Comprehensive Analysis of a lncRNA-miRNA-mRNA Competing Endogenous RNA Network in Heart Failure

Background: Acute heart failure caused by progressive heart failure is a common disease in intensive care units (ICU). The growing incidence rate of heart failure and its high mortality rate result are very important sociosanitary problems. Therefore, it is important to identify the molecular mechanism by which heart failure occurs and to identify treatment for this mechanism. Recently, the mechanism of ceRNA has attracted increasing attention. The aim of the present study was to identify the candidate ceRNA network in the progression of heartfailure. Method: Microarray datasets GSE9128, GSE61741 and GSE77399 were downloaded from Gene Expression Omnibus (GEO) database. The differentially expressed genes (DEGs) were identied, and function enrichment analyses were performed. The protein-protein interaction network (PPI) was constructed and identication of hub-genes was performed using STRING and Cytoscape. Furthermore, according to the ceRNA theory, network of ceRNA was constructed. Result: In the present study, based on the ceRNA theory and above series of analyses, network of ceRNA which include 7 mRNAs (BCL2A1, DUSP1, EGR1, MYC, NR4A2, PTGS2 and RAC2), 3 miRNAs (miR-20a, miR-129-59 and miR-185-5p) and 3 lncRNAs (GAS5, H19 and PCGEM1) were obtained. Conclusion: In conclusion, these ndings can be used to carrying on further study to identify the important roles of the ceRNA, biological function, appropriate treatment targets and biomarkers in the progression of heart failure. DUSP1, EGR1, MYC, NR4A2, PTGS2 and RAC2), 3 miRNAs (miR-20a, miR-129-59 and miR-185-5p) and 3 lncRNAs (GAS5, H19 and PCGEM1) were obtained. These ndings can be used to carrying on further study to identify the biological function, appropriate treatment targets and biomarkers in the progression of heart failure.


Introduction
Acute decompensated heart failure is a growing major public health problem all through the world [1]. A variety of diseases including coronary artery disease, hypertension, cardiac valves disease and dilated cardiomyopathy contribute to the progression and occurrence of heart failure [2]. With the development of these diseases, cardiac remodeling and regeneration cause the development of heart failure [3]. Moreover, ventilator support, continuous renal replacement treatment and positive vasoactive agents are needed when the symptoms of acute heart failure occur in intensive care units. This is a huge economic burden on the health of the population. Some previous studies have showed that the mechanism of cardiac remodeling and regeneration [4]. However, some further investigations are needed to understand the molecular mechanism of heart failure.
LncRNA participated in the regulation of many physiological and pathological processes through a variety of mechanisms [5]. LncRNA can not only directly regulate the expressions of target genes, but also affect the expression of miRNA by binding miRNA, further affecting the expression level of target gene mRNA of miRNA [6]. Competing endogenous RNA (ceRNA) is widely involved in the regulation process of vital activities such as in ammation, apoptosis and differentiation. This regulatory mechanism also occurs in many cardiovascular diseases [7].
In this study, GSE9128, GSE61741 and GSE77399 were used form the Gene Expression Omnibus database to identify expressed mRNAs, miRNAs and lncRNAs to construct a ceRNA network. Therefore, the purpose of this study was to screen out differentially expressed lncRNA, miRNA and mRNA in the circulation of patients with heart failure, and to construct ceRNA network, so as to provide bases for nding molecular markers and therapeutic targets related to acute heart failure.

Materials And Methods
Data collection and identi cation of differentially expressed mRNAs, miRNAs and lncRNAs. The database of GEO (http://www.ncbi.nlm.nih.gov/geo) is a public functional genomics data repository of high throughout gene expression data, chips and microarrays [8]. In GEO, we found three datasets including GSE9128 [9], GSE61741 [10] and GSE77399 [11] that met our retrieval requirement for study. The differentially expressed mRNAs, miRNAs and lncRNAs between healthy controls and heart failure were identi ed by an excellent web tool named GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r). The adjusted P-values (adj. P) and Benjamini and Hochberg false discovery rate were applied to provide a balance between discovery of statistically signi cant genes. LogFC (fold change) >1 and adj. P-value <0.05 were considered statistically signi cant.
Functional and pathway enrichment analysis of DEGs. The biological function and signaling pathways analysis of differential genes obtained by the above methods were analyzed by the web database of DAVID. The web database of DAVID (http://david.ncifcrf.gov) is a common tool which is used to GO annotation and KEGG pathways enrichment of DEGs [12]. It can provide a comprehensive set of functional annotation information of genes and proteins. GO annotation is a main bioinformatics tool to annotate genes and analyze biological process of DEGS [13]. P<0.05 was considered statistically signi cant.
Identi cation of hub-genes and PPI network construction. In order to identify hub-genes in these DEGs, the database of STRING (Search Tool for the Retrieval of Interacting Genes http://string-db.org) was used to obtain the predicted interactions (version 10.0) [14]. The software of Cytoscape (version 3.6.1) is an open source bioinformatics software platform that can display visualization of PPI (protein-protein interaction) network [15]. There is an APP named MCODE (Molecular Complex Detection) in Cytoscape.
MCODE can identify dense areas and most signi cant module from the network analyzed by STRING.
The hub-genes were selected with degrees > 9.
Prediction and construction network of mRNAs and miRNAs. The mRNA targets of miRNA were predicted using EVmiRNA (http://bioinfo.life.hust.edu.cn/EVmiRNA) [16]. Top 15 signi cantly differentially expressed miRNAs of GSE61741 were selected. The target genes of top 15 signi cantly differentially expressed miRNAs of GSE61741 and differentially expressed mRNAs of GSE9128 were intersected. Based on the negative regulation relationship between miRNAs and mRNAs, the miRNA-mRNA network was constructed and visualized using the software of Cytoscape (version 3.6.1).
Prediction and construction of the ceRNA network. The lncRNA targets of the miRNAs were predicted using DIANA-LncBase v2(http://carolina.imis.athena-innovation.gr/) [17]. Top 15 signi cantly differentially expressed miRNAs of GSE61741 were selected. The target lncRNAs of top 15 signi cantly differentially expressed miRNAs of GSE61741 and differentially expressed lncRNAs of GSE77399 were intersected. Based on the association of lncRNA, miRNA and mRNA, the alluvial of ceRNA was constructed and visualized using the web tool (http://www.bioinformatics.com.cn), an online platfrom for data analysis and visualization.

Results
Identi cation of differentially expressed mRNAs, miRNAs and lncRNAs in patietns with heart failure. A total of three datasets including GSE9128, GSE61741 and GSE77399 were downloaded from the GEO database. GEO2R was used to screen for the differentially expressed mRNAs, miRNAs and lncRNAs between heart failure patients and healthy controls. miRNA predicted target lncRNA analysis. Based on the above data, target lncRNAs of above 12 DEMis and differentially expressed lncRNAs of GSE77399 were intersected. According to negative regulation relationship between miRNAs and lncRNAs, only three pairs of lncRNA-miRNA including H19-has-miR-20a, PCGEM1-has-miR-129-5p and GAS5-has-miR-185-5p were obtained.
Hub-gene selection and ceRNA relationship construction. The most signi cant module of PPI network was obtained and visualized using Cytoscape. The hub genes were selected with degrees > 9. A total of 30 genes were identi ed as hub genes ( Figure 6). According to above data, target mRNAs of -has-miR-20a, has-miR-129-5p, has-miR-185-5p and hub-genes were intersected. Based on the ceRNA theory, miRNAs negatively regulate the expression of mRNAs and lncRNAs, seven ceRNA relationships were obtained (Figure 7).

Discussion
Acute heart failure caused by progressive heart failure is a common disease in ICU. These patients often require mechanical ventilation and renal replacement therapy, which results in an increased economic burden for patients and their families [18]. Therefore, it is important to identify the molecular mechanism by which heart failure occurs and to identify treatment for this mechanism. Recently, the mechanism of ceRNA has attracted increasing attention. The relationship among lncRNA, miRNA and mRNA play an important role in the regulation of multiple processes. At present, numerous studies have been conducted on the ceRNA molecular mechanism of heart failure. Zhang et al have found that lncRNA-CHAR/miR-20b/PTEN play an important role in the progression of cardiac hypertrophy [19]. The research of Liang et al showed that lncRNA PFL contributes to cardiac brosis by acting miRNA-let-7d [20].
In the present study, we analyzed the datasets of GSE9128, GSE61741 and GSE77399 to nd more possible molecular mechanism of ceRNA in heart failure. Based on the ceRNA theory and above series of analyses, network of ceRNA which include 7 mRNAs (BCL2A1, DUSP1, EGR1, MYC, NR4A2, PTGS2 and RAC2), 3 miRNAs (miR-20a, miR-129-59 and miR-185-5p) and 3 lncRNAs (GAS5, H19 and PCGEM1) were obtained. MYC played critical roles in heart failure development progress [21]. NR4A2 is a member of the NR4A orphan nucleus receptor family. NR4A2 has protective function for cardiomyocytes against myocardial infarction [22]. DUSP1 can regulate cardiac metabolism. Overexpress DUSP1 can alleviate the fatal mitochondrial ssion and provide a survival advantage to myocardial tissue [23]. The level of EGR1 is related to effectiveness of percutaneous coronary intervention. If the level of EGR1 is signi cant decreased in the early postoperative period, the patient may be suspected of having no-re ow [24]. PTGS2 (cyclooxyfenase-2) represents a key enzyme in arachidonic acid metabolism in health and disease. It is expressed in several human tissues and induced in various cell types in response to in ammatory cytokine [25]. BCL2A1 and RAC2 have not been reported in the progression of heart failure. BCL2A1 is one of B-cell lymphoma2 (BCL2) proteins which are important cell death regulators. BCL2A1 is overexpressed in a variety of cancer cells, including hematological malignancies and solid tumors, and may contribute to tumor progression [26]. RAC2 is a GTpase that is exclusively expressed in hematopoietic cells. Mutations in RAC2 is associated with immunode ciencies in some patients [27].
In previous study, the expression of miR-20a-5p is associated with the degree of left ventricular dilation [28]. Downregulation of miR-129-5p was observed in the serum of chronic heart failure patients.
miR-129-5p mimic improved heart function and hemodynamic parameters [29]. At present, the study of miR-185-5p associated with heart failure is rare. There are some researches about miR-185-5p in the eld of cancer. miR-185-5p was proved that it can inhibit cell metastasis of HCC by suppressing ROCK2 [30].
Zhang et al found that expression of plasma H19 was high and it was independent predictors for coronary artery disease [31]. Li et al have found that the expression of GAS5 was low in peripheral blood of chronic heart failure [32]. These researches and ndings are consistent with our results. At present, the study of PCGEM1 in heart failure was rare. There were some researches of PCGEM1 about cancers. In the research of Ho et al, they found that PCGEM1 is often upregulated in prostate cancer [33]. PCGEM1 also promotes cell proliferation, migration and invasion in cervical cancer [34]. Therefore, further study about the function and expression of PCGEM1 in heart failure is needed to identify. These ndings indicate that these lncRNAs and miRNAs may have profound function in the progression of heart failure. Therefore, further studies about these mRNAs, lncRNAs and miRNAs are needed to identify their relationship and mechanisms.
There are several limitations in the present study. Firstly, we only have a result of bioinformatics analysis. Therefore, future in vitro and in vivo experiments are required to verify these results in heart failure pathology. Secondly, studies with larger cohorts of patients with heart failure are required to con rm the diagnostic and therapeutic value of the identi ed ceRNAs.
In conclusion, in the present study, we have performed a bioinformatics analysis to identify ceRNA network that may be involved in the progression of heart failure. In the present study, based on the ceRNA theory and above series of analyses, network of ceRNA which include 7 mRNAs (BCL2A1, DUSP1, EGR1, MYC, NR4A2, PTGS2 and RAC2), 3 miRNAs (miR-20a, miR-129-59 and miR-185-5p) and 3 lncRNAs (GAS5, H19 and PCGEM1) were obtained. These ndings can be used to carrying on further study to identify the biological function, appropriate treatment targets and biomarkers in the progression of heart failure.

Declarations
Funding sources: This research did not receive any speci c grant from funding agencies in the public, commercial, or notfor-pro t sectors.

Con ict of interest
There was no con ict of interest between each author

Author contributions
Qinghui Fu contributed to the conception of the study; Jun Hu and Enjiang Chen performed the bioinformatic analysis; Rufang Jiang contributed signi cantly to analysis and manuscript preparation; Xiaoqian Luo and Weina Lu performed the data analyses and wrote the manuscript; Meiling Weng helped perform the analysis with constructive discussions.

Consent to participate
All the authors have consented to participate in this study.

Consent to publish
All the authors have consented to publish this study       Sankey diagram for the ceRNA network in heart failure. Each rectangle represents a gene, and the connection degree of each gene is visualized based on the size of the rectangle.