Identifying a Serum Exosomal-Associated lncRNA/circRNA-miRNA-mRNA Network in Coronary Heart Disease

Background Accumulating evidence supports the importance of noncoding RNAs and exosomes in coronary heart disease (CHD). However, exosomal-associated competing endogenous RNA- (ceRNA-) mediated regulatory mechanisms in CHD are largely unexplored. The present study aimed to explore exosomal-associated ceRNA networks in CHD. Methods Data from 6 CHD patients and 32 normal controls were downloaded from the ExoRBase database. CHD and normal controls were compared by screening differentially expressed mRNAs (DEMs), lncRNAs (DELs), and circRNAs (DECs) in serum exosomes. MicroRNAs (miRNAs) targeting DEMs were predicted using the Targetscan and miRanda databases, and miRNAs targeted by DELs and DECs were predicted using the miRcode and starBase databases, respectively. The biological functions and related signaling pathways of DEMs were analyzed using the David and KOBAS databases. Subsequently, a protein-protein interaction (PPI) network was established to screen out on which hub genes enrichment analyses should be performed, and a ceRNA network (lncRNA/circRNA-miRNA-mRNA) was constructed to elucidate ceRNA axes in CHD. Results A total of 312 DEMs, 43 DELs, and 85 DECs were identified between CHD patients and normal controls. Functional enrichment analysis showed that DEMs were significantly enriched in “chromatin silencing at rDNA,” “telomere organization,” and “negative regulation of gene expression, epigenetic.” PPI network analysis showed that 25 hub DEMs were closely related to CHD, of which ubiquitin C (UBC) was the most important. Hub genes were mainly enriched in “cellular protein metabolic process” functions. The exosomal-associated ceRNA regulatory network incorporated 48 DEMs, 73 predicted miRNAs, 10 DELs, and 15 DECs. The LncRNA/circRNA-miRNA-mRNA interaction axes (RPL7AP11/hsa-miR-17-5p/UBC and RPL7AP11/hsa-miR-20b-5p/UBC) were obtained from the network. Conclusions Our findings provide a novel perspective on the potential role of exosomal-associated ceRNA network regulation of the pathogenesis of CHD.


Introduction
Coronary heart disease (CHD) is a complex biological process accompanied by wide transcriptional changes, the mechanism of CHD is still complex and unclear [1].
Noncoding RNAs mainly comprise microRNAs (miRNAs/miRs), long noncoding RNAs (lncRNAs) and circular RNAs (circRNAs). MiRNAs are a class of small noncoding RNAs, which could block protein translation or induce degradation with the combination of speci c region of target messenger RNA (mRNA) [2] LncRNAs have more diverse functions acting as epigenetic regulators, molecular scaffolds, or decoys [3]. CircRNAs can function as templates for viroid and viral replication, as intermediates in RNA processing reactions, as regulators of transcription in cis, as small nucleolar RNAs, and as miRNA sponges [4]. With the development of sequencing technology and bioinformatics, it has been found that noncoding RNAs were involved in the pathophysiology of cardiovascular diseases [5]. Ahmadi, R et al [6] demonstrated that miR-342-5p could be a biomarker for diagnosis of CHD associated with in ammatory cytokines. Wang, H et al [7] revealed that lncRNA BRAF-activated noncoding RNA is associated with the occurrence of CHD. Moreover, circRNAs, or lncRNAs have been found to interact with miRNAs as competitive RNAs (ceRNAs) to regulate target mRNAs activity and participate in CHD. For instance, circ YOD1 Deubiquitinase might be a novel target for diagnosing CHD from lncRNA/circRNA-miRNA-mRNA ceRNA network [8].
However, serum RNAs might often be degraded by RNA enzyme and may not accurately re ect the pathological differences. While exosomes could protect them from being degraded [9]. Exosomes are small vesicles with a diameter of approximately 30-150 nm containing proteins, nucleic acids, and lipids [10], which were related to diverse regulation processes of cardiovascular disorders, including myocardial injury, repair, regeneration and so on [11].
To better understand the underlying molecular regulatory mechanisms of CHD, we aimed to identify the differentially expressed exosomal-related lncRNAs, circRNAs, miRNAs and construted the ceRNA network to discover more accurate and reliable candidate diagnostic biomarkers and therapeutic targets of CHD.

Ethics approval
This study was approved by the Ethics Committee of The First A liated Hospital of Nanjing Medical University. All methods were performed in accordance with the relevant guidelines and regulations.

Data collection:
The study owchart is shown in Fig. 1. Currently, the exoRBase database (http://www.exorbase.org/) is a repository of circRNAs, lncRNAs and mRNAs derived from RNA-seq data including human blood exosomes analysis. These samples come from different biological conditions, including normal persons (NP), CHD, colorectal cancer, hepatocellular carcinoma, pancreatic adenocarcinoma and breast cancer [12]. In the Study, data of NP and CHD blood samples were downloaded, including 6 patients with CHD and 32 normal controls.
Identi cation of differently expressed mRNAs, lncRNAs and circRNAs: The lists of differentially expressed circRNAs (DECs), lncRNAs (DELs), and mRNAs (DEMs) between controls and patients with CHD were generated using the LIMMA package in the R software. The values of |log 2 Fold Change (FC)| > 0 and P-value < 0.05 were selected as the cut-off criteria.

Integration of PPI network and module analysis
The PPI network of DEMs was constructed by STRING (https://string-db.org) and visualized with Cytoscape software [13]. Furthermore, the Molecular Complex Detection (MCODE) application in Cytoscape was applied to select the PPI network modules, with a cut-off=2, node score cut-off=0.2, k-core=2 and maximum depth=100 as the selection criteria. In addition, the nodes with degree ≥ 5 were identi ed as hub nodes in the PPI network.

Functional enrichment analyses
Gene ontology (GO) analysis was used to annotate the DEMs and hub genes based on biological processes (BP), cellular components (CC) and molecular functions (MF) [14]. To investigate the biological function of DEMs and hub genes, the database for annotation, visualization, and integrated discovery (DAVID) online tool (version 6.8; david.abcc.ncifcrf.gov) was utilized to perform GO analysis [15]. In addition, the KOBAS 3.0 online analysis database was used to perform pathway enrichment analysis [16]. The signi cant enrichment for GO and KEGG analyses threshold was p-value <0.05 and count ≥2.
Construction of the lncRNA/circRNA-miRNA-mRNA ceRNA network CeRNA regulation has been reported to serve important roles in human disease, the circRNA or lncRNA-miRNA-mRNA interactions network was constructed to explore the association among circRNA, lncRNA, miRNA, and mRNA [21]. Finally, the lncRNA/circRNA-miRNA-mRNA network was built to visualize the interactions using Cytoscape.

Statistical analysis
All data were expressed as the mean ± standard error. Statistical analyses were performed using GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA, USA). P <0.05 was considered to indicate a statistically signi cant difference.  Table S1-S3. Finally, based on the pre-set criteria of P-value < 0.05 and | log2(fold change, FC) | > 0.5, we plot the heat maps for DEMs, DELs and DECs, respectively, as shown in Fig. 2a-c.

Functional enrichment analyses
Functional enrichment analyses illustrated that the DEMs were mainly enriched "chromatin silencing at rDNA", "telomere organization", "negative regulation of gene expression, epigenetic" for the BP terms. CC analysis showed that the DEMs were signi cantly enriched for the "nucleosome", "nuclear chromosome", "nuclear chromosome". For the MF category, the DEMs were enriched in "histone binding", "protein heterodimerization activity", "poly(A) RNA binding" (Fig. 4 and Additional le 1: Table S4). With the enrichment analyses for hub genes, we found biological function of UBC was enriched in "cellular protein metabolic process", the location of UBC was abundantly enriched in "extracellular exosome", and the molecular function comprised "poly(A) RNA binding" (Fig. 5 and Additional le 1: Table S5).
LncRNA/circRNA-miRNA-mRNA ceRNA network Accordingly, ceRNAs network analyses were performed to unravel the functions of identi ed differently expressed exosomal-associated ceRNA network in CHD patients. This ceRNA network consisting of 48 DEMs, 72 predicted miRNAs, 10 DELs, and 15 DECs. From the ceRNA network, we identi ed lncRNA RPL7AP11 competed for binding to hsa-miR-20a-5p and hsa-miR-17-5p, thereby affecting UBC expression (Fig. 6). These results suggested that the ceRNA networks we predicted in this paper might be a key factor underlying the pathogenesis of CHD.

Discussion
Atherosclerotic disease and its thrombotic complication may lead to the development of CHD, and if untreated, it progresses into myocardial infarction. Exosomes are a type of extracellular vesicle that contain constituents (protein, DNA, and RNA) of the cells that secrete them [22].
Despite years of research, the underlying pathogenesis of coronary artery disease has not been fully de ned. Recently, dysregulated expression of RNAs (lncRNAs, circRNAs, miRNAs, mRNAs) have been partially found to be associated with CHD [8,23]. However, serum RNAs might often be degraded by RNA enzyme and may not accurately re ect the pathological differences. While exosomes could protect them from being degraded [9]. Hence, we identi ed serum exosomal-associated RNAs and constructed the ceRNA network in CHD, revealing a new targeting axis in the pathogenesis of CHD. To our knowledge, this was the rst to explore exosomal-associated ceRNA network in CHD.
In this study, we rst identi ed 312 DEMs, 85 DECs and 43 DELs involving in the pathogenesis of coronary heart disease. Enrichment analysis and PPI network were subsequently performed, of which, UBC (ubiquitin C) was one of the most important hub genes. After the prediction of miRNAs targeting mRNA, exosomal-associated circRNA/lncRNA-miRNA-mRNA ceRNA network was constructed. Our results suggest speci c ceRNA axes in the pathogenesis of CHD, which may be promising targets for CHD diagnosis.
UBC (Ubiquitin C), belonging to the ubiquitin family, is associated with protein degradation, DNA repair, kinase modi cation, autophagy, regulation of in ammation and regulation of other cell signaling pathways [24,25]. Our enrichment analysis both in DEMs and hub genes showed UBC participated in cellular protein metabolic process. Ji Y and his colleagues [26] proved that the expression level of ubiquitin was signi cantly higher in CHD patients than healthy individuals and the levels of ubiquitin were consistent with the severity of different classes of CHD. Our study further con rmed the function of UBC in the pathogenesis of CHD, which could be a non-invasive biomarker.
MiR-17-5p was reported to regulate cell cycle, proliferation, apoptosis. Board evidence has elucidated its profound function in regulating cardiovascular diseases. The de ciency of miR17 in neonatal mice is lethal and the over-expression of miR-17-5p could extend the life span of mice [27]. Liu G et al [28] con rmed the up-regulation of miR-17-5p could contribute to hypoxia-induced proliferation of human pulmonary artery smooth muscle cells, leading to pulmonary hypertension. Yang S et al [29] found miR-17-5p silencing protects heart function after AMI through decreasing the rate of apoptosis and repairing vascular injury. Moreover, recent studies have shown circulating miR-17-5p could be a novel biomarker for diagnosis of acute myocardial infarction [30].
MiR-20b-5p was found to attenuate hypoxia-induced apoptosis in cardiomyocytes [31]. Also, Zhen W et al [32] found the overexpression of miR-20b-5p could increase cell viability and repress autophagy and apoptosis in human umbilical vein endothelial cells with hypoxia-reoxygenation injury. Both hypoxia and hypoxia-reoxygenation models are similar to patients with MI and revascularization of MI, we considered its vital role in regulating CHD. However, there has been little research about the function of miR-20b-5p in CHD patients, which needs to be further validated.
The lncRNA RPL7AP11 (ribosomal protein L7a pseudogene 11) is a pseudogene of ribosomal protein L7a (RPL7A). Lou W et al [33] found RPL7A was down-regulated in high density lipoprotein induced vascular endothelial cell line ECV 304. Pseudogenes, abundant in the human genome, were considered as nonfunctional "junk genes traditionally [34]. Recent studies have proved its function in various diseases. However, there has been limited research on RPL7AP11. More evidence needs to be con rmed.
Our study demonstrated that RPL7AP11 could sponge hsa-miR-17-5p and hsa-miR-20b-5p to upregulate UBC, thus regulating the pathogenesis of CHD through cellular protein metabolic process.
There are some limitations to the present study. Firstly, the sample scale was not large. An additional validation cohort should be included in further studies to analyze the expression of these identi ed lncRNAs, circRNAs, miRNAs and mRNAs. Secondly, how these novel exosomal-associated ceRNA axes participate in the process of CHD development is still unclear. Further cell and animal experiments are needed to verify these ndings.