Research on the circular RNA bioinformatics in patients with acute myocardial infarction

Abstract Objective Through the detection of circular RNA (circRNA) using expression profiling chips, we searched for circRNAs related to acute myocardial infarction (AMI) and explored their relationship and possible mechanisms with AMI. Method The study subjects included 3 AMI patients and 3 controls, and circRNA expression profiling analysis was performed using a microarray gene chip to identify circRNAs with large differences in expression between groups and to construct a circRNA‐miRNA network. Results Compared with the control group, there were 650 differentially expressed circRNAs found in AMI patients (P < .05, fold change > 2), including 535 up‐regulated circRNAs, such as hsa_circ_0050908, hsa_circRNA4010‐22, hsa_circ_0081241, hsa_circ_0010551, hsa_circRNA4010‐20, hsa_circRNA14702, hsa_circ_0115392, has_circRNA1825‐44, has_circRNA8493‐7, and hsa_circ_0025097. Furthermore, there were 115 down‐regulated circRNAs, such as hsa_circ_0066439, hsa_circ_0054211, hsa_circ_0095920, hsa_circ_0122984, hsa_circ_0113067, hsa_circ_0039155, hsa_circRNA4014‐45, hsa_circ_0122979, hsa_circ_0059665, and hsa_circ_0009319. The circRNAs hsa_circ_0066439, hsa_circ_0081241, and hsa_circ_0122984 can regulate multiple signal pathways to participate in the AMI process through hsa‐miR‐1254, hsa‐miR‐328‐5p, and other miRNAs. In addition, the expression of circRNA‐miRNA in peripheral blood is related to the network. Differentially expressed circRNAs are involved in chromatin organization, chromatin‐modifying enzymes, signal transduction, lysine degradation, the mitogen‐activated protein kinase (MAPK) signaling pathway, focal adhesion, and a variety of other pathways, such as myocardial infarction, coronary heart disease, hypertension, and other diseases. The gene ontology analysis results show that molecular function mainly involves binding and molecular structural activity, whereas the biological process mainly involves a single biological process, a cellular component for organization, and a cellular process, and the cellular component mainly involves a protein complex, an extracellular matrix, and a membrane. Conclusion circRNA and microRNA interact to participate in the development of AMI. circRNA may be involved in the pathogenesis of AMI.


| INTRODUC TI ON
In recent years, the incidence, disability, and mortality in cardiovascular diseases (CVDs) have increased dramatically, and the main cause of death in CVDs is acute myocardial infarction (AMI). 1,2 Although significant progress has been made in the diagnosis and treatment of AMI, effective therapeutic targets for the protection of myocardial cells from apoptosis remain limited. Therefore, there is an urgent need to understand the pathogenesis of AMI at the molecular level. Most important is the need for the discovery of novel molecular entities for AMI-related apoptosis. Recently, noncoding RNA (ncRNA) has been suggested as a biomarker for AMI. 3 Circular RNA (circRNA) is a type of ncRNA that exists in the form of a covalently closed continuous loop and is stably expressed in many types of cells, and its ability to regulate gene expression comes mainly from its binding or inhibiting of microRNA (miRNA). 4,5 Several studies have demonstrated the key role of circRNA in heart development and physiology. 6 The abnormal expression of circRNA has been linked to CVDs such as heart failure, myocardial infarction, and atherosclerosis, suggesting the potential importance of circRNA in these pathological conditions. [7][8][9] However, there have been few studies on the correlation between circRNA and AMI. In this study, a microarray gene chip was used to analyze the circRNA expression profile of patients with AMI, and the circRNA with a large expression difference from the control group was analyzed. By searching for candidate circRNA related to AMI, a new promising breakthrough point was provided for AMI diagnostic markers or targeted therapy, and key information was provided for revealing the complex regulatory mechanism of A-MI.

| Subject selection and sample collection
Selection of study subjects: The study selected 3 patients diagnosed with AMI and 3 healthy subjects.
Inclusion and exclusion criteria：The diagnostic criteria for AMI were as follows： 1. Typical severe sternum pain with a duration of more than 30 minutes: The clinical manifestations are dull pain, squeezing pain in the posterior sternum or precordial area, and radioactive pain lasting more than 30 minutes in the neck, shoulder, and back, accompanied by sweating and dying. Furthermore, there may be clinical manifestations of heart failure or cardiogenic shock.

Typical dynamic changes of the patient's electrocardiogram:
The patient's relevant signs were dynamically monitored by electrocardiogram, and the results of the electrocardiogram showed pathological abnormal Q waves and ST segment elevation. Although the electrocardiogram did not have pathological Q waves or ST segment elevation, it showed that the T wave and ST segment had an ischemic and dynamically changing performance.
3. Increased levels of creatinine kinase isoenzyme MB, myoglobin, and cardiac troponin I.

Emergency coronary angiography or percutaneous coronary in-
tervention (PCI), confirming that at least one of the three main coronary arteries had a luminal stenosis greater than 50%.
By meeting 2 of the 4 criteria above, the diagnosis of AMI can be made.
Exclusion criteria were as follows：previous old myocardial infarction or PCI; acute or chronic infection; hematological or systemic immune diseases; severe heart, kidney, liver, and lung dysfunction;

| RNA extraction
The total RNA from the each patient's plasma was extracted using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. In accordance with the manufacturer's procedures, a mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA) was used for purification. The purity and concentration of the RNA were determined from a 260/280 reading by using a spectrophotometer (NanoDrop Nd-1000, Thermo Fisher Scientific, Waltham, MA, USA). RNA integrity was determined with a lab-on-a-chip kit using an RNA 6000 nano chip and Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA) to assess RNA quality comprehensively.

| RNA amplification, labeling, and hybridization
Fluorescent dye Cy3 dCTP-labeled cDNA was prepared using Eberwine linear RNA amplification and enzymatic reaction. 10 In addition, we used a labeling kit (CapitalBio, Beijing, China) and CapitalBio-cRNA amplification to produce higher yields of labeled cDNA according to the manufacturer's instructions.

| Microarray imaging and data analysis
Data normalization and difference analysis were performed on the circRNA array data using GeneSpring software V13.0 (Agilent Technologies, Santa Clara, CA, USA). The data results were analyzed for data summarization, normalization, and quality control. These circRNA target sequences were obtained from Circbase, Deepbase, and Rybak-Wolf (2015). 11 In order to select differentially expressed genes, thresholds of > 2 and <−2-fold changes were used in this study, and the P-value corrected by the Benjamini-Hochberg procedure was 0.05. Using the adjust data function of Cluster 3.0 (Stanford University, Stanford, CA, USA) software, the data were log2-transformed, and the median location was centered on the gene, and then, hierarchical clustering with average linkage was used for further analysis. Finally, Java TreeView (Stanford University School of Medicine, Stanford, CA, USA) was used for tree visualization.

| Construction of the circRNA-miRNA network
After the differentially expressed circRNAs were screened, StarBase (Sun Yat-sen University, China) software was used to predict the target miRNAs of the circRNAs screened in the AMI group to obtain a list of miRNAs. circRNA plays an important role in miRNA function and transcriptional control by acting as a competitive endogenous RNA or positive regulator on its parent-encoded gene. A circRNA-miRNA network was constructed based on miRanda database prediction (http://mirdb.org/). These circRNA-miRNA pairs were selected to construct the network using the open source bioinformatics software Cytoscape (National Institute of General Medical Sciences, USA). In network analysis, degree centrality is defined as the link number between one node and another node. Degree is the simplest and most important measure of gene centrality in a network of relative importance. 12

| Enrichment analysis
The NHGRI GWAS Catalog (http://www.genome.gov/gwast udies/), KEGG DISEASE (http://www.genome.jp/kegg/disea se/), and OMIM (http://www.ncbi.nlm.nih.gov/omim) bioinformatics databases were used to search for the disease enrichment analysis of the genes that were significantly expressed in the whole blood samples of the patients with AMI, and P < .05 was considered significant. The differentially expressed circRNAs in the whole blood of subjects were analyzed by pathway enrichment analysis using the Reactome, KEGG PATHWAY, BioCyc, and PANTHER databases, and P < .05 was considered as a meaningful analysis.
Similarly, gene ontology analysis was performed on the linear mRNA transcripts corresponding to the 650 differentially expressed circRNAs selected from the specimens of the AMI group.
The analysis included mainly molecular functions, biological processes, and cellular components.

| Statistical analysis
Basic data were gathered regarding the age, weight, height, blood pressure, and blood lipids of the patients in the control and AMI groups. Statistical calculations were performed using SPSS (version 23.0, IBM, USA) software, and continuous data were expressed as the mean + standard deviation (x + s). The t test was used to compare the continuous variables between the two groups, and the categorical variables were expressed as counts, while the chi-square test was used to compare the categorical variables between the two groups. P < .05 was considered statistically significant.

| Population parameters of the subjects
The population parameters of the subjects are shown in Table 1. It can be seen from this table that there were no statistically significant differences for age, body mass index, systolic blood pressure, diastolic blood pressure, or blood lipid biochemistry between the AMI group and control group (P > .05).

| Microarray gene chip analysis identified significantly different circRNAs
In order to understand the molecular mechanisms involved and to search for AMI biomarkers, a microarray gene chip was used to screen and analyze the circRNA expression profiles of our study.

F I G U R E 1
Profiling of circular RNAs in the plasmas from AMI patients and normal controls. (a) Cluster analysis of gene expression between AMI patients and control samples. Each column represents the expression profile of a tissue sample, and each row corresponds to a circRNA. High expression level is indicated by red and lower levels by green. (b) Volcano plot shows the up-regulated and downregulated circRNAs in AMI patients vs control group. Higher expression levels are indicated by red, lower expression levels are indicated by green, and no significant difference is indicated by black. (c) Circos diagram of the difference circRNA between the control group and the AMI group. The length of the bar indicates the multiple of the differential gene, the red bar in the inner circle indicates that the differential gene is upregulated, and the green bar indicates that the differential gene is down-regulated

| Screening of target miRNAs for the differentially expressed circRNAs
A total of 475 out of 650 circRNAs with different expressions could bind more than two miRNAs. Among the top 10 differentially expressed circRNAs, six had more than two target miRNAs, and their circRNA-miRNA network relationship is shown in Figure 2A. pathways through miRNA hsa-miR-1254, hsa-miR-328-5p, and other target miRNAs to participate in the AMI process. By further scrutinizing more stringent parameters, such as processed signal variation between repetitions and referencing to the established circRNA databases and publications, six new circRNA candidates were selected.
The miRNA software was used to predict the target miRNAs of the six differentially expressed circRNAs. The results showed that the target miRNAs of the six differentially expressed circRNAs were larger than two. The circRNA-miRNA network of these six circRNAs is shown in Figure 2B. It can be seen from this figure that RNA hsa_circ_0043563, hsa_circ_0119137, hsa_circ_0106804, and hsa_circ_0085214 can regulate multiple signal pathways through miRNA hsa-miR-4763-3p, hsa-miR-328-5p, etc Thus, it can participate in the occurrence of AMI.
Subsequent hsa_circ_0025097 and hsa_circ_0028302 can also regulate multiple signal pathways through hsa-miR-328-5p. Therefore, there are network relationships between circRNA-miRNA that may be the key to their role in the pathogenesis and progression of AMI.

| Enrichment analysis results for disease
Disease, pathway, and GO analysis suggest that these differentially expressed circRNAs are relevant to several vital biological processes, cellular components, molecular functions, and critical signaling pathways ( Figure 3). By searching bioinformatics databases such as KEGG DISEASE, NHGRI GWAS Catalog, and OMIM, the significantly expressed genes in the whole blood samples of the patients with AMI were subject to enrichment analysis, and significance was marked by P < .05. The results are shown in Table 3. As can be seen from this table, after using 3 databases to perform a disease enrichment analysis of characteristic genes, 3 diseases closely related to the cardiovascular system were found, which were myocardial infarction, coronary heart disease, and hypertension. From the results, it can be determined that the target genes are closely related to coronary heart disease.

| Pathway analysis
To study the differentially expressed circRNAs in the whole blood of  Table 4. Based on the results of the pathway enrichment analysis performed using the various databases above, it is not difficult to discover that the target genes of these differentially expressed

| Gene ontology analysis
Similarly, gene ontology analysis was performed on the linear mRNA transcripts corresponding to 650 differentially expressed circRNAs selected from the specimens of the AMI group. The analysis included mainly molecular functions, biological processes, and cellular components.
The gene ontology analysis demonstrated that the differentially expressed circRNAs play regulatory roles in cells through various biological processes, such as cellular, single-organism, and cellular component organization ( Figure 4); cellular components, such as protein complex, extracellular matrix, membrane, and extracellular matrix component ( Figure 5); and molecular functions, such as binding and structural molecule activity ( Figure 6).

| D ISCUSS I ON
Based on whether they can be translated, circRNAs were divided into noncoding circRNAs and coding circRNAs that have a closed F I G U R E 4 GO functional hierarchical network diagram (biological process). The first line of each node describes the GO node name, the second line describes the node number in GO, and the third line describes the corrected P-value. The node color indicates the significance of the corrected P-value, and the darker the color indicates the more significant the corrected P-value circular structure and are not affected by exonucleases. 13 Its expression is relatively more stable, and it is not easily degraded. In recent years, circRNA has emerged as a new member of the RNA family for that has attracted attention. 14  Furthermore, HRCR inhibits cardiac hypertrophy and heart failure, Cdrlas induces myocardial infarction, Circ-Fox03 promotes heart aging, and cANRIL is associated with atherosclerosis. 17 Because the circular structure of circRNA is more stable, it is easier to use as a potential new biomarker. miRNA is also a type of ncRNA, and it has been one of the most studied of these RNAs in recent years. It has been shown to play a very important role in the development of various diseases, and the role of miRNAs in CVDs has also gradually been discovered. Studies have found that circulating miRNA-134, miRNA-22, miRNA-328, and miRNA-499 have abnormal expression levels in the plasma of patients with AMI, suggesting that they may be potential biomarkers of AMI. [18][19][20][21] There are many miRNA response elements on circRNA. The sponge adsorption effect of cir-cRNA on miRNA occurs mainly through response elements binding miRNA to block the inhibition of miRNA on its target gene expression. When circRNA is highly expressed, the target gene expression of miRNA is up-regulated, and when circRNA is lowly expressed, the target gene expression of miRNA is down-regulated. 22 It can be said that the miRNA sponge function performed by circRNA plays an important role in the occurrence of diseases, competitively inhibiting the expression of miRNA and blocking the expression of miRNA on target genes. Therefore, the relationship between circRNA and AMI has great research potential. Since, as compared to linear structures, circRNA does not contain a poly A tail, it is not easily cleaved by exonuclease and exists more stably in an organism. 23,24 It also has high conservation, expression, and tissue specificity. 25 Therefore, the F I G U R E 5 GO functional hierarchical network diagram (cellular component process). The first line of each node describes the GO node name, the second line describes the node number in GO, and the third line describes the corrected P-value. The node color indicates the significance of the corrected P-value, and the darker the color indicates, the more significant the corrected P-value regulation efficiency of circRNA is higher than that of linear structure RNA. 26 These prior studies have indicated that circRNA is likely to have an important relationship with the occurrence of CVDs.
In this study, the circRNA expression in patients with AMI was analyzed using a microarray gene chip, and the circRNAs with significantly different expressions from controls were analyzed. The results showed that, as compared to the control group, 650 differentially expressed circRNAs were screened out for the AMI group, of which 535 were up-regulated and 115 were down-regulated (Table 2, Figure 1C). Such a large number of circRNA differential expression results indicate that circRNA is likely to play a positive role in the occurrence and development of AMI. The circRNA-miRNA network was constructed after screening out the differentially expressed circRNAs, and it showed that 475 differentially expressed circRNAs could bind > 2 miRNAs. Among the 10 circRNAs with the largest difference, three of them can bind to miR-328-5p-related cardiovascular diseases, including one up-regulated hsa_circ_0081241 and two down-regulated hsa_circ_0066439 and hsa_circ_0122984 ( Figure 2A). In addition, the 6 novel circRNA candidates were able to bind to miR-328-5p related to myocardial infarction ( Figure 2B F I G U R E 6 GO functional hierarchical network diagram (molecular function process). The first line of each node describes the GO node name, the second line describes the node number in GO, and the third line describes the corrected P-value. The node color indicates the significance of the corrected P-value, and the darker the color indicates the more significant the corrected P-value that levels of miR-328 in the plasma and whole blood of AMI patients were significantly increased: 10.9 times and 16.1 times as compared to a control group, respectively. The level of miR-133 increased 4.4 times, suggesting that miR-328 and miR-133 may represent novel biomarkers of AMI.
Among the previous studies, Ruan et al 27  The results of the current study have shown that multiple circRNAs can bind to miR-328-5p, which is associated with myocardial infarction. The possibility of a network correlation between the circRNA and miRNA was first confirmed, and then, it was suggested that this network correlation pathway is likely to have a certain regulatory mechanism during the occurrence and progression of AMI. Thus, this regulatory mechanism will affect the occurrence, development, and prognosis of AMI. This is of great significance for identifying circRNA as a possible biomarker for AMI and revealing the complex regulatory mechanisms in the process of AMI. According to the results of enrichment analysis and gene ontology analysis, it can be found that the circRNA-miRNA pathway is likely to exist in the occurrence and development of AMI, such as lysine degradation, MAPK signaling pathway, and focal adhesion (Table 4). Combined with the screened circRNAs and bound miRNAs, this pathway is likely to participate in the process of regulating the occurrence and development of AMI.
There are some limitations in this study. First, the sample size was small, and the sample was only obtained from six subjects.
Second, there is no verification of RT-PCR for the top circRNA candidates expressing differences. In the future, circRNAs, such as hsa_circ_0066439, hsa_circ_0043563, hsa_circ_0119137, and miR-328-5p, should be verified to analyze the levels and interactions of these new circRNAs with miR-328-5p.

| CON CLUS ION
There were 650 circRNAs that were differentially expressed in AMI disease, and an interaction between circRNA and miRNA is involved in the occurrence and development of AMI. By combining the results of the disease enrichment analysis, pathway enrichment analysis, and gene ontology analysis, it can be found that the circRNA-miRNA interaction pathway very likely participates in regulating the occurrence and development of AMI.

ACK N OWLED G M ENTS
The author thanks all participants for their contributions.

CO M PE TI N G I NTER E S TS
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
YH Tang involved in protocol/project development and final approval of the version to be submitted. MH Jiang collected or managed the data. LL Yin analyzed the data, and wrote the study.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.