Therapeutic potential of omaveloxolone in counteracting muscle atrophy post-denervation: a multi-omics approach

Abstract

Background

Muscle atrophy caused by denervation is common in neuromuscular diseases, leading to loss of muscle mass and function. However, a comprehensive understanding of the overall molecular network changes during muscle denervation atrophy is still deficient, hindering the development of effective treatments.

Method

In this study, a sciatic nerve transection model was employed in male C57BL/6 J mice to induce muscle denervation atrophy. Gastrocnemius muscles were harvested at 3 days, 2 weeks, and 4 weeks post-denervation for transcriptomic and proteomic analysis. An integrative multi-omics approach was utilized to identify key genes essential for disease progression. Targeted proteomics using PRM was then employed to validate the differential expression of central genes. Combine single-nucleus sequencing results to observe the expression levels of PRM-validated genes in different cell types within muscle tissue.Through upstream regulatory analysis, NRF2 was identified as a potential therapeutic target. The therapeutic potential of the NRF2-targeting drug Omaveloxolone was evaluated in the mouse model.

Result

This research examined the temporal alterations in transcripts and proteins during muscle atrophy subsequent to denervation. A comprehensive analysis identified 54,534 transcripts and 3,218 proteins, of which 23,282 transcripts and 1,852 proteins exhibited statistically significant changes at 3 days, 2 weeks, and 4 weeks post-denervation. Utilizing multi-omics approaches, 30 hubgenes were selected, and PRM validation confirmed significant expression variances in 23 genes. The findings highlighted the involvement of mitochondrial dysfunction, oxidative stress, and metabolic disturbances in the pathogenesis of muscle atrophy, with a pronounced impact on type II muscle fibers, particularly type IIb fibers. The potential therapeutic benefits of Omaveloxolone in mitigating oxidative stress and preserving mitochondrial morphology were confirmed, thereby presenting novel strategies for addressing muscle atrophy induced by denervation. GSEA analysis results show that Autophagy, glutathione metabolism, and PPAR signaling pathways are significantly upregulated, while inflammation-related and neurodegenerative disease-related pathways are significantly inhibited in the Omaveloxolone group.GSR expression and the GSH/GSSG ratio were significantly higher in the Omaveloxolone group compared to the control group, while MuSK expression was significantly lower than in the control group.

Conclusion

In our study, we revealed the crucial role of oxidative stress, glucose metabolism, and mitochondrial dysfunction in denervation-induced muscle atrophy, identifying NRF2 as a potential therapeutic target. Omaveloxolone was shown to stabilize mitochondrial function, enhance antioxidant capacity, and protect neuromuscular junctions, thereby offering promising therapeutic potential for treating denervation-induced muscle atrophy.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-024-05810-7.

Introduction

Skeletal muscle, comprising 40% of body weight, is integral to the maintenance of posture and movement, as well as serving as a reservoir for structural proteins (e.g. actin and myosin), enzymes, and metabolism-related proteins. Skeletal muscle system plays a crucial role in regulating energy metabolism and various metabolic activities involving carbohydrates, lipids, and proteins within the body [1, 2]. Muscle mass is influenced by the number of fibers and the cross-sectional area of each fiber, with fiber area being determined by the number of myofibrils composed of myosin and actin filaments arranged into sarcomeres. The maintenance of skeletal muscle mass necessitates motor nerve innervation, a vascular network, energy production from anaerobic and aerobic metabolism, and the regulation of protein synthesis and degradation processes [3].

Motor nerves establish connections with target muscle fibers through neuromuscular junctions [4]. Following the loss of nerve innervation, there is a decrease in the synthesis rate of muscle protein, leading to the activation of protein degradation pathways such as the ubiquitin–proteasome system and autophagy-lysosome pathway, ultimately resulting in a reduction in muscle fiber volume [5]. Additionally, the transition of muscle fiber type from fast-twitch to slow-twitch impacts muscle contractile characteristics and endurance, accompanied by metabolic changes, decline in mitochondrial function, and reduced energy metabolism efficiency, further compromising muscle strength and endurance.

Current studies have identified elevated oxidative stress, inflammation and decreased mitochondrial function as significant upstream factors contributing to skeletal muscle atrophy across diverse etiologies [6]. More than 10% of genes associated with muscle atrophy are directly linked to energy metabolism [1]. In the pathological progression of muscle atrophy, the activation of the miR-142a-5p/MFN1 axis induces mitochondrial network disruption, dysfunction, autophagy, and cell apoptosis, thereby exacerbating muscle wasting [7]. The inflammatory response is also a critical component in the development of muscle atrophy. The accumulation of fibro-adipogenic progenitors (FAPs) was conformed contribute to muscle fiber atrophy and fibrosis via the STAT3-IL-6 signaling pathway [8]. Furthermore, mTORC1 activity is heightened in denervated muscles, leading to muscle atrophy through the activation of FoxO and E3 ubiquitin ligases, underscoring the intricate involvement of mTORC1 in neurogenic muscle atrophy [9]. The activation of the IL-6/JAK/STAT3 pathway is also significant in facilitating the denervation of skeletal muscle atrophy. [10] However, current research primarily examines individual signaling pathways and molecular functions, lacking a comprehensive understanding of the overall molecular network changes during muscle denervation atrophy. The process of muscle atrophy induced by denervation is characterized by dynamic changes, necessitating a thorough investigation of the molecular mechanisms and pathological alterations occurring at various stages. The utilization of omics technologies, which are fundamental to systems medicine, and the integrated use of multi-omics approaches are essential for a comprehensive exploration of disease etiology, identification of potential drug targets, and advancement of precision medicine [11]. For instance, Wang M et al.'s research elucidated the primary mechanisms underlying the progression of silicosis through integrated omics investigations, leading to the identification of gefitinib and fostamatinib as targeted drugs against EGFR and SYK, respectively, thereby enhancing the pathological progression of silicosis [12]. Liang C et al. utilized bioinformatics and network pharmacology analyses to predict the efficacy of triptolide in treating silicosis by modulating multiple signaling pathways, with subsequent experimental validation confirming its notable anti-inflammatory and anti-fibrotic properties [13].

This study investigated the temporal alterations in transcriptional and protein profiles within the gastrocnemius muscle following denervation, utilizing a sciatic nerve transection model and obtaining muscle tissue at various time intervals (3 days, 2 weeks, 4 weeks). Integration of transcriptomic and proteomic data enabled the identification of pivotal genes and proteins exhibiting consistent expression patterns during denervation-induced muscle atrophy. The identified proteins predominantly functioned in oxidative stress, endoplasmic reticulum stress, and glucose metabolism pathways. Through integrative multi-omics analysis, a set of genes that play a key role in the disease process were identified, and upstream regulatory analysis was conducted. Ultimately, NRF2 was selected as a potential therapeutic target to delay denervation-induced muscle atrophy. Utilizing the FDA-approved Omaveloxolone, we targeted the activation of NRF2 to mitigate muscle denervation atrophy by reducing oxidative stress and stabilizing energy metabolism. This investigation not only enhanced comprehension of the molecular mechanisms underlying muscle denervation atrophy but also laid the groundwork for the development of novel therapeutic interventions.

Method and material

Animal model preparation and drug treatment

SPF-grade C57BL/6 J male mice, aged 8 weeks and weighing between 20-30 g, were sourced from the Xinjiang Medical University Animal Experiment Center of Xinjiang Medical University (License No: SYXK(Xin)-2018–0003), and were carefully selected and housed in a controlled environment with a 24-h light–dark cycle. The denervation model was established by anesthetizing the mice with 0.5% pentobarbital sodium, making a skin incision on the right hindlimb to expose and sever the sciatic nerve, creating a 0.5 cm nerve defect. Successful model creation was confirmed by observing right foot drooping during walking. The gastrocnemius muscles were chosen as the subject for proteomic and transcriptomic analysis and categorized into 4 groups (3 biological replicates for each group): normal group, 3 days post-denervation group, 2 weeks post-denervation group, and 4 weeks post-denervation group.

Omaveloxolone (MCE, China, HY-12212) powder is dissolved in DMSO and formulated as 5% DMSO + 40%PEG-300 + 5% Tween80 + 50% saline. Omaveloxolone was administered to the mice at a dose of 10 mg/kg daily for 28 days starting 30 min after model establishment, while the control group was treated with drug-free formula solution. All animal experiments were conducted under the ethical approval of the Xinjiang Medical University Animal Experiment Committe. (Approval number: IACUC-KT-20231010–31).

Masson stain

The Masson staining technique is employed to visualize collagen fibers in blue and muscle fibers in red, as well as to demonstrate the fibrosis of denervated muscles. The Masson's trichrome stain kit (ServicBio, G1006) is utilized for the processes of dewaxing, dehydrating, and staining paraffin sections. The samples undergo incubation, heating, multiple staining steps, and washing, followed by differentiation and dehydration using glacial acetic acid and ethanol. The percentage area of collagen fibers is calculated by dividing the collagen fiber area by the sum of collagen and muscle fiber areas.

Bioinformatics analysis

The bioinformatics analysis involved identifying differential expression signatures from each omics dataset through database searches and data management. Principal component analysis (PCA) was then conducted using all identified differentially expressed proteins and genes to investigate differences within and between groups of each sample in each omics dataset. The fuzzy c-means algorithm, implemented by the R package (v. 4.3) via Mfuzz (v. 2.62.0), was utilized to cluster the differential proteins and genes of the transcriptome and proteome for spatiotemporal expression simulation [14, 15]. Gene Ontology (GO) annotations, categorized into Biological Process, Cellular Component, and Molecular Function, and KEGG (Kyoto Encyclopedia of Genes and Genomes) database were important tools used for bioinformatics analysis. The GO and KEGG enrichment analysis of differentially expressed genes was performed using the clusterProfiler (4.4.4) R package, with the Benjamini & Hochberg method to adjust the P value. Terms with a corrected P value less than 0.05 were considered significantly enriched GO terms and KEGG pathways. The statistical enrichment of differentially expressed genes in the KEGG pathway was analyzed using the clusterProfiler R software. Protein–protein interaction (PPI) analysis of differentially expressed genes was conducted based on the STRING database, and hubgenes and key subnetworks in the PPI network were further identified using cytoHubba and MCODE plug-ins in Cytoscape software. Additionally, the key genes were predicted as upstream regulated transcription factors using TRRUST v2 (http://www.grnpedia.org/trrust/).Search the corresponding target drugs in the Drugbank database (http://go.drugbank.com). Choose 251 proteins that exhibit differential expression patterns in both the transcriptome and proteome data sets between the 2-week and normal groups, for subsequent analysis using Ingenuity Pathway Analysis (IPA version number 111725566) to predict potential upstream regulators and regulatory pathways influencing the expression of these genes. We performed gene set enrichment analysis using software GSEA and MSigDB [16] to identify whether a set of genes in specific KEGG pathways terms shows significant differences in two groups. Briefly, we input geneexpression matrix and rank genes by SignaltoNoise normalization method. Enrichment scores and p-value was calculated in default parameters.

Protein extraction and protesome analysis

The muscle tissue was cut, treated with SDT lysis solution, homogenized, boiled, and centrifuged to remove insoluble matter. The supernatant was filtered and collected for protein quantification using the BCA method. 20 µg of protein from each sample was mixed with loading buffer, boiled, and run on SDS-PAGE. Protein solution was then treated with DTT, cooled, and centrifuged with UA buffer in a 30kD ultrafiltration tube. The process was repeated once, discarding the filtrate. Then, 100 μL of IAA buffer was added and allowed to react in the dark at room temperature for 30 min. After centrifugation and washing with UA buffer twice, 40 μL of Trypsin buffer was added and incubated at 37 °C for 16–18 h. Finally, 40 μL of 50 mM NH4HCO3 solution was added, centrifuged, and the filtrate was collected. The peptides were purified, dried, and dissolved in formic acid. They were then separated using nanoflow chromatography and analyzed with a mass spectrometer. Data processing and quantitative analysis were done using Maxquant software.

RNA extraction and transcriptome analysis

Muscle tissue samples were ground into powder in liquid nitrogen and RNA was extracted using Buffer RLS and a specific kit. RNA quality was checked with an Agilent bioanalyzer before library construction for sequencing, which included enriching mRNA, fragmenting it, synthesizing cDNA, and preparing the library with Illumina's NEBNext Ultra RNA Library Prep Kit. The process involved amplifying cDNA and selecting 200 bp fragments. The library was then quantified and validated before sequencing on an Illumina platform. Clean reads were aligned to a reference genome and gene expression levels were quantified. DESeq2 and edgeR were used for differential expression analysis of samples with and without biological replicates, respectively. Genes were deemed significantly differentially expressed based on adjusted P values and log fold changes.

Parallel reaction monitoring(PRM) validation

We employed the 4D-PRM technique, a high-resolution mass spectrometry method that incorporates Trapped Ion Mobility Spectrometry (TIMS), to enhance ion selectivity and sensitivity for targeted protein analysis in denervated gastrocnemius muscles. Proteins were extracted and analyzed using a nanoflow NanoElute system and a timsTOF Pro mass spectrometer, with a 60-min gradient chromatographic separation on a 1.6 μm C18 column at flow rates of 300 nL/min. PRM-PASEF mode was utilized for precise quantification, scanning precursor ions within a range of 100–1700 m/z and employing controlled ion fragmentation. The mass spectrometry data was analyzed utilizing Biognosys' SpectroDive software against a Mus musculus-specific Uniprot database, yielding comprehensive information on temporal alterations in protein expression within the muscle tissue.

WGA stain

WGA binds to cell surfaces and was used to measure myofiber diameter in muscle tissue. Tissue sections were treated to remove paraffin and dehydrated before antigen retrieval was done with citrate buffer in a microwave. A circle delineating the tissue was sketched using a histochemistry pen and subsequently rinsed with phosphate-buffered saline (PBS). A solution of WGA dye was concocted by diluting 5% bovine serum albumin (BSA) (4240GR250, Biofroxx) at a 1:20 ratio, then subjected to incubation at either 37 °C for 1 to 1.5 h or overnight at 4 °C. Following the incubation period, the tissue sections underwent three washes with PBS, after which DAPI was introduced for nuclear staining (D8417-1MG, Sigma) and allowed to incubate at room temperature in the absence of light for 20–30 min before being rinsed with PBS. Ultimately, the slides were affixed using an anti-fluorescence quenching mounting medium (V900155-25G, Sigma). Utilize the Image-Pro Plus software for the calculation of the average cross-sectional area and average diameter of muscle fibers.

Immunohistochemistry

Paraffin sections were treated with citrate buffer and heated in a microwave to retrieve antigens. Tissue outlining was done with a histochemistry pen, followed by washing with PBS and blocking with hydrogen peroxide solution. Subsequently, the sections were incubated overnight with primary antibodies at 4 °C, followed by three more PBS washes. The secondary antibody was then added dropwise and incubated at 37 °C for 30 min, followed by three final PBS washes. The primary antibodies used were Sod2 (Proteintech, China, Cat No.66474-1-lg, 1:600), NRF2 (Proteintech, China, Cat No.16396-1-AP, 1:150), Sdha (Proteintech, China, Cat No.14865-1-AP, 1:300), and PKM (Proteintech, China, Cat No.15822-1-AP, 1:300). The secondary antibodies used were HRP-conjugated goat anti-rabbit (Aspen Biology, China, AS1107, 1:200) and HRP-conjugated goat anti-mouse (Aspen Biology, China, AS1106, 1:200). A DAB working solution was prepared and used to observe a brown color reaction under a microscope. The reaction was stopped with tap water, and the sections were dehydrated and made transparent with alcohol, ethanol, and xylene before being mounted with neutral gum. Image-Pro Plus software was used to quantify the IHC staining intensity of the sections by average optical density (Intensity Optical Density/Area).

Western blot

Skeletal muscle samples were lysed with cold RIPA lysis buffer for 30 min on ice. Centrifuge the tissue lysate at 12,000 g for 20 min, and transfer the supernatant to a new centrifuge tube. Determine the protein concentration of each sample using the BCA protein quantification kit. Take the same amount of protein sample (30 μg) and separate the protein on an 8% or 12% SDS-PAGE gel. The protein is then transferred to a PVDF membrane. Block the PVDF membrane with 5% skim milk powder in TBST buffer (100 mM NaCl, 10 mM Tris–HCl, pH 7.5, 0.1% Tween-20) for 1 h at room temperature. Specific primary antibodies were then added and incubated overnight at 4 °C. Primary antibodies include, muscle specific kinase (MuSK, Affbiotech, China, AF7607, 1:500), glutathione reductase (GSR, Proteintech, China, 18257–1-AP, 1:1000), β-actin (Tdybio, China, TDY051, 1:10000), wash the membrane three times with TBST buffer, add horseradish peroxidase (HRP)-labeled secondary antibody, and incubate at room temperature for 2 h. Secondary antibodies included HRP-Goat anti Rabbit (ASPEN, China, AS1107, 1:10000). Develop using ECL chemiluminescence detection kit (ASPEN, AS1059). Visualize specific protein bands with the ChemiDoc MP Imaging System. Image Lab 5.1 software was used to quantify the protein band intensity in Western blot, and express the relative protein abundance of each sample relative to the reference protein signal.

GSH/GSSH ratio measurement

Glutathione (GSH) /glutathione disulfide(GSSG) ratio measurement uses GSH and GSSG Assay Kit (Beyotime,China, S0053). Prepare the required reagents according to the instructions, including GSSG stock solution, DTNB stock solution, protein removal reagent M solution, NADPH stock solution and Diluted glutathione reductase, etc. Add samples and standards to a 96-well plate according to the instructions, then add total glutathione detection working solution, incubate at 25 °C for 5 min, and add 0.5 mg/ml NADPH solution. Finally, use a microplate reader to measure the OD value at a wavelength of 412 nm, record the absorbance change every 5 min, measure for a total of 25 min, and calculate the GSH and GSSG content in the sample through the standard curve.

Transmission electron microscopy

Muscle tissues were cut into 1mm³ pieces and fixed in a solution with 2.5% glutaraldehyde in 0.1 M cacodylate buffer at pH 7.4. They were then post-fixed with 1% osmium tetroxide in the same buffer to enhance contrast. The samples were dehydrated with graded ethanol solutions to prepare for embedding. After dehydration, the muscles were embedded in epoxy resin, sectioned into 70 nm thick slices, and treated with uranyl acetate and lead citrate for contrast. The stained grids were analyzed with a transmission electron microscope (NEC, JEM-1200EX, Japan) to examine muscle fiber ultrastructure and identify mitochondrial pathology.

Single-nucleus cell sequencing data analysis

The GSE183802 dataset, containing six samples (three normal samples and three Denervation samples), was obtained from the GEO database [17]. Using the R package "Seurat," we created a Seurat object from the matrices of these six samples and set the threshold to nfeature > 200 && nfeature < 6000 && nCount > 500. This resulted in a total of 29,736 cells. Data normalization was performed using the NormalizeData function, and highly variable genes between cells were identified using the FindVariableFeatures function (feature selection). The “Variance Stabilizing Transformation” method was employed to select 2000 highly variable genes, which were used for subsequent cell type identification and linear dimensionality reduction of single-cell data to determine the usable dimensions of the data.

After completing PCA dimensionality reduction, we performed unsupervised clustering analysis on the 29,736 cells using the FindNeighbors and FindClusters functions in the Seurat package, setting the resolution to 2, and clustering the cells into 32 groups. The results of these clusters were annotated using markers from the original dataset publication. The 23 key genes validated by PRM were traced back to the annotated cell types, and violin plots depicting intergroup differences were generated, with significance calculated using the Wilcoxon test.。

Statistics

The normality of quantitative data was initially assessed using GraphPad Prism 9.0.0 (GraphPad Software, Inc). Comparisons among multiple groups in experimental data were conducted through a one-way ANOVA followed by Bonferroni's multiple comparisons test. Pairwise comparisons between two groups were carried out using Student’s t-test. Notations used for statistical significance include: ns for no significance, * for p < 0.05, ** for p < 0.01, *** for p < 0.001, and **** for p < 0.0001.

Result

Changes in gene expression and protein levels in the gastrocnemius muscle following denervation

This study employed dynamic assessment to investigate transcriptional and protein-level changes in the gastrocnemius muscle following denervation. A denervation mouse model was utilized to examine RNA and protein alterations at 3 days, 2 weeks, and 4 weeks post-sciatic nerve disconnection. Gastrocnemius muscle tissue was collected and subjected to analysis for changes in wet weight ratio and muscle fibrosis using Masson staining (Fig. 1A). A comprehensive analysis identified a total of 54,534 transcripts and 3,218 proteins through transcriptomic and proteomic approaches at various time points. Statistical analysis identified 23,282 transcripts and 1,852 proteins exhibiting statistically significant variances across the three time points (Fig. 1B, C). Principal component analysis delineated discernible expression patterns between denervated muscle tissue and normal muscle tissue, with resemblances noted between the 2-week and 4-week post-injury time points (Fig. 1B). Additionally, proteomic analysis corroborated substantial distinctions in expression patterns between the denervated and control groups throughout the three time points (Fig. 1C). Heatmap analysis visually depicted both similarities and heterogeneity within the transcriptome and proteome profiles. (Fig. 1D, E).

Fig. 1.

Transcriptomic and proteomic analysis of the gastrocnemius muscles in a denervated mouse model at various time point. A Appearance and wet weight ratio of gastrocnemius muscles at 3 days, 2 weeks, and 4 weeks post-denervation. Bars represent the average ± S.D. of 3 groups (n = 10). Muscle tissues show significant atrophy at 2 weeks and 4 weeks post-denervation. Masson's trichrome staining indicates a progressive increase in muscle fibrosis as denervation progresses. B RNA-seq analysis identified a total of 54,534 transcripts, with 23,282 transcripts common across the three time points; Principal component analysis (PCA) revealed that the first principal component explained 34.83% of the variance in the data, while the second component accounted for 31.29% of the variance. C 4D-Label free proteomics identified a total of 3,218 proteins, with 1,852 proteins common across the three time points. PCA results revealed that the first principal component explained 51% of the data variance, and the second component explained 11% of the data variance. D Heatmap of differentially expressed genes in the transcriptome at 3 days, 2 weeks, and 4 weeks compared to normal. E Heatmap of differential protein expression in proteomics at 3 days vs. normal, 2 weeks vs. normal, and 4 weeks vs. normal

Filtering key genes for muscle atrophy using multi-omics

A multi-omics approach was employed to ascertain pivotal genes linked to denervation-induced muscle atrophy, utilizing a threshold of padj < 0.05 |log2FoldChange|≥ 1.0 to discern differential genes and padj < 0.05 |log2FoldChange|≥ 0.26 to differential proteins. This screening criterion is consistent with the criteria for subsequent integrated analysis of proteomics and transcriptomics. Transcriptome analysis unveiled 2730 significantly altered genes (1284 up-regulated, 1436 down-regulated) following 3 days of denervation in contrast to normal muscle tissue, 2796 genes (1392 up-regulated, 1404 down-regulated) after 2 weeks, and 3053 genes (1664 up-regulated, 1389 down-regulated) after 4 weeks. (Fig. 2A) The proteomic analysis revealed differential expression of 244 proteins (166 up-regulated, 78 down-regulated) following 3 days of denervation, 1095 proteins (705 up-regulated, 390 down-regulated) after 2 weeks, and 1065 proteins (705 up-regulated, 360 down-regulated) after 4 weeks. (Fig. 2B) Integration of the proteomic data with gene expression analysis identified 97 genes with consistent expression patterns after 3 days, 251 genes after 2 weeks, and 223 genes after 4 weeks, underscoring the significance of cross-referencing omics datasets. The study revealed 24 genes, consisting of 15 up-regulated and 9 down-regulated genes, that displayed consistent expression patterns at three distinct time points following denervation: 3 days, 2 weeks, and 4 weeks. (Fig. 2C–E) These genes demonstrated alterations at both the protein and transcriptional levels, indicating their potential significance in the development of muscle denervation atrophy.

Fig. 2.

Multi-omics analysis to identify differentially expressed genes with consistent expression trends. A With a threshold of padj < 0.05 and |log2FoldChange|> 1.0, transcriptome analysis unveiled 2,730 significantly altered genes (1,284 up-regulated, 1,436 down-regulated) after 3 days of denervation compared to normal muscle tissue, 2,796 genes (1,392 up-regulated, 1,404 down-regulated) after 2 weeks, and 3,053 genes (1,664 up-regulated, 1,389 down-regulated) after 4 weeks. B With padj < 0.05 and |log2FoldChange|> = 0.26, the proteomic analysis revealed differential expression of 244 proteins (166 up-regulated, 78 down-regulated) following 3 days of denervation, 1,095 proteins (705 up-regulated, 390 down-regulated) after 2 weeks, and 1,065 proteins (705 up-regulated, 360 down-regulated) after 4 weeks; C The quadrant chart depicting the correlation analysis between proteins and transcripts illustrates the differential expression observed at the proteomic and transcriptomic levels. The horizontal and vertical coordinates correspond to log2-transformed fold changes in the transcriptome and proteome, respectively, demonstrating instances in which expression patterns are concordant, discordant, or uniquely altered at either the protein or transcript level. For non-expressed results in proteins or transcripts, expression values are set to zero by default in the plot, thus points on the Y-axis represent expressions only found in the transcriptome, and points on the X-axis represent expressions only found in the proteome. D The comparison of Venn diagrams illustrating the counts of differentially expressed proteins and transcripts reveals that there are 97 genes exhibiting consistent expression trends in the 3d.vs.normal group, 251 genes in the 2w.vs.normal group, and 223 genes in the 4w.vs.normal group. Notably, 24 genes are found to be consistently expressed across all three time points. E A heatmap was generated through hierarchical clustering analysis utilizing differential protein and associated transcriptomic data. Each row in the heatmap corresponds to a differentially intersected gene, with the tree structure on the left illustrating the clustering of expression patterns of these intersected genes

Spatiotemporal dynamic analysis of differential proteins and genes

The R software package Mfuzz was employed for bioinformatics analysis to investigate the temporal changes of differential proteins and genes in denervation-induced muscle atrophy in a specific model. A total of 1127 differential genes and 566 proteins were identified using the screening criteria of at least one time point meeting a significance level of P < 0.05 and Fold Change > 2. (Supplementary file1-2)The analysis classified the 566 proteins and 1127 genes into 6 clusters according to their spatiotemporal expression patterns (Fig. 3A, B). 24 common gene-proteins and transcriptional data are categorized into four groups according to their temporal expression patterns. (Fig. 3C).

Fig. 3.

Analysis of the spatiotemporal trends in the progression of muscle denervation. A 566 proteins and (B) 1127 mRNAs were respectively clustered into 6 distinct expression clusters using fuzzy c-means clustering to demonstrate the dynamic expression changes at the transcriptional and protein levels during denervation progression. Each cluster highlights the top 3 significantly enriched pathways. In the multi-omics dataset, proteins or genes exhibiting persistent upregulation during muscle atrophy caused by denervation are highlighted in red, whereas those displaying persistent downregulation are highlighted in blue. C 24 common gene-proteins and transcriptional data are clustered into 4 categories based on their temporal expression patterns. D The Mfuzz analysis method was utilized to identify key genes and proteins that exhibit continuous upregulation or downregulation. The results revealed that proteins showing continuous upregulation were enriched in 18 KEGG pathways, while genes showing continuous upregulation were enriched in 17 KEGG pathways, with two pathways being common to both. Conversely, proteins exhibiting continuous downregulation were enriched in 32 KEGG pathways, and genes showing continuous downregulation were enriched in 38 KEGG pathways, with 12 pathways being common to both. Notably, among the genes and proteins enriched in common pathways, 18 genes (9 upregulated marked in red and 9 downregulated marked in green) displayed consistent expression patterns at both the protein and transcriptional levels

Notably, The proteins that are consistently up-regulated in denervation-induced muscle atrophy are predominantly linked to pathways involving cellular stress response, cytoskeletal dynamics, signal transduction, and intercellular communication. These pathways encompass processes such as protein processing in the endoplasmic reticulum, regulation of the actin cytoskeleton, and RNA transport and processing. At the transcriptional level, denervated muscle atrophy is characterized by the continuous up-regulation of differential genes enriched in pathways such as NF-κB signaling and apoptosis. This up-regulation indicates the presence of an inflammatory state and heightened apoptosis during the process of denervation-induced muscle atrophy, resulting in increased apoptosis, activation of the immune response, and disruption of signaling pathways.

Proteins and genes that exhibit sustained down-regulation in denervated muscle atrophy are enriched in critical pathways related to energy metabolism, carbohydrate metabolism, amino acid biosynthesis, glucose regulation, hypoxic response, and nucleotide metabolism. These extensive perturbations in metabolic pathways indicate inadequate energy provision, diminished capacity for protein synthesis and repair, aberrant glucose metabolism, compromised DNA and RNA synthesis and repair. These findings imply that reducing oxidative stress and stabilizing energy metabolism could play a significant role in mitigating or reversing the progression of the disease.

After conducting trend analysis, we focused on one type of continuously up-regulated gene set (C1, C3) and one type of continuously down-regulated gene set (C2, C4) at both protein and transcription levels. Enrichment analysis revealed that the C1 gene set was enriched in 18 KEGG pathways, while the C3 gene set was enriched in 17 pathways, with two pathways overlapping between C1 and C3, involving 17 gene symbols. On the other hand, the C2 gene set showed enrichment in 32 KEGG pathways, and the C4 gene set in 38 KEGG pathways, with 12 pathways common to both gene sets, covering 71 gene symbols. Ultimately, we identified a total of 18 genes exhibiting consistent expression trends at both protein and transcription levels, with 9 genes up-regulated and 9 genes down-regulated (Fig. 3D).

Filter Hubgene by PPI network diagram

Utilizing cytoscape software, we constructed a PPI network diagram incorporating 24 differential genes identified from the multi-omics joint analysis and 18 genes from the Mfuzz analysis, totaling 42 key genes (Fig. 4A). CytoHubba was then used to filter critical nodes and subnetworks, revealing the top 6 genes as Hsp90aa1, Eno3, Tpi1, Pkm, Sdha, and Sod2. Subsequently, the MCODE plug-in identified 3 key subnetworks (Fig. 4B, D). The top 30 proteins were analyzed using the TRRUST website to predict upstream regulated transcription factors, resulting in the prediction of Nfe2l2(also name as NRF2), Rela, and Nfkb1. (Fig. 4C).

Fig. 4.

Selection of hub genes through the PPI network diagram. A 24 differential genes were identified through a multi-omics combined analysis, and 18 differential genes were identified by Mfuzz analysis; both sets were integrated into a collection of 42 key genes. B A PPI network diagram was constructed using the 42 key genes, and the top 30 hub genes were selected based on degree values. C In the TRRUST network, transcription factors such as Nfe2l2, Rela, and Nfkb1 are predicted to regulate the top 30 hub genes upstream. D NRF2 was chosen for inclusion in the group of 42 essential genes, which were subsequently categorized into three distinct sub-networks. The protein–protein interaction (PPI) network visualization illustrates the interactions between NRF2 and Fn1, Sod2, Hsp90aa1, and Pdg. E, G Significantly activated and inhibited upstream pathways and top15 regulatory molecules in IPA analysis. F The map of Mitochondrial Dysfunction pathway

A total of 251 genes exhibiting consistent expression patterns in both transcriptomic and proteomic data across the 2-week and normal groups were chosen for Ingenuity Pathway Analysis (IPA). (Supplementary file 3) The investigation focused on the overlap between molecules within the dataset and those associated with specific pathways to determine the potential activation or inhibition of these pathways. Pathway activation following denervation was evaluated utilizing a z-score metric, with a z-score ≥ 2 signifying substantial activation and a z-score ≤ -2 indicating significant inhibition. (Fig. 4E, Supplementary file 4) Based on the results obtained from the Ingenuity Pathway Analysis (IPA), the three pathways that exhibited the highest levels of activation were Mitochondrial Dysfunction (z-score = 3), Synaptogenesis Signaling Pathway (z-score = 2.714), and mTOR (z-score = 2.646). Conversely, the pathways that were most significantly inhibited included Electron Transport, ATP Synthesis, and Heat Production by Uncoupling Proteins (z-score = −4.69), Oxidative Phosphorylation (z-score = −4.359), and The Citric Acid (TCA) Cycle and Respiratory(z-score = −3.606). IPA results indicate that reduced energy metabolism and mitochondrial dysfunction are the primary upstream regulatory factors for muscle denervation atrophy, suggesting that stabilizing mitochondria to enhance energy metabolism levels is essential for delaying muscle atrophy. In the map of Mitochondrial Dysfunction pathway, key activating factors identified include oxidative stress and endoplasmic reticulum stress, with green indicating activation and red indicating inhibition. In the map, we identified the transcription factor NRF2 as a crucial node for alleviating oxidative stress and stabilizing mitochondrial function. In the prediction of upstream transcription factors within the hubgenes set, NRF2 also showed significant enrichment. This further emphasizes the potential of NRF2 as an important target for delaying muscle denervation atrophy. (Fig. 4F).

The Activation z-score algorithm was employed to forecast the activation or inhibition of upstream regulators, thereby reducing the occurrence of significant predictions arising from random data. In this investigation, metribolone (activation z-score = −5.395), PPARGC1A (activation z-score = −4.953), and mono-(2-ethylhexyl) phthalate (activation z-score = −4.372) emerged as the most notably inhibited upstream regulators. Conversely, Torin1 (activation z-score = 4.025), CLPP (activation z-score = 3.9), and WNT3A (activation z-score = 3.713) were identified as the most significantly activated upstream regulators. (Fig. 4G, Supplementary file 5).

Hubgene parallel reaction monitoring (PRM) verification

The top 30 hubgenes were selected to investigate their protein expression differences. A denervated muscle atrophy model was created using C57 mice, and PRM testing was conducted at various time points. Results showed that 12 genes exhibited expression variances compared to normal tissues across all three time points, while 11 genes showed differences at each time point. Notably, 2 proteins had significant expression disparities at 2 weeks and 4 weeks post-injury, while 8 proteins did not display differences across the three time points. Consequently, these 23 differentially expressed genes were identified as core targets in denervated muscle atrophy. (Fig. 5A-F) GO and KEGG analyses revealed that these key proteins primarily function in the pathways associated with carbohydrate catabolic process, glycolytic process, monosaccharide biosynthesis process. These are all biological processes closely related to energy metabolism. (Fig. 5G–H, Supplementary file 6).

Fig. 5.

4D-PRM confirms protein level changes in the top 30 hub genes during denervation, with 23 genes showing significant differences. A ENO3, Fbp2, GMPR, Pkm, and Pgam2 consistently demonstrate downregulation during the progression of denervation-induced muscle atrophy. These genes primarily participate in glycolysis and gluconeogenesis, suggesting a notable reduction in the functionality of these metabolic pathways following denervation. The persistent upregulation of Pgd and Tpi1 reflects compensatory adjustments made by muscle tissue in response to oxidative stress. Bars represent the average ± S. D of 3 independent experiments (n = 3) (B) Eef1a1, Eif4a1, and Rpl3, key components in the protein synthesis pathway, exhibit sustained upregulation following denervation, indicative of the muscle tissue's adaptive response to stress. Bars represent the average ± S. D of 3 independent experiments (n = 3) (C) Art3 is primarily involved in intracellular signal transduction, while Auh is a mitochondrial protein associated with fatty acid metabolism; both proteins exhibit consistent downregulation. Additionally, Klc1, a constituent of the kinesin motor complex responsible for material transport, demonstrates persistent upregulation. Bars represent the average ± S. D of 3 independent experiments (n = 3) (D) The sustained elevation of hspb8, hspb7, and hsp90aa1 gene expression signifies stress-induced injury in muscle tissue subsequent to denervation. The persistent upregulation of Hsp90b1, associated with endoplasmic reticulum oxidative stress, implies the occurrence of endoplasmic reticulum stress. Bars represent the average ± S. D of 3 independent experiments (n = 3) (E) Flnc, Klhl41, Sync, and Tubb6, proteins related to maintaining muscle structural integrity, are persistently upregulated during the pathological process. Bars represent the average ± S. D of 3 independent experiments (n = 3) (F) The continuous upregulation of Fn1 and Tnc reflects the development of muscle fibrosis following denervation. Bars represent the average ± S. D of 3 independent experiments (n = 3) * for p < 0.05, ** for p < 0.01, *** for p < 0.001, and **** for p < 0.0001. G–H GO/KEGG analysis of the 23 gene

The results above indicate that oxidative stress damage, endoplasmic reticulum stress, and disruptions in energy and glucose metabolism play significant roles in the development of muscle denervation atrophy, ultimately resulting in atrophy and fibrosis. Furthermore, mitochondrial dysfunction is identified as the primary factor contributing to these pathological changes. Therefore, stabilizing mitochondrial function could be a crucial therapeutic approach to delay muscle denervation atrophy.In our study, we identified NRF2 as a potential therapeutic target. Activation of NRF2 alleviates cellular oxidative stress and stabilizes mitochondrial dysfunction, thereby enhancing energy metabolism pathways to improve muscle denervation atrophy. To confirm this hypothesis, we assessed the therapeutic effects of the targeted activator omaveloxone on denervation muscle atrophy in our subsequent research.

Expression of 23 genes in different cell types of muscle tissue

After a series of processes including standardization, PCA dimensionality reduction, and cell clustering, mononuclear cell data were ultimately annotated into 13 cell types (Fig. 6A, C): Type I (n = 185), Type IIa (n = 1276), Type IIx (n = 1376), Type IIb (n = 15,204), NMJ (n = 130), MTJ (n = 374), MuSCs (n = 239), FAPs (n = 4428), Macrophage (n = 435), Endothelial (n = 534), Pericyte (n = 356), Adipocyte (n = 157), and Others (n = 5042). Cells annotated as Others were excluded (Fig. 6B), and the distribution of different cells between disease and normal groups was displayed (Fig. 6D, E). The results showed a significant increase in FAPs cells and a significant decrease in Type IIx cells in the Denervation group. A violin plot was created to trace the expression of 23 key validated genes across 12 cell types between groups (Fig. 6F). The results revealed that Eno3, Gmpr, Pkm, Pgam2, and Tpi1 showed the most significant differential expression in Type II muscle cells, especially in Type IIb fibers, followed by FAPs cells. Eno3, Gmpr, and Pgam2 showed a downward trend in expression in Type II fibers and FAPs, consistent with PRM results. Hspb7 and Hspb8 genes showed the most significant differential expression in Type II fibers. Klc1, Eef1a1, Eif4a1, and Tubb6 showed significant differential expression in Type IIb cells. Flnc and Klhl41 showed significant differential expression in Type IIa, IIx, and IIb muscle fibers, with an upward trend consistent with PRM results. Art3 and Auh showed differential expression in almost all cell types, with downregulated expression in Type IIa and IIb cells, consistent with PRM results. In summary, Type II muscle fibers, especially Type IIb fibers, are the most affected cell types in denervation-induced muscle atrophy.

Fig. 6.

A Annotation of cells into 13 subpopulations; B Visualization of 12 cell subpopulations after excluding unannotated cell types; C Bubble plot of different cell markers corresponding to various cell types; D, E Comparison of the distribution proportions of different cell types between the Denervation and normal groups; F Violin plots showing the expression levels of 23 PRM-validated genes in different cell types, with intergroup comparisons calculated for significance using the Wilcoxon test. * for p < 0.05, ** for p < 0.01, *** for p < 0.001, and **** for p < 0.0001

Omaveloxolone reduces oxidative stress and endoplasmic reticulum stress

Omaveloxolone, a targeted activator sourced from the DrugBank database, received FDA approval in 2023 for treating Friedreich's Ataxia. While studies have examined the role of NRF2 in various neurodegenerative diseases, research on its specific impact on denervated muscle atrophy remains scarce. To address this gap, a mouse denervation model was utilized, with omaveloxolone administered intraperitoneally at 10 mg/kg dosage 30 min post-membrane creation. Gastrocnemius muscles were collected after 4 weeks of continuous injection. Results indicated a higher wet weight ratio and muscle fiber area in the omaveloxolone group compared to the control group. (Fig. 7A–C) The mean cross-sectional area and diameter of muscle fibers in the omaveloxolone group exhibited a statistically significant increase compared to those in the control group. (Fig. 7D) Immunohistochemistry revealed increased levels of Sod2, Sdha, and NRF2 proteins in the omaveloxolone group, suggesting a reduction in oxidative stress damage. (Fig. 7E–I) Additionally, up-regulation of glycolysis-related proteins PKM in the omaveloxolone group hinted at enhanced energy supply through the glycolysis pathway. TEM analysis of muscle tissue revealed significant mitochondrial matrix edema in the gastrocnemius muscle of the control group, characterized by the presence of vacuoles, disappearance of mitochondrial cristae shape, and damage to the mitochondrial membrane integrity. (Fig. 7J) Conversely, in the Omaveloxolone experimental group, a less pronounced degree of mitochondrial swelling was observed compared to the control group, with the mitochondrial morphology largely preserved. Although the mitochondrial cristae were disordered, the fundamental shape was still discernible. This indicates that Omaveloxolone plays a role in maintaining the stability of mitochondrial morphology.

Fig. 7.

A Comparison of gross photographs, Masson staining, and WGA staining of the gastrocnemius muscles between the omaveloxolone group and the control group. B The wet weight of the gastrocnemius muscle treated with omaveloxolone was significantly higher than that of the control control group (n = 11). C In the omaveloxolone group, the proportion of collagen fibers (blue) to muscle fibers (red) was lower than in the control group, indicating a lower degree of fibrosis (n = 5). D The average area and diameter of muscle fibers in the omaveloxolone group were greater than those in the control group (n = 5). E ICH was performed on the affected side of the gastrocnemius muscle to detect the expression of Sod2, Sdha, NRF2, and PKM proteins (n = 5). F-I The results showed that the average optical density values (IOD/Area) of Sod2, Sdha, NRF2, and PKM in the omaveloxolone group were significantly higher than those in the control group. J TEM revealed significant mitochondrial swelling and damage in the control group, while the Omaveloxolone group showed less severe swelling and maintained better mitochondrial integrity. Green arrows show mitochondria

RNA-seq and GSEA analyses reveal the protective mechanism of Omaveloxolone in denervation muscle atrophy model

RNA-seq sequencing was performed on the affected gastrocnemius muscles of mice in the Omaveloxolone and Control groups. Using a threshold of p-adj < 0.05 and |log2(FC)|> 1, a total of 474 differentially expressed genes were identified (188 significantly upregulated and 286 significantly downregulated) (Fig. 8A). GSEA analysis, with significance thresholds of |NES|> 1, NOM p-val < 0.05, and FDR q-val < 0.25, revealed 17 significantly activated pathways, including Autophagy, Glutathione metabolism, Mitophagy, and PPAR signaling pathway (Fig. 8B). These findings suggest that Omaveloxolone may enhance tissue antioxidant capacity by improving glutathione metabolism, and maintain mitochondrial quality and energy metabolism through promoting mitophagy and activating the PPAR signaling pathway. Although denervation-induced muscle atrophy is not classified as a neurodegenerative disease, the functional degradation of the neuromuscular junction plays a key role in disease progression. The pathways of Huntington's disease and Parkinson's disease were significantly inhibited, suggesting that Omaveloxolone may help maintain neuromuscular junction function. (Fig. 8C) The suppression of the Natural killer cell mediated cytotoxicity and Antigen processing and presentation pathways may reflect Omaveloxolone's role in modulating immune responses, reducing inflammation and oxidative stress, thereby protecting neuromuscular junctions and delaying muscle atrophy. (Fig. 8C) Western blot analysis of GSR and MuSK expression showed that GSR levels and the GSH/GSSH ratio were significantly higher in the Omaveloxolone group compared to the control group (Fig. 8D, F, G), indicating that the drug intervention significantly enhanced the antioxidant capacity of muscle tissue. In normal muscle tissue, MuSK is highly concentrated in the specialized motor endplate region of the sarcolemma and exhibits a granular distribution, which is crucial for NMJ development and maintenance. After denervation, MuSK expression in the target muscle significantly increases and shows a diffuse distribution on the sarcolemma. The expression of MuSK was significantly lower in the Omaveloxolone group suggesting that Omaveloxolone has a protective effect on maintaining the motor endplate in denervated muscle. (Fig. 8D, F).

Fig. 8.

A 474 differentially expressed genes were identified with188 significantly upregulated and 286 significantly downregulated. The volcano plot and heatmap to visualize DEGs distribution and expression; B GSEA analysis indicated that Autophagy, Mitophagy, Glutathione metabolism, PPAR signaling pathway were activated; C Huntington disease, Parkinson disease, Natural killer cell mediated cytotoxicity, Antigen processing and presentatior were inhibited; D Western blot analysis of the expression trends of GSR and MuSK proteins; E, F results show a significant increase in GSR expression in the Omaveloxolone group and a significant decrease in MuSK expression in the Omaveloxolone group; G GSH/GSSH is significantly higher in the Omaveloxolone group than in the Control group, indicating a higher antioxidant capacity in muscle tissue

Discussion

In this study, we constructed a time-resolved multi-omics atlas following muscle denervation and identified a series of key gene sets crucial for disease progression through comprehensive comparison and mutual validation of multi-omics data. Furthermore, predicting the upstream transcription factors of these gene sets provided new directions for the discovery of therapeutic drugs. There are complex post-transcriptional regulatory mechanisms between mRNA and protein levels, such as RNA splicing, post-translational modifications of mRNA, and regulation by non-coding RNAs, leading to a relatively low correlation between transcription and protein levels (approximately 0.3–0.4) [18–20]. This means that changes at the RNA level do not always reflect changes at the protein level. We focused on genes that showed consistent expression trends at both the transcriptional and protein levels at each time point. This consistency indicates that the regulation of these genes during transcription and translation is highly coordinated, ensuring that regulation of mRNA directly reflects on protein expression, thereby effectively regulating the biological phenotype. From the perspective of omics high-throughput data analysis, selecting genes with consistent expression trends helps reduce the risk of false positives, ensuring that the selected genes are biologically significant, not just statistical coincidences. Therefore, genes with consistent expression trends at the transcription and protein levels are considered key to understanding the mechanisms of disease onset and progression.

Denervation-induced muscle atrophy is a complex pathological process involving multiple interacting factors such as calcium (Ca²⁺) signaling, oxidative stress (ROS/RNS), and inflammatory responses. Ca²⁺ overload in the cytoplasm and mitochondria is a major cause in chronic oxidative stress and increased pro-inflammatory cytokines [21, 22]. After peripheral nerve injury, the innervated skeletal muscle loses its contractile function, and blood flow decreases, leading to muscle hypoxia, which in turn increases the formation of ROS [23]. Excessive ROS can activate FOXO3a through multiple pathways, thereby upregulating the expression of atrogin-1 and MuRF1, promoting protein degradation [24]. Excessive ROS can also activate various pathways, including NF-κB, to induce inflammation. Particularly on the third day after denervation, NF-κB levels significantly increase, not only enhancing the release of pro-inflammatory factors but also suppressing the skeletal muscle's antioxidant system, leading to reduced antioxidant capacity in muscle tissue [25]. Oxidative stress can also increase the production of myostatin (MSTN), a protein that triggers the MSTN-Smad2/3 pathway, reducing protein synthesis. Additionally, it elevates IL-6 levels through p38MAPK activation, ultimately leading to skeletal muscle atrophy.The regeneration of muscle tissue largely depends on muscle satellite cells, which are muscle stem cells [26]. Persistent elevation of ROS reduces the levels of glutathione (GSH), glutathione peroxidase (GPx), and nuclear factor erythroid 2-related factor 2 (NRF2), disrupting NRF2 signaling. This leads to impaired skeletal muscle recovery and aging of muscle satellite cells [22]. PGC-1α plays a pivotal role in orchestrating mitochondrial biogenesis and the regulation of antioxidant enzymes [27]. Although antioxidant enzymes respond to the increased levels of ROS, high levels of ROS/RNS can still modify and inactivate these antioxidant enzymes and PGC-1α. Therefore, previous studies suggest that denervation-induced oxidative stress and inflammation are the main factors in skeletal muscle atrophy and cell apoptosis.

We found 30 important genes and identified major changes in muscle atrophy after denervation: oxidative stress, decreased energy/glucose metabolism which interact and promote each other. Loss of innervation causes mitochondrial dysfunction and increased autophagy, leading to an imbalance between oxidants and antioxidants in cells [28, 29]. This imbalance can result in oxidative stress damage, which is a key factor in muscle atrophy and tissue fibrosis [30–32]. Hspb8, a member of small heat shock protein family, can help reduce oxidative stress by enhancing mitochondrial membrane potential [33]. PRM results revealed increased Hspb8 and decreased Sod2 levels (although not statistically significant) in denervated gastrocnemius muscle, suggesting oxidative stress damage. Pgd, involved in the pentose phosphate pathway, showed continuous upregulation, indicating a greater need for NADPH to combat oxidative stress [34]. Continued increase in Fn1 and Tnc proteins indicates excessive buildup of extracellular matrix proteins, leading to muscle fibrosis [35, 36].

Excessive ROS directly affects the redox balance of the ER, causing oxidative damage to cells' lipids, proteins, and DNA. The aggregation of oxidized proteins increases, disrupting the normal folding and modification processes of proteins in the ER, leading to the accumulation of misfolded proteins in the ER and further triggering ER stress responses [37–39]. Hspb7, Hsp90b1, and Hsp90aa1 assist in proper protein folding, maintain protein function, and enhance cell tolerance to stress conditions. Hsp90b1 (also known as GRP94) is a paralog of HSP90 in the ER and a key factor in the ER stress response [40, 41]. The sustained increase in expression of heat shock proteins reflects cells' compensatory regulation to oxidative and ER stress.The ER has three sensors: PERK, ATF6, and IRE1 [42]. PERK activation can control antioxidant gene expression by phosphorylating NRF2 [43]. If NRF2 response is disrupted, damaged proteins can build up in the ER, leading to PERK-driven cell death [44]. Excessive ER stress can worsen oxidative stress through PERK/NRF2 crosstalk [45]. This is a primary factor in our decision to select NRF2 as a therapeutic target.

Human skeletal muscle fibers are classified into three types: slow oxidative (Type I), fast oxidative (Type IIa), and fast glycolytic (Type IIb), with mice having an additional Type IIx [46]. Type I and IIa fibers have higher activities in oxidative phosphorylation, the TCA cycle, and lipid metabolism, while Type IIx and IIb fibers primarily rely on glycolysis and gluconeogenesis [6]. Enolase 3 (Eno3), pyruvate kinase M (Pkm), and phosphoglycerate mutase 2 (Pgam2) are essential enzymes involved in the glycolysis pathway. Eno3, or β-enolase, is a key enzyme in skeletal muscle that converts 2-phosphoglycerate to phosphoenolpyruvate in glycolysis, aiding in energy production and balance especially during high-energy demand muscle contraction [47–49] PKM (Pyruvate Kinase, M Isoform) is an isozyme of pyruvate kinase involved in the final step of glycolysis, converting PEP to pyruvate and generating ATP.Pgam2(Phosphoglycerate Mutase 2) is a muscle-specific enzyme in the phosphoglycerate mutase family, converting 3-PG to 2-PG in the 8th step of glycolysis [50, 51] Fbp2(Fructose-1,6-bisphosphatase2) is a crucial enzyme in gluconeogenesis, breaking down fructose-1,6-bisphosphate into fructose-6-phosphate [52]. The decreased activity of Eno3, PKM, Pgam2, and Fbp2 in denervated muscle suggests a state of stress and low energy demand. The expression changes of these genes are particularly significant in type IIb cells. This metabolic reprogramming reduces glycolytic enzyme activity, leading to decreased energy production and potentially worsening muscle atrophy [53].

We chose 251 proteins with consistent expression patterns at the 2-week time point for IPA analysis due to its significance in muscle atrophy research. The results obtained from the IPA analysis indicate that the Mitochondrial dysfunction pathway exhibits the most significant level of activation. Examination of the pathway diagram reveals that oxidative stress and endoplasmic reticulum stress play pivotal roles in activating this pathway.The significant inhibition of the three pathways of Electron transport, ATP synthesis, and heat production by uncoupling proteins pathway (z-score = −4.69), Oxidative phosphorylation (z-score = −4.359), and The citric acid (TCA) cycle and respiratory electron transport indicates that muscle denervation results in inadequate energy provision as a consequence of mitochondrial dysfunction, thereby intensifying degenerative alterations in skeletal muscle. This finding is further supported by the analysis of the top 30 genes. IPA analysis identified metribolone and PPARGC1A as highly inhibited upstream regulatory factors, and torin1 as the most activated factor. PPARGC1A, also known as PGC-1α, is a transcriptional coactivator that regulates energy metabolism by controlling mitochondrial dynamics and bioenergetics, specifically enhancing mitochondrial oxidative phosphorylation to increase energy production [54, 55]. The inhibition of PGC-1α indicates that mitochondrial function is severely impacted during the process of denervation in cells. Torin1 is a potent inhibitor of the mTOR signaling pathway, which is crucial for maintaining skeletal muscle homeostasis. IPA analysis results show that the mTOR pathway is significantly activated. Within the map of mitochondrial dysfunction, it is observed that activation of mTOR inhibits the expression of Sirt1, which in turn suppresses the expression of PGC-1α and its downstream regulators NRF1 and NRF2. This exacerbates oxidative stress and leads to mitochondrial dysfunction. NRF2 is a key molecule in mitigating oxidative stress and maintaining mitochondrial function.

After analyzing multi-omics data, we suggest targeting oxidative stress and boosting energy metabolism pathways to delay muscle denervation atrophy, with NRF2 as a key therapeutic target. Nrf2 plays a vital role in maintaining cellular redox homeostasis by regulating the biosynthesis, utilization, and regeneration of the three primary antioxidants—glutathione, thioredoxin, and NADPH. Additionally, Nrf2 modulates reactive oxygen species production via mitochondrial pathways and NADPH oxidase [56]. Omaveloxolone is a small molecule compound that activates the NRF2 pathway, approved by the FDA in 2023 for clinical treatment of Friedreich’s Ataxia. Additionally, Omaveloxolone has also shown promising applications in neurodegenerative diseases and mitochondrial dysfunction-related diseases [57–59]. Omaveloxolone inhibits the binding of NRF2 to Keap1 (Kelch-like ECH-associated protein 1), leading accumulation of NRF2 in the nucleus. The translocated NRF2 protein interacts with the antioxidant response element (ARE) to stimulate the transcription of detoxification and antioxidant enzymes, ultimately contributing to antioxidant function [60]. Besides enhancing antioxidant activity, Omaveloxolone improves antioxidant activity and enhances mitochondrial function, biosynthesis, and substrate turnover [58]. This study conducted a preliminary investigation into the therapeutic efficacy of Omaveloxolone in mitigating denervation-induced muscle atrophy. Our results suggest that Omaveloxolone effectively delays muscle atrophy and diminishes muscle fibrosis. Immunohistochemical analysis demonstrated a notable increase in the expression levels of SOD2, Sdha, NRF2, and PKM proteins in the gastrocnemius muscle tissues treated with Omaveloxolone compared to the control control group. The results also indicate that Omaveloxolone can activate the NRF2 signaling pathway, leading to increased expression of the antioxidant enzyme Sod2 in muscle cells. Furthermore, the upregulation of key enzymes Sdha and PKM involved in the TCA cycle and glycolysis, respectively, suggests that Omaveloxolone treatment can simultaneously enhance glycolytic activity and energy metabolism in muscle tissues, ultimately improving energy supply. Results of TEM revealed that the mitochondrial morphology in the denervated gastrocnemius muscles of mice treated with Omaveloxolone exhibited greater completeness compared to the control group. Furthermore, through RNA-seq and GSEA analysis of the affected gastrocnemius muscle in the treatment and control groups, we found that the mitophagy pathway, glutathione metabolism pathway, and PPAR signaling pathway were significantly activated in the treatment group. In contrast, inflammation-related pathways and neurodegenerative disease-related pathways were significantly inhibited. MuSK is primarily distributed in a granular pattern on the surface of the sarcolemma in skeletal muscle. It maintains the stability of the NMJ by regulating the aggregation and stabilization of AchR and MuSK is an indispensable molecule in the formation and maintenance of the NMJ [61]. MuSK expression is significantly upregulated after denervation, particularly between the 2nd and 7th weeks, with a diffuse distribution on the sarcolemma [62]. In the sciatic nerve crush model, MuSK also shows a significant upregulation in expression, which gradually returns to normal followed by nerve regeneration and muscle reinnervation occur, and its distribution reverts to a granular pattern [63]. Therefore, effective intervention during denervation can reduce the high levels of MuSK expression. For example, Fujun Chen et al. demonstrated that L-type calcium channels (dihydropyridine receptors, DHPRs) play a crucial role in NMJ formation and maintenance. Blocking DHPRs leads to a significant upregulation of MuSK expression, while the application of DHPRs agonist Bay k8644 results in a marked decrease in MuSK expression [64]. In the Omaveloxolone group, the expression of GSR and the GSH/GSSH ratio were significantly higher than in the control group, while MuSK expression was significantly lower. This indicates that Omaveloxolone can not only enhance the antioxidant stress capacity of muscle tissue, reduce inflammatory responses, and maintain mitochondrial function, but it may also protect neuromuscular junctions, thereby staving off muscle atrophy.

Our study still have some limitations. In this study, we focused solely on the efficacy of intraperitoneal administration, whereas Omaveloxolone is primarily administered orally in clinical settings. Future research should explore the effects of various administration methods, dosages, and frequencies on delaying neurogenic muscle atrophy to better reflect real-world clinical scenarios. In this study, we conducted a preliminary investigation into the mechanisms by which Omaveloxolone mitigates denervation-induced muscle atrophy. In subsequent studies, the protective mechanism of Omaveloxolone on the neuromuscular junction, especially motor endplate, should be thoroughly investigated. Our research has demonstrated that NRF2 is an ideal target for the treatment of denervation-induced muscle atrophy. Future research should also focus on improving the muscle-targeting specificity of drugs to ensure that NRF2 can be effectively and precisely activated in denervated skeletal muscle.

Conclusion

In conclusion, our study demonstrated a dynamic molecular landscape of muscle atrophy caused by denervation by analyzing the transcriptome and proteome of the gastrocnemius muscle in a denervated mouse model. We emphasize the importance of oxidative stress, glucose and energy metabolism related genes in muscle denervation atrophy and mitochondrial dysfunction is the major source. NRF2 is a promising potential therapeutic target in the treatment of denervation-induced muscle atrophy. Omaveloxolone can maintain mitochondrial function stability by activating autophagy and PPAR signaling pathways, enhance antioxidant stress capacity of muscles by upregulating glutathione metabolism pathway, protect the neuromuscular junction by downregulating inflammation-related and neurodegenerative disease-related signaling pathways. Presently, there is a lack of research on the utilization of Omaveloxolone in delaying denervation-induced muscle atrophy. Thus, this investigation validates the therapeutic efficacy of Omaveloxolone in addressing muscle atrophy, offering a crucial empirical foundation and theoretical rationale for the advancement of novel treatment modalities.

Author contributions

Sulong Wang: Writing-original draft, Conceptualization; Formal analysis, Data curation; Xin Yang: Data curation, Investigation,Visualization, Writing-review&editing; Kai Liu: Investigation, Data curation; Debin Xiong: Investigation, Data curation; Ainizier Yalikun: Investigation, Formal analysis; Yimurang Hamiti: Investigation; Aihemaitijiang Yusufu: Funding acquisition, Conceptualization; Writing-review&editing, Supervision.

Funding

This work was supported by Key Project of the Natural Science Foundation of Xinjiang Uygur Autonomous Region (2021D01D20).

Availability of data and materials

Data will be made available on request.

Declarations

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper