Muscle miRNAome shows suppression of chronic inflammatory miRNAs with both prednisone and vamorolone

Abstract

Corticosteroids are highly prescribed and effective anti-inflammatory drugs but the burden of side effects with chronic use significantly detracts from patient quality of life, particularly in children. Developing safer steroids amenable to long-term use is an important goal for treatment of chronic inflammatory diseases such as Duchenne muscular dystrophy (DMD). We have developed vamorolone (VBP15), a first-in-class dissociative glucocorticoid receptor (GR) ligand that shows the anti-inflammatory efficacy of corticosteroids without key steroid side effects in animal models. miRNAs are increasingly recognized as key regulators of inflammatory responses. To define effects of prednisolone and vamorolone on the muscle miRNAome, we performed a preclinical discovery study in the mdx mouse model of DMD. miRNAs associated with inflammation were highly elevated in mdx muscle. Both vamorolone and prednisolone returned these toward wild-type levels (miR-142-5p, miR-142-3p, miR-146a, miR-301a, miR-324-3p, miR-455-5p, miR-455-3p, miR-497, miR-652). Effects of vamorolone were largely limited to reduction of proinflammatory miRNAs. In contrast, prednisolone activated a separate group of miRNAs associated with steroid side effects and a noncoding RNA cluster homologous to human chromosome 14q32. Effects were validated for inflammatory miRNAs in a second, independent preclinical study. For the anti-inflammatory miRNA signature, bioinformatic analyses showed all of these miRNAs are directly regulated by, or in turn activate, the inflammatory transcription factor NF-κB. Moving forward miR-146a and miR-142 are of particular interest as biomarkers or novel drug targets. These data validate NF-κB signaling as a target of dissociative GR-ligand efficacy in vivo and provide new insight into miRNA signaling in chronic inflammation.

Keywords: Duchenne muscular dystrophy; inflammation; miRNA; muscle; steroids.

Figures

Fig. 1.

Summary of muscle miRNA changes discovered in response to dystrophy and its treatment. Expression of the miRNAome was quantified in diaphragm muscle of mice from a discovery set of mice (n = 5 mice per group). Groups included WT (vehicle), mdx (vehicle), mdx treated with prednisolone (5 mg/kg), and mdx treated with vamorolone (15 mg/kg), with mice treated from 2 to 8 wk of age in a prophylactic trial design. A: a Venn diagram illustrates the proportion of miRNAs that are significantly different than untreated mdx muscle in the WT, prednisolone, and vamorolone groups. The nine miRNAs that were significantly different in all three groups vs. untreated mdx, highlighted here, were chosen as a focus set of efficacy-associated miRNAs. B: heat map visualization of the expression of the nine efficacy associated miRNA markers within each individual mouse. C: bar graph showing the number of miRNAs that significantly increased or decreased in response to either of the drug treatments. D: heat map of the eight unique miRNAs that were increased by prednisolone. Heat map: red, increased; green, decreased. Pred, prednisolone; Vam, vamorolone; WT, wild type.

Fig. 2.

Behavior of efficacy miRNA signature is maintained in an independent validation trial. A validation set of samples was obtained from a second, independent mdx trial performed at a different stage of the mdx disease. Mice received daily oral vehicle, prednisolone (5 mg/kg), or vamorolone (45 mg/kg) for 4 mo, with treadmill running to unmask mdx phenotypes and muscle harvested at 6 mo of age. The nine miRNAs identified as associated with efficacy in the TLDA arrays were quantified in diaphragm muscle using quantitative RT-PCR in this second set of mice. (Values are graphed as % of untreated mdx expression levels; 1 outlier removed from miR-455-5p after significant Grubb’s test; n = 5 per group; ANOVA with post hoc comparison to mdx vehicle; *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.0005). TLDA, Taqman low-density array.

Fig. 3.

Promoter analysis of miRNAs indicates NF-κB signaling is a shared target of effective drugs. Transcription factor (NF-κB, GR) binding sites and histone (H3) modifications that mark regulatory regions were examined using ChIP-seq data from ENCODE. DNA-binding motifs for each transcription factor were identified through the Factorbook repository. A: schematic of the gene locus for miR-142, illustrating the binding site of 13 neighboring DNA loci that are bound directly by NF-κB. Corresponding epigenetic modification maps are provided showing the location of histone modifications associated with active promoters (H3K4me3) and poised/active enhancers (H3K4me1 and H3K27Ac) in the immediate vicinity of miR-142. B: sequence logo pictogram of base frequency at NF-κB binding sites, with the consensus NF-κB motif provided immediately below. Also provided are five representative NF-κB binding site sequences near miR-142, listed in order from the 5′ to 3′ direction. C: summary of promoter analysis and literature data indicating each miRNA and known factors or conditions associated with its transcriptional regulation. ChIP-seq, chromatin immunoprecipitation sequencing; COL27A1, collagen type 27 alpha 1 chain; ENCODE, Encyclopedia of DNA Elements; GR, glucocorticoid receptor; IL4, interleukin 4; MoDCs, monocyte-induced dendritic cells; φ, macrophage; NKRF, NF-κB-repressing factor; STAT6, signal transducer and activator of transcription 6; TWEAK, TNF-like weak inducer of apoptosis.

Fig. 4.

Prednisolone increases miRNAs associated with GR regulation and the 14q32 mega cluster. For the conserved miRNAs that were specifically elevated in prednisolone-treated mouse muscle, we analyzed the genomic loci, transcription factor binding sites (GR and NF-κB), and histone (H3) modifications using ChIP-seq data from ENCODE. DNA motifs bound by the GR were identified through Factorbook. A: all six of the miRNAs are transcribed from a well-conserved noncoding RNA cluster (mouse 12F1) which is extensively characterized in humans at the 14q32 locus; that human cluster is depicted here along with corresponding GR-binding sites, as well as histone modifications that correspond to active gene promoters (H3K4me3) and poised or active gene enhancer elements (H3K4me1, H3K27Ac). No NF-κB-binding sites are found at this locus. B: sequence logo pictogram of base frequency at GR binding sites, with the consensus GR motif sequence provided immediately below. Also provided are three representative GR-binding site sequences from this locus numbered from the 5′ to 3′ direction. C: summary of bioinformatic analyses for each miRNA, with a list of conditions that are associated with increased levels of each miRNA. †encoded by the 14q32 cluster of miRNAs with GR-bound enhancers. ChIP-seq, chromatin immunoprecipitation sequencing; ENCODE, Encyclopedia of DNA Elements; GR, glucocorticoid receptor; lncRNA, long noncoding RNA; MEG, maternally expressed gene; snoRNA, small nucleolar RNA.

Fig. 5.

Proposed model of NF-κB and GR-regulated miRNAs in the treatment of muscular dystrophy. In DMD, inflammatory signaling promotes the chronic activation of NF-κB. This, in turn, activates NF-κB gene targets, including miRNAs that regulate the expression of proteins in the NF-κB signaling pathway, creating a chronic inflammatory feedback loop. Here we show that the NF-κB-regulated miRNAs miR-142-3p, miR-142-5p, miR-146a, miR-301a, miR-455-3p, miR-455-5p, miR-497, and miR-652 are all elevated in dystrophic muscle. These NF-κB-regulated miRNAs are all effectively decreased by both vamorolone (Vam) and prednisone (Pred) treatment, via the GR. miR-324-3p, a miRNA that activates NF-κB in a positive feedback loop, is also decreased by both drugs. Acting through a separate pathway which can be selectively avoided by dissociative steroid chemistries, prednisone also directly causes GR-mediated transactivation of gene transcription. This results in elevated levels of a miRNA cluster located on chromosome 14q32. These microRNAs are associated with steroid side effects such as insulin resistance, hypertension, stress, and mood disturbances. Blue lines, pathways affected by vamorolone; red lines, pathways affected by prednisolone. DMD, Duchenne muscular dystrophy; GR, glucocorticoid receptor.