Biodata Mining of Differentially Expressed Genes between Acute Myocardial Infarction and Unstable Angina Based on Integrated Bioinformatics

Acute coronary syndrome (ACS) is a complex syndrome of clinical symptoms. In order to accurately diagnose the type of disease in ACS patients, this study is aimed at exploring the differentially expressed genes (DEGs) and biological pathways between acute myocardial infarction (AMI) and unstable angina (UA). The GSE29111 and GSE60993 datasets containing microarray data from AMI and UA patients were downloaded from the Gene Expression Omnibus (GEO) database. DEG analysis of these 2 datasets is performed using the “limma” package in R software. DEGs were also analyzed using protein-protein interaction (PPI), Molecular Complex Detection (MCODE) algorithm, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. Correlation analysis and “cytoHubba” were used to analyze the hub genes. A total of 286 DEGs were obtained from GSE29111 and GSE60993, including 132 upregulated genes and 154 downregulated genes. Subsequent comprehensive analysis identified 20 key genes that may be related to the occurrence and development of AMI and UA and were involved in the inflammatory response, interaction of neuroactive ligand-receptor, calcium signaling pathway, inflammatory mediator regulation of TRP channels, viral protein interaction with cytokine and cytokine receptor, human cytomegalovirus infection, and cytokine-cytokine receptor interaction pathway. The integrated bioinformatical analysis could improve our understanding of DEGs between AMI and UA. The results of this study might provide a new perspective and reference for the early diagnosis and treatment of ACS.


Introduction
Acute coronary syndrome (ACS) is a complex syndrome of clinical symptoms, which is characterized by acute myocardial ischemia with severe coronary stenosis or occlusion caused by coronary plaque rupture and subsequent thrombosis, mainly including acute myocardial infarction (AMI) and unstable angina (UA) [1,2]. AMI can be divided into acute ST-segment elevation acute myocardial infarction (STEMI) and acute non-ST-segment elevation acute coronary syndrome (NSTEMI) due to the different stratification of early diagnosis and treatment [3]. According to the statistics of the American Heart Association (AHA) and the National Institutes of Health in 2019, there were about 633,000 patients with acute coronary syndrome in 2014, including about 609,000 AMI patients and about 24,000 UA patients [4]. In addition, a study showed that the incidence of AMI had been increasing in younger populations and decreasing in the elderly and that improvement in the in-hospital mortality of AMI may have reached a plateau in all age groups in Japan [5].
Based on clinical epidemiology, the mechanism of the relationship between UA and AMI was studied, and great progress has been made [6][7][8][9]. Studies have shown that some UA patients will develop into AMI if they are not treated properly [10,11]. Therefore, the identification of differential genes (DEGs) between UA and AMI is essential for the development of new targeted drugs to improve clinical efficacy.
With the development of high-throughput gene chip and transcriptome sequencing methods, more and more gene chip technologies are used in the research of cancer and cardiovascular diseases, which enriches the biological markers for early diagnosis, classification, and prognosis of diseases and also provides references for the targeted therapy of diseases [12][13][14][15]. Microarray technology is conducive to understanding gene association, location, expression, and linkage studies and is widely used to compare DEGs in patients with different diseases [16]. At present, although there have been some bioinformatics-based studies on the miRNA, DEGs, and mRNA of AMI, the role of DEGs between AMI and UA remains unclear [17,18]. In this study, the microarray data containing both AMI and UA were downloaded from the GEO database, and the integrated bioinformatics method was adopted to screen the DEGs, protein-protein interaction (PPI), and related key pathways between AMI and UA. This research may provide new insights for understanding the functional regulation mechanism of DEGs between AMI and UA and help to develop relevant targeted therapies in further in-depth work. The flow chart of integrated bioinformatics research is shown in Figure 1.

Materials and Methods
2.1. Gene Expression Profile Data Collection. The GEO database (https://www.ncbi.nlm.nih.gov/gds/) developed and maintained by the National Center for Biotechnology Information (NCBI) not only collects and shares publicly available transcriptome data but also facilitates biomedical research [19]. In the current study, "Acute myocardial infarction" and "Unstable angina" were used as keywords to obtain gene expression microarray datasets related to AMI and UA. The criteria of the chip included are as follows: (1) the sample source of the dataset is human, (2) the dataset includes AMI and UA samples, and (3) there are no less than 20 samples per dataset.
2.2. Screening of DEGs. First, the gene expression dataset is translated into common gene symbols by using annotation information from the platform. The "limma" package of R software was utilized to screen the original dataset for expression normalization and differential expression analysis [20]. In addition, the actual situation determines whether the normalized data is log2 converted. The heatmap and volcano map show data results. The threshold of DEG screening was |log 2 FC | ≥1 and P < 0:05 (FC: fold change).

Protein Interaction Analysis and Module Analysis of
DEGs. The interaction between proteins is crucial for understanding the metabolic and molecular mechanisms of diseases. The number of organisms covered by the STRING database (http://string-db.org, version: 11.0) is more than double to 5090 compared with the previous version. It collects, scores, and integrates all public sources of proteinprotein interaction information and supplements this information through computational predictions, including direct and indirect correlations between proteins [21]. In this study, DEGs were input into STRING, and the species was set as "Homo sapiens." In addition, the data with a confidence score greater than 0.7 was selected (low confidence: score > 0:15, moderate confidence: score > 0:4, and high confidence: score > 0:7). The results were imported into Cytoscape 3.7.1 (https://cytoscape.org/) to establish a protein-protein interaction (PPI) network model [22], and the MCODE (Molecular Complex Detection) plug-in was used for module analysis [23]. The parameters were set to the default values (degree cutoff = 2, node score cutoff = 0:2, K-core = 2, and Max:depth = 100). Then, the Maximal Clique Centrality (MCC) algorithm in "cytoHubba" was applied to screen the hub genes with high connectivity in PPI networks [24].
2.4. Enrichment Analysis of DEGs. GO provides a framework for describing the function of gene products in organisms, which can identify the biological characteristics of highthroughput genome and transcriptome data, including biological process (BP), cellular component (CC), and molecular function (MF) [25]. KEGG can be used in the functional interpretation and practical application of DEG information [26]. In order to reveal the biological significance of DEGs, DAVID v6.8 (https://david.ncifcrf.gov/), "clusterProfiler" [27], and "GOplot" packages [28] in R software were utilized for GO function and KEGG pathway enrichment analysis, and the results with FDR < 0:05 were selected for analysis and visualization.

Expression Level Analysis and Correlation
Analysis of Key DEGs. According to the MCC algorithm, 10 genes with the most stable and highest score in the network were selected as the key genes. The differences in hub gene expression between AMI and UA were shown using the boxplot tool in ImageGP (http://www.ehbio.com/ImageGP/). In addition, in order to better understand the key DEGs, the Pearson correlation coefficient was used for correlation analysis, and the "Corrplot" package in R software was used for visualization [29,30].

Results
3.1. Screening of Differentially Expressed Genes. The present study finally included 2 datasets (GSE29111 and GSE60993), each of which contained AMI and UA samples, including 53 AMI samples and 25 UA samples (Table 1). DEG analysis of these 2 datasets is performed using the "limma" package in R software. A total of 242 differential genes were obtained from GSE29111, including 128  Table 2). Figure 2 shows the DEG volcano plots and heatmaps of the included dataset.

Enrichment Analysis of DEGs.
GO and KEGG pathway enrichment analysis was performed on PPI data of GSE29111 differentially expressed genes using DAVID and "GOplot" package in R software. According to FDR < 0:05, 4 GO entries with significant enrichment were selected (Figures 5(a) and 5(b), Table 4). It contains 2 BP items and 2 CC items: inflammatory response, G-proteincoupled receptor signaling pathway, integral component of plasma membrane, and plasma membrane. According to FDR < 0:05, 3 significantly enriched KEGG pathways were screened: neuroactive ligand-receptor interaction, calcium signaling pathway, and inflammatory mediator regulation of TRP channels ( Figure 5(c)).
In the current study, GO and KEGG pathway enrichment analysis was performed on PPI data of GSE60993 differentially expressed genes through the "clusterProfiler" package. According to P adjust < 0:01, a total of 52 significantly enriched GO entries were screened ( Figure 6(a) and Table 5). There are 41 BP items, mainly related to neutrophil activation, immune response, and inflammatory response; 9 CC items, primarily relevant to particles and plasma membrane; and 2 MF items, basically associated with cytokine receptor activity and cytokine binding. According to P adjust < 0:05, 3 significantly enriched KEGG pathways were screened: viral protein interaction with cytokine and cytokine receptor, human cytomegalovirus infection, and cytokine-cytokine receptor interaction ( Figure 6(b)).

Discussion
Many factors affect the occurrence and development of ACS, such as gender, fatigue, and frailty [31][32][33]. According to "China cardiovascular disease report 2018," the mortality rate of AMI showed an overall upward trend from 2002 to 2016, and the mortality rate in rural areas was higher than that in urban areas many times. Although a lot of research has been conducted on the molecular mechanisms related to AMI and UA, the understanding of the differentially expressed genes between AMI and UA is insufficient. Therefore, it is still a challenge to diagnose the disease types of ACS patients quickly and accurately. Microarray analysis is a powerful tool for screening hub genes, which can effectively reduce candidate genes associated with multifactorial diseases and identify potential molecular mechanisms of disease and biomarkers for diagnosis and prognosis [34,35]. In this study, we obtained the microarray datasets of AMI and UA patients from the GEO database to clarify their DEGs and candidate biomarker.
The integrated bioinformatics analysis method was performed to screen the potential differential genes related to AMI and UA. Through retrieval, 2 datasets (GSE29111 and GSE60993) from the GEO database were included. Although datasets as GSE29111 and GSE60993 were already used and published by some authors, the main comparison is the DEGs between AMI and healthy controls. The difference is that this study used these two datasets to explore the DEGs between AMI and UA. A total of 242 DEGs were obtained from GSE29111, including 128 upregulated genes and 114 downregulated genes. There were 44 DEGs obtained from GSE60993, including 4 upregulated genes and 40 downregulated genes. According to the MCC algorithm, the 10 most stable and highest score genes in the PPI network were selected as core genes. The hub genes involved in GSE29111 included BDKRB1, BDKRB2, CCL25, HRH1, KISS1, NPBWR1, GRPR, TACR3, HTR2B, and ORM1. The core genes involved in GSE60993 include IFI44, FFAR2, GNG10, CXCR2, MCEMP1, FPR2, CLEC4D, IFIT3, RSAD2, and PLSCR1. These identified core genes act as a whole and may play an important role in the early diagnosis of ACS.
GO enrichment analysis showed that the DEGs screened from GSE29111 were mainly enriched in the inflammatory response, G-protein-coupled receptor signaling pathway, integral component of plasma membrane, and plasma membrane. The GO entries of the DEGs screened from the GSE60993 were mainly related to neutrophil activation, immune response, inflammatory response, particles, plasma membrane, cytokine receptor activity, and cytokine binding. Over the past decades, the critical role of immune inflammatory processes in the occurrence, development, and deterioration of many cardiovascular diseases has received    [39]. Moreover, it triggers acute proinflammatory responses through a synergistic effect of multiple processes, which leads to the release of a variety of proinflammatory mediators and induces the recruitment of inflammatory cells to the infarction area [40]. In the animal model of AMI, the injured myocardial cells released interleukin-1 alpha (IL-1α), while interleukin-1 beta (IL-1β) increased after infarction [41]. The findings of our study suggest that the DEGs between AMI and UA are related to the inflammatory response, which provides a reference for their treatment.
The enrichment analysis of the KEGG pathway showed that the main enriched pathways of DEGs screened from the GSE29111 dataset were related to the interaction of neuroactive ligand-receptor (hsa04080), calcium signaling pathway (hsa04020), and inflammatory mediator regulation of TRP channels (hsa04750). Previous studies have shown that the nervous system regulation of cardiac activity has nothing to do with the physiological state of the heart [42]. However, the autonomic balance between sympathetic and parasympathetic systems plays an important role in the regulation of the cardiovascular system [43]. The facts show that the interruption of cardiac parasympathetic (vagus) activity is a common feature of various cardiovascular diseases including AMI [44]. In addition, some animal experiments have shown that increased vagus nerve activity has a protective effect on the heart of myocardial ischemia/reperfusion injury [45,46]. Wang et al. showed that in patients undergoing coronary artery bypass grafting, the cardioprotective effect of sevoflurane was related to the interaction of neuroactive ligand-receptor [47]. Chen et al. showed that the neuroactive ligand-receptor interaction pathway was closely related to human arrhythmogenic right ventricular cardiomyopathy [48]. Although the correlation between AMI and neuroactive ligand-receptor interaction has not been reported, this pathway may be a new mechanism of AMI, which is worth further study.
Calcium, as a common intracellular second messenger, is involved in regulating fertilization, gene transcription, secretion, and myocardial cells [49]. Ruiz-Meana et al. showed that calcium increased during myocardial infarction [50]. Garcia-Dorado et al. proposed to develop effective and reliable treatments to inhibit Ca 2+ -mediated myocardial cell       ROS and intracellular Ca 2+ concentration, and these characteristics make TRP channels effectively regulate inflammation [52]. Although neuroactive ligand-receptor interaction, calcium signaling pathway, and inflammatory mediator regulation of TRP channels have little research on ACS, they may be new ways of occurrence and development of AMI and UA, which is worthy of further study.
In the current study, the core genes enriched in the neuroactive ligand-receptor interaction with the smallest FDR value were discussed and analyzed in detail. HRH1, NPBWR1, BDKRB2, HTR2B, TACR3, BDKRB1, and GRPR are 7 core genes enriched in the neuroactive ligand-receptor interaction. Histamine receptor H1 (HRH1) belongs to the G-protein-coupled receptor family. Its gene products are expressed in many tissues such as smooth muscle and neurons [53,54]. Bhuiyan et al. found that HRH1 is mainly expressed in the neurons of the solitary tract nucleus, and histamine in the solitary tract nucleus may play a role in regulating cardiovascular homeostasis by activating H1 receptors expressed in the neurons of the solitary tract nucleus [55]. One study showed that the activation of HRH1 drove the formation of atherosclerotic lesions by increasing the vascular permeability of low-density lipoprotein [56]. Neuropeptide W (NPW) and neuropeptide B (NPB) are 2 structural and functional regulatory peptides. They are highly expressed in some brain regions and peripheral tissues (such as the heart, kidney, and stomach), but their distribution in tissues is not similar [57,58]. In addition, they act on target tissues through neuropeptide B/W receptor 1 (NPBWR1) and neuropeptide B/W receptor 2 (NPBWR2). NPB activates NPBWR1, while NPW stimulates the 2 receptors with similar potency and has different binding affinities [59,60]. NPW increases the calcium flow into the cells by acting on NPBWR1 and plays a role in regulating vascular myofibrillar tension, so it may be involved in the development of hypertension [61]. In a rat model, the results suggest that central NPW30 can increase sympathetic outflow and affect cardiovascular function [62]. In the past few years, a large number of literatures have shown that B1 bradykinin receptor (BDKRB1) and B2 bradykinin receptor (BDKRB2) are involved in a variety of hemodynamics and metabolism [63]. Marketou et al. found that early use of selective BDKRB2 agonist in the treatment of AMI in mice could   [64]. Therefore, the therapeutic value and potential risk strategy of BDKRB2 agonist in AMI deserve further study. As a neurotransmitter, 5-hydroxytryptamine (5-HT) not only plays a signal transduction role in immune cells but also regulates mood, behavior, and cardiovascular and gastrointestinal functions [65,66]. A study showed that 5-HT could regulate macrophage-mediated angiogenesis by reducing the expression of matrix metalloproteinase 12 in tumor-infiltrating macrophages [67]. In addition, it plays a role in the development of pulmonary hypertension depending on the expression of 5-hydroxytryptamine receptor 2B (HTR2B) in bone marrow progenitor cells [68]. de Las et al. showed that 5-HT could regulate macrophage polarization and maintain an anti-inflammatory state through 5-HT2B and 5-HT7 [69]. TACR3 is an important receptor protein, is expressed in the central nervous system, and activated the phosphatidylinositol-calcium second messenger system [70,71]. In addition, the involvement of TACR3 in human reproductive systems such as normal hypogonadism is evident [72]. It has been reported that neurokinin B (endogenous ligand of TACR3) is involved in the pathogenesis of Parkinson's disease [73]. Schooling's findings suggest that drugs with TACR3 antagonists can regulate the reproductive axis to reduce cardiovascular morbidity and mortality [74]. Additionally, in our study, we observed the TACR3 upregulated in the AMI patients in comparison to UA controls. Gastrin-releasing peptide (GRP) is a neuropeptide whose receptor GRPR is expressed in a variety of cell types and can activate signaling pathways that induce neutrophil migration in vitro and in vivo [75]. In addition, GRPR was found to be expressed in immune cells [76]. Recent studies have shown that selective GRPR antagonist (RC-3095) has anti-inflammatory effects on arthritis and sepsis models [77,78]. Interestingly, our research identified that HRH1, NPBWR1, BDKRB2, HTR2B, TACR3, BDKRB1, and GRPR are closely related to inflammation. This suggests that anti-inflammatory measures may be needed to treat AMI and UA.
The KEGG pathway enriched by the DEGs screened from the GSE60993 dataset is related to viral protein interaction with cytokine and cytokine receptor (hsa04061), human cytomegalovirus infection (hsa05163), and cytokinecytokine receptor interaction (hsa04060). Cytokines are extracellular molecules that transmit intercellular signals, mainly involved in cell differentiation and inflammation through binding to specific receptors on the cell surface [79]. Inflammation plays an important role in different stages of coronary atherosclerosis, such as plaque formation, progression, stability, rupture, and acute thrombosis [80]. The combined action of neutrophils and mononuclear macrophages and inflammatory factors promotes the occurrence and development of atherosclerosis and acute coronary events [81]. It has been reported that some functions of mononuclear macrophages and neutrophils are closely related to many cytokines expressed on their surface and corresponding receptors, such as adhesion, chemotaxis, and phagocytosis [82,83]. Many cytokines related to inflammation are involved in the development of ACS, such as TNF-α and IL-6 [84]. Boekholdt et al. showed that the increase of the IL-8 level was associated with the increased risk of coronary artery disease [85]. Therefore, we speculate that the role of the cytokine-cytokine receptor interaction signaling pathway in the progression of AMI and UA may be achieved by regulating the inflammatory response.
In this study, the hub genes enriched in these 3 pathways are C-X-C chemokine receptor type 2 (CXCR2), the receptor of interleukin-8 (IL-8). Neutrophils play an important role in innate immunity [86]. CXC chemokines and related receptors have strong effects on chemotaxis and neutrophil activation, allowing them to migrate to infected or injured sites [87]. Yan et al. detected and analyzed the differential gene expression between AMI patients and stable angina patients and found that the mRNA expression related to CXCR2 in AMI patients was significantly upregulated compared with stable angina patients [88]. A number of studies have shown that the occurrence and development of coronary atherosclerotic heart disease are closely related to monocyte chemoattractant protein 1 (MCP-1), IL-8, and other major chemokines [89][90][91]. This suggests that the vital DEGs play essential roles in the development and progression of AMI and UA via different signaling pathways. Collectively, most of the DEGs were inflammation-related genes. Expression analysis and the correlation of key DEGs will undoubtedly aid in the understanding of the roles of such genes in the development of AMI and UA. Notably, in order to get more precise correlation reports, we need to conduct a series of experimental studies to prove.
However, some limitations of our study should be considered. For instance, the differential genes analyzed by the two chips did not overlap, which may be caused by various reasons such as the source of the included samples, the annotation analysis platform, and the time of blood sampling. In addition, this study is based on the analysis of chip data; further experiments are needed to verify.

Conclusion
In summary, the integrated bioinformatical analysis could improve our understanding of DEGs between AMI and UA. The integrated bioinformatical analysis could improve our understanding of DEGs between AMI and UA.