Deletion of miR-146a enhances therapeutic protein restoration in model of dystrophin exon skipping

Abstract

Duchenne muscular dystrophy (DMD) is a progressive muscle disease caused by the absence of dystrophin protein. One current DMD therapeutic strategy, exon skipping, produces a truncated dystrophin isoform using phosphorodiamidate morpholino oligomers (PMOs). However, the potential of exon skipping therapeutics has not been fully realized as increases in dystrophin protein have been minimal in clinical trials. Here, we investigate how miR-146a-5p, which is highly elevated in dystrophic muscle, impacts dystrophin protein levels. We find inflammation strongly induces miR-146a in dystrophic, but not wild-type myotubes. Bioinformatics analysis reveals that the dystrophin 3'UTR harbors a miR-146a binding site, and subsequent luciferase assays demonstrate miR-146a binding inhibits dystrophin translation. In dystrophin-null mdx52 mice, co-injection of miR-146a reduces dystrophin restoration by an exon 51 skipping PMO. To directly investigate how miR-146a impacts therapeutic dystrophin rescue, we generated mdx52 with body-wide miR-146a deletion (146aX). Administration of an exon skipping PMO via intramuscular or intravenous injection markedly increases dystrophin protein levels in 146aX versus mdx52 muscles; skipped dystrophin transcript levels are unchanged, suggesting a post-transcriptional mechanism-of-action. Together, these data show that miR-146a expression opposes therapeutic dystrophin restoration, suggesting miR-146a inhibition warrants further research as a potential DMD exon skipping co-therapy.

Keywords: Becker Muscular Dystrophy (BMD); Duchenne Muscular Dystrophy (DMD); RNA therapeutics; antisense oligonucleotides; dystrophin; exon skipping; miR-146a-5p; microRNA.

Figure 1.. miR-146a is significantly elevated in BMD and DMD patients and model mice.

(A) The sequence of miR-146a-5p is conserved across species. Note that in mammals there is 100% conservation. (B) Left; miR-146a levels are elevated in BMD patients with low dystrophin (<20% dystrophin of unaffected individuals). n=5 normal, 6 “BMD high” and 4 “BMD low”; ANOVA, **p<0.001, adapted from . Right; Dystrophin protein and miR-146a levels are inversely correlated in BMD patients. Spearman correlation, r=−0.6365, **p<0.01. Adapted from. (C) miR-146a levels are significantly elevated in both BMD and DMD patients (taken from raw data in Eisenberg et. al) n=9 unaffected (UA), 6 BMD and 8 DMD biopsies; ANOVA, #p<0.1, *p<0.05. (D) miR-146a levels are significantly increased in mdx52 mouse diaphragm (D), tibialis anterior (TA), quadriceps (Q), gastrocnemius (G), and triceps (Tri). n=6/group; Two-way ANOVA, **p<0.01, *p<0.05. (E) miR-146a levels in bmx mouse diaphragm (D), tibialis anterior (TA), quadriceps (Q), gastrocnemius (G), and triceps (Tri). miR-146a is significantly increased in the latter 3 muscles. n=6/group, Two-way ANOVA, *p<0.05. (F) miR-146a is more highly expressed and induced in mdx vs. WT H2K myotubes treated with 10 ng/mL TNF-α (n=4 replicates per group) Student’s t-test, ****p<0.0001. Data are represented as mean ± S.E.M. See also Figure S1.

Figure 2.. miR-146a targets the dystrophin 3′UTR and reduces exon skipping-mediated dystrophin restoration.

(A-B) Predicted miR-146a-5p binding sites in the human DMD 3’UTR. (A) Scan-miR was used to search for putative miR-146a-5p binding sites within the human dystrophin 3′UTR sequence. The y axis denotes the dissociation constant (−log Kd), and the x axis denotes the location on the dystrophin 3′UTR in the 5′ to 3′ direction. (B) List of 17 putative miR-146a binding sites in the DMD 3’ UTR; orange denotes the strongest binding site and the one that was mutated in the luciferase construct in (C). The predicted binding sites include: a 7mer-m8 where there is perfect complementarity of the seed sequence with the UTR and this complementarity extends to nucleotide 8; a 6mer-A1 where there is seed sequence complementarity at nucleotide positions 3–7 and additional complementarity at nucleotide 8; a 6mer-A1, where the seed sequence shows complementarity to the UTR at positions 2–6 and additionally possesses an adenosine (A) at position 1; an wobbled 8mer, where there is an adenosine position 1, seed sequence complementarity at position 2–7 that is interrupted by a single mismatch and additional complementarity at position 8 and several non-canonical binding sites . (C) Schematic showing the binding site of miR-146a on the DMD 3’ UTR (top) and the base pairing of miR-146a with the DMD 3’ UTR when the binding site is mutated (bottom). (D) A luciferase reporter assay was used to examine miR-146a binding to the WT and mutant DMD 3’ UTR in HEK-293T cells. Transfection with miR-146a significantly reduces reporter activity in cells with WT DMD 3’ UTR whereas there is no reduction in reporter activity in cells with the mutated DMD 3’ UTR. n=3–4, ANOVA, *p<0.05. (E) The TAs of mdx52 mice were injected with 2μg of exon skipping PMO and 10μg of either control or miR-146a. TA cross sections were stained for dystrophin (red) and wheat germ agglutinin (white). (E’) Quantification of dystrophin positive fibers; significantly fewer dystrophin-positive myofibers were observed in muscles co-injected with miR-146a compared to control. n=9–10, Student’s t-test, *p<0.05. Data represented as mean ± S.E.M. See also Figure S2.

Figure 3.. Generation of dmd^−/−; 146a^−/− (146aX mice).

(A) Breeding schematic to develop 146aX mice. (B) qRT-PCR showing miR-146a levels in the gastrocnemius muscles of WT, mdx, and 146aX mice. 146aX mice have no detectable levels of miR-146a via qRT-PCR. n=6/group, ANOVA, ****p<0.0001. (C)Left; Representative dystrophin immunofluorescence (red) and laminin staining in wild-type (WT), mdx52 and 146aX quadriceps showing strong dystrophin staining in WT muscle and no staining except for a few revertant fibers in mdx52 and 146aX muscles (white boxes with asterisks). Right; Quantification of revertant fibers in mdx52 and 146aX muscles shows that they represent only a small fraction of all fibers. There is no significant difference in revertant dystrophin-positive fibers between genotypes. n=6–7/group, students t-test, ns p=0.6830.

Figure 4.. Deletion of miR-146a significantly increases dystrophin restoration in mdx52 mice treated with intramuscular PMO.

(A) Schematic of intramuscular PMO injection experimental design. mdx and 146aX tibialis anterior muscles (TAs) were injected with 2μg of an exon 51 skipping PMO. 2 weeks post-injection muscles were harvested for analysis. n=10 muscles/group and 5 mice per group. Wes = Western capillary immunoassay, IF = immunofluorescence. (B) qRT-PCR was performed to quantify % exon skipping in PMO-injected mdx52 and 146aX TAs as previously described. Exon skipping efficiency at the RNA level is not significantly between mdx52 and 146aX TAs. (C) Capillary-based immunoassay (Wes) was used to quantify levels of dystrophin protein. (C’) Dystrophin protein is significantly increased in 146aX mice injected with PMO compared to mdx mice. (D) Immunofluorescence showing dystrophin staining in the TA of mdx and 146aX mice with intramuscular PMO. (i) shows a zoomed in area of high dystrophin rescue and (ii) shows a zoomed in area of low dystrophin rescue. Bars = 500μm and 50 μm respectively. (D’) Quantification of dystrophin-positive myofibers. The percentage of dystrophin-positive myofibers following PMO injection is significantly increased in 146aX mice compared to mdx mice. Images were blinded during quantification. Data represented as mean ± S.E.M. ns p>0.05, *p≤0.05, **p≤0.01, ****p≤0.0001. See also, Figure S3.

Figure 5.. Deletion of miR-146a significantly increases dystrophin protein in mdx52 mice treated with systemic PMO.

(A) Schematic of experimental design for systemic PMO injections in mdx and 146aX mice. mdx52 and 146aX mice were administered systemic PMO via the retroorbital sinus (400 mg/kg, exon 51 skipping PMO) at 15 weeks of age. One week later a second 400mg/kg PMO injection was performed. Muscle function tests were performed one week after the second injection and mice were sacrificed 2 weeks after the second injection. Muscles were analyzed by qRT-PCR to determine extent of skipped Dmd mRNA and by Western capillary immunoassay (Wes) and immunofluorescence (IF) to determine dystrophin protein levels (IF). n=4 mice and 5 muscles per group. (B) Exon skipping efficiency at the RNA level via systemic injection of PMO in mdx and 146aX mice is not significantly different. Student’s t-test. (C) Capillary-based immunoassay (Wes) was used to quantify dystrophin protein levels in skeletal muscle of mdx and 146aX mice systemically treated with exon-skipping PMO via retroorbital injection. Depicted is a virtual Wes blot . (C’) Wes quantification of dystrophin protein levels in skeletal muscle. Dystrophin protein is significantly increased in 146aX mice versus mdx52 mice. Student’s t-test, *p<0.05. Note: % Dystrophin was calculated by normalizing to the average mdx intensity for each muscle and setting it to 100%. (D) Dystrophin protein levels from Wes normalized to % Dmd exon skipped transcripts. There is a higher ratio of dystrophin restoration in 146aX mice compared to mdx mice. Data represented as mean ± S.E.M. Student’s t-test, *p<0.05.

Figure 6.. Deletion of miR-146a significantly increases dystrophin positive fibers in mdx52 mice.

(A) Dystrophin immunofluorescence (red) in PMO-injected mdx and 146aX diaphragm and triceps muscles. Bar= 200mM and 100 mM, respectively. Muscles were counterstained with DAPI (nuclei) and laminin (green, not shown) to count total muscle fibers). (A’) Quantification of dystrophin-positive myofibers, (n=4 mice and 5 muscles per group). The percentage of dystrophin-positive myofibers following PMO injection is significantly increased in 146aX mice compared to mdx mice. (B)The percentage of dystrophin positive fibers normalized to the percentage of Dmd exon skipped transcripts. There is a higher ratio of dystrophin restoration in 146aX mice compared to mdx mice. Data represented as mean ± S.E.M Student’s t-test, p<0.05. (C) Visual standard for classification of dystrophin high, low, and negative fibers. (D) Left; Quantification of “high” intensity dystrophin-positive fibers. Right; quantification of “low” intensity dystrophin positive fibers. Images were blinded during quantification. Student’s t-test, #p<0.1, *p<0.05. All data represented as mean ± S.E.M. Dia, diaphragm; TA, tibialis anterior; Quad, quadriceps; Gastroc, gastrocnemius; Tri, triceps. See also Figure S4.

Figure 7.. Functional and histological analysis of PMO-treated mdx52 and 146aX mice.

(A) Muscle function tests in mdx52 and 146aX mice. Left; Hindlimb grip strength was performed and normalized to body weight. Right; Wire hang-time assay. n=4 mice per group; Student’s t-test. (B) Terminal tissue masses of triceps (n=8) and diaphragms (n=4). Student’s t-test, #p<0.1, *p<0.05. All data represented as mean ± S.E.M.(C) Immunofluorescence of muscle sections using laminin (green), dystrophin (red) and DAPI (blue) staining of treated mdx52 versus 146aX muscles to visualize both dystrophin restoration and myofiber size. Bar=100μm. (C’) Quantification of myofiber size variability via the variance coefficient of minimal Feret’s diameter. (D) To visualize how dystrophin restoration affects localization of a dystrophin-associated protein (α-sarcoglycan), laminin (red), a-sarcoglycan (green), and dystrophin (blue) were visualized via immunofluorescence in mdx52 and 146aX diaphragms. Note that α-sarcoglycan is brighter in treated 146aX diaphragms which is apparent by the yellow laminin/a-sarcoglycan co-localization (merge) image versus the more orange hue in the laminin/α-sarcoglycan image in treated mdx52 diaphragms. Bar = 50μm. See also Figure S5.