and 0.55 N, respectively, p < 0.05) (Figure S6A). A similar trend was observed for all four limbs' grip strength (Figure S6B) and at a longer follow‐up of 18 weeks (Figure S6C,D).
To confirm that the phenotypic rescue is due to restoration of dystrophin expression, we performed immunostaining for DGC complex proteins in TA muscles, 24 weeks post‐treatment. The expression of dystrophin and other DGC complex proteins (Figure S7A) was higher in the AAV9K51Q‐μDys and AAV9N57Q‐μDys mutant vector administered group than in the AAV9WT‐μDys treated mice. Further, quantification of dystrophin positive myofibers in the TA muscle sections (n = 3 TA muscles and n ≥ 3 sections per muscle) revealed a significant increase in the percentage of dystrophin positive fibres in AAV9K51Q‐μDys vector group when compared to mock (Mean: 45.71% vs 0.82%, p < 0.0001) and AAV9WT‐μDys (45.71% vs 27.72%, p < 0.0001) vector treated cohort of mice (Figure S7B). AAV9N57Q‐μDys vector‐treated mice also demonstrated higher dystrophin‐positive fibres when compared to mock‐treated (36.23% vs 0.82%, p < 0.0001) or AAV9WT‐μDys vector‐treated mice (36.23% vs 27.72%, p < 0.0001) (Figure S7B). Considering the consistent and improved performance of AAV9K51Q‐μDys vectors in vitro, and in mitigating the phenotype in vivo, we utilised only this capsid for further studies.
Codon‐Optimised μDys Transgene Demonstrated Superior Efficacy in vitro and in vivo Intramuscularly
3.3
We reasoned that codon‐optimisation of the μDys cDNA sequence for human skeletal tissues may improve the translational efficiency and the transgene expression. We constructed pssAAV‐CAG‐Kozak‐CoμDys (CAG‐Koz‐CoμDys) and packaged it in the AAV9K51Q capsid to generate AAV9K51Q‐CAG‐Kozak‐CoμDys vector (AAV9K51Q‐CoμDys) (Table S1). The efficiency of the vector carrying the codon‐optimised transgene was improved, with a 1.3‐fold (p < 0.05) higher microdystrophin expression in the C2C12 cells when compared to cells transduced with AAV9K51Q‐μDys vector (Figure 1A,B).
Validation of codon‐optimised microdystrophin transgene in vitro. Transduction of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys vectors in C2C12 cells (A), at an MOI: 1 × 105 and its quantification (B) are shown. Representative images were obtained in confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). Semi‐quantification was performed from images (n ≥ 8 per group) obtained in ZOE‐Fluorescent cell imager. The data is represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys treated conditions with respect to the cell control and Hash (#) refers to statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys. # p < 0.05, ***p < 0.001, ***p < 0.0001. For statistical comparison, a one‐way ANOVA was performed (GraphPad Prism 8.0.2 software).
We further evaluated the efficiency of AAV9K51Q‐CoμDys vector in mdx mice by intramuscular administration of the vectors at a dose of 2 × 10^11^ vgs/leg in TA muscles. Three groups of mice (n = 9–10 per group) were used, including AAV9K51Q‐μDys, AAV9K51Q‐CoμDys, and the control group (PBS). The phenotypic response was assessed 8 weeks later. Hind limb grip strength of AAV9K51Q‐μDys vector and AAV9K51Q‐CoμDys treated mice was found to be significantly higher when compared with mock‐treated animals (Mean: 0.68 N vs 0.28 N and Mean 0.89 N vs 0.28 N, p < 0.0001, respectively) (Figure 2A). Hind limbs grip strength of AAV9K51Q‐CoμDys vector administered mice was significantly higher than that of AAV9K51Q‐μDys vector treated mice (Mean: 0.89 N vs. 0.68 N, p < 0.0001) (Figure 2A). A similar finding was observed for all four limbs (Figure 2B). Further follow‐up at 12 and 16 weeks, showed a similar pattern of increased grip strength (Figure 2C–F). These data confirm that AAV9K51Q‐CoμDys vector has a prolonged impact in improving the muscle strength of DMD mice.
Muscle strength of mice administered with engineered AAV9 capsid and codon‐optimised μDys vector. Data obtained after 8 weeks of gene transfer for hind limbs (A), all four limbs (B) followed by 12 weeks follow‐up for hind limbs (C) and four limbs (D) are depicted. A further long‐term follow‐up of 16 weeks for hind limbs (E) and four limbs (F) of mice treated with AAV9K51Q‐μDys (n = 10) and AAV9K51Q‐CoμDys (n = 9) vectors showed significantly higher grip strength than mock‐treated group (n = 10). Each data point represents the average value of 5 readings measured per mouse. The data obtained are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys with respect to the mock, and Hash (#) refers to statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys. # p < 0.05, **p < 0.01, ***/#### p < 0.0001. For statistical comparison, a one‐way ANOVA was performed (GraphPad Prism 8.0.2 software).
Subsequently, immunostaining of TA muscles 4 months after gene therapy revealed that the expression of dystrophin and other DGC proteins like dystrobrevin and dystroglycan was higher in the vector‐treated groups, and these proteins were largely localised to the sarcolemma (Figure 3A). The AAV9K51Q‐CoμDys‐treated TA muscles had a higher percentage of dystrophin‐positive fibres (AAV9K51Q‐CoμDys vs AAV9K51Q‐μDys: 57.86% vs 46.78%, p < 0.0001) (Figure 3B). Additionally, MT staining of TA muscle sections from treated mice showed a visible decrease in fibrosis (Mock vs AAV9K51Q‐μDys: 1.29‐fold, p < 0.0001; Mock vs AAV9K51Q‐CoμDys: 1.62‐fold, p < 0.0001), 4 months after intramuscular gene therapy (Figure S8A,B). Similarly, a significantly lesser collagen deposition (~1.25‐fold, p < 0.0001) was also noticed (Figure S8B). H&E staining of TA muscles in mice administered with AAV9K51Q‐CoμDys demonstrated ~2‐fold lesser centralised nuclei than mice treated with AAV9K51Q‐μDys vectors (Mean: 16.21% vs 35.87%; p < 0.0001), signifying the rescue of skeletal muscle characteristic features (Figure S9A,B). Vector genome (vg) quantification was performed to detect the AAV‐specific genome in the vector‐treated muscles using qPCR. Our data reveal that AAV9K51Q‐CoμDys‐treated muscles had twice the number of vector copies when compared to AAV9K51Q‐μDys (Figure S10). These data highlight the benefit of utilising codon‐optimised μDys in improving dystrophin expression.
Immunofluorescence staining exhibits increased levels of DGC proteins in TA muscles of optimised AAV9 vector‐administered animals. The expression of dystrophin‐glycoprotein complex (DGC) proteins in TA muscle of mdx mice, 17 weeks after administration of AAV9K51Q‐μDys, AAV9K51Q‐CoμDys vectors (A), and quantitative analysis of dystrophin‐positive fibres (B) are shown. Images are representative of each group. The sections were imaged with the confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). The quantification of dystrophin‐positive fibres was done from images (n ≥ 15/group from n = 3 sections/muscle from n = 3 animals/group) captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) in 20X magnification. The co‐staining of TA muscle sections (n = 3 animals) was performed with laminin and dystrophin antibodies. The laminin‐stained images were used to quantify the total muscle fibre number. The number of dystrophin‐positive fibres and total fibre number was manually counted and marked in the images using ImageJ software, and their percentage is denoted in (B). The data are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys with respect to the Mock, and Hash(#) refers to statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys. ***/#### p < 0.0001. For statistical comparison, a one‐way ANOVA was performed (GraphPad Prism 8.0.2 software).
Systemic Gene Transfer of the Most Optimal AAV9 Vector Rescued the Dystrophic Phenotype in mdx Mice
3.4
Based on the improved function of AAV9K51Q‐CoμDys vectors observed during muscle gene transfer, we further evaluated the therapeutic potential of this vector alone, by systemic administration. Animals were randomised into 2 groups (n = 8 per group), including the mock (PBS‐administered) and the vector‐treated (AAV9K51Q‐CoμDys, at a dose of 4 × 10^12^ vgs/mouse) group. Muscle strength was assessed periodically. At 12 weeks after systemic gene transfer, the hind limb grip strength of vector treated mice was significantly higher than the mock‐treated animals (Mean: 0.85 N vs 0.34 N, p < 0.0001) (Figure 4A). Similarly, ~2‐fold (p < 0.0001) higher four limb grip strength was observed in vector‐treated mice when compared to mock animals (Figure 4B). A similar trend was observed during follow‐up evaluations at 25 weeks, 35 weeks, 45 weeks, and 55 weeks after gene therapy (Figure 4A,B). Animals were further evaluated for the maximum twitch and tetanic force produced by the TA muscle, which is the major agonist of dorsiflexors during isometric contraction. At 58 weeks after gene transfer, AAV9K51Q‐CoμDys‐treated group had improved twitch force (AAV9K51Q‐CoμDys vs Mock, Mean force 0.34mN/g vs 0.21mN/g, p < 0.0001) (Figure 4C) and tetanic isometric contraction (Mean:1.22mN/g vs 0.70mN/g, p < 0.0001) (Figure 4D).
Phenotypic rescue in DMD mice after systemic gene transfer. Muscle grip strength of hind limbs (A) and all four limbs (B) exerted by mock administered (n = 8) and AAV9K51Q‐CoμDys vector (n = 8) administered groups are shown. Grip strength was measured at 12 weeks, 25 weeks, and 55 weeks after gene therapy. At all the time points, both hind limbs and four limbs grip strength of vector‐treated mice were significantly higher than that of mock‐treated animals. Each data point represents the average value of 5 readings measured per mouse. To assess the complete functional benefit of systemic gene transfer, 60 weeks after vector administration, maximal twitch force (C) and maximal tetanic force (D) were measured from mock‐treated (n = 8) or AAV9K51Q‐CoμDys‐treated (n = 8) animals. The agonist of dorsiflexors, the TA muscle of vector‐treated mice, was stimulated by an electric stimulus and the maximum twitch and tetanic isometric muscle contraction were recorded, and it was found to be significantly higher in AAV9 vector‐treated mice than in mock‐treated mice. The data was recorded individually for right and left limbs of each mouse, then an average was taken and normalised to the bodyweight of the animal. The data obtained are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐CoμDys with respect to the mock group. ***p < 0.0001. For statistical comparison, an unpaired Student t‐test was used (GraphPad Prism 8.0.2 software).
Cardiac dysfunction is observed in mdx mice at a later phase of their life [39]. We performed the echocardiographic analysis of systemic vector‐treated mice at 60 weeks after gene transfer. The representative echocardiographic images are shown in Figure 5A. In treated mice, we observed a significant improvement in left ventricular end systolic diameter (AAV9K51Q‐CoμDys, Mean: 1.63 mm vs Mock, Mean: 2.51 mm, p < 0.05) (Figure 5B) and diastolic diameter (Mean: 2.82 mm vs 3.59 mm, p < 0.05) (Figure 5C), marked by improvement in systolic (2.5‐fold, p < 0.05) (Figure 5D) and diastolic volumes (1.75‐fold, p < 0.05) (Figure 5E). Overall, cardiac function was enhanced in the vector‐treated mice as evident from the improvements in ejection fraction (Mean: 74.50% vs 58.13%, p < 0.05) (Figure 5F) and fractional shortening (Mean: 43.54% vs 30.86%, p < 0.05) (Figure 5G).
Systemic gene transfer of AAV9K51Q‐CoμDys vector improves cardiac function in mdx mice. Echocardiography of mock (n = 8), AAV9K51Q‐CoμDys (n = 8) was performed 60 weeks after systemic gene transfer. Representative M‐mode transthoracic echocardiographic tracings of different vector‐treated and control animals are shown (A). Different cardiac performance parameters, such as systolic diameter (B), diastolic diameter (C), systolic volume (D), diastolic volume (E), ejection fraction (F), and fractional shortening (G), were significantly improved in mice that received AAV9K51Q‐CoμDys vector. The data was analysed in Vevo Lab 5.6.0 software. The data obtained are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐CoμDys with respect to the mock. p < 0.05. For statistical comparison, an unpaired Student t‐test was performed (GraphPad Prism 8.0.2 software).
To corroborate the functional rescue, we performed immunostaining of different skeletal muscles, 62 weeks after gene therapy (Figure 6A). A quantification of dystrophin‐positive fibres in TA muscle (Mean: 50.59% vs 2.35%, p < 0.0001), GAS muscle (Mean: 59.31% vs 1.26%, p < 0.0001), and DIA muscle (Mean: 52.13% vs 3.45%, p < 0.0001) demonstrated higher dystrophin‐positive fibres when compared to mock‐treated mice (Figure 6B). The GAS muscle demonstrates individual myofibers positive for dystrophin separated by an interstitium layer (Figure S11). The DGC proteins were prominently detected in TA muscle sections of vector–treated mice (Figure S12). The improvement in cardiac function is due to the robust expression of dystrophin in the cardiac tissue (Mean: 38.23% vs 0.05%, p < 0.0001) (Figure S13A,B). Further MT staining showed a significant reduction in collagen deposition in TA (~2.7‐fold, p < 0.0001), GAS (~2‐fold, p < 0.0001), and cardiac muscle (~4.24‐fold, p < 0.0001) of vector‐treated mice (Figure S14A,B). H&E staining also demonstrated a decrease in central nucleation of myofibers (Figure 7A,B), signifying the therapeutic benefit of optimised AAV9K51Q‐CoμDys. A biodistribution analysis revealed that the vectors were present in different organs such as brain, heart, lungs, kidney, liver and in skeletal muscles (TA, GAS, soleus, extensor digitorum longus (EDL), and quadriceps (QD)) (Figure S15). Interestingly, no significant changes in B cell and T cell populations were noted in treated vs control animals (Figure S16A–C), and further assessment of capsid‐specific T cell response using ELISPOT assay revealed significantly reduced interferon‐γ secretion in the spleenocytes of vector‐treated mice (Mock vs AAV9K51Q‐CoμDys, Mean: 122 vs 37 spots, p < 0.05) (Figure S16D). These data highlight the safety of the AAV9K51Q‐CoμDys vectors during systemic administration. Finally, creatine kinase (CK) levels were reduced in vector‐treated mice (AAV9K51Q‐CoμDys vs. Mock: Mean: 478.6 U/L vs 760.7 U/L), which underscores the integrity of sarcolemma in vector‐treated mice (Figure S17). Taken together, our data demonstrate the efficacy and safety of AAV9K51Q‐CoμDys vector with a marked improvement in phenotype of mdx mice, up to 1 year after vector administration.
Sustained and robust expression of dystrophin protein in skeletal muscles of mdx mice that received systemic gene transfer. Representative immunofluorescent images of laminin and dystrophin protein in the tibialis anterior (TA), gastrocnemius (GAS), and diaphragm (DIA), of control or vector‐treated mice, 62 weeks post gene transfer (A) and quantitative analysis of dystrophin‐positive fibres (B) are shown. The sections were imaged with the confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). The quantification of dystrophin‐positive fibres was performed from images (n = 9–10/group from n = 3 animals/group) captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) at 20X magnification. The TA, GAS, and DIA sections (n = 3 animals) were co‐stained with laminin and dystrophin, and the laminin was used to quantify the total muscle fibre number. The number of dystrophin‐positive fibres and total fibre number was manually counted and marked in the images using ImageJ software. These were used to quantify the percentage of dystrophin‐positive fibres, as denoted in (B). The data are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐CoμDys with respect to the Mock. ***p < 0.0001. For statistical comparison, an unpaired Student t‐test was performed (GraphPad Prism 8.0.2 software).
Improvement of muscle pathology upon systemic gene transfer in mdx mice. Haematoxylin and Eosin (H&E) staining of Tibialis anterior (TA), gastrocnemius (GAS), and diaphragm (DIA) muscle sections from Mock‐treated or Vector‐treated mice after 62 weeks of systemic gene transfer (A) and quantification of central nucleation in myofibers (B) are shown. Green arrows refer to the centrally located nucleus in myofibers, and yellow arrow denotes the fibrotic area in (A). Representative images were captured from paraffin‐processed or cryosections of H&E‐stained tissue sections. Images were captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) (Scale bar: 75 μm). Quantification of centrally nucleated myofibers in H&E images was performed using Image J analysis as described in the methods section (B). H&E images (n = 8–9 images/group) from n = 3 tissues per group were used for quantification analysis. The data are represented as mean ± SD. Asterisk () represents statistical comparison of AAV9K51Q‐CoμDys with respect to the mock. ***p < 0.0001. For statistical analysis, an unpaired Student t‐test was performed (GraphPad Prism 8.0.2 software).
Discussion
4
The phase I/II clinical trial in DMD patients showed an increased expression of dystrophin and enhanced muscle function in all patients; however, its effects were limited in the phase III trial [9]. Another clinical trial reported fatality following administration of AAV based minidystrophin [15]. These efforts are however very significant, as they have thoroughly evaluated the safety and efficacy of AAV vectors in this difficult‐to‐cure disease, and also underscore the difficulties in obtaining body‐wide expression of AAV‐dystrophin during its clinical translation. These data also highlight the need for robust dystrophin expression at preferably low doses utilising optimal promoter/enhancer combinations, transgenes, capsids, and other additive strategies. The present study is a preclinical evaluation of engineered AAV9 vectors incorporating some of these features to overcome some of the current limitations in DMD gene therapy.
One of the most significant challenges in DMD gene therapy is the limited persistence of transgene in the transduced muscles [40]. Moreover, high doses of AAV vectors are required for delivering the transgene in the cardiac and skeletal muscles that make up an extensive muscle tissue in the body [41]. Influx of high dose of AAV vectors has the potential to elicit immune response [42, 43]. Another aspect is that the AAV capsid proteins can acquire post‐translational modifications (PTMs) during production, and this can influence the infectivity [24]. Thus, studying the PTMs of AAV capsids may be useful for achieving better clinical properties [44]. Since, AAV9 is known to be highly efficient in transducing the skeletal muscles [45], we have used engineered AAV9 vectors modified at their PTM sites in the present study. It is noteworthy that different AAV serotypes, including AAV1, AAV8, AAV9, and AAVrh74, are in clinical trials for efficient muscle transduction [40]. However, cardiac fibre transduction by AAV9 vectors surpasses that of serotypes AAV4, AAV6, AAV7, and AAV8 significantly [46, 47]. Considering this property of AAV9 in skeletal and cardiac muscles, we generated PTM site‐modified AAV9 vectors and have demonstrated its enhanced gene transfer profile in mdx mice. It has to be noted that other PTM‐modified vectors with enhanced transduction efficacy have been employed to achieve superior transduction in multiple tissues [24, 30, 48]. Similarly, MyoAAV engineered through directed evolution for muscle‐targeted gene therapies has shown significantly better transduction following intravenous injection in mice and non‐human primates [49].
While muscle‐specific promoters are an excellent choice for dystrophin overexpression, the systemic DMD pathology on tissues beyond muscles, such as brain [50], retina [35], kidney [51] suggests a potential benefit of incorporating a ubiquitous promoter like CAG. The ubiquitous promoter driven (CAG) SMN1 gene improved survival and phenotypic rescue in spinal muscular atrophy patients [52]. Additionally, the inclusion of a Kozak sequence can enhance protein expression and significantly improve the outcome of gene transfer. Studies with Kozak‐driven microdystrophin variants like ΔAB/R3‐R18/ΔCT [53] or ∆R4‐R23/∆CT [35] have shown improved protein expression in mdx mice. We combined both of these strategies of introducing a CAG promoter and a Kozak sequence in our transgene vector. Our data demonstrates increased microdystrophin expression by CAG‐Kozak‐μDys in vitro and better transduction marked with DGC protein localisation in sarcolemma (in vivo) when compared to muscle specific vectors. Nonetheless, it is also important to acknowledge that the safety profile of μDys expression under a muscle specific promoter such as MHCK7 is likely to be better in higher animal models, but these require additional validation experiments.
Codon‐optimisation is a simple but effective approach to reduce rare tRNA availability and codon bias. AAV‐based gene therapy using codon‐optimised variants of human Factor 8 yielded much higher human factor VIII expression in mice (2% vs 19%, p = 0.0003) than mice that received modified B domain‐deleted version of the human Factor 8 [54]. Similarly, a strong therapeutic protein expression and behavioural improvement have been reported in mice treated with codon‐optimised utrophin delivered using AAV9 [55]. Our study focused on codon‐optimising the microdystrophin transgene to improve therapeutic mRNA stability and protein expression [55]. We observed an improved muscle grip strength of the limbs in animals treated with AAV9K51Q‐CoμDys up to 4 months after muscle‐directed gene transfer. Mice treated with the CoμDys transgene also showed higher levels of dystrophin‐positive fibres, with a concomitant decrease in the number of centralised nuclei and collagen deposition.
In the present study, we have utilised a lower dose ~2 × 10^11^ vg/TA muscle for local muscle gene transfer, with AAV9WT and AAV9K51Q vectors encoding CoμDys that revealed efficient dystrophin expression and phenotypic rescue. Similar findings have been reported, with a AAV9‐mini dystrophin vector administration in TA muscles at a dose of 2.5 × 10^11^ vg/leg that led to a significant increase in isometric force generation, signifying phenotypic rescue [56]. Subsequently, we performed systemic gene transfer of the most optimal vector (AAV9K51Q‐CoμDys) at a low dose of 4 × 10^12^ vg/mouse. A single intravenous administration of these vectors mitigated the dystrophic phenotype of mdx mice by robust and sustained dystrophin expression in both skeletal (~50%, p < 0.0001) and cardiac muscles (~38%, p < 0.0001). It has been previously reported that dystrophin levels between 29%‐57% are sufficient to prevent dystrophy in humans [57], underscoring the translational potential of these vectors. Our findings are also consistent with previous studies, which reported ~65% dystrophin expression in TA muscle of mice that received a single administration of AAVrh74‐MHCK7‐μDys (Dose:6 × 10^12^ vg/mouse), 6 months post gene therapy [32].
Similar to DMD patients, mdx mice also develop myocardial necrosis and fibrosis, leading to cardiomyopathy at 12–21 months of age. We observed a significant improvement in the cardiac parameters and cardiac performance, as evident from improvement in ejection fraction (~1.2‐fold, p < 0.05) and fractional shortening (~1.4‐fold, p < 0.05) in AAV9 vector‐treated mice in comparison to the control mice. These data are consistent with earlier reports, that demonstrated that systemic administration of AAV9‐DWORF vectors (6 × 10^12^ vg/mouse) in mdx mice rescued dilated cardiac myopathic phenotype [58]. Our data also revealed comparable levels of B and T cell populations in the vector‐treated and mock animals upto 62 weeks after systemic gene therapy, highlighting the safety of the vector system. This is consistent with our earlier studies involving PTM‐modified AAV2 vectors, which had shown reduced host immune reaction [24, 30].
Our study has a few limitations. We have not assessed the immunological response generated by the therapeutic protein, dystrophin, which might impact the long‐term efficacy and safety of gene therapy. We also recognise that further dose titration of these vectors by systemic gene transfer and in comparison to muscle specific promoters is needed to evaluate their safety thoroughly [32].
Conclusions
5
Our study, utilising a combination of engineered AAV9 vectors containing an optimised promoter and transgene, has shown substantial phenotypic rescue in a pre‐clinical mouse model of DMD, up to 4 months after intramuscular gene therapy and 1 year after systemic gene therapy. These optimised vectors open additional avenues for their evaluation in higher models of DMD and other neuromuscular diseases such as SMA [59], with a potential for clinical application.
Author Contributions
Mohankumar B. Senthilkumar: investigation, methodology, validation, formal analysis, writing – original draft. Sanya Sharma: investigation, methodology, validation, formal analysis, writing – original draft. Navaneeth Srinivasan: investigation. Anila Varghese: investigation. Pratiksha Sarangi: investigation. Vijayata Singh: investigation. Narendra Kumar: investigation. Devyani Yenurkar: investigation. Sudip Mukherjee: validation, resources. Sonal Amit: investigation. Sameer Bhatia: investigation, validation. Santosh K. Misra: validation, resources, funding acquisition. Ratna Dua Puri: validation, resources, funding acquisition. Jeffrey Chamberlain: writing – review and editing. Giridhara R. Jayandharan: conceptualization, funding acquisition, resources, supervision, writing – original draft, writing – review and editing.
Funding
This study was supported by a Team Science grant (IA/TSG/22/1/600401) awarded by DBT‐Wellcome Trust India Alliance and by IIT Kanpur.
Conflicts of Interest
IIT Kanpur has applied for patents on AAV technology for gene therapy, and a few technologies have been licensed. Other authors declare no competing interests.
Supporting information
Table S1: AAV9 vector quantification. AAV9 vectors packaged with μDys/CoμDys transgene (s) were quantified using a quantitative (qPCR) with polyadenylation (PolyA) signal‐specific primers. The average vector titers obtained from replicate analysis (n = 3) are presented. *AAV9 vectors packaged with plasmid carrying microdystrophin transgene regulated by CAG‐Kozak. Figure S1: Designing of microdystrophin construct. The plasmid pssAAV‐CAG‐Kozak‐μDys was cloned from pLV‐hsa‐μDys to pAAV‐CAG‐eGFP with an additional Kozak sequence upstream of μDys. Figure S2: Expression analysis of microdystrophin constructs in vitro. Transfection data from microdystrophin (μDys) vectors in HeLa (A), C2C12 (B) and H9C2 (C) cells are depicted. Similarly, the viral vectors were packaged in AAV9WT capsid, and transduction was performed in HeLa (D), C2C12 (E) and H9C2 (F) cells. A quantitative (q)PCR was performed to detect microdystrophin expression. The μDys expression levels from plasmid vector constructs MHCK7 vs CAG promoter are shown in panels A–C. The comparison of μDys expression with AAV9WT vectors containing the CAG promoter (AAV9WT‐μDys) vs the muscle‐specific promoter (AAV9WT‐MHCK7‐μDys) is shown in panels D–F. The expression levels were compared to the MHCK7‐μDys/AAV9WT‐MHCK7‐μDys and normalised to 18S rRNA expression. The data are represented as mean ± SEM. *Represents statistical comparisons with respect to MHCK7‐μDys/AAV9WT‐MHCK7‐μDys experimental group. *p < 0.05, **p < 0.01, ***p < 0.001. The data is representative of one independent experiment (total, n = 2 biological replicates) with three technical replicates per group. qPCR data was analysed by the Biorad CFX Manager 3.1. Figure S3: Intramuscular administration of AAV9 vector in TA muscle of DMD mice. For vector administration, 6–10 weeks old mdx mice were anaesthetised and the TA muscle was exposed by making a small incision. Vectors were administered using a Hamilton syringe and the incision was sutured following the administration. Image created using Biorender.com. Figure S4: AAV9 vector containing a ubiquitous promoter showed enhanced dystrophin expression. Immunohistochemistry of the TA muscle of vector‐treated mice was performed, and representative images of dystrophin/laminin‐stained sections at 33 weeks after AAV9WT‐CAG‐Koz‐μDys (AAV9WT‐μDys) and AAV9WT‐MHCK7‐μDys vector administration are shown (A). The sections were imaged with the confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). Representative images of dystrophin‐glycoprotein complex (DGC) protein expression in TA muscle after 33 weeks of vector administration are shown (B). Representative images were captured at a magnification of 20X with Advanced High Sensitive Spectral Confocal and Multiphoton Microscope LSM780NLO, CarlZeiss GmbH, Wein, Austria. Scale bar is 50 μm. The quantification of dystrophin‐positive fibres was performed from images (n = 29–35 images/group from n = 4 animals/group) captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) in 20X magnification (C). To quantify the number of total muscle fibres in a particular field, TA muscle sections were costained with laminin along with dystrophin. These laminin‐stained images were used to quantify the number of total muscle fibres. The number of dystrophin‐positive fibres and total fibre number was manually counted and marked in the images using ImageJ software and their percentage is denoted in (C). The data are represented as mean ± SD *Represents statistical comparison of AAV9WT‐MHCK7‐μDys and AAV9WT‐μDys with respect to the mock and # refers to statistical comparison of AAV9WT‐μDys with respect to AAV9WT‐MHCK7‐μDys experimental group. ****^/####^ p < 0.0001. For statistical analysis, a one‐way ANOVA was performed with GraphPad Prism 8.0.2 software. Figure S5: Rationally engineered AAV9 vectors demonstrate enhanced transduction efficiency. Transduction assay was performed to assess the efficiency of engineered AAV9 vectors packaged with CAG‐Kozak ‐microdystrophin. Infection of HeLa (A) and C2C12 cells (B) was performed at an MOI of 1 × 10^5^ vgs/cell. Quantitative (q) PCR was performed to detect microdystrophin expression. The expression levels were compared to the cell control group and data was normalised to 18S rRNA levels. The data obtained is represented as mean ± SEM. *Represents statistical comparison of AAV9WT‐μDys, AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to the cell control and # refers to statistical comparison of AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to AAV9WT‐μDys. ns‐ non‐significant, *^/#^ p < 0.05, **^/##^ p < 0.01, ***p < 0.001. The data is representative from an independent experiment (total, n = 2 experiments) with three technical replicates per group. qPCR data was analysed by the Biorad CFX Manager 3.1. Figure S6: Muscle grip strength is enhanced in mdx mice treated with engineered AAV9 vectors. Muscle grip strength of hind limbs (A) and all four limbs (B) is represented at 12 weeks post gene therapy for mock (n = 8), AAV9WT‐μDys (n = 8), AAV9K51Q‐μDys (n = 7), and AAV9N57Q‐μDys (n = 8). Follow‐up of 18 weeks post vector administration for grip strength of hind limbs (C) and four limbs (D) in mock (n = 6), AAV9WT‐μDys (n = 7), AAV9K51Q‐μDys (n = 6), and AAV9N57Q‐μDys (n = 7) are represented. At 12 weeks after vector administration, one animal from each group was withdrawn for IHC analysis (data not shown), and one animal from the mock group died. Each data point represents the average value of 5 readings measured per mouse. The data are represented as mean ± SD. *Represents statistical comparison of AAV9WT‐μDys, AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to the Mock, and # refers to statistical comparison of AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to AAV9WT‐μDys. ns‐ non‐significant, *^/#^ p < 0.05, ***p < 0.001, ****^/####^ p < 0.0001. For statistical comparison, a one‐way ANOVA was performed (GraphPad Prism 8.0.2 software). Figure S7: Localisation of DGC proteins in TA muscle sections of mice administered with PTM modified AAV9 vectors. Representative immunofluorescence images for the expression of dystrophin costained with laminin and other dystrophin‐glycoprotein complex (DGC) proteins like dystrobrevin and β‐dystroglycan in TA muscle of mdx mice, 24‐weeks after vector administration (A) are shown. Laminin/dystrophin representative images are taken in a confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). Quantification of dystrophin‐positive fibres in treated mice (B). The quantification of dystrophin‐positive fibres was done from images (n = 14–17/group) captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) in 20X magnification. For each group, n = 3 animals and n = 3 sections per muscle were chosen for imaging analysis. The laminin and dystrophin co‐stained images were used to quantify the total muscle fibre number and dystrophin positive fibres, respectively. The number of dystrophin‐positive fibres and total fibre number was manually counted and marked in the images using ImageJ software and their percentage is denoted. The data obtained are represented as mean ± SD. *Represents statistical comparison of AAV9WT‐μDys, AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to the Mock and # refers to statistical comparison of AAV9K51Q‐μDys, AAV9N57Q‐μDys with respect to AAV9WT‐μDys. ****^/####^ p < 0.0001. For statistical comparison, a one‐way ANOVA was performed (GraphPad Prism 8.0.2 software). Figure S8: Collagen deposition in vector administered DMD mice. Masson's trichrome (MT) staining of TA muscle sections from Mock‐treated or Vector treated mice (after 17 weeks) was performed. Yellow arrow highlight collagen deposition (A). The top panel, A (i–iii) represents the whole TA muscle section images (Magnification: 4X) from Mock or AAV9K51Q vector‐treated mice (Scale bar: 500 μm). The magnified images (20X magnification) of the same MT‐stained TA muscle sections from the above panel (i–iii) are depicted in the bottom panel, A (iv–vi) (Scale bar: 75 μm). Representative MT stained images were captured from paraffin or cryosections of the tissues. The quantification was performed as described in the methods section. Images were captured in an inverted light microscope (Dmi8, Leica Microsystems, Germany). Collagen‐deposited areas in MT images were quantified using ImageJ analysis as described in the methods section (B). Three tissues per group were analysed (n ≥ 10 images/group). The data are represented as mean ± SD. *Represents statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys with respect to the mock and # refers to the statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys experimental group. ****^/####^ p < 0.0001. For statistical analysis, a one‐way ANOVA was performed with GraphPad Prism 8.0.2 software. Figure S9: Morphological analysis of vector‐administered DMD mice. Haematoxylin and Eosin (H&E) staining of TA muscle sections from vector‐treated mice (after 17 weeks) was performed. Green arrow refers to the centrally located nucleus in myofibers, and yellow arrow denotes the fibrotic areas (A). The top panel, A (i–iii) represents the whole TA muscle section images (Magnification: 4X) from Mock‐ or AAV9K51Q vector‐treated mice (Scale bar: 500 μm). The magnified images (20× magnification) of the same H&E‐stained TA muscle sections (i–iii) are depicted in the bottom panel (iv–vi). (Scale bar: 75 μm). Representative images were captured from paraffin‐processed or cryosections of H&E stained tissue sections. A semi‐quantification was performed as described in the methods section. Images were captured in an inverted microscope (Dmi8, Leica Microsystems, Germany). Quantification of centrally nucleated myofibers in H&E images was performed using ImageJ analysis as described in the methods section (B). H&E images (n ≥ 15 images/group) from n = 2–3 tissues per group were used for quantification analysis. The data are represented as mean ± SD. *Represents statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys with respect to the mock, and # refers to statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys experimental group. ****^/####^ p < 0.0001. For statistical analysis, a one‐way ANOVA was executed with GraphPad Prism 8.0.2 software. Figure S10: Vector genome quantification of TA muscles from mice upon intramuscular gene transfer. Vector copy numbers in the TA muscles were analysed using quantitative PCR. Briefly, total DNA (n = 4 animals/group) was isolated and 20 ng of DNA was amplified using ITR specific primer pair. The quantification was performed using suitable plasmid standards. The data is represented as mean ± SD. *Represents the statistical comparison of AAV9K51Q‐μDys and AAV9K51Q‐CoμDys with respect to the mock and # refers to the statistical comparison of AAV9K51Q‐CoμDys with respect to AAV9K51Q‐μDys experimental group. ****^/####^ p < 0.0001. For statistical analysis, a one‐way ANOVA was performed with GraphPad Prism 8.0.2 software. Figure S11: Restored dystrophin expression in discrete myofibers of mdx that received systemic gene transfer. Representative images of laminin and dystrophin‐stained gastrocnemius muscle of mock or vector‐treated mice 1 year after systemic gene transfer. The immunostained images depict the co‐localisation of laminin and dystrophin in different myofibers separated by an interstitium. Representative images were captured using Fluoview FV4000 confocal laser scanning microscope (Evident (Olympus), Tokyo, Japan) at 60X magnification (Scale bar: 25 μm). Figure S12: Restoration of DGC proteins in the sarcolemma of TA muscle of mice after systemic gene transfer. Representative immunofluorescent images of dystrophin glycoprotein complex (DGC) proteins like dystrobrevin and β‐dystroglycan in the tibialis anterior (TA) muscle of mock or vector‐treated mice, 62 weeks post gene transfer. The sections were imaged with the confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). Figure S13: Increased dystrophin expression in cardiac muscle of vector‐treated dystrophic mice. Representative images of dystrophin‐stained heart tissue sections of vector‐treated mice after 62 weeks of systemic vector administration (A) and quantification of dystrophin positive fibres (B). The sections were imaged with the confocal microscope (AXR Nikon) with a 40X objective (Scale bar: 50 μm). The quantification of dystrophin‐positive fibres was done from images (n = 12–14 images/group from n = 3 animals/group) captured in an inverted microscope (Dmi8, Leica Microsystems, Germany) in 20X magnification. The heart muscle sections (n = 3 animals) were stained with wheat germ agglutinin to label the sarcolemma and imaged in an inverted microscope (Dmi8, Leica Microsystems, Germany) at 20X magnification. These wheat germ agglutinin‐stained images of heart sections (n = 15 images from 3 different animals) were used to quantify the total muscle fibre number. The number of dystrophin‐positive fibres and total fibre number was manually counted and marked in the images using ImageJ software. The total muscle fibres in the heart (~162) (data not shown) were used to quantify the percentage of dystrophin‐positive fibres as denoted in B. The data are represented as mean ± SD. *Represents statistical comparison of AAV9K51Q‐CoμDys with respect to the Mock. ****p < 0.0001. For statistical comparison, an unpaired Student t‐test was performed (GraphPad Prism 8.0.2 software). Figure S14: Reduced fibrosis in DMD mice upon systemic gene transfer. Masson's trichrome (MT) staining of Tibialis anterior (TA), gastrocnemius (GAS), and Heart muscle sections from Mock‐treated or AAV9K51Q vector treated mice after 62 weeks of systemic gene transfer (A) and quantification of collagen deposition in myofibers (B) were performed. Yellow arrow indicates the collagen deposition in A. MT staining and quantification of fibrotic areas was performed from paraffin‐processed tissue sections of experimental animals. Images were captured in an Inverted microscope (Dmi8, Leica Microsystems, Germany) (Scale bar:75 μm). Quantification of collagen deposition in MT images was performed using Image J analysis as described in the methods section (B). MT images (n > 15 images/group) from n = 3 tissues per group were used for quantification analysis. The data are represented as mean ± SD. *Represents statistical comparison of AAV9K51Q‐CoμDys with respect to the mock. ****p < 0.0001. For statistical analysis, an unpaired Student t‐test was performed with GraphPad Prism 8.0.2 software. Figure S15: Biodistribution of vectors in mice that received systemic gene transfer. Vector copy number was assessed by quantitative PCR in the skeletal muscles, like tibialis anterior, gastrocnemius, quadriceps, extensor digitorum longus, soleus, diaphragm, and in major organs like brain, heart, lungs, liver and kidney. Briefly, total DNA (n = 4 animals/group and technical replicates n = 2/tissue) was isolated from about 300 mg tissue of each mice, except for soleus and extensor digitorum longus, in which the whole tissue was used. Approximately, 200 ng of DNA was amplified using CoμDys‐specific primers. The data is represented as mean ± SD. *Represents the statistical comparison of AAV9K51Q‐CoμDys with respect to the mock. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. For statistical analysis, an unpaired Student's t‐test was performed with GraphPad Prism 8.0.2 software. Figure S16: Immunological safety of optimised AAV9‐CoμDys vector after systemic gene therapy. Blood was collected at 62 weeks after gene transfer from mice that received mock or vector treatment, and were assessed for T cell and B cell markers by flow cytometry. Data for helper T cells (CD3^+^CD4^+^) (A), cytotoxic T cells (CD3^+^CD8^+^) (B) and B cells (CD19^+^) (C) from the peripheral blood of mock‐treated (n = 8), and AAV9K51Q‐CoμDys‐treated (n = 8) mice are shown. Spleenocytes (1 × 10^6^ cells/well) from experimental animals were stimulated with concanavalin A (2 μg/mL; positive control) or an AAV capsid–specific peptide (2 μg/mL). The IFN‐γ ELISPOT assay was performed as per the manufacturer's protocol (Mabtech, Nacka Strand, Sweden). Representative images of the spots generated from mock‐treated (n = 4), and AAV9K51Q‐CoμDys treated (n = 4), and their quantification data are provided in (D). The data are represented as mean ± SD. *Represents statistical comparison of AAV9K51Q‐CoμDys with the mock group, *p < 0.05. ns‐ non‐significant. For statistical analysis, an unpaired Student t‐test was performed with GraphPad Prism 8.0.2 software. Figure S17: Level of creatine kinase (CK) in mdx after 1 year of systemic vector administration. Blood was collected at 62 weeks after gene transfer from mice that received mock or vector treatment, and plasma was isolated. The level of creatine kinase (CK) was measured in a Fully Automated Clinical Chemistry Analyser (EM 200) as per the manufacturer's protocol. Data represented is from n = 7 animals from each mock and vector treated animals. For statistical analysis, an unpaired Student's t‐test was performed with GraphPad Prism 8.0.2 software.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1N. Deconinck and B. Dan , “Pathophysiology of Duchenne Muscular Dystrophy: Current Hypotheses,” Pediatric Neurology 36, no. 1 (2007): 1–7.17162189 10.1016/j.pediatrneurol.2006.09.016 · doi ↗ · pubmed ↗
- 2M. Chang , Y. Cai , Z. Gao , et al., “Duchenne Muscular Dystrophy: Pathogenesis and Promising Therapies,” Journal of Neurology 270, no. 8 (2023): 3733–3749.37258941 10.1007/s 00415-023-11796-x · doi ↗ · pubmed ↗
- 3C. Angelini , E. Pegoraro , E. Turella , M. Intino , A. Pini , and C. Costa , “Deflazacort in Duchenne Dystrophy: Study of Long‐Term Effect,” Muscle & Nerve 17, no. 4 (1994): 386–391.8170484 10.1002/mus.880170405 · doi ↗ · pubmed ↗
- 4L. Merlini , A. Cicognani , E. Malaspina , et al., “Early Prednisone Treatment in Duchenne Muscular Dystrophy,” Muscle & Nerve: Official Journal of the American Association of Electrodiagnostic Medicine 27, no. 2 (2003): 222–227.10.1002/mus.1031912548530 · doi ↗ · pubmed ↗
- 5V. Ricotti , D. A. Ridout , E. Scott , et al., “Long‐Term Benefits and Adverse Effects of Intermittent Versus Daily Glucocorticoids in Boys With Duchenne Muscular Dystrophy,” Journal of Neurology, Neurosurgery, and Psychiatry 84, no. 6 (2013): 698–705.23250964 10.1136/jnnp-2012-303902 · doi ↗ · pubmed ↗
- 6D. J. Birnkrant , K. Bushby , C. M. Bann , et al., “Diagnosis and Management of Duchenne Muscular Dystrophy, Part 2: Respiratory, Cardiac, Bone Health, and Orthopaedic Management,” Lancet Neurology 17, no. 4 (2018): 347–361.29395990 10.1016/S 1474-4422(18)30025-5PMC 5889091 · doi ↗ · pubmed ↗
- 7R. Goossens , N. Verwey , Y. Ariyurek , F. Schnell , and A. Aartsma‐Rus , “DMD Antisense Oligonucleotide Mediated Exon Skipping Efficiency Correlates With Flanking Intron Retention Time and Target Position Within the Exon,” RNA Biology 20, no. 1 (2023): 693–702.37667454 10.1080/15476286.2023.2254041 PMC 10481881 · doi ↗ · pubmed ↗
- 8K. Bushby , R. Finkel , B. Wong , et al., “Ataluren Treatment of Patients With Nonsense Mutation Dystrophinopathy,” Muscle & Nerve 50, no. 4 (2014): 477–487.25042182 10.1002/mus.24332 PMC 4241581 · doi ↗ · pubmed ↗