Repurposed Nrf2 activator dimethyl fumarate rescues muscle inflammation and fibrosis in an aggravated mdx mouse model of Duchenne muscular dystrophy

Abstract

In inherited neuromuscular disease, Duchenne muscular dystrophy (DMD), glucocorticoids significantly slow disease progression yet impart side effects severe enough to preclude use in a significant proportion of patients. Extending our findings that acute treatment with FDA approved multiple sclerosis drug, dimethyl fumarate (DMF), rescues muscle pathology in juvenile mdx mice, we aimed to conduct tiered pre-clinical testing toward translation. To aggravate disease phenotype in adult mdx muscles that usually lack human equivalent muscle pathology, we used bi-weekly treadmill running for 4 weeks which increased plasma DMD biomarker, creatine kinase, by 2-fold and quadriceps fibrosis by ∼30 %. Using this model, we screened DMF for 5 weeks in a head-to-head comparison, and in combination, with standard-of-care prednisone (PRED), to model the most likely clinical trial scenario. We show comparable efficacy between DMF and PRED at reducing inflammation via NF-κB suppression and CD68⁺ macrophage infiltration. Moderate term DMF monotherapy had additional anti-fibrotic and anti-lipogenic effects on skeletal and cardiac muscle beyond those seen with PRED treatment, although combination therapy exacerbated fibrosis in quadriceps. Our study supports DMF as a repurposing candidate for DMD, especially for patients who cannot tolerate chronic glucocorticoid treatment. We also highlight the importance of evaluating combination therapy to identify potential off-target effects between emerging therapeutics and glucocorticoids towards better designed clinical trials.

Keywords: Dimethyl fumarate; Duchenne muscular dystrophy; Muscle pathology; Therapeutics.

Graphical abstract

Fig. 1

Twice-weekly forced treadmill running aggravates the mdx phenotype. (A) Schematic of treatment and treadmill protocols. Mdx aggravation was proved using clinically compatible parameters of (B) plasma creatine kinase and (C) forelimb grip strength. H&E staining assessed the (D) unhealthy/healthy tissue ratio (black arrows) and (E) centronucleated fibre proportion (orange arrows) of the quadriceps (F–H). Pan macrophage marker, CD68, and Masson's trichrome staining assessed (I–L) immune infiltration and (M–P) fibrosis (orange arrows) of quadriceps. Glycoprotein and marker of ECM remodelling, fibronectin-1, was assessed in the quadriceps (black arrows) (Q–T). Extent of DMD phenotype was calculated (mdx AGG/mdx SED∗100) to indicate extent of disease aggravation [13]. Data in B-E, I, M and Q are presented as mean ± SEM and n are indicated by individual data points. Statistical analysis used one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. F–H, J-L, N–P, and R-T scale bar = 50 mm.

Fig. 2

DMF treatment improves grip strength and modulates transcriptional control of myofibril assembly. Functional parameters including (A) maximal forelimb grip strength and (B) minimal holding impulse (whole body grip strength). Anthropometric measures (C) fat mass index and associated transcriptional pathways (D) the fatty acid metabolism pathway, as well as (E) lean mass index was assessed. Absolute force was measured ex vivo in (F) EDL and (G) soleus and specific force was calculated (H–I). The transcriptomics dataset was probed for (J) muscle contractile associated genes. Data in D and J are based on log2 fold change from WT for mdx VEH and mdx VEH for treatment groups (DMF, PRED and DMF + PRED). Data in A-C, and E-I are presented as mean ± SEM and n are indicated by individual data points. Statistical significance was tested via one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

Fig. 3

DMF reduces inflammation and macrophage infiltration of quadriceps. The (A) unhealthy/healthy tissue ratio and (B) percentage of centronucleated fibres are shown in the quadriceps (C–F; yellow arrows indicate centronucleated fibres; black arrows indicate infiltrate). (G) Transcriptomic signature of myogenesis. (H) Pan macrophage marker CD68 was immunohistochemically assessed in the quadriceps and (I–L) representative images show CD68⁺ macrophages (indicated by black arrows). Protein expression of (M) phosphorylated and (N) total NF-κB in the gastrocnemius was quantified via Western blot. (O) The quadriceps transcriptomic dataset was probed for inflammatory markers commonly upregulated in DMD. (P) Degenerating area of the heart was also assessed via H&E and (Q–T) representative images show active cardiomyocyte degeneration (indicated by white arrows). (U) Clinically compatible biomarker creatine kinase was assessed at the experimental endpoint. Data in heatmaps (G and O) are based on log2 fold change from WT for mdx VEH and mdx VEH for treatment groups (DMF, PRED and DMF + PRED). Data in A-B, H, M-N, P and U are presented as mean ± SEM and n are indicated by individual data points. Statistical significance was tested via one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. C–F, I-L and Q-T scale bar = 50 mm.

Fig. 4

DMF reduces fibrosis of quadriceps and heart. Expression of the proposed (A) fibrosis-associated DMD seed genes [24] and (B) DMD molecular signature [25] was probed using our transcriptomic dataset. (C) Quadriceps fibrosis was quantified using Masson's trichrome staining and (D–G) representative images are shown. (H) Extracellular matrix remodelling was assessed via staining for the glycoprotein fibronectin 1 and (I–L) representative images are shown. Fibrotic area of the heart was also assessed and (M–Q) representative images are shown. Data in heatmaps (A–B) are based on log2 fold change from WT for mdx VEH and mdx VEH for treatment groups (DMF, PRED and DMF + PRED). Data in C, H and M are presented as mean ± SEM and n are indicated by individual data points. Statistical significance was tested via one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. D-G, I-L and N-Q scale bar = 50 mm.

Fig. 5

DMF decreases muscle lipid content through modulation of adipogenesis and sphingolipid metabolism pathways in quadriceps. (A–E) Percentage of perilipin-1+ lipid droplets in quadriceps. Lipid associated transcriptional pathways were probed including (F) FAPs, (G) TNFR mediated ceramide production, which was a significantly altered Reactome pathway in mdx aggravated (AGG) relative to WT quadriceps, (H) sphingolipid metabolism and (I) adipogenesis. Data in all heatmaps (F–I) are based on log2 fold change from WT for mdx VEH and mdx VEH for treatment groups (DMF, PRED and DMF + PRED). Data in (A) is presented as mean ± SEM and n are indicated by individual data points. Statistical significance was tested via one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. B-E Scale bar = 50 mm.

Fig. 6

The impact of the treatments on the mdx mouse. The recovery score is used to express the effect of a treatment, not just by the difference between treated and untreated mdx mice, but relative to the extent of the deficiency between WT and mdx mice. A recovery score of 100 % indicates the parameter is equal to that of the WT while a score of 0 % indicates no improvement has been made. This was calculated as (mdx treated)-(mdx untreated)/(WT)-(mdx untreated) x 100 [13].

Fig. 7

DMF treatment tempers expression of detoxification enzymes in muscle. (A) DMFs MOA involves dual anti-inflammatory and -oxidative function through transcription of Nrf2. The muscle transcriptome was probed for molecular pathways involved in DMF's MOA including the (B) Keap1-Nrf2 pathway and (C) Nrf2 regulation of antioxidant/detoxification enzymes pathway. Protein expression of (D) Nrf2, (E) Keap1, (F) NQO1 and (G) SOD1 was quantified via western blot. The muscle transcriptome was also probed for (H) non-receptor antioxidant response element (ARE) repressors and ARE repressors via receptors. A diagrammatic representation of Nrf2-ARE inhibition pathway demonstrates [1] GSK-3b mediated Nrf2 phosphorylation and degradation in the cytoplasm and nucleus [2]; competitive inhibition of Nrf2 binding with small Maf proteins (sMAF) and [3] competitive inhibition of Nrf2-sMAF with the Bach1/2-sMAF heterodimer which prevents Nrf2 binding to the ARE [4]; interaction of RARα with Nrf2 inhibiting binding to the ARE. Heatmap data (B–C and H) are based on log2 fold change from WT for mdx VEH and mdx VEH for treatment groups (DMF, PRED and DMF + PRED). Data in D-G are presented as mean ± SEM and n are indicated by individual data points. Statistical significance was tested via one-way ANOVA: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.