Screening and verification of the lncRNA-miRNA-mRNA regulatory network in muscle atrophy after spinal cord injury

LIU Fu-chun, LI Xiao-lu, LIU Qing-Qing, GUI Yu-chang, ZHANG Yin-wei, XU Jian-wen?

1.Guangxi University of Traditional Chinese Medicine Graduate School, Nanning 530001, China

2.Department of Rehabilitation Medicine, First Affiliated Hospital of Guangxi Medical University, Nanning 530000, China

Keywords:

ABSTRACT Objective: To find the key targets of muscle atrophy after spinal cord injury (SCI)were excavated, to construct the lncRNA-miRNA-mRNA regulatory network based on bioinformatics analysis, and to verify the expression changes of key regulatory networks in muscle atrophy after SCI by animal experiments, so as to seek new research directions for the pathogenesis and treatment of muscle atrophy after SCI.Methods: The GSE21497 data set was downloaded from the GEO database for differential expression gene screening and WGCNA treatment.Combined with the online prediction database, key mRNAs were screened out.GO and KEGG enrichment analyses of key mRNAs were performed using the DAVID database to construct the lncRNA-miRNA-mRNA regulatory network.The key regulatory genes were selected and then verified by RT-qPCR.Results: A total of 1405 differentially expressed genes were screened, and 30 key mRNAs were predicted by the WGCNA and online database.GO and KEGG enrichment analyses showed that it was mainly enriched in the functions of neuron regeneration, protection, signal transmission, the HIF signaling pathway, PD-L1 expression and the PD-1 checkpoint pathway.Four key regulatory networks were identified (LINC00410/miR-17-5p/KCNK10, LINC00410/miR-17-5p/PCDHA3, LINC00410/miR-20b-5p/KCNK10,LINC00410/miR-20b-5p/PCDHA3).The results of RT-qPCR showed that, compared with the control group, the expression of miR-17-5p and miR-20b-5p in the observation group increased, and the expression of KCNK10 and PCDHA3 decreased.Conclusions: MiR-17-5p, miR-20b-5p, KCNK10, and PCDHA3 may play an important regulatory role in the regeneration, protection, and signal transmission of neurons, which is expected to become a new target for the diagnosis and treatment of muscle atrophy after SCI.

1.Introduction

Spinal cord injury (SCI)) refers to a neurological disease in which the spinal cord is subjected to direct or indirect violence,causing motor and sensory dysfunction in the injured plane and the following innervation area.Even pose a huge threat to the lives of patients, there is still a lack of effective treatment.According to a Meta-analysis, there are about 225,500 people per 1 million people.The incidence of SCI in men is much higher than that in women, especially in young and middle-aged people, which leads to the loss of patient labor.It not only brings huge treatment costs and a serious physiological burden to patients’ families but also causes huge economic losses to society[1,2].After SCI, skeletal muscle loses its innervation and nutrition, and the muscle is disused for a long time, resulting in muscle atrophy, which has a great impact on the rehabilitation and quality of life of patients.Although there is evidence that electrical stimulation[3,4], exercise training[5,6], and other means have a certain therapeutic effect, the effect is still unsatisfactory, and the mechanism of muscle atrophy after SCI is not yet clear, which is a major problem to be solved in the rehabilitation of SCI patients.In recent years, bioinformatics has been widely used to mine valuable genes in high-throughput sequencing data, which can provide certain theoretical bases and research directions for researchers.Based on this, it is expected to save a lot of research time and energy by selecting valuable genes with purpose and pertinence[7,8].GSE21497 is a data set for studying the gene expression profile of skeletal muscle in patients on the 5th day after SCI compared with the 2nd day after SCI.In this study, the data set was downloaded from the GEO database, and the lncRNAs and mRNAs in the data set were analyzed for differential expression.The lncRNA-miRNA-mRNA regulatory network was constructed, and the key genes in the network were experimentally verified, aiming to provide a theoretical basis and target selection for the diagnosis and treatment of muscle atrophy after SCI.

2.Materials and Methods

2.1 Gene chip

GSE21497 is derived from the GEO (gene expression omnibus,https:// www.ncbi.nlm.nih.gov/ geo/) database, whose gene sequencing platform is GPL570 [HG-U133 _ Plus _ 2 ] Affymetrix Human Genome U133 Plus 2.0 Array.The database uses a selfcontrol method to sample the vastus lateralis muscle biopsy samples of 10 patients (9 males and 1 female, with an average age of 44 years, including 6 quadriplegia and 4 paraplegia) with SCI on the 2nd and 5th day after SCI, respectively.

2.2 Data processing and differential expression gene screening

R software (R.4.2.0) was used to process the data, and the exprs function was used to obtain the expression matrix.The Affy package was used to perform quality control, background correction, and standardized preprocessing on the original data.According to the annotation information on the platform, the probe was annotated with gene markers.The genes that hybridized to the same probe were standardized to eliminate redundant probes.The Limma package was used to screen differentially expressed mRNAs (Differentially Expressed mRNAs, DEmRNAs) and differentially expressed lncRNAs (DElncRNAs).The filtering condition was P＜0.05 and |logFC |＞0.585.Finally, the volcano and heat maps of DEmRNAs and DElncRNAs were drawn using the heatmap package.

2.3 Weighted correlation network analysis (WGCNA)

WGCNA was performed on DEmRNAs to analyze the Pearson correlation between each pair of genes and generate a relationship matrix.According to the scale-free topology fitting index (R2 )value and the average connectivity, the optimal soft threshold (β) is determined, and the WGCNA is constructed.Then, the following methods are used to determine the key modules related to disease progression.Firstly, similar genes are merged according to the high topological overlap between genes to construct multiple gene modules, and gene modules are clustered according to the coexpression between modules to merge modules with high similarity.Then, the correlation between the feature vector genes of each module and disease progression was analyzed (The absolute mean of the correlation between gene expression and disease progression in each module is considered the correlation between the module and disease progression).

2.4 Construction of lncRNA-miRNA-mRNA regulatory network

Firstly, the miRcode database was used to predict miRNAs interacting with DElncRNAs to obtain lncRNA-miRNA interactions.Then, the miRWalk, Targetscan, and miRDB databases were used to predict the target genes of miRNAs (inding probability: 0.95; binding site location: 3’ UTR).Then, the predicted mRNA was matched with the genes in the key co-expression module to obtain the key mRNA,and the miRNA-mRNA interaction was determined.Finally, the lncRNA-miRNA-mRNA regulatory network was constructed using Cytoscape 3.8 software, and the results were graphically displayed.

2.5 Functional enrichment analysis

Based on the above analysis, the key mRNA was identified as a candidate gene for functional and pathway enrichment analyses.The DAVID online database (https://david.ncifcrf.gov/) was used for gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes(KEGG) signaling pathway enrichment analysis.GO annotated the function of differential genes from three aspects: biological process(BP), cellular component (CC) and molecular function (MF).

2.6 Experimental verification of key regulatory networks

2.61 Screening of key regulatory networks

From the lncRNA-miRNA-mRNA regulatory network constructed in 1.4, the key regulatory networks were selected for experimental verification by consulting the literature.

2.62 Experimental animals

8 healthy SPF adult SD rats, female, weighing 200 ± 20 g, were purchased from the Experimental Animal Center of Guangxi Medical University (production license: SYXK (Gui) 2020-0004),raised at 25 ℃, light/dark: 12/12 h.

2.63 Animal model preparation

The random number method was used to group eight animals:four in the observation group and four in the control group.The experimental animals were anesthetized with pentobarbital sodium(concentration: 3 %, 30 mg/kg), and the SCI model was prepared by the classical Allen’s method.(Strike rod: 10 g; strike rod diameter:2.5 mm; strike height: 5 cm).Penicillin was injected intraperitoneally for 3 days to prevent infection, and manual massage of the abdomen of the rat was performed every day to assist its urination and defecation.The control group only exposed the spinal cord without striking.The experimental protocol has been approved by the Animal Experimental Ethics Committee of Guangxi Medical University.

2.64 Sampling and RNA extraction

7 days after modeling, excessive anesthetics were given to the rats until they were euthanized.After that, normal saline was immediately perfused into the heart, and the biceps femoris of the rats was sampled on ice.After the completion of the sampling, the PBS buffer was used to rinse three times, the residual liquid on the surface was extracted with filter paper and placed in a clean grinding tube, and the muscle tissue was cut as much as possible with sterile scissors.Add grinding magnetic beads and NucleoZol RNA reagent(Macherey-Nagel, Germany) for full grinding, and then complete the total RNA extraction according to the NucleoZol RNA reagent instructions.

2.65 Real-time quantitative PCR detection

The total RNA was reverse transcribed into cDNA (Takara,RR047A, and RR036A, Japan), and then cDNA amplification was performed using a 7 500 fluorescence quantitative PCR instrument.The reference gene used for miRNA was U6, and the mRNA was GAPDH.The gene primer sequence used is shown in Table 1.The fluorescence quantitative kit was TB Green Premix Ex Taq Ⅱ(RR820A, Japan).The relative expression of cDNA was calculated by the 2-ΔΔCtmethod.

Tab 1 Primer sequence table

2.7 Statistical analysis

This study will use SPSS Statistics 17 software for statistical analysis.The statistical method is the independent sample t test.When t ＜ 0.05, the difference is considered statistically significant.

3.Results

3.1 Differentially expressed genes

After P＜0.05, and |logFC|＞0.585 screening, 1405 differentially expressed genes on the 5 th day after SCI were obtained.Among them, there were 262 lncRNAs and 1143 mRNAs.The screening results were shown in volcanogram and heat map, as shown in Fig.1 a-d.

3.2 WGCNA

Fig 1 Differentially expressed genes, volcano plot, and heat map of SCI muscle atrophy

WGCNA was used to analyze DEmRNA.First, the soft threshold(β) was screened, and there was a good connection when β was 7 ( Fig.2).Then, the genes were modularly analyzed, and the hierarchical clustering tree and co-expression module of the WGCNA network were constructed.Finally, four main gene modules (turquoise, blue, brown, and grey) were obtained, among which the turquoise module-characteristic correlation coefficient score was up to 0.89 (Fig.3).

Fig 2 Screening of a soft threshold (β) of a scale-free network

3.3 Construction of the lncRNA-miRNA-mRNA network

A total of 211 miRNAs were predicted according to the miRcode database, and 4922 mRNAs were predicted by the miRWalk,Targetscan, and miRDB databases.The predicted mRNAs were then intersected with the genes in the turquoise module to obtain a total of 30 key mRNAs (Fig.4).According to the lncRNAmiRNA interaction pair and the miRNA-mRNA interaction pair,the lncRNA-miRNA-mRNA regulatory network was constructed.The final results were visualized by Cytoscape 3.8 (Fig.5).Through literature review, four regulatory networks (LINC00410/miR-17-5p/KCNK10, LINC00410/miR-17-5p/PCDHA3, LINC00410/miR-20b-5p/KCNK10, and LINC00410/miR-20b-5p/PCDHA3) were selected for experimental verification (four genes of miR-17-5p, 3miR-20b-5p, KCNK10, and PCDHA were verified).

3.4 Functional enrichment analysis

Fig 3 Hierarchical clustering tree and co-expression module of the WGCNA network

The 30 intersection mRNAs obtained in 2.3 were analyzed for GO and KEGG signaling pathway enrichment using the DAVID online database.The top ten results of BP, CC, MF, and KEGG pathway enrichment were plotted and displayed, respectively (Fig.6).

Fig 4 Key mRNA

Fig 5 lncRNA-miRNA-mRNA regulatory network

3.5 RT-qPCR results

Compared with the control group, the expression of miR-17-5p (t=0.002038) and miR-20b-5p (t=0.000535) in the biceps femoris muscle tissue of the observation group increased, while the expression of KCNK10 mRNA (t=2.0743E-7) and PCDHA3 mRNA(t=2.1378E-8) decreased (Fig.7).

Fig 6 GO, KEGG enrichment analysis

Fig 7 RT-qPCR results of key genes

4.Discussions

As a serious neurological disease, SCI seriously endangers the lives of patients, and a series of symptoms after SCI also seriously affect the quality of life of patients, but no effective treatment has been found.After SCI, due to the nerve injury, the muscles cannot be controlled and nourished, resulting in muscle atrophy.Muscle atrophy further aggravates the damage to motor function, resulting in the nerves and muscles in the region being unable to receive timely and effective nutrition, forming a vicious circle and leading to further aggravation of muscle atrophy.Therefore, it is of great significance to explore the pathological mechanism and treatment of muscle atrophy after SCI.Severe SCI can lead to impaired nerve signal transduction pathways, atrophy of motor neurons, and pathological changes in neuromuscular junctions, resulting in decreased muscle fiber formation, muscle atrophy, and motor dysfunction[9, 10].Compared with most disuse muscle atrophy, the signal and pattern of muscle atrophy after SCI may be more serious, and the speed of muscle atrophy may be faster.Although atrophy signals lead to muscle fiber loss after severe SCI, the reduction of anabolic signals may be one of the causes of persistent muscle loss[11].The downstream protein of the IGF-1/PI3K/ Akt signaling pathway,mammalian target of rapamycin (mTOR), is a key protein involved in anabolic metabolism.The phosphorylation of mTOR will target and start its downstream proteins involved in translation, thereby increasing protein synthesis.If the pathway can be continuously activated, it is possible to reverse muscle atrophy[12].Studies have shown that the content of muscle fibers and the expression of phosphorylated PI3K, AKT, and mTOR decreased after SCI, which may be one of the reasons for the long-term decrease in protein synthesis and muscle atrophy in SCI patients[13-15].It has been reported that SCI trauma or subsequent surgical intervention can lead to a more severe inflammatory response and aggravate muscle atrophy, while the use of high-dose glucocorticoids can inhibit the inflammatory response of secondary injury after SCI in the acute and subacute phases[16, 17].On the contrary, studies have shown that high doses of methylprednisolone increase the expression of FOXO1, MAFbx, MuRF1, and Redd1, and the increased expression of these proteins will inhibit the phosphorylation of mTOR, reduce protein synthesis, resulting in muscle fiber loss[18].Although there have been many reports on muscle atrophy after SCI, the signal patterns of muscle atrophy and anabolism are still unclear, and there is no effective treatment.This study investigates the pathological mechanism of muscle atrophy after SCI at the molecular level and provides a theoretical foundation for further research into more effective treatment methods.

In this study, bioinformatics analysis was used to analyze the differences in gene expression profiles of vastus lateralis muscle biopsy on the 2nd and 5th day after SCI, and the lncRNA-miRNAmRNA regulatory network was constructed to screen out key regulatory genes.RT-qPCR experiments were performed in rats to provide new targets and directions for the treatment of SCI muscle atrophy.

GO enrichment analysis showed that key mRNAs were mainly involved in the regulation of cell components such as dendrites,neuronal cell bodies, neuronal perinuclear bodies, cell membranes,nuclear membranes, and plasma membranes, indicating that they were closely related to neuronal remodeling and signal transmission after SCI.At the molecular function level, it is mainly involved in regulating the activity of protein kinase, ligase, transmembrane transporter, and receptor and participating in the binding of actin,suggesting that it may be closely related to the synthesis and signal transduction of muscle cells.At the level of biological processes, it is mainly involved in cell proliferation, angiogenesis, amino acid transport, positive regulation of nitric oxide synthase synthesis,and positive regulation of PI3K signaling, suggesting that it may be closely related to cell proliferation, promotion of autophagy,and inhibition of apoptosis.In summary, GO enrichment analysis suggested that key mRNAs may affect muscle atrophy after SCI mainly by affecting neuronal generation and signal transduction.

KEGG pathway enrichment analysis showed that key genes were mainly involved in the regulation of the hypoxia-inducible factor(HIF-1) signaling pathway, programmed death ligand 1 (PD-L1)expression, and programmed death protein 1 (PD-1) checkpoint pathway in tumors.HIF-1 is a transcription factor composed of HIF-1α and HIF-1β subunits.Under sufficient O2conditions, proline residue 402 and/or proline residue-564 use O2and α-ketoglutarate as substrates to hydroxylate HIF-1α through proline hydroxylase domain protein 2, resulting in E3 ubiquitination and proteasome subunit degradation.In contrast, under hypoxic conditions, HIF-1α binds to p300/CBP and acts as a major regulator of many hypoxiainducible genes.Studies have shown that ferroptosis plays an important role in SCI, and five genes (Stat3, Tlr4, Hmox1, Hif1α ,and Cybb) in the HIF-1 signaling pathway play an important regulatory role in ferroptosis[19].Protein deacetylation plays an important role in the secondary injury of SCI, which will further aggravate it.Triacetin may promote protein acetylation through the HIF-1 signaling pathway, inhibit neuronal inflammation, and induce apoptosis[20].Similarly, other studies have also shown that the HIF-1 signaling pathway is of great significance in secondary injury after SCI[21, 22], suggesting that this pathway may become a new target for the treatment of SCI and muscle atrophy.PD-1 is an immunosuppressive molecule.PD-1 and its ligand PD-L1 can maintain the self-tolerance of the organism to the immune system by regulating the response of T cells and B cells to human cells and inhibiting the inflammatory response of T cells.Studies have shown that PD-1 and PD-L1 are significantly expressed on microglia after SCI and are important in the inflammatory process.Knockdown of PD-1 expression will aggravate the inflammatory response after SCI[23, 24].Dexmedetomidine can reduce neuroinflammation by up-regulating PD-1 expression[25], suggesting that the pathway may affect the process of SCI mainly by regulating neuronal inflammation.In summary, some key mRNAs may be involved in the occurrence of muscle atrophy after SCI by affecting the inflammatory response of neuronal cells through the HIF-1 signaling pathway, PD-L1 expression, and PD-1 checkpoint pathway.These mechanisms provide new directions and goals for the treatment of muscle atrophy after SCI.

In this study, the DEmRNA, DElnRNA, and miRcode, miRWalk,Targetscan, and miRDB databases were used to predict the lncRNAmiRNA-mRNA regulatory network.After further literature review,four key regulatory networks were finally screened for experimental verification.The results showed that, compared with the control group, the expression levels of miR-17-5p and miR-20b-5p in the observation group increased, and the expression levels of KCNK10 and PCDHA3 decreased, which was consistent with the results of the bioinformatics prediction.Previous studies have shown that the expression level of miR-17-5p is up-regulated after SCI, and the synthesis of miR-17-5p mimics can rescue the proliferation defects of Dicer1-null astrocytes, while the inhibitor of miR-17-5p can block the proliferation of astrocytes induced by lipopolysaccharide, and the proliferation of reactive astrocytes may be regulated by the JAK/STAT3 pathway[26].Mesenchymal stem cells (MSCs) can alleviate SCI, which is one of the research hotspots in SCI treatment.Yue Xiao-Hua et al[27].found that VEGF-A mRNA can specifically bind to miR-17-5p.After knocking down VEGF-A, the intact spinal cord tissue of SCI mice decreased, while knocking down miR-17-5p showed the opposite results, indicating that inhibiting the expression of miR-17-5p can up-regulate the expression of VEGF-A in MSCs and promote the repair of spinal cord injury by MSCs.Zhang Liang et al[28].found that down-regulation of miR-17-5p expression promoted the growth of neuronal axons, and the STAT3 phosphorylation inhibitor AG490 reversed this promotion.It was verified that miR-17-5p can regulate the growth of cortical neuronal axons by targeting the STAT3/GAP-43 pathway, indicating that miR-17-5p plays a key regulatory role in the growth of neuronal axons.In this study, the expression of miR-17-5p in the observation group was up-regulated, which was consistent with other scholars’reports, suggesting that muscle atrophy after SCI may be related to neurological dysfunction and abnormal axonal growth.Wang Yi-Min et al[29].considered that miR-20b-5p may be a potential biomarker for SCI patients after bioinformatics analysis.Studies have shown that miR-20b-5p can interfere with calcium homeostasis and axon growth by binding to amyloid β precursor protein (APP), which is of great significance in neurological diseases[30].Similarly, another study showed that inhibition of miR-20b-5p expression can target RhoC, attenuate Aβ25-35-induced apoptosis in PC12 cells, and play an important role in the regulation of neuronal apoptosis[31].In this study, the expression of miR-20b-5p was up-regulated, which was consistent with related research reports, suggesting that muscle atrophy after SIC may be related to the apoptosis of nerve cells.KCNK10 is a member of the series pore domain potassium channel family and plays an important role in stabilizing the negative resting membrane potential and balancing depolarization.Sierk Haenisch et al[32].found that KCNK10 is the target gene of miRNA-187-3p, and miRNA-187-3p can negatively regulate the expression of KCNK10.The up-regulation of KCNK10 expression has a protective effect and can have a beneficial effect on the homeostasis of potassium and glutamate in astrocytes; Zhao Guang-Chao et al[33].also found that the neuroprotective effect of isoflurane preconditioning on ischemia-reperfusion injury was related to the increase in KCNK10 channel activation.In this experiment, the expression of KCNK10 was down-regulated, which may be due to insufficient activation of the KCNK10 channel after nerve injury, resulting in a nerve injury that cannot be controlled in time, thereby aggravating the degree of disease.PCDHA3 is a member of the protocadherinα gene cluster.It is an unusual genomic structure similar to the B-cell and T-cell receptor gene clusters.It is likely to play a key role in the establishment and function of specific cell-cell connections and neuronal cell adhesion in the brain[34, 35].In this experiment,the expression of PCDHA3 was down-regulated, suggesting that muscle atrophy may be related to the dysfunction of intercellular junction construction.In summary, the research on the regulation and expression of miR-17-5p, miR-20b-5p, KCNK10, PCDHA3,etc., may be one of the directions worthy of further discussion in the mechanism of muscle atrophy after SCI.

In summary, significant differential expression of miR-17-5p, miR-20b-5p, KCNK10, and PCDHA3 was found in muscle tissue after SCI, suggesting that these genes may be the key genes of muscle atrophy after SCI, which is expected to provide new research targets and directions for the diagnosis, treatment, and rehabilitation of muscle atrophy after SCI.However, there are still some limitations to this study.First of all, due to the relatively small amount of data on muscle atrophy after SCI in the GEO database and the limited sample size, the research results may have some bias; secondly, the constructed lncRNA-miRNA-mRNA regulatory network of muscle atrophy after SCI lacks effective evidence to prove the regulatory relationship between regulatory networks.Further experiments such as dual luciferase and pull-down are needed to further verify the biological mechanism of the lncRNA-miRNA-mRNA regulatory model in vitro and in vivo.Furthermore, because the lncRNAs in the lncRNA-miRNA-mRNA regulatory network are all humanderived, no corresponding gene sequence was found on the mouse source, and it was not possible to verify whether the expression changes were consistent with bioinformatics predictions.However,in general, this experiment provides a certain reference direction for further exploring the mechanism and therapeutic targets of muscle atrophy after SCI.

Author’s contribution and conflict of interest

LIUFU Chun participated in thesis writing, experimental design,and data analysis; XU Jian-Wen participated in guiding and revising the thesis; LI Xiao-Lu and LIU Qing-Qing participated in some data collation and chart production; ZHANG Yin-Wei and Gui Yu-Chang participated in animal modeling and sampling.All the above authors have no conflict of interest.