Lysosomal damage is a therapeutic target in Duchenne muscular dystrophy

Abstract

Duchenne muscular dystrophy (DMD), a muscle degenerative disease affecting young boys, arises from the loss of dystrophin. Current gene therapy approaches aim to restore a shortened form of dystrophin (microdystrophin) via adeno-associated vector delivery. While recent clinical studies show promise, therapeutic efficacy remains incomplete, emphasizing the need for improved approaches. Here, we identified lysosomal perturbations in myofibers of patients with DMD and animal models, an overlooked mechanism of cellular damage in muscular dystrophies. These were notably marked by the up-regulation and recruitment of Galectin-3, a biomarker of lysosomal membrane permeabilization, to lysosomes, alongside alterations in lysosome number, morphology, and function. Microdystrophin therapy in Dmd^mdx mice fails to fully correct these damages. However, combining it with trehalose, a lysosome-protective disaccharide, substantially improves the outcome, enhancing muscle function, myopathology, and transcriptome. These findings highlight lysosomal damage as an important pathomechanism in DMD and suggest that combining trehalose with gene therapy could enhance therapeutic efficacy.

Fig. 1.. LGALS3 up-regulation in dystrophic muscle correlates with LMP.

(A) Western blot analysis of TA, GA, and diaphragm muscles and serum samples of 7-week-old WT and Dmd^mdx-4Cv mice (n = 4 to 6) for LGALS3. Total protein staining was used as the loading control. (B) Relative Lgals3 mRNA expression in the TA, GA, and diaphragm, normalized to Rplp0 (FC shown, average ± SD, n = 6). (C) Band intensity quantification of LGALS3 relative to total protein staining in TA, GA, and diaphragm muscles (FC shown, average ± SD, n = 4 to 6). (D) Band intensity quantification of LGALS3 relative to total protein staining in serum (FC shown, average ± SD, n = 6). (E) Representative confocal images of TA muscle cross sections immunostained for LAMP2 (green) and LGALS3 (red). Scale bars, 100 or 25 μm (zoomed-in images). (F) Confocal images of serial transversal sections of TA muscles from a Dmd^mdx-4Cv mouse immunostained for LAMP2/LGALS3 and LAMININ/CD11b. Zoomed-in images show the inflammatory area of dystrophic muscle. Scale bars, 200 or 50 μm (zoomed-in images). (G) Quantification of double-positive puncta (LAMP2⁺LGALS3⁺) in the myofibers of TA muscles of WT and Dmd^mdx-4Cv mice (average ± SD, >400 myofibers analyzed from n = 6). (H) Representative confocal images of transversal sections from muscles of a patient with DMD and a healthy control (2-year-old) and a GRMD dog and a WT control (6-month-old), immunostained for LAMP2 and LGALS3. Left panels show H&E labeling from the same muscle biopsies. Scale bars, 25 or 10 μm (zoomed-in images) and 50 μm (H&E images). An unpaired two-tailed t test was used for statistical comparisons of WT and Dmd^mdx-4Cv groups. t test: *P < 0.05, **P < 0.01, and ****P < 0.0001.

Fig. 2.. Lysosomal stress in dystrophic muscle.

(A) Western blot analysis of GA muscle for LAMP2 (n = 6). (B) Band density quantification of LAMP2 relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (average ± SD, n = 6). (C) Reconstructed 3D images of isolated myofibers from the FDB muscle of Dmd^mdx-4Cv and WT mice, immunostained for LAMP2 and LGALS3. Scale bars, 10 or 20 μm. (D) 3D modeling of LAMP2 spots. Green spots show average sized lysosomes (diameter <1.5 μm), and purple spots show enlarged lysosomes (diameter >1.5 μm). Scale bars, 10 μm. (E) FC of the number of lysosomes per myofiber (average ± SD, n = 4). (F) Analysis of five nearest spots to each spot, represented as a density plot (with a nonlinear fit Gaussian curve). (G) Quantification of LAMP2 spot diameter (min to max points). (H) Density plot showing the distribution of LAMP2 spot diameter. (I) Ultrastructural transmission electron micrograph of Dmd^mdx-4Cv and WT TA muscles. The left panel shows low-magnification images, and red circles indicate larger and denser lysosomes in Dmd^mdx-4Cv muscles compared to WT muscles (lysosome indicated by a blue circle). The right panel shows a high-magnification image highlighting a large (390 nm) lysosome found in the Dmd^mdx-4Cv muscle. Scale bars, 3 μm (upper left panel), 1 μm (lower left panel), and 300 nm (zoomed-in image). (J) Quantification of lysosome diameter from electron micrographs (nm; min to max values shown) (n = 2). (K) Representative confocal images show the fluorescence of DQ-Red BSA after an uptake of 4 hours in primary myoblasts isolated from WT and Dmd^mdx-4Cv muscles. Bright fluorescence puncta of DQ-Red BSA indicate delivery to acidic lysosomes of the cells. Scale bars, 50 and 10 μm (zoomed-in images). (L) Quantification of MFI per cell in each condition (n > 158 cells). An unpaired two-tailed t test was used for statistical comparisons of WT and Dmd^mdx-4Cv groups. t test: *P < 0.05, ***P < 0.001, and ****P < 0.0001.

Fig. 3.. The lysosomal damage assay in primary myoblasts reveals early defects.

(A) Schematic representation of the experimental setup. Cells were treated with LLOMe (2.5 mM) for 30 min to induce LMP and deacidification, resulting in decreased LysoTracker fluorescence. Washout steps indicate LLOMe removal and incubation with fresh media, enabling lysosomal repair and reacidification, reflected by the recovery of LysoTracker red fluorescence. (B) Representative confocal images of WT and Dmd^mdx-4Cv primary myoblasts immunostained for LAMP2 and LGALS3. Scale bars, 30 or 10 μm. (C to E) Violin plots showing the number of LGALS3-positive puncta per cell (normalized to cell surface area). At least 65 cells per condition were quantified from two independent experiments. (C) Comparison between untreated WT and Dmd^mdx-4Cv cells. LLOMe treatment and washout results are shown for (D) WT and (E) Dmd^mdx-4Cv cells. (F) Representative confocal images of WT and Dmd^mdx-4Cv primary myoblasts stained with LysoTracker Red DND-99 after 30 min of incubation. Scale bars, 10 μm. (G to I) Violin plots showing the number of LysoTracker-positive puncta per cell (normalized to cell surface area). (G) Comparison between untreated WT and Dmd^mdx-4Cv cells. LLOMe treatment effects are shown in (H) WT and (I) Dmd^mdx-4Cv cells. At least 70 cells per condition were analyzed. (J to L) MFI per LysoTracker dot, represented as density plots with nonlinear Gaussian curve fitting. (J) Baseline comparison between WT and Dmd^mdx-4Cv cells (unpaired two-tailed t test, *P = 0.0049). LLOMe treatment effects are shown in (K) WT and (L) Dmd^mdx-4Cv cells. For statistical analysis, unpaired two-tailed t tests were used for comparisons between WT and Dmd^mdx-4Cv. ****P < 0.0001. A one-way ANOVA with Tukey’s multiple comparisons test was used to assess differences across treatment conditions within the same genotype. P values: *P < 0.05 and ****P < 0.0001; ns, nonsignificant.

Fig. 4.. Evaluation of lysosomal biogenesis and repair pathways in Dmd^mdx-4Cv mice.

(A) Confocal images of serial transversal sections of GA muscle immunostained for TFEB and laminin. Scale bars, 100 or 25 μm. (B) Quantification of TFEB-positive nuclei (average ± SD, n = 5). (C) RNA sequencing analysis of GA muscle from 7-week-old Dmd^mdx-4Cv and control mice, displaying the expression of certain genes under the control of TFEB (average FC shown, n = 5). (D) Representative confocal images of WT and Dmd^mdx-4Cv primary myoblasts immunostained for TFEB. Scale bars, 40 or 10 μm. (E) Western blot analysis of nuclear and cytosolic fractions from WT and Dmd^mdx-4Cv primary myoblasts for TFEB. Α-tubulin and lamin B were used as loading controls for cytosolic and nuclear fractions, respectively. (F) Band intensity quantification of TFEB relative to Α-tubulin (for cytosolic fractions) and lamin B (for nuclear fractions) (average ± SD). (G) Representative confocal images of TA transversal sections immunostained for ALIX and LAMP2. Scale bars, 25 μm. (H) Quantification of LAMP2⁺ALIX⁺ puncta per myofiber (average ± SD, >692 myofibers analyzed from n = 6). (I) Western blot of TA muscle for ALIX and (J) band density quantification of ALIX relative to GAPDH (average ± SD, n = 6). An unpaired two-tailed t test was used for statistical comparisons of WT and Dmd^mdx-4Cv groups. t test: *P < 0.05, **P < 0.01, and ****P < 0.0001.

Fig. 5.. Assessment of autophagy and lysophagy in Dmd^mdx-4Cv muscle.

(A) Western blot analysis of TA muscle for LC3, with GAPDH and total protein staining used as loading controls. Graphs show the band density quantification of LC3-I and LC3-II/LC3-I ratio, normalized to GAPDH (average ± SD, n = 6). (B) Western blot analysis of TA muscle for SQSTM1, with GAPDH and total protein staining used as loading controls. The graph shows the band density quantification of SQSTM1, normalized to GAPDH (average ± SD, n = 6). (C) Representative confocal images of TA transversal sections immunostained for LAMP2 and LC3B. White particles in merged images depict LAMP2⁺LC3B⁺ particles representing autolysosomes. Scale bars, 100 or 25 μm. (D) Quantification of LAMP2⁺, LC3B⁺, and LAMP2⁺LC3B⁺ particles in the myofibers (average ± SD, >1045 myofibers analyzed from n = 6). (E) Lysosome and autolysosome ratios were estimated by the quantification of the percentages of LAMP2⁺LC3B⁻ (lysosomes) and LAMP2⁺LC3B⁺ (autolysosomes) particles of total LAMP2⁺ particles per myofiber (average ± SD). (F) Representative confocal images of TA transversal sections immunostained for LC3B and LGALS3. Scale bars, 25 μm. (G) Scatterplots (average ± SD) showing ratios of active lysophagy and non–lysophagy-related autophagy, estimated from the quantification of the percentages of LC3B⁺LGALS3⁺ (lysophagy) and LC3B⁺LGALS3⁻ (non–lysophagy-related autophagy) particles of total LC3B⁺ particles per myofiber (average ± SD). (H) Representative confocal images of TA transversal sections immunostained for LC3B, LGALS3, and LAMP2. LAMP2⁺LC3B⁺LGALS3⁺ puncta represent damaged lysosomes sequestered by autophagosomes. Scale bars, 25 μm. An unpaired two-tailed t test was used for statistical comparisons of WT and Dmd^mdx-4Cv groups. t test: *P < 0.05, **P < 0.01, and ****P < 0.0001.

Fig. 6.. HCD exacerbates lysosomal cholesterol accumulation.

(A) Schematic representation of the study setup. Three-week-old WT and Dmd^mdx-4Cv mice were fed with a standard chow diet or an HCD/high-fat diet for 4 weeks. (B) Cross sections of livers stained with Oil Red O, labeling nonpolar lipids. (C) Sirius red (SR) staining of diaphragm and GA muscles cross sections. Scale bars, 200 or 100 μm (diaphragm) and 500 or 100 μm (GA). (D and E) Quantification of fibrosis (Sirus red–positive area) (average ± SD, n = 6). (F) Free cholesterol quantification in whole GA muscle lysates (average ± SD, n = 4). (G) Representative confocal images of serial transversal sections of GA muscle labeled with filipin (free cholesterol) and Syto13 (nuclei) and immunostained with LAMP2. Scale bars, 25 μm. (H) Colocalization analysis of filipin signal overlapping LAMP2, calculated by Manders’ overlap coefficient (average ± SD, n = 3). An ANOVA test with Tukey’s multiple comparisons test was used for statistical comparisons. *P < 0.05, **P < 0.05, ***P < 0.001, and ****P < 0.0001.

Fig. 7.. HCD increases lysosomal damage in skeletal muscle.

(A) Representative confocal images of transversal sections of GA muscle immunostained for LAMP2 and LGALS3 (upper panel), LC3B and LGALS4 (middle panel), and SQSTM1 (lower panel). Scale bars, 25 μm. (B) Quantification of double-positive puncta (LAMP2⁺LGALS3⁺) in the myofibers of GA muscles (scatterplots with median, >784 myofibers analyzed from n = 6). (C) Relative Lgals3 mRNA expression in the GA normalized to Rplp0 (average ± SD, n = 5 or 6). (D) Western blot analysis of GA muscles for LGALS3 (n = 6). Total protein staining is used as the loading control. (E) Band intensity quantification of LGALS3 relative to total protein staining (average ± SD, n = 6). (F) Quantification of SQSTM1⁺ in the myofibers of GA muscles (n > 119 myofibers). (G) Scatterplots (average ± SD) showing ratios of active lysophagy and non–lysophagy-related autophagy, estimated from the quantification of the percentages of LC3B⁺LGALS3⁺ and LC3B⁺LGALS3⁻ particles of total LC3B⁺ particles per myofiber (average ± SD, >539 myofibers). An ANOVA test with Tukey’s multiple comparisons test was used for statistical comparisons. *P < 0.05, **P < 0.05, ***P < 0.001, and ****P < 0.0001.

Fig. 8.. Evaluation of gene therapy strategies for DMD and LGMDR5 in corresponding mouse models.

(A) Schematic representation of the study setup. Six-week-old Dmd^mdx mice were injected intravenously with a rAAV encapsulating a μDys encoding sequence, under the control of an spc512 promoter, at two doses of 5 × 10¹² vg/kg (low dose) and 1 × 10¹³ vg/kg (high dose). Four-week-old Sgcg^−/− mice were injected intravenously with a rAAV9 encapsulating the human sequence of SGCG, under the control of a desmin promoter, at a dose of 2 × 10¹³ vg/kg. IV, intravenous. (B) Dosage of CK in serum of mice at the end of the study (average ± SD). (C) ELISA quantification of MYOM3 in the serum of the mice before euthanasia. Data are presented as the scatterplot (average ± SD). (D) Global force evaluation by an escape test (force in newton normalized to body weight in grams) before euthanasia of mice (average ± SD). (E) Relative Lgals3 mRNA expression in the GA normalized to Rplp0 (average ± SD). (F) Representative confocal images of transversal sections of GA muscle immunostained for LAMP2 and LGALS3. Scale bars, 25 μm. (G) Quantification of double-positive puncta (LAMP2⁺LGALS3⁺) in GA myofibers (median and range shown). An ANOVA test with Tukey’s multiple comparisons test was used for statistical comparisons. *P < 0.05, **P < 0.05, ***P < 0.001, and ****P < 0.0001. Gray asterisks represent relative comparison to the “WT_Phy Ser” group, and red asterisks represent comparison to the “mdx_Phy Ser” group.

Fig. 9.. TR partially restores dystrophic phenotypes and potentiates μDys gene therapy in Dmd^mdx-4Cv mice.

(A) Study setup. Dmd^mdx-4Cv mice were started on TR treatment (2% dilution in drinking water) at 3 weeks. For groups 4 and 5, mice were injected intravenously with a rAAV9 encapsulating the μDys sequence at a dose of 7 × 10¹² vg/kg. (B) Seric CK analysis before the escape test (average ± SD, n = 5). (C) ELISA quantification of MYOM3 before the escape test (average ± SD, n = 5). (D) Global force evaluation (normalized to mice body weight, N/g) by the escape test (average ± SD, n = 5). (E) Histological characterization of muscles. Top panel: H&E labeling of GA muscle cross sections. Middle panel: Sirius red labeling of diaphragm and GA muscles. Bottom panel: Representative images of GA muscle serial cross sections immunostained for mouse immunoglobulin (IgG), laminin, and CD11b. Scale bars, 200 μm (H&E) and 500 and 100 μm (zoomed-in images). (F) Analysis of myofiber size distribution in GA muscles represented by the variance coefficient calculated as [SD of the muscle fiber size/mean of muscle fiber size] * 1000 (median with min-max values). (G) Centronucleation index in the GA muscle (median with min-max values). (H and I) Fibrosis analysis of GA and diaphragm by the quantification of Sirius red–positive areas on transversal sections (average ± SD). (J) Quantification of IgG uptake by myofibers (positive myofiber number normalized to muscle cross-sectional area). (K) Evaluation of CD11b⁺ cell infiltration in the muscle by the quantification of CD11b area. An ANOVA test with Tukey’s multiple comparisons test was used for statistical comparisons (compared to WT control and mdx control). ANOVA: *P < 0.05, **P < 0.05, ***P < 0.001, and ****P < 0.0001. Black asterisks represent relative comparison to the “WT control” group, and red asterisks represent comparison to the “mdx control” group. An unpaired two-tailed t test was used for statistical comparisons between mdx-μDys and mdx_μDys + TR groups. t test: *P < 0.05 and **P < 0.01.

Fig. 10.. TR alleviates lysosomal damage and enhances transcriptomic correction in μDys-treated Dmd^mdx-4Cv mice.

(A) Representative confocal images of TA muscle sections immunostained for LAMP2 and LGALS3 (upper panel), LC3B and LGALS3 (middle panel), and SQSTM1 (lower panel). Scale bars, 25 μm. (B) Quantification of LAMP2⁺LGALS3⁺ puncta in TA myofibers (average ± SD, >538 myofibers analyzed from n = 5). (C and D) Density plot with a nonlinear fit Gaussian curve, showing the distribution of average spot size (C) and lysosome number per myofiber (D) (>538 myofibers analyzed from n = 5). (E) Quantification of SQSTM1+ puncta in TA myofibers (average ± SD, >552 myofibers analyzed from n = 5). (F) Ratios of active lysophagy and non–lysophagy-related autophagy, estimated from the quantification of LC3B⁺LGALS3⁺ and LC3B⁺LGALS3⁻ percentages (average ± SD, >529 myofibers analyzed from n = 5). (G to I) Comparison of global transcriptomic changes in GA muscle. (G) PCA plot using the expression level of 10000 genes with the highest variance. Gene expression was first normalized by size factors and transformed by variance-stabilizing transformation using the DESeq2 RStudio package. (H) Heatmap presenting log₂FC in comparison to WT muscle for all 8625 DEGs found in mdx muscle. The log₂FC values are illustrated in row z-scores, colored from blue to red, arranged from the lowest to highest. (I) Volcano plots of multiple comparisons illustrating transcriptomic changes before and after different treatments. Down-regulated and up-regulated DEGs are colored in blue and red, respectively, in all volcano plots. The right upper panel shows volcano plots comparing mdx control to treated groups, in which DEGs are correctly restored after treatment. The lower right panel shows treated groups compared to WT control, in which significant DEGs are the genes that are not or incompletely restored after treatment. An ANOVA test with Tukey’s multiple comparisons test was used for statistical comparisons in (B) to (E). ANOVA: ****P < 0.0001. Black asterisks represent relative comparison to the “WT control” group, and red asterisks represent comparison to the “mdx control” group.