Immunoproteasome Inhibition Positively Impacts the Gut-Muscle Axis in Duchenne Muscular Dystrophy

Abstract

Background: Duchenne Muscular Dystrophy (DMD) features immune-muscle crosstalk, where muscle fibre degeneration enhances pro-inflammatory macrophage infiltration, worsening inflammation and impairing regeneration.

Methods: We investigated the impact of immunoproteasome (IP) inhibition on the gut-muscle axis in mdx mice, a well-established model of DMD. We employed microbiota perturbation models, including broad-spectrum antibiotic treatment (ABX) and faecal microbiota transplantation (FMT) from IP-inhibited mdx mice. IP inhibition effects were assessed by analysing gut microbiota composition, intestinal inflammation, muscle integrity and associated metabolic and inflammatory pathways.

Results: IP inhibitor ONX-0914 significantly impacted the intestinal inflammatory microenvironment and gut microbiota of mdx mice. ONX-0914 treatment increased gastrointestinal transit (increased wet/dry faecal weights, p = 0.0486 and p = 0.0112, respectively) and partially restored intestinal barrier integrity (reduced FITC-dextran leakage, p = 0.0449). JAM-A was significantly upregulated (p < 0.0001). Colonic CD206+ M2 macrophages increased, while CD68 + M1 cells partially decreased. ONX-0914 downregulated IP isoforms in macrophages (PSMB8: p = 0.0022; PSMB9: p = 0.0186) as well as FOXO-1 (p = 0.0380) and TNF-α (p = 0.0487). Antibiotic-induced microbiota depletion abrogated these effects. Metagenomic analysis revealed significant differences in microbiota composition between C57Bl controls and mdx mice (PERMANOVA p < 0.001), with ONX-0914 inducing enrichment of stachyose degradation pathways. Metabolomic analysis showed enrichment of bacterial metabolites, fatty acid and sugar metabolism pathways, with increased glutathione, galactose, glycerol, glyceraldehyde and TCA cycle intermediates. ONX-0914 improved mitochondrial activity in skeletal muscle, as increased expression of ETC complexes (mdx vs. mdx+ONX: Complex II, p = 0.0338; Complex IV, p = 0.0023) and TCA enzymes (mdx vs. FTMmdx+ONX: IDH p = 0.0258; FH p = 0.0366). This led to a shift towards oxidative muscle fibres and improved muscle morphology (increased fibre size, p < 0.0001 mdx vs. mdx+ONX and mdx vs. FTMmdx+ONX). Muscle performance was enhanced with reduced CPK levels (p = 0.0015 mdx vs. mdx+ONX) and fibrosis (decreased TGFβ: mdx vs. mdx+ONX, p = 0.0248; mdx vs. FTMmdx+ONX, p = 0.0279). ONX-0914 reduced CD68+ (mdx vs. mdx+ONX, p = 0.0024; mdx vs. FTMmdx+ONX, p < 0.0001) and increased CD206+ (mdx vs. FTMmdx+ONX: p = 0.0083) macrophages in muscle, downregulated inflammatory genes (mdx vs. mdx+ONX: ccl2 p = 0.0327, vcam-1p = 0.0378) and reduced pro-inflammatory proteins (MCP1, mdx vs. mdx+ONX, p = 0.0442). Inflammatory cytokines and endothelial vessel density in ONX-0914 treated mdx were restored to wild type mice. These data demonstrate that ONX-0914 enhances muscle function through microbiota-dependent mechanisms.

Conclusions: Our study advances the understanding of the role of dysbiosis in DMD disease and identifies IP inhibition as a potential therapeutic strategy to modulate the dystrophic gut-muscle axis, offering new perspectives for microbiota-targeted therapies.

Keywords: Duchenne muscular dystrophy; immunoproteasome; macrophages; muscle metabolism.

FIGURE 1

Colonic features of 3 m mdx+ONX‐0914 mice. (A) Mucus layer thickness and crypt length were quantified for n = 5 mice per group (with pooled samples of n = 10 each). (B) The number of pellets and the weight of faeces from 3 m C57Bl, mdx and mdx+ONX mice (n = 5 each). (C) Evaluation of FITC‐dextran concentration in the faecal stools of 3 m C57Bl, mdx and mdx+ONX mice (n = 6 each) (two independent experiments). (D) Evaluation of Lipocalin‐2 concentration in the stool of 3 m C57Bl (n = 5), mdx and mdx+ONX mice (n = 6 each) (two independent experiments). (E) The data are showed in graph as the percentage of JAM‐A+ pixel intensity per mm². (F) The data are showed in graphs as the number of CD68+ and CD206 + macrophages by the area of colonic sections. Data information: data are presented as mean ± SD (*p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001; One‐Way ANOVA Kruskal–Wallis test for evaluation of images and One‐Way ANOVA with Tukey's multiple comparisons test for WB experiments).

FIGURE 2

Dysbiotic microbiota of 3 m mdx mice is modulated by ONX‐0914. (A) Species‐level genome bin (SGB) richness in 3 m C57Bl (n = 8), mdx (n = 8) and mdx + ONX (n = 8). Mann–Whithey tests were performed (** = p < 0.01) (B) Heatmap of the top 50 SGBs based on their prevalence in 3 m C57Bl (n = 8), mdx (n = 8) and mdx + ONX (n = 8). Samples were clustered based on Bray–Curtis dissimilarity on the SGB relative abundans. (C) MDS of beta‐diversity of 3 m C57Bl (n = 8), mdx (n = 8) and mdx+ONX (n = 8) as measured by Bray–Curtis dissimilarity on the SGB‐level relative abundances. PERMANOVA tests (1000 iterations) were performed. (D) MDS of beta‐diversity of 3 m C57Bl (n = 8), mdx (n = 8) and mdx+ONX (n = 8) as measured by Bray–Curtis dissimilarity on the microbial pathway relative abundances. (E) Relative abundance of the SGBs significantly associated with 3 m C57Bl (n = 8)—upper panel –; mdx (n = 8)—middle panel –; mdx+ONX (n = 8)—lower panel—in the LefSe analysis (FDR < 0.05). Linear Discriminant Analysis (LDA) score and FRD (q = Benjamini/Hochberg correction) are reported. (F) Relative abundance of the microbial pathways significantly associated with mdx (n = 8) and mdx + ONX (n = 8)—stachyose degradation—in the LefSe analysis (FDR < 0.05). LDA score and FRD (q = Benjamini/Hochberg correction) are reported.

FIGURE 3

Microbial community analysis in 3‐month‐old C57Bl, mdx and mdx + ONX mice. (A) Alpha diversity differences between groups. SGB richness, Shannon and Simpson diversity were computed. Significance was assessed using Mann–Whitney tests. Only significant comparisons are shown. (*: 0.01 < p < = 0.05; **: 0.001 < p < = 0.01). (B) MDS based on beta diversity for taxonomic composition. Bray–Curtis and Jaccard distances were computed on the arcsine square root‐transformed relative abundances. Significant differences were computed using PERMANOVA tests (1000 iterations). (C) MDS based on beta diversity for microbial pathways. Bray–Curtis and Jaccard distances were computed on the arcsine square root‐transformed relative abundances. Significant differences were computed using PERMANOVA tests (1000 iterations). (D) Linear Discriminant Analysis (LDA) score of the top 15 SGBs significantly associated with 3 m C57Bl (n = 8), mdx (n = 8) and mdx+ONX (n = 8) in the LefSe analysis (FDR < 0.05).

FIGURE 4

Evaluation of metabolome of 3 m mdx+ONX‐0914 serum. (A) SCFA faecal quantification of 3 m C57Bl (n = 3), mdx (n = 4) and mdx + ONX (n = 4) mice. (B) Partial Least Square Discriminant Analysis score plot of CTRL (red, n = 4), mdx (green, n = 4) and mdx + ONX (blue, n = 3) serum samples. In brackets are reported the amount of explained variance for each latent component. (C) Concentration of the metabolites with a variable importance in projection (VIP) score higher than 1.5 isolated from serum samples. Heatmap of the concentration of the top 25 metabolites according to ANOVA. (D) Metabolic pathway showing the interplay of the VIP selected metabolites.

FIGURE 5

The muscle response to ONX‐0914 is dependent on the gut microbiota. (A) Graph portrays the percentage of myofibers expressing different MyHC isoforms in TAs of C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice and mdx^FTMmdx and mdx^FTMmdx+ONX mice (n = 3 each and n = 10 images per animal) (two independent experiments). (B) Quantification of the percentage of SDH+ myofibers of TAs from C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 5 each and n = 8 images per animal) and mdx^Fmdx and mdx^Fmdx+ONX mice (n = 4 each and n = 10 images per animal) (two independent experiments). (C) Quantification of myofiber area and relative frequency of the myofiber CSA expressed as the frequency distribution of the TA muscles of C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice and mdx^FTMmdx and mdx^FTMmdx+ONX mice (n = 4 each; pooled samples per group: n = 5971 for C57Bl, n = 6301 for mdx, n = 7386 for mdx+ONX, n = 2535 for mdx+ABX; n = 9737 for mdx+ABX+ONX; n = 11 575 for mdx^Fmdx; n = 10 687 for mdx^Fmdx+ONX). (D) Evaluation of tetanic force in TA of C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 6 each) and mdx^Fmdx and mdx^Fmdx+ONX mice (n = 4 each). (E) Sera CPK analysis of C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 6 each) and mdx^Fmdx and mdx^Fmdx+ONX mice (n = 4 each). (F) Quantification of the percentage of fibrosis of TAs from C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 3 each and n = 10 images per animal) and mdx^Fmdx and mdx^{Fmdx + ONX} mice (n = 3 each and n = 10 images per animal) (two independent experiments). Data information: data are presented as mean ± SD (*p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001; One‐Way ANOVA Kruskal–Wallis test).

FIGURE 6

Muscle‐resident pro‐inflammatory macrophages are diminished in ONX‐0914 mdx treated mice. (A) CD68+ M1 and CD206+ M2 macrophages were quantified in TA of 3 m C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 6 each and n = 8 images per animal) and in mdx^FTMmdx and mdx^FTMmdx+ONX mice (n = 4 each and n = 8 images per animal). CD206 staining is shown in green, CD68 in red and DAPI in blue. CD68+ cells are indicated by broken arrow in left panel, while cells co‐expressing CD68 and CD206 are indicated in the right panel. Scale bars: 50 μm. CD31+, αSMA+ and isolectin+ vessels were quantified in TA of 3 m C57Bl, mdx, mdx+ABX, mdx+ONX and mdx+ABX+ONX mice (n = 3 each and n = 5 images per animal) and in mdx^FTMmdx and mdx^FTMmdx+ONX mice (n = 4 each and n = 4 images per animal). CD31 staining is shown in purple, αSMA in green, isolectin in red and DAPI in blue. αSMA+ vessels are indicated by broken arrow. Scale bars: 100 μm. (B) The data are showed in graphs as the number of CD68+ and CD206+ macrophages by the area of muscular sections. (C) RT‐PCR analysis of two independent experiments showed the modulation of genes normally expressed in M1‐ (ccl2, icam‐1, vcam‐1) and M2‐ (IL‐10) macrophages in muscles of 3 m C57Bl, mdx and mdx+ONX (n = 4 each) mice. (D) The data are showed in graphs as the number of CD31+, αSMA+ and isolectin+ vessels by mm² of muscular sections. Data information: data are presented as mean ± SD (*p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001; One‐Way ANOVA Kruskal–Wallis test).