Analyzing the multi-target pharmacological mechanism of folium Artemisia argyi acting on breast cancer: a network pharmacology approach

Background Folium Artemisia argyi (FAA) is a traditional Chinese herbal medicine that is widely used in the clinic. However, the underlying mechanisms of its anticancer effects have not been fully elucidated. Methods In this study, we applied a network pharmacology approach to identify the potential mechanisms of FAA against breast cancer. To be specific, we screened the active ingredients and potential targets of the FAA through the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database. Meanwhile, we employed the oral bioavailability (OB) and drug-likeness (DL) to search for potential bioactive compounds of FAA. Breast cancer-related target genes data were gathered from the GeneCards and Online Mendelian Inheritance in Man (OMIM) databases, and the protein-protein interaction (PPI) data were acquired from the Search Tool for the Retrieval of Interacting Genes (STRING) database. In addition, we constructed the network and performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Enrichment Analysis. Results We obtained a total of nine active ingredients and 236 potential targets from FAA to construct a network, which showed that quercetin served as the major ingredient in FAA. AKT1 (RAC-alpha serine/threonine-protein kinase), MYC (Myc proto-oncogene protein), CASP3 (Caspase-3), EGFR (Epidermal growth factor receptor), JUN (Transcription factor AP-1), CCND1 (G1/S-specific cyclin-D1), VEGFA (Vascular endothelial growth factor A), ESR1 (Estrogen receptor), MAPK1 (Mitogen-activated protein kinase 1), and EGF (pro-epidermal growth factor) were identified as key targets of FAA in the treatment of breast cancer. The PPI cluster demonstrated that AKT1 was the seed in this cluster, indicating that AKT1 played a crucial role in connecting other nodes in the PPI network. This enrichment demonstrated that FAA was highly related to signal transduction, endocrine system, replication and repair, as well as cell growth and death. The enrichment results also verified that the underlying mechanisms of FAA against breast cancer might be attributed to the coordinated regulation of several cancer-related pathways, such as the MAPK and mammalian target of rapamycin (mTOR) signaling pathways, among others. Conclusions This study identified the potential targets and pathways of FAA in the treatment of breast cancer using a network pharmacology approach, and systematically elucidated the mechanisms of FAA in the treatment of breast cancer.


Introduction
Breast cancer is the most frequent malignancy occurring in women with 1.38 million new cases each year and nearly 0.46 million related deaths globally [1,2]. According to current projections, there will be approximately 3.2 million new cases per year by 2050 [3]. As we all know, breast cancer is a heterogeneous disease with differences in occurrence, development, treatment and prognosis [4]. At present, appropriate comprehensive adjuvant treatments containing chemotherapy, radiotherapy, endocrine and targeting HER2 therapy are widely put into use, according ve major molecular subtypes [5]. However, these treatments are costly and usually make a series of short-and long-term side effects, such as febrile neutropenia [6], alopecia [7], peripheral neuropathy [8] and cardiotoxicity [9], all of which signi cantly decrease the patient's quality of life. Not to mention those older patients in the terminal stages, the more intolerant these adverse reactions.
Folium artemisiae argyi (FAA), commonly called wormwood, is a traditional Chinese herb medicine which has antipyretic, analgesic and hemostatic effects [10]. It is used internally for warming channels, arresting bleeding, dispelling cold and relieving pain, while externally for eliminating dampness and relieving itching [11] for thousands of years. Recently, growing attention has been paid to the role of Traditional Chinese medicine in the prevention and treatment of cancer because of various limitations of Western medicine, such as the toxicity and adverse side effects [12]. At the same time, Chinese herbs can target multiple points to achieve synergistic actions [13,14]. According to pharmacology research, FAA contains multiple active chemical constituents, such as avonoids, terpenoids, phenolic acids and volatile oils [15,16] and exhibits a variety of effects, such as anticancer, anti-in ammatory and antioxidative [17,18]. For example, Sha et al. suggested that FAA inhibited proliferation and promoted apoptosis in breast cancer cells through Bcl-2 family proteins and the MEK/ERK pathway [10]. It was also reported that FAA had dose-dependent inhibitory effect on hepatoma cells [11]. However, although many studies veri ed that FAA showed remarkable antitumor functions, the underlying mechanisms have not yet been absolutely understood.
It is widely known that herbal medicines include multi-component, multi-target and multi-pathway features [19,20]. Traditional Chinese medicine network pharmacology is a systematic research method based on the interaction network of herbs, compounds, targets, diseases and genes [21]. This approach lays emphasis on the integration of bio-informatics, systems biology and pharmacology, which not only explains the complex interactions between herbs and diseases at a systematic level, but also conforms to the systematic and holistic perspective of the Traditional Chinese medicine theory [22,23]. Thus, we utilized a network pharmacology approach to explore the pharmacological mechanisms of FAA as a treatment for breast cancer in this study. At rst, we screened for active ingredients of FAA by estimating their oral bio-availability (OB) and drug-likeness (DL) [24]. Next, we picked out the common targets shared by the FAA compound targets and the breast cancer related targets via two databases (Genecards and the Online Mendelian Inheritance in Man database (OMIM)) and then constructed the network by investigating the potential interactions between the various target nodes. In addition, protein-protein interaction (PPI) data were obtained from the Search Tool for the Retrieval of Interacting Genes (STRING) database, and enrichment analyses (gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG)) were performed to explore the potential mechanisms of FAA against breast cancer. To summarize, this study aimed to identify the potential targets and pathways of FAA as a treatment against breast cancer using the network pharmacology approach, and systematically elucidate the mechanisms of FAA in the treatment of breast cancer.

Active Ingredients and Targets against Breast Cancer in FAA
The ingredients in FAA were acquired from the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP). It serves as a systematic platform to study herbs, including the identi cation of compounds and the screening of compound targets [25]. In addition, to identify the corresponding targets of FAA compounds against breast cancer, the TCMSP database was used to dig out potential targets.
Finally, 9 active herbal ingredients of FAA were selected (Table 1) by linking the active ingredients of FAA to the targets against breast cancer. A total of 236 targets of FAA compounds were obtained in total (speci c targets were not shown).

Pharmacokinetic Predictions
In pharmaceutical study, ADME (absorption, distribution, metabolism and excretion) is a critical pattern to identify [24]. Therefore, we employed two major ADME-related properties, namely, the OB and DL to search the potential bio-active compounds of FAA in this study. Ingredients with OB≥30% and DL≥0.18 were considered as suggested drug screening criteria. The detailed information for all ingredients before screening was listed in Table S1.

Potential Target Genes of Breast Cancer
The data for the breast cancer related target genes were gathered from Genecards and the OMIM. The species was set to Homo sapiens. Genecards is an extensive platform that provides insight into predicted and annotated human genes. All of the gene-centric data are collected from 150 web resources, including genetic, genomic, proteomic, transcriptomic, and functional information [26]. Search strategy: Set the keyword as "breast cancer" and the score >30 after logging in to Genecards. The detailed information was listed in Table S2. The OMIM is a comprehensive, authoritative, and timely knowledgebase which links and catalogues all known diseases with a genetic component and provides further references to genomic analyses of catalogued genes [27]. Search strategy: Choose gene map at the website, and then set the keyword as "breast cancer". The detailed information was listed in Table S3.

PPI Data
We acquired the PPI data from the STRING database. It de nes PPI with con dence ranges for data scores (high >0.7; medium >0.4; low >0.15) [28]. In our study, we selected a con dence score of >0.4 to construct our PPI network.

Network Construction
The PPI network has been widely applied to show lots of different interactions of proteins in the context of complex diseases [23,29], including breast cancer, prostate cancer, lung cancer, gastric cancer, etc. In this study, we constructed the network as follows: (1) we acquired the targets shared by the FAA compound targets and the breast cancer related targets, (2) We put these targets into STRING database and obtained the FAA against breast cancer targets PPI network, and (3) we exported the PPI results as a simple tabular text output (.tsv) and then imported the .tsv le into Cytoscape (version 3.6.1) to reconstruct the network in order to achieve better visualization and understanding for further analysis [30].

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Enrichment Analysis
In this study, we used the ClusterPro ler package of R3.5.2 to perform GO enrichment analysis of the targets. The higher the score, the greater the importance of the genes represented in the list [31]. The ClusterPro ler package of R3.5.2 was also used to analyze KEGG pathway enrichment of overlapping target genes. KEGG analysis was used to explore the biological pathways and potential biological functions on the basis of the enrichment analysis of functional items [32].

Active Ingredients
In this study, we acquired a total of 9 active ingredients of FAA after ADME identi cation. Detailed information was shown in Table 1 (all Mol IDs could be tracked in the TCMSP database). All the FAA compounds before screening were presented in Table S1.

FAA Compound-Target Network
To further uncover the potential pharmacological mechanisms of FAA against breast cancer, target genes common to both active ingredients of FAA and breast cancer were selected in different databases. A total of 75 genes (Table S4) belonging to both the FAA target gene and breast cancer target gene networks were screened via Venn analysis (Figure 1a). The compound-target network was presented in Figure 1b, including 82 nodes and 171 edges, with a network density of 0.051 and a network diameter of 3. The detailed information of this network was depicted in Table 2. As could be seen from gure 1b, quercetin was the most critical component of FAA, which was connected to the most targets. It was one of the avonoids from natural plant and exhibited a variety of activities such as antioxidant, anti-in ammatory, antiviral and antimicrobial through multiple signal transduction pathways [33,34]. Several studies had validated that quercetin could inhibit various tumor progression, including breast cancer [35], prostate cancer [36], gastric cancer [37], ovarian cancer [38] and colorectal cancer [39]. Moreover, there were also some literatures reported that it not only had a synergistic effect when combined with chemotherapeutic or radiotherapy agents, but also could reduce an expected adverse side effect and a toxic reaction [40].
In addition, the results showed that many targets were affected by two or more compounds. For instance, Prostaglandin G/H synthase 1 (PTGS1) and Prostaglandin G/H synthase 2 (PTGS2) were both modulated by quercetin, stigmasterol, mandenol, etc. Constitutive PTGS1 and inducible PTGS2 belonged to two isozymes of PTGS which had pivotal effects both as a peroxidase and a dioxygenase [41]. Other study suggested that PTGS2 might inversely control the breast cancer metastasis and chemoresistance through the regulation of EMT, apoptosis and senescence [42]. And nuclear receptor coactivator 2 (NCOA2), a member of the p160 family, performed key roles in many different physiological and pathological processes, including cell growth, energy metabolism, endocrine regulation, and circadian rhythm [43]. More importantly, the expression of the NCOA2 gene played crucial roles in the development, progression, and metastasis of malignant tumors, such as breast cancer [44]. In prostate cancer patients, high expression of NCOA2 was more likely to relapse after androgen-deprivation therapy [45].
Similarly, Beta-2 adrenergic receptor (ADRB2), Gamma-aminobutyric acid receptor subunit alpha-1 (GABRA1), Heat shock protein HSP 90, Progesterone receptor (PGR) and Sodium channel protein type 5 subunit alpha (SCN5A) could also be regulated by more than two active ingredients. Not only was obtained an approximate observation of the relationship between these active ingredients and targets, but also discovered the potential pharmacological effects of FAA from this network of Figure 1b.

PPI Network
To explore the underlying mechanisms of FAA as a therapy against breast cancer, a PPI network of the FAA compound targets against breast cancer was constructed by connecting the targets of FAA compound and the breast cancer. First, we obtained a total of 75 target genes which belonged to both the FAA target gene and breast cancer target gene and got targets symbol names by uniprot. Next, all of these 75 target genes were imported into the STRING database to generate the PPI results (settings: Homo sapiens and con dence >0.4). The original STRING PPI network was presented in Figure S1. Then, we imported the PPI data generated in STRING into Cytoscape (version 3.6.1).
As Figure 2a showed, this PPI network included 75 nodes and 1247 edges, with a network diameter of 3, a clustering coe cient of 0.733 and an average number of 33.253 neighbors. The average node degree was 33.3 (both the different colors and the size of the circles indicated the degree). The detailed information of this network was displayed in Table 3. All target degrees were calculated using this network. In Figure 2b, the 10 targets with the greatest degrees were AKT1 (degree = 67), MYC (degree = 65), CASP3 (degree = 63), EGFR (degree = 62), JUN (degree = 61), CCND1 (degree = 60), VEGFA (degree = 60), ESR1 (degree = 59), MAPK1 (degree = 57) and EGF (degree = 55). As shown in Figure 3c, the cluster consisted of 68 nodes and 1155 edges. The average node degree was 34 and the clustering coe cient was 0.77. The red diamond in Figure 3c, AKT1, was the seed in this cluster and interacted with other FAA targets. This gure intuitively indicated that AKT1 played an important role in connecting other nodes in this PPI network. It was well-known that the serine/threonine kinase AKT1, one of the three isoforms in the Akt family, had emerged as a downstream effector of PI3K [46]. AKT inhibited apoptosis by suppressing the actions of BAD and caspase-9 [47]. In breast cancer, AKT1 activation accelerates cell proliferation, whereas Akt1 inhibition promotes Epithelial-to-Mesenchymal Transition [48].

GO Enrichment
To further discuss the multiple mechanisms of FAA as a treatment against breast cancer, we conducted GO enrichment analysis on the 75 common targets shared by the FAA compound targets and the breast cancer related targets [31]. To be more speci c, the top 30 targets are as follows (Figure 2b Table 4. Thus, we speculated that FAA probably executed its pharmacological effects on breast cancer by simultaneously involving these molecular functions.

KEGG Enrichment
Meanwhile, we further carried on KEGG [32] enrichment analysis on the 75 common targets in order to clarify the integral regulation of FAA for the treatment of breast cancer. We obtained 74 pathways in total which belonged to several categories, including human diseases, cellular processes, and drug resistance, among others, of which the top 30 signi cantly enriched KEGG targets were presented (adjusted p-value <0.001) in Figure 4. In the cancer-related disease, prostate cancer (hsa05215), bladder cancer (hsa05219), pancreatic cancer (hsa05212), breast cancer (hsa05224), colorectal cancer (hsa05210), non-small cell lung cancer (hsa05223), small cell lung cancer (hsa05222), gastric cancer (hsa05226), endometrial cancer (hsa05226), renal cell carcinoma (hsa05211), thyroid cancer (hsa05216), and small cell lung cancer (hsa05222) data were processed using KEGG enrichment analysis. Detailed KEGG information was shown in Table 5. This result showed that FAA had the highly potential to treat a wide range of cancers, such as breast cancer [10], prostate cancer [49], bladder cancer [50], colorectal cancer [51], and gastric cancer [52], which were con rmed by other researchers. Further, the results in Figure 5 also veri ed that these signaling pathways remarkably enriched by the potential targets of FAA in breast cancer were strongly associated with signal transduction, endocrine system, replication and repair, cell growth and death, most of which played a essential role in the development and progression of cancers, such as pathways in PI3K/AKT signaling pathway (hsa04151), Estrogen signaling pathway (hsa04915), MAPK signaling pathway (hsa04010), mammalian target of rapamycin (mTOR) signaling pathway (hsa04150), apoptosis (hsa04210) and cell cycle (hsa04110). Therefore, we speculated that the underlying mechanism of FAA against breast cancer might be attributed to coordinated regulation of several cancerrelated pathways.

Conclusions
So far, though many researches veri ed that FAA exhibited arresting antitumor activities, the underlying mechanisms of its antitumor activities had not yet been totally understood. The network pharmacology emphasized the integration of bioinformatics, systems biology and pharmacology, which not only explained the complex interactions between diseases and Chinese herbs at a systematic level, but also demonstrated to the systematic and holistic perspective of the traditional Chinese medicine theory [12]. To better explore the pharmacological mechanisms of FAA as a treatment for breast cancer, we applied the network pharmacology approach to dig out the potential mechanisms of FAA as a treatment against breast cancer by compound-target network construction, PPI network, GO and KEGG enrichment analysis. We took advantage of OB and DL to explore the potential active ingredients of FAA. Until now, the studies on the pharmacokinetics of FAA were few. Eunjeong et al. [17] found that FAA had the anticancer activities through the inhibition of cell growth and induction of apoptosis in breast cancer cells. In this study, we acquired 9 active ingredients and 236 potential targets from FAA in total, and validated a synergistic herb strategy featuring multi-component, multi-target and multi-pathway characteristics. The compound-target network con rmed that quercetin served as the major ingredient in FAA. Moreover, the PPI network provided information concerning the source of the interactions. The PPI analysis indicated that FAA had a signi cant effect on breast cancer by in uencing the whole biological network, including targets such as AKT1, MYC, CASP3, EGFR, JUN, CCND1, VEGFA, ESR1, MAPK1, and EGF. The PPI cluster demonstrated that AKT1 was the seed, suggesting that AKT1 played a crucial role in connecting other nodes in the PPI network. Next, the enrichment analysis presented that FAA was strongly related to signal transduction, endocrine system, replication and repair, cell growth and death. The enrichment results also showed that the underlying mechanism of FAA against breast cancer might be attributed to coordinated regulation of several cancer-related pathways, such as MAPK signaling pathway, mammalian target of rapamycin (mTOR) signaling pathway, among of others.
In conclusion, this study applied a network approach demonstrating how FAA compounds alter different pathways against breast cancer, which was supplementary to other researches on drugs against breast cancer. Furthermore, we con rmed that FAA substantially in uenced lots of breast cancer related targets, a nding which was consistent with present cancer study trends showing that the occurrence and development of breast cancer was the gradual accumulation of distinct genome modi cations in cancer cells [53,54]. We fully expect that our research can help to promote the employment of network pharmacology in uncovering the potential mechanisms of anticancer Chinese herbs, and provide clues to assess the synergy of herbs in the treatment of other complex diseases, especially cancer.