Network Pharmacology to Explore Active Components and Pharmacological Mechanism of Action of Compound Longmaining for the Treatment of Myocardial Infarction

Jia Li Shanxi University of Traditional Chinese Medicine Wenfei Wang Shanxi University of Traditional Chinese Medicine Diaodiao Bu Shanxi University of Traditional Chinese Medicine Junbo Zou Shanxi University of Traditional Chinese Medicine Yao Wang Shanxi University of Traditional Chinese Medicine Yanzhuo Jia Shanxi University of Traditional Chinese Medicine Shangshang Yu Shanxi University of Traditional Chinese Medicine Jiahui Zheng Shanxi University of Traditional Chinese Medicine Xiaofei Zhang (  2051028@sntcm.edu.cn ) Shanxi University of Traditional Chinese Medicine https://orcid.org/0000-0002-5261-6101 Changli Wang Shanxi University of Traditional Chinese Medicine


Introduction
Cardiovascular disease is the leading cause of death among urban and rural residents in China, a trend that has increased recently [1]. In 2014, there were approximately 290 million patients with cardiovascular disease, which included 2.5 million myocardial infarction cases. In the United States, there are approximately 1.5 million cases of MI annually with a yearly incidence rate of approximately 600 cases per 100,000 people [2]. Although the prognosis of MI has been greatly improved, it is still the main cause of morbidity and mortality in the world [3].Therefore, it is important to continue to develop new therapeutics based upon the use of novel strategies to interrogate mechanisms involved in MI. Following MI, there is cardiomyocyte necrosis, the result of ischemia that causes an in ammatory response.
Oxidative stress mediated by in ammatory factors damages vascular endothelial cells, and leads to plaque rupture and hemorrhage [4]. These triggering events cause serious clinical manifestations.
Traditional Chinese medicines (TCMs) have the potential advantage of low toxicity and broad spectrum activity to prevent and treat a wide variety of diseases including MI. CLMN is composed of the following natural ingredients: Pueraria lobata, Rhizoma Dioscoreae Nipponicae, Ligusticum chuanxiong, propolis.
Active molecules identi ed in CLMN are avonoids, puerarin, alkaloids, lactones, phenols, organic acids, steroidal saponins, avonoids and terpenes [5]. It is an in-hospital preparation developed by Professor Tao Genyu of the a liated Hospital of Shaanxi University of Chinese Medicine. Since previous studies have con rmed that CLMN decoction has a signi cant therapeutic effect on MI [6], we used network pharmacology combined with GEO chip analysis and molecular docking to explore which components, targets and pathways are regulated by CLMN.
Network pharmacology can provide insight into "component-target-pathways" of drugs, and can help predict the correlation between drugs and diseases [7]. Microarray analysis has been widely used to identify differentially expressed genes (DEGs) and functional pathways involved in disease progression.
Lastly, molecular docking is a technique to visualize drug interactions with target proteins [8].
In this study, we used network pharmacology, microarray analysis and molecular docking to study and identify the pharmacological actions of CLMN for MI. In addition, we used docking analysis, with docking scores to characterize molecular interactions of compounds with target proteins. Finally, HE staining and immunohistochemistry were used to study e cacy and mechanisms of action of CLMN in a mouse model of MI. The ow chart of the study is shown in Fig. 1.

Acquisition of disease targets
In order to obtain gene data related to MI, we search GeneCards (https://www.genecards.org/) [14] using the keyword "Myocardial Infarction". Microarray technology and bioinformatics analysis allowed us to identify differentially expressed genes (DEGs). In this study, we used two chips, GSE48060 and GSE66360 from the GEO (https://www.ncbi.nlm.nih.gov/geo/) database [15] and analyzed them with R language "limma" installation package. The screening conditions were adj.P.Val ≤ 0.05, | logFC | ≥ 0.5 to screen out the DEGs of diseases and draw volcano maps. Using the "pheatmap" package, the rst 20 genes identi ed by the most signi cant up-regulation or down-regulation were obtained and heat maps were drawn.

Gene mapping
The target genes obtained from the GeneCards database and the DEGs obtained from the GEO database were intersected with the target genes corresponding to the active components of CLMN, resulting in an MI disease gene dataset. The Venn diagram between "compound-target-disease" was constructed by using the "Venn diagram" package of R language, and the common target genes were screened by mapping.
Construction and Analysis of the protein-protein interaction (PPI) Network The target genes of CLMN components were introduced into the STRING (https://string-db.org/) database [16], and the research species was set as "Homo sapiens". In the setting, the highest con dence level was set to 0.9, the free target genes were hidden, and the TSV format le of the relationship diagram was derived after obtaining the target-protein interaction relationship. We imported the TSV le into Cytoscape3.7.2 [17], adjusted the color of the nodes according to the degree value, in order to more intuitively show the important CLMN targets.

Construction of drug active ingredient-core target network
The target genes of the active components of CLMN were imported into Cytoscape3.7.2, the node representing the drug active component was set to "source node", the node representing the disease target gene was set to "target node", and the network attribute was set to "interaction type" to draw the "drug active ingredient-core target" network.

GO and KEGG enrichment analysis
The core target genes of drug-disease were transformed from Gene Symbol to entrezID, by R language, and the data was processed by "BiocManager". The bar chart and bubble chart of GO and KEGG enrichment analysis were obtained under the condition of p < 0.05.

Molecular docking
The protein structure of the key target was downloaded from the Protein Date Bank (PDB, https://www.rcsb.org/) database [18], and we download the sdf le for the ligand (active component and positive drug corresponding to the target) from the Pubchem database. The LibDock tool of Discovery Studio4.0 software was used to dock the key target protein with its corresponding active component and positive drug, and the score was analyzed.

Preparation of CLMN decoction
The CLMN decoction was prepared as described previously. 18g of Pueraria lobata, 12g of Chuanxiong, and 6g of propolis, were added 14 volumes of water, heated, boiled twice and the ltrate was concentrates to 0.32g/mL [19].

Experimental animals
Male Balbc mice weighing 22-28 g and 10-14 weeks old were purchased from Chengdu Dashuo Experimental Animal Co., Ltd., Chengdu, China.The experimental animal license number is SYXK (chuan) 2020-030. Animals were maintained under standard laboratory temperature and humidity conditions, and had free access to food and water throughout the study. This experiment was approved by the Animal Ethics Committee of Shaanxi University of Chinese Medicine.
Sixty mice were randomly divided into 5 groups: sham operation group, MI model group, low-dose, medium-dose and high-dose CLMN groups. The mice in the sham operation group and MI model group were given normal saline, and the mice in the CLMN group were given CLMN decoction 3.26 g/kg/d (low dose), 6.52 g/kg/d (medium dose) and 13.04 g/kg/d (high dose) for 7 days, respectively.

The MI mouse model
The MI model was established as described previously. Brie y, we opened the pericardium and located the left anterior descending coronary artery (LAD), and inserted a needle, then the coronary artery of the left anterior descending branch of the left atrium was ligated with 8/0 mono lament polypropylene suture. This resulted in permanent ischemia of the artery below the ligation line. For the sham group, the needle was inserted into the left anterior descending branch of the coronary artery without ligation after thoracotomy [20,21].

Tissue collection & HE staining
Following MI, mice were anesthetized and the fresh heart tissue was quickly taken out and frozen in liquid nitrogen for pathological examination. Myocardial tissue was then xed in 4% paraformaldehyde and embedded in para n. Thereafter, the tissue was cut into 5um thick sections and stained with hematoxylin-eosin (HE), and histopathological examination was performed under a light microscope (OLUMPUS Japan) at 200 magni cation.

Immunohistochemical analysis
The myocardial tissue was cut into 5um slices, placed in xylene, and hydrated with an ethanol gradient (100% 95%, 80%, 70%, pure water). After antigen repair, slices were incubated in H 2 O 2 for 15 min in the dark, and the primary antibody diluted by 5%BSA (dissolved in PBS) was added, and slices were incubated at 4℃ overnight. The following day, the second antibody was added and incubated at room temperature for 30 min, and then washed with PBS for 3 times (5 minutes each). DAB-H 2 O 2 was then used until the color develops. Slices were then incubated in hematoxylin for 2 minutes, washed with water, differentiated with hydrochloric acid and ethanol for 3 s, washed with running water for 15 min, and dehydrated. The expression of TNF-α, TRAF2 and IκBα was evaluated by quantitative grayscale scanning with a medical image analysis system after taking pictures under high magni cation visual eld (400×).

Statistical analysis
Data was analyzed using SPSS Statistics 26 software, and the results are expressed as mean ±SD. When the variance was homogeneous, one-way analysis of variance was used to analyze the mean between groups, and P <0.05 was statistically signi cant.

Collection of CLMN active ingredients and targets
In this study, 33 active compounds were collected from the TCMSP, PubChem and Swiss Target Prediction databases, targeting 2035 proteins, including 9 active compounds in Pueraria lobata targeting 506 proteins, 3 active compounds in Rhizoma Dioscoreae Nipponicae targeting 94 proteins, 12 active compounds in Ligusticum chuanxiong targeted 443 proteins, and 9 active compounds in propolis targeting 992 proteins.
Screening core targets 3652 targets were retrieved from the GeneCards database, and the data from GSE48060 and GSE66360 chips and platform were downloaded from the GEO database. 898 signi cant differential genes were obtained. We drew both DEGs volcano and heat maps, as shown in Fig Protein interaction network (PPI) analysis The 29 targets described above were imported into the STRING database and were view in the network diagram "Analysis" of the database. The PPI network consisted of 29 nodes and 34 edges. The average node degree was 2.34 and the average clustering coe cient was 0.521 < 2.83e ~ 14, which indicated that the target genes had potential interactions. The TSV le data of protein interaction were obtained, and then the data was imported into the Cytoscape software to construct the protein interaction network diagram, as shown in Fig. 3a. The "Network Analyzer" tool was used to score the proteins. In the Figure, the top 10 proteins are represented by a red oval, and the rest are represented by a blue oval. TNF, JUN, IL1B, VEGFA, CCR5, MMP9, NFKBIA, proteins with high scores play an important role in the regulation of the network and be the key targets of CLMN for the treatment of MI.
Construction of "active ingredient-core target" network Using the "merge" tool within the Cytoscape3.7.2 software, the network diagram for 32 drug active components and 29 MI targets was constructed (Fig. 3b). Among them, the ovals of different colors were used to represent each active ingredient from different TCMs, and the triangle represents the disease targets. The network consists of 61 nodes and 123 edges. From the graph, it is apparent that a single active component corresponds to multiple targets, and one target can also correspond to multiple active components. Taken together, CLMN therapy for MI has characteristics of multi-components and multitargets. Using the Network Analyzer analysis tool, the top components were puerarin, Kaempferol, apigenin, daidzein and top targets were PTGS2, ABCB1, MMP9, JUN and CZCR1, suggesting that all may play an important role in the mechanism of CLMN in the treatment of MI.

GO and KEGG enrichment analysis
Using R language, 29 common targets were selected for GO and KEGG enrichment analysis with signi cance of p<0.05. We correlated CLMN treatment with neuroin ammation, bacterial molecules, lipopolysaccharide and cells to external stimuli and regulation of the production of immunoreactive cytokines. KEGG pathway enrichment of the rst 20 signal pathways indicates that CLMN treatment involves NF-kappa B, TNF, IL-17. Figure 4 represents the drawing of a GO analysis bubble chart of CLMN (Fig. 4a-c), KEGG pathway analysis bar chart of CLMN (Fig. 4d) and NF-kappa B signal pathway map enriched in the rst place (Fig. 4e). Based on the above analysis, the data suggests that CLMN results in a multi-modal pharmacological pro le for CLMN

Molecular docking analysis
Five important target proteins, TNF, IL1B, PTGS2, VCAM1 and NFKBIA, which are ranked rst in the PPI network and enriched in the rst NF-kappa B pathway, were docked with their corresponding compounds by Discovery Studio4.5 software, and compared with the positive drugs of the corresponding target proteins. The docking results are shown in Table 1, and the interaction between them is shown in Fig. 5.

HE staining results
The results of HE staining are shown in Fig. 6. The structure of cardiomyocytes in the sham operation group was normal and orderly, with no pathological changes. However, in the MI model group, the cardiomyocytes were injured and disordered, with a large number of in ammatory cell in ltrates, and large areas of myocardial infarction and myocardial brosis. With the low dose of CLMN, more in ammatory cell in ltrates were observed, but with the middle and high dose groups of CLMN, in ammation was alleviated, the area of myocardial infarction was signi cantly reduced, and cardiomyocyte brosis was inhibited.

TNF-α, TRAF-2 and IkB α proteins in Myocardial Infarctions
The results of immunohistochemical staining (Fig. 7) showed that the expression of TNF-α and TRAF-2 proteins were increased by MI, while IkBα protein decreased. The middle and high dose groups of CLMN signi cantly decreased the expression of TNF-α and TRAF-2 proteins and increased the expression level of IkBα protein.

Discussion
MI causes an in ammatory response mediated by cytokines [22], which may decrease myocardial contractile function and lead to myocardial hypertrophy, brosis and remodeling. In this paper, based upon network pharmacology combined with GEO chip, 29 core targets of CLMN were screened, and the results of docking between the core targets and active components selected by molecular docking software were stable.
The MI mouse model was established to verify important targets of the NF-kappa B signal pathway. TNF, TRAF-2 and NFKBIA mere assayed by immunohistochemical methods, and the mechanisms of action of CLMN for the treatment of MI were analyzed. CLMN can regulate multiple target and pathways through Daidzein-4,7-diglucoside, rutin, puerarin. The components mainly regulate the expression of NF-kappa B signal pathway and the expression of TNF, IL1B, PTGS2, VCAM1, NFKBIA. We showed that the TNF-α and TRAF-2 were signi cantly decreased and the expression level of IkBα (NFKBIA) protein was signi cantly increased in the myocardium of mice treated with CLMN.
PPI network analysis showed that TNF, JUN, IL1B, PTGS2, MMP9, NFKBIA genes have a large degree of freedom. Target proteins appear to be related and combine to play a synergistic effect of CLMN. Blocking IL1β signal transduction can reduce cardiac brosis and MMP expression by reducing leukocyte recruitment and in ammation [23]. MMP9 accelerates the removal of necrotizing cardiomyocytes by degrading collagen, bronectin and other extracellular matrix components, but MMP9 can also prolong in ammation by preventing neutrophil apoptosis. In turn, collagen and extracellular matrix components are further degraded, which may damage tissue integrity and lead to infarction expansion, which is not conducive to left ventricular remodeling [24]. PTGS2 has been shown to have anti-in ammatory properties and regulate vascular reactivity [25].
There are 32 components in CLMN that may be responsible for the therapeutic effect of CLMN, including puerarin, kaempferol and apigenin that have high scores in active ingredient-core target interactions. Studies have found that puerarin can protect cardiomyocytes by reducing the in ammatory response in cardiomyocytes, and puerarin can promote the release of vasoactive substance NO in cardiomyocytes, increase blood ow in cardiac vessels, and attenuate myocardial injury caused by continuous ischemia [26]. Kaempferol is a avonoid, which not only has anti-atherosclerosis, anti-in ammatory, anti-oxidative and antibacterial effects, but can also reduce hypertrophy, interstitial brosis and cardiomyocyte apoptosis [27][28][29]. Ren et al. [30] previously showed that apigenin can signi cantly inhibit the expression of miR-33 to promote the expression of ABCA1 and increase cholesterol e ux, reduce blood lipids and alleviate in ammation.
The biological process enrichment analysis of CLMN shows that target-related biological processes are mainly involved in the regulation of DNA binding transcription factor activity and response to lipopolysaccharide, bacterial molecules, neuroin ammation and cell response to external stimuli. The enrichment pathway of KEGG suggests important involvement of NF-kappa B, TNF, and IL-17, which supports the observation that the pathogenesis of MI is related to in ammation. TNF-α is considered to be one of the most important in ammatory cytokines [31]. Mature cardiomyocytes have the ability to produce TNF-α and its proteins under some stress-related conditions, and TNF-α can directly affect cardiac function and participate in a variety of cardiovascular pathological processes [32]. NF-kappa B pathway is the most deeply studied signal transduction pathway in the development of in ammation. NFkappa B signal pathway is triggered by activated TNF-α. TNF-α activates NF-KB by binding to the TNF-R1 receptor and the TNF receptor to the dead zone, TRAF2 and serine /threonine kinase RIP. In the activated NF-kappa B signal pathway, after IkBα is phosphorylated, the nuclear localization signal on NF-KB is exposed, and NF-kB enters the nucleus from the cytoplasm and binds to the KB site on the regulatory gene promoter to induce the transcription of the target gene, thus inducing a large number of in ammatory cytokines to participate in the in ammatory response [33]. IkBα is a key protein in the NFkappa B pathway, which is necessary for the activation of NF-KB mediated by TNF-α. It is the main regulatory inhibitory protein of NF-KB (p50-p65) in vivo. There is a KB site in the promoter of IkB α gene, so its synthesis is also regulated by NF-KB, thus forming a negative feedback regulation of NF-KB.        Effects of CLMN Pretreatment on the expression of TNF-α, TRAF-2, IkBα protein in Myocardial tissue of mice with MI(Immunohistochemistry, x400) and the changes of TNF-α, TRAF-2, IkBα protein expression in Myocardial tissue of mice.