Proteomics-based evaluation of AAV dystrophin gene therapy outcomes in mdx skeletal muscle

Abstract

Duchenne muscular dystrophy (DMD) is a fatal genetic muscle-wasting disease characterized by loss of dystrophin protein. Therapeutic attempts to restore a functional copy of dystrophin to striated muscle are under active development, and many utilize adeno-associated viral (AAV) vectors. However, the limited cargo capacity of AAVs precludes delivery of full-length dystrophin, a 427 kDa protein, to target tissues. Recently, we developed a method to express large dystrophin constructs using the protein trans-splicing mechanism mediated by split inteins and myotropic AAV vectors. The efficacy of this approach to restore muscle function in mdx4cv mice was previously assessed using histology, dystrophin immunolabeling, and Western blotting. Here, we expand our molecular characterization of dystrophin constructs with variable lengths using a mass spectrometry-based proteomics approach, providing insight into unique protein expression profiles in skeletal muscles of wild-type, dystrophic mdx4cv, and AAV-treated mdx4cv mice. Our data reveal several affected cellular processes in mdx4cv skeletal muscles with changes in the expression profiles of key proteins to muscle homeostasis, whereas successful expression of dystrophin constructs results in an intermediate to complete restoration. This study highlights several biomarkers that could be used in future preclinical or clinical studies to evaluate the effectiveness of therapeutic strategies.

Keywords: Biomarkers; Gene therapy; Genetics; Muscle biology; Proteomics.

Figure 1. Schematic representation of dystrophin clones tested, split intein approach to express large constructs, and proteomics workflow.

(A) Structural organization of full-length dystrophin (muscle isoform Dp427), μDys currently evaluated in clinical trials, and midi-dystrophin (ΔSR5–15). (B) Dual vector strategy to express a midi-dystrophin using split intein gp41.1. (C) Triple vector strategy to reexpress full-length dystrophin using 2 orthogonal split inteins Nrdj1 and gp41.1. (D) Workflow for characterization of the protein expression profile in mdx^4cv skeletal muscle. Gastrocnemius muscles were isolated from WT, saline-treated mdx^4cv, or systemically injected mdx^4cv with low doses of AAVMYO1. Total proteins from 6 muscles per group were extracted and labeled with TMT isobaric tags before protein quantification using LC-MS/MS. Fl-Dys, full-length dystrophin.

Figure 2. Detection of dystrophin expression and quantification of peptide-specific abundance using proteomics.

Dystrophin peptide abundances were quantified using TMT proteomics in gastrocnemius muscle samples of WT, mdx^4cv treated with saline or AAVMYO1 to express μDys-mdx^4cv, midiDys-mdx^4cv, and full-Dys-mdx^4cv. (A) Quantified abundance of AAV-mediated dystrophin expressed in mdx^4cv mice versus endogenous full-Dys in WT mice. (B) Abundance of peptides present only in full-Dys (endogenous dystrophin in WT and full-Dys construct via triple AAVMYO1 treatment). (C) Abundance of peptides specific to transgenic/human dystrophins (shared between μDys5, midi-Dys, and full-Dys constructs) delivered by AAVMYO1. (D) Abundance of peptides specific to large dystrophins (endogenous WT dystrophin, or AAV-delivered midi-Dys and full-Dys constructs). The non-zero value for dystrophin peptides in saline-treated mdx^4cv is most likely due to coisolation interference common to TMT proteomics analyses. Bar graphs depict mean ± SEM of n = 6 mice/group, except peptide L7 and L10, which were n =3. Comparisons between groups were made using 1-way ANOVA with Tukey’s multiple-comparison test. ***P < 0.001 versus WT; ^$$$P < 0.001 versus mdx^4cv saline. μDys: micro-dystrophin, mDys and midi-Dys: midi-dystrophin, fDys and full-Dys: full-length dystrophin.

Figure 3. Histology analysis of gastrocnemius muscle cross sections showing improvements with dystrophin constructs.

(A) Representative images of gastrocnemius muscle cross sections stained with H&E or trichrome (top rows, scale bars: 50 μm), or immunolabeled with antibodies specific for periostin (scale bars: 50 μm) or dystrophin-glycoprotein elements (lower panel, scale bars: 100 μm). These images were acquired in RGB colors but inverted to black and white for better visualization. The original panel is presented in Supplemental Figures 2 and 3. (B) Percentage of dystrophin-positive fibers; 600–1000 myofibers were counted per sample, with n = 6 analyzed per group. (C) The collagen area of the gastrocnemius muscle was measured using trichrome-stained cross sections. n = 5 samples per group. (D) Gastrocnemius myofiber area and (E) minimal Feret’s diameter. More than 700 myofibers per sample from n = 6 per group were analyzed. The average values are shown on top of the violin bars. The solid line represents the median, while the dashed lines show the quartiles. (F) Periostin area measured from cross-section muscle sections immunolabeled with specific antibodies against periostin. n = 6 samples per group. (G) Periostin abundance level detected from proteomics analysis of gastrocnemius muscles. (H) Abundance levels of different collagens were measured using the proteomics method from gastrocnemius samples. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT; ^$P < 0.05, ^$$P < 0.01, ^$$$P < 0.001 versus saline group; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus μDys group; ^&&P < 0.01 versus midi-Dys group using 1-way ANOVA followed by Tukey’s post hoc test. Dys⁺: dystrophin-positive. H&E, hematoxylin and eosin; μDys, micro-dystrophin; mDys and midi-Dys, midi-dystrophin; fDys and full-Dys, full-length dystrophin.

Figure 4. Comparison of protein expression profiles between experimental groups.

Protein expression profiles in gastrocnemius muscle were compared between WT mice and (A) saline-mdx^4cv, (B) mdx^4cv injected with single AAVMYO1 μDys, (C) mdx^4cv injected with dual AAVMYO1 to express midi-dystrophin, or (D) mdx^4cv injected with triple AAVMYO1 vector to express full-length dystrophin. A 2-tailed, unpaired Student’s t test was used to calculate P values for pairwise fold changes, and the Benjamini-Hochberg method was used to control the false discovery rate (FDR). Corrected P values were log-transformed and plotted against log-transformed fold change values to obtain volcano plots, and a minimum corrected P-value cutoff of 0.05 and minimum relative fold change cutoff of ±1 were applied to identify differentially expressed proteins (DEPs) in pairwise comparisons. Data were collected from a sample size of n = 6 per group.

Figure 5. Analysis of protein expression profile demonstrates proteomic rescue by dystrophin constructs.

(A) Heatmap depicting the top upregulated and downregulated proteins between WT and mdx^4cv muscle. Gene ontology (GO) enrichment analysis was performed using GOrilla and g:Profiler to determine the molecular function (MF), biological process (BP), and cellular compartment (CC) enrichment of significantly (B) upregulated and (C) downregulated proteins in mdx^4cv gastrocnemius muscle compared with WT muscle.

Figure 6. Dystrophin delivery alleviates DGC protein defects in mdx^4cv mice.

Relative abundance of (A) sarcoglycans, (B) dystroglycan, dystrobrevin, (C) syntrophins, (D) utrophin, and (E) protein-arginine deiminase type-2, myoglobin, and tubulin β6 class V measured by the proteomics method. Bar graphs depict mean ± SEM from n = 5–6 mice/group. Comparisons between groups were made using 1-way ANOVA with Tukey’s multiple-comparisons test. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT; ^$P < 0.05, ^$$P < 0.01, ^$$$P < 0.001 versus saline group; ^##P < 0.01, ^###P < 0.001 versus μDys group; ^&P < 0.05, ^&&P < 0.01 versus midi-Dys. μDys, micro-dystrophin; mDys, midi-dystrophin; fDys, full-length dystrophin.

Figure 7. Amelioration of altered membrane repair and myogenesis pathway markers in mdx^4cv muscle mediated by AAV-dystrophin constructs.

(A) Heatmap showing elevated expression of various proteins implicated in membrane trafficking and repair in mdx^4cv gastrocnemius muscle and partial restoration with μDys5, midi-dystrophin, or full-length dystrophin delivered by AAV vectors. (B) Annexin (A1, A4, and A5) abundance in WT, dystrophic, or AAV-treated muscles. (C) Abundance of proteins involved in muscle repair. (D) Expression of key proteins involved in membrane trafficking and remodeling. (E) Galectin-1 and (F) galectin-3 abundance in mdx^4cv and WT muscles. Bar graphs represent mean ± SEM of n = 6 mice/group. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT; ^$P < 0.05, ^$$P < 0.01, ^$$$P < 0.001 versus saline group using 1-way ANOVA followed by Tukey’s post hoc test. μDys, micro-dystrophin; mDys, midi-dystrophin; fDys, full-length dystrophin.

Figure 8. Proteins with unrestored expression in mdx^4cv mice treated with various dystrophin constructs.

Heatmaps displaying proteins that did not display significant restoration to WT levels in (A) μDys5-mdx^4cv, (B) midi-Dys-mdx^4cv, or (C) full-Dys-mdx^4cv gastrocnemius muscles. Exemplary proteins with unrestored levels in AAV-Dys construct groups include (D) carboxylesterase 1D (Ces1d), (E) spermine oxidase (Smox), (F) tRNA methyltransferase 10 homolog C (Trmt10c), (G) adenosylmethionine decarboxylase (Amd1), (H) histone H1.2 (H1-2), (I) myosin light chain 6B (Myl6b), (J) heme-binding protein (Hebp1), (K) nicotinamide nucleotide transhydrogenase (Nnt), and (L) eukaryotic translation initiation factor 2D (Eif2d). Bar graphs depict mean ± SEM from n = 5–6 mice/group. Comparisons between groups were made using 1-way ANOVA with Tukey’s multiple-comparison test. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT; ^$P < 0.05, ^$$$P < 0.001 versus saline; ^#P < 0.05, ^##P < 0.01 versus μDys group. μDys, micro-dystrophin; mDys, midi-dystrophin; fDys, full-length dystrophin.