Identification of potential microRNAs and KEGG pathways in denervation muscle atrophy based on meta-analysis

The molecular mechanism of muscle atrophy has been studied a lot, but there is no comprehensive analysis focusing on the denervated muscle atrophy. The gene network that controls the development of denervated muscle atrophy needs further elucidation. We examined differentially expressed genes (DEGs) from five denervated muscle atrophy microarray datasets and predicted microRNAs that target these DEGs. We also included the differentially expressed microRNAs datasets of denervated muscle atrophy in previous studies as background information to identify potential key microRNAs. Finally, we compared denervated muscle atrophy with disuse muscle atrophy caused by other reasons, and obtained the Den-genes which only differentially expressed in denervated muscle atrophy. In this meta-analysis, we obtained 429 up-regulated genes, 525 down-regulated genes and a batch of key microRNAs in denervated muscle atrophy. We found eight important microRNA-mRNA interactions (miR-1/Jun, miR-1/Vegfa, miR-497/Vegfa, miR-23a/Vegfa, miR-206/Vegfa, miR-497/Suclg1, miR-27a/Suclg1, miR-27a/Mapk14). The top five KEGG pathways enriched by Den-genes are Insulin signaling pathway, T cell receptor signaling pathway, MAPK signaling pathway, Toll-like receptor signaling pathway and B cell receptor signaling pathway. Our research has delineated the RNA regulatory network of denervated muscle atrophy, and uncovered the specific genes and terms in denervated muscle atrophy.

Specific genes in denervated muscle atrophy. By retrieving, we got three data sets of disuse muscle atrophy caused by unloading or casting (Table 6), genes only differentially expressed in denervated muscle atrophy were regarded as denervation specific genes, called den-genes. We got 187 up-regulated den-genes and 180 down-regulated den-genes (Fig. 5), then performed KEGG enrichment analysis on these genes and got top10 most enriched pathways (Fig. 6). We searched and found the expression of neurotrophic factors (NGF, BDNF, NT-3, NT-4, CNTF, Neuregulin-1 and Neuritin) changed in nerves or muscles after nerve injury and there was evidence that Insulin signaling pathway, MAPK signaling pathway, Neurotrophin signaling pathway, T cell receptor signaling pathway, Toll-like receptor signaling pathway, and B cell receptor signaling pathway were regulated by the above neurotrophic factors (Table 7). Furthermore, most of these terms in skeletal muscle showed different expression after electrical stimulation and Insulin signaling pathway, MAPK signaling pathway, Toll-like receptor signaling pathway, Neurotrophin signaling pathway, and VEGF signaling pathway show regulatory effects on ion channels and connexins (Table 8).

Discussion
In this study, we collected and aggregated gene expression data sets of denervated muscle atrophy, and obtained 429 up-regulated mRNAs and 525 down-regulated mRNAs. We drew a PPI network, and sorted the nodes according to the interaction degree. The top10 proteins were Gapdh, Mapk14, Jun, Cat, Casp3, Vegfa, Decr1, Cs and Suclg1. Mapk14, Jun, and Casp3 have been confirmed to be important in denervated muscle atrophy. (1) Mapk14: the three main categories of mitogen-activated protein kinase (MAPK) family proteins are as follows: extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK) and p38 MAPK (Mapk14). The members of the p38 MAPK family (p38α, p38β, p38γ and p38δ MAPK) act as transducers of cellular stress and various non-stress-related stimuli. Therefore, the p38 MAPK pathway has multiple functions and is involved in various cellular processes, including aging, apoptosis, cell cycle arrest, inflammation, and tumorigenesis 63 . The p38 MAPK pathway can mediate the expression of MuRF1 and Atrogin1 64,65 , and knockdown of p38αMAPK can inhibit muscle atrophy caused by denervation 66 . (2) Jun: the protein deacetylase HDAC4 is strongly induced in muscles affected by motor neuron diseases such as ALS 19 . Denervation-induced HDAC4 activates AP1 (Jun) transcription factor by stimulating MAPK signaling to promote denervated muscle atrophy 67 . (3) Casp3: in the ubiquitin-proteasome pathway and apoptosis process, the activation of Caspase-3 (Casp3) is a common phenomenon. Caspase-3 regulates denervation-induced signal transduction through the mitochondrial-related cell death/apoptosis pathway, resulting in the loss of muscle mass. The lack of caspase-3 has a protective effect on denervated muscle atrophy 68 . The other proteins Gapdh, Cat, Vegfa, Decr1, Cs, and Suclg1 are mostly important links in redox and energy metabolism pathways [69][70][71][72][73] , and they may be potential targets in denervated muscle atrophy for treatment.   www.nature.com/scientificreports/ We included the differentially expressed miRNA expression data of denervated muscle atrophy in previous studies as background information 31,32 , identified the potential key microRNAs and eight microRNA-mRNA interactions of which the mRNAs are the top 10 denervation-related mRNAs identified in this research (miR-1/ Jun, miR-1/Vegfa, miR-497/Vegfa, miR-23a/Vegfa, miR-206/Vegfa, miR-497/Suclg1, miR-27a/Suclg1, miR-27a/ Mapk14). Most of the Key miRNAs we obtained showed important regulatory effects on skeletal muscle (Table 7). After 7-day hind limbs unloading in mice, the serum levels of muscle-specific miRNAs such as miR-1, miR-23a, miR-206 increased significantly, and could induce severe muscle atrophy 74 . MiR-1 has been confirmed to play an important role in skeletal muscle development 75,76 . MiR-23a can target Atrogin-1 and MuRF1 and inhibit their translation, and the ectopic expression of miR-23a can protect muscles from atrophy in vitro and in vivo, indicating that miR-23a is a key regulator of muscle atrophy 77 . Overexpression of miR-27a in mice with chronic kidney disease attenuated muscle loss, improved grip strength and reduced the expression of FoxO1, MuRF1and Atrogin1 40 .
Denervation, long-term bed rest, unloading of hind limbs, immobilization or microgravity can cause disuse muscle atrophy. Current research focuses on the general signaling pathways and common key regulators that mediate disuse muscle atrophy [26][27][28] , but obviously, denervation is also accompanied by the loss of neurotrophic factors and a series of changes brought about by nerve injury. These denervation-specific pathological changes might be accompanied by specific transcriptome changes which remain to be discovered.
The nervous system controls skeletal muscle through two mechanisms: (1) neuromotor control, which causes muscle contraction through excitation and contraction coupling. (2) Neurotrophic control, which regulates muscles by releasing soluble factors from the nerve endings of motor neurons on the NMJ (Neuromuscular junction) 78 . NGF, BDNF, NT-3, NT-4, CNTF, Neuregulin-1 and Neuritin are neurotrophic factors differentially expressed in nerves or skeletal muscle after nerve injury 79 , and they were found to regulate Insulin signaling pathway, MAPK Signaling pathway, Neurotrophin signaling pathway, T cell receptor signaling pathway, Toll-like receptor signaling pathway and B cell receptor signaling pathway which enriched by den-genes ( Table 8). The specific changes in skeletal muscle after nerve transection include increased membrane permeability, decreased membrane potential and increased membrane excitability. Most of these are caused by changes in the expression of ion channels and the insertion of connexins 39, 43, and 45 into the muscle membrane, which mediate skeletal muscle atrophy 78,80 . Insulin signaling pathway, MAPK signaling pathway, Toll-like receptor signaling pathway, Neurotrophin signaling pathway and VEGF signaling pathway showed the potential to regulate ion channels and connexins. Insulin signaling pathway and Toll-like receptor signaling pathway can affect the formation of neuromuscular junctions (Table 9). Therefore, Insulin signaling pathway and Toll-like receptor signaling pathway, as metabolic or inflammation-related terms, have been seldom studied in the neuromuscular systerm, but this study found that they may have important research value in the denervated muscle atrophy.
Current research focuses on the general signaling pathways and common key regulators that mediate disuse muscle atrophy, but obviously, denervation is also accompanied by the loss of neurotrophic factors and a series of changes brought about by nerve injury. These denervation-specific pathological changes must be accompanied by specific transcriptome changes which remain to be discovered. We systematically searched and integrated the gene expression data of denervated muscle atrophy to construct the microRNA-mRNA regulatory network. By comparing denervated muscle atrophy with disused muscle atrophy caused by other reasons, we have discovered the genes and terms that are specifically differentially expressed in denervated muscle atrophy.

Methods
Inclusion criteria for gene expression data. We performed a meta-analysis following the PRISMA Statement 97 , and found the gene expression data of denervated muscle atrophy by searching NCBI-GEO (http:// www. ncbi. nlm. nih. gov/ geo/) and PubMed (http:// www. ncbi. nlm. nih. gov/ pubmed). The keywords used were: "Denervated atrophy", "Denervation AND muscle", "nerve AND muscle", and "Denervation AND muscle AND sequencing". These meta-analysis searches comprised studies published between 2010 and 2020. Our inclusion criteria were (1) gene expression data in skeletal muscle with peripheral nerve injury, (2) the number of samples in each group should be greater than two, (3) the duration of denervation was 7 to 14 days, (4) all types of skeletal muscle were considered, (5) the inclusion of normal tissues for comparison, and (6) all gene expression analysis platforms were considered. Our exclusion criteria were (1) non-muscle samples, (2) non-mRNA datasets, and (3) review studies. And we also searched NCBI-GEO (http:// www. ncbi. nlm. nih. gov/ geo/) and PubMed (http:// www. ncbi. nlm. nih. gov/ pubmed) to find the gene expression data of disuse muscle atrophy. The keywords used were: "Disuse atrophy", "Disuse AND muscle", "unloading AND muscle", and "cast AND muscle". These meta-analysis searches comprised studies published between 2010 and 2020. Our inclusion criteria were (1) gene expression data in skeletal muscle with disuse atrophy, (2) the number of samples in each group should be greater than two, (3) the In this study, each individual dataset was processed using edgeR packages to compare control samples and experiment samples to screen DEGs. The Benjamini & Hochberg false discovery rate method was used for p-value adjustment, and the default adj-p-value significance level cut-off is 0.05.

Meta-analysis of global gene expression data in denervated muscle atrophy. Meta-analysis was
performed on the five datasets of denervated muscle atrophy using Vote counting generic ways of combining information 98 , and the DEGs that were common between all datasets were identified. The results were visualized using Venn diagrams by a graphing software (ORIGIN2019; OriginLab, Northampton, MA).

GO and KEGG enrichment analysis of DEGs.
The DEGs were used to identify over-represented gene ontology categories and KEGG pathways by using DAVID 6.7 (https:// david-d. ncifc rf. gov/). The p value has been corrected using FDR (false discovery rate). GO/KEGG enrichment with p < 0.05 was regarded as statistically significant and those with number of genes < 10 were removed. Bar chart of GO terms was drawn by Microsoft Excel 2016. KEGG network was visualized by Cytoscape software v. 3.7.2 54.

Protein-protein interaction (PPI) network construction for common DEGs.
The DEGs were subjected to STRING v.11.0 database 99 analysis to construct PPI networks. Minimum required interaction score was 0.4 (medium confidence). PPI networks were visualized by Cytoscape software v. 3.7.2 100 and the degree which indicates the number of interactions of each node was calculated by cytoHubba 101 . Identification of microRNAs as potential modulators of DEGs. The DEGs identified in our metaanalysis were used for microRNA prediction by DIANA-Tarbase 102 to identify potential regulators in denervated muscle atrophy. Visualization of microRNA-gene interaction networks were generated using Cytoscape v3.7.2 100 and the degree which indicates the number of interactions of each node was calculated by cytoHubba 101 . www.nature.com/scientificreports/ www.nature.com/scientificreports/      www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.