Identification of Active Compounds of Mahuang Fuzi Xixin Decoction and Their Mechanisms of Action by LC-MS/MS and Network Pharmacology

The decoction is an important dosage form of traditional Chinese medicine (TCM) administration. The Mahuang Fuzi Xixin decoction (MFXD) is widely used to treat allergic rhinitis (AR) in China. However, its active compounds and therapeutic mechanisms are unclear. The aim of this study was to establish an integrative method to identify the bioactive compounds and reveal the mechanisms of action of MFXD. LC-MS/MS was used to identify the compounds in MFXD, followed by screening for oral bioavailability. TCMSP, BindingDB, STRING, DAVID, and KEGG databases and algorithms were used to gather information. Cytoscape was used to visualize the networks. Twenty-four bioactive compounds were identified, and thirty-seven predicted targets of these compounds were associated with AR. DAVID analysis suggested that these compounds exert their therapeutic effects by modulating the Fc epsilon RI, B-cell receptor, Toll-like receptor, TNF, NF-κB, and T-cell receptor signaling pathways. The PI3K/AKT and cAMP signaling pathways were also implicated. Ten of the identified compounds, quercetin, pseudoephedrine, ephedrine, β-asarone, methylephedrine, α-linolenic acid, cathine, ferulic acid, nardosinone, and higenamine, seemed to account for most of the beneficial effects of MFXD in AR. This study showed that LC-MS/MS followed by network pharmacology analysis is useful to elucidate the complex mechanisms of action of TCM formulas.


Introduction
Allergic rhinitis (AR), an immunoglobulin E-(IgE-) mediated inflammatory disease, seriously impairs the quality of life of patients [1]. Epidemiological surveys show that AR affects more than 20% of the world's population and its incidence has progressively increased in developing countries [2]. Currently, treatments based on Western medicine alleviate the symptoms of AR but do not cure. Once drug treatments stop, the disease relapses [3]. Traditional Chinese medicine (TCM) has long been used as effective therapeutic interventions in Asia, particularly China [4]. TCM is primarily based on the use of compound formulas, utilizing a combination of herbs or their extracts for improved efficacy.
Research into the active and effective compounds of TCM promotes the development and design of new therapeutic drugs. e decoction is the main form of TCM administration. e identification of the chemical components in a decoction underpins research on the mechanism of action of TCM. Mahuang Fuzi Xixin decoction (MFXD), a classical Chinese herbal formula, is widely used to treat AR. MFXD consists of Mahuang (Herba Ephedrae), Fuzi (Radix aconiti lateralis praeparata), and Xixin (Radix et Rhizoma Asari), which are boiled in water at specific proportions before administration. Pharmacological studies have shown that MFXD exerts anti-inflammatory and antiallergic effects by preventing the release of mediators from macrophages and

Materials and Methods
2.1. Materials. HPLC-grade acetonitrile and methanol were procured from Merck (Darmstadt, Germany). Formic acid was procured from Sigma-Aldrich (MO, USA). Other chemicals were of analytical grade. e decocting pieces of Mahuang, Fuzi, and Xixin were purchased from Kangmei Pharmaceutical Co., Ltd. (Guangzhou, China) and met the standards of the Chinese Pharmacopeia. Voucher specimens (no. 20191117) were deposited at the authors' laboratory at Southern Medical University.

Preparation of MFXD.
MFXD consisted of Mahuang, Fuzi, and Xixin at a weight ratio of 2 : 3 1. Mahuang was immersed in distilled water (15 times the total weight) for 30 min and boiled for 20 min. Fuzi and Xixin were then added to the suspension, which was simmered for another 90 min. e liquid extract obtained was concentrated to 1.52 g/mL under reduced pressure.

Construction of the Chemical Ingredient Database.
We processed the chromatogram by matching the data with an in-house Orbitrap Traditional Chinese Medicine Library (OTCML), which provided a list of compounds with their formulas and MS/MS fragment modes. Based on errors less than 5 ppm and MS/MS fragment matching, we identified compounds in decoction prepared from MFXD.

Oral Bioavailability (OB)
Screening. OB is an important pharmacokinetic property used to assess the rate and percentage of an orally administered drug that has been absorbed into the blood circulation to produce pharmacological effects. Important parameters of the identified compounds, such as Caco-2 cell permeability, human intestinal absorption (HIA), and OB limits F (F-20%, F-30%), were screened using ADMETlab, a platform for systematic ADMET evaluation based on a comprehensively collected ADMET database [15].

Potential AR-Associated Targets of the Compounds in MFXD.
Predicting whether a compound interacts with intended targets is a critical phase of drug discovery [16]. e targets of compounds in MFXD were obtained from the Traditional Chinese Medicine System Pharmacology (TCMSP, http://lsp.nwu.edu.cn/tcmsp.php) and BindingDB (https://www.bindingdb.org/bind/index.jsp) databases. After the deletion of redundant hits, the remaining protein targets were standardized by their ID and gene symbols in the UniProtKB database (http://www.uniprot.org/) [17].

Construction of the Target Protein-Protein Interaction (PPI) Network.
To determine whether the therapeutic targets of MFXD are associated with AR, we intersected the drug targets and disease targets to obtain the common targets, which were considered potential therapeutic targets. e common target proteins were used to construct the PPI network on the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) 11.0 platform (https:// string-db.org/) using the minimum required interaction score of 0.400. Subsequently, the topological property of the PPI network was analyzed using plugins in Cytoscape 3.7.2. Genes were ranked to define hub genes using cytoHubba with the maximal clique centrality (MCC) algorithm. e module was extracted by MCODE using a node score cutoff of 0.2 and K-core of 2.

Target Pathway and Enrichment Analysis.
To analyze the gene ontology (GO) functional annotations and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of genes and their roles in signal transduction, the Database for Annotation, Visualization, and Integrated Discovery (DAVID 6.8) was employed. DAVID is an online analytical program that provides a comprehensive set of functional annotation tools to explore the biological functions of genes from gene lists [18]. DAVID can identify and describe the biological processes, cellular components, molecular functions, and pathways that are associated with genes of interest.

Construction of Compound-Target-Pathway Networks.
To understand the underlying mechanism of MFXD in AR, a large-scale combinatorial network was established by integrating data obtained on drug information, drug-target interaction, and target-related pathway interactions.
rough the hierarchical network, we characterized the relationships and pathways targeted by MFXD in AR.

Analysis of Chemical Compounds from MFXD.
e total ion chromatogram of MFXD is shown in Figure 1. is was matched with the OTCML database, jointly constructed by ermo Scientific and Tsinghua University as a reference for compound identification. Using electrostatic field orbital trap high-resolution MS for fragment MS acquisition, more than 1,200 reference compounds of TCM have been collected in this database. More than 7,000 second-order MS images of the highest quality are available, enabling the rapid and accurate characterization of TCM components and natural products. A total of 24 compounds in MFXD were identified through chromatogram matching. Table 1 shows retention time, experimental and calculated m/z values, molecular formulas, errors in parts per million (ppm), and the major MS/MS fragments. e mass error of all identified compounds was less than 5 ppm.
If a compound is poorly absorbed following oral administration, the pharmacological effects may not be realized, even if the compound has potent effects on the pharmacological target in vitro [19]. In this study, Caco-2, OB limits F (20% and 30% bioavailability), and HIA were used to assess the components identified in MFXD. e results from OB and MS were combined to select compounds that were considered to be orally bioactive in the decoction prepared from MFXD. Table 2 presents the ADMET properties of the 24 components identified in MFXD. Most of the compounds showed favorable absorption properties.

Construction and Analysis of the Component-Target Interaction Network.
e TCMSP database contains 499 Chinese herbs with 19,384 compounds, 3,311 targets, and 837 associated diseases [20]. BindingDB is an experimental protein-small molecule database for virtual compound screening based on maximal chemical similarity and support Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine vector machine methods [21]. We screened the targets from the TCMSP and BindingDB databases for the 24 active compounds. A total of 198 targets were obtained after the deletion of redundant hits (113 from TCMSP and 138 from BindingDB).
To visualize the relationship between compounds and their targets, we constructed a compound-target interaction network ( Figure 2). is network had a density of 0.022 with a characteristic path length of 3.396 and an average number of 4.918 neighbors. In the compound-target network, multiple compounds could act on the same target protein and a single compound could be associated with multiple target proteins. For instance, prostaglandin G/H synthase 2 (PTGS2-P35354) and sodium-dependent noradrenaline transporter (SLC6A2-P23975) represented the hub nodes with a high degree of distribution, whereas peripheral nodes, such as gallic acid, interacted with the progesterone receptor (PGR-P06401) and represented a lower degree of distribution.
ese results were consistent with the common characteristics of TCM, in which multiple compounds and multiple targets are observed.

Construction and Analysis of the Target PPI Network.
rough the comparative analysis of 198 component targets and AR-related disease targets, we identified 37 common potential targets for MFXD in AR (Table S1). e STRING database is a commonly used tool for predicting PPI and producing integrated and objective association networks [22]. Using the common protein targets as input for network visualization, a diversified PPI network was created (Figure 3(a)). e network systematically summarized the interactions of MFXD targets associated with AR treatment. PSIP1 was not analyzed in the PPI network, as it does not interact with other proteins. e network shows viable protein target nodes (n � 36) connected by edges (n � 135) with an average node degree of 7.5 and average local clustering coefficient of 0.548. e PPI enrichment p value was less than 1.0 e −16 ; thus, proteins have more interactions among themselves than would be expected for a random set of proteins of similar size drawn from the genome. Such a significant enrichment indicated that the proteins are at least partially biologically connected as a group. e genes ranked by the MCC algorithm were selected using the cytoHubba plugin. e predicted top 10 contributing hub genes were TNF, PPARG, ALB, PTGS2, IL-10, ACTB, NR3C1, ACE, SERPINE1, and AHR, which were considered crucial targets of MFXD against AR. Two significant modules were selected from the PPI network by MCODE (Figures 3(b) and 3(c)). For module 1, the genes were significantly enriched in the nuclear factor-κB (NF-κB) signaling pathway and tumor necrosis factor (TNF) signaling pathway. Cluster 2 genes were significantly enriched in the T-cell receptor signaling pathway. ese data suggested that MFXD likely acts on AR through all three signaling pathways.

DAVID Pathway Analysis.
To further investigate the multiple mechanisms of MFXD at a systematic level, 37 common genes were uploaded into DAVID 6.8. Functional association clustering analysis discovered 14 annotation clusters, and the highest enrichment score was 4.07. e top 20 terms in molecular function, cellular component, and biological process are presented in Figures 4(a)-4(c), respectively.
GO enrichment analysis showed that the target genes were expressed in the plasma membrane, cell surface, cell membrane, and other cell compartments. At the molecular level, the target genes were involved in drug binding, enzyme binding, G-protein acetylcholine receptor activity, and steroid binging. At the cellular level, they were related to cell apoptosis, proliferation, and migration. Moreover, biological processes were also enriched with inflammation-related terms. AR is a chronic inflammatory disease that involves the release of various inflammatory mediators [1]. erefore, the inhibition of the inflammatory response is a therapeutic strategy for AR.
KEGG pathway annotation showed that 31 of the 37 (86.1%) potential target genes were enriched and involved in 153 pathways associated with the immune system, cardiovascular system, cancer, inflammation, and diabetes. As AR is an allergic disease, pathways that are associated with the immune system and inflammation were selected (Figure 4(d)). For example, the Fc epsilon RI, B-cell receptor, T-cell receptor, and Toll-like receptor signaling pathways are   Evidence-Based Complementary and Alternative Medicine involved in the regulation of the immune system and the TNF and NF-κB signaling pathways are involved in inflammation. We also identified the PI3K/AKT, cAMP, and AMPK signaling pathways in this analysis. e PI3K/AKT signaling pathway is important for cell growth, differentiation, metabolism, survival, and apoptosis [23]. Recent research has revealed that the PI3K-AKT signaling pathway regulates mast cell activity and is modulated in AR [24,25]. Moreover, the central and peripheral regulation of energy homeostasis relies on the cAMP signaling pathway [26]. KEGG pathway analysis showed that the active components of MFXD act on the gene nodes within the PI3K/AKT and cAMP signaling pathways ( Figure 5). ese data are consistent with the multiple effects that TCM has on different signaling and cellular pathways [27]. Comparative analysis of these KEGG pathways revealed that, among the 37 potential target genes identified, genes that were repeatedly associated with these pathways were PIK3CG, TNF, PTGS2, CHRM2, CHRM1, GSK3B, ADRB1, CHRM5, CHRM3, and HSP90AB1.

Compound-Target-Pathway Network Analysis of MFXD.
Chinese herbal formulas can contain dozens, or even hundreds, of compounds, and each compound can act on one or multiple targets to exert synergistic therapeutic effects [28]. To directly explore potential synergistic relationships, a compound-target-pathway combination network was constructed ( Figure 6). is network revealed that MFXD contains multiple compounds that have multiple targets within multiple pathways that are associated with AR treatment. For example, quercetin acted on 14 targets (such as P48736, P22301, Q75475, and P08684) and gallic acid acted on three targets (P23219, P35354, and P48736). P48736 was associated with both quercetin and gallic acid. Using network topological analysis, the top 10 compounds that may make major contributions to AR treatment are presented in descending order: quercetin, pseudoephedrine, ephedrine, β-asarone, methylephedrine, α-linolenic acid, cathine, ferulic acid, nardosinone, and higenamine. Quercetin can stimulate the immune system, inhibit the release of histamine, decrease the production of proinflammatory cytokines, increase the synthesis of leukotrienes, restrain the formation of IgE antibody, and improve the 1/ 2 balance by participating in multiple signal pathways [29]. All these mechanisms of action contribute to the antiinflammatory and immunomodulatory properties of quercetin, which can be effectively utilized in the treatment of AR, as shown in an animal model [30]. Gallic acid alleviates nasal inflammation via the activation of 1 and inhibition of 2 and 17 cells and has immunomodulatory effects [31]. α-Linolenic acid dampens AR through the eosinophilic production of 15-hydroxyeicosapentaenoic acid [32], and ephedrine has been used as a nasal wash for AR treatment [33].
Most of the compounds, such as kaempferitrin and pseudoephedrine, interact with PTGS2 (P35354). PTGS2 is responsible for the production of inflammatory      Evidence-Based Complementary and Alternative Medicine prostaglandin E2, which inhibits T regulatory cell differentiation to induce AR-related inflammation [34]. We also discovered compounds that acted together on the targets P01375 and P48736 to regulate the Fc epsilon RI signaling pathway. Gallic acid was observed to interact with PIK3CG (P48736), a key protein that activates the B-cell receptor, T-cell receptor, and Toll-like receptor signal pathways. Finally, targets such as P48736, P01375, P48736, P01375, and P49841, which may provide protection against AR, interacted with most of the compounds.

Conclusion
Unlike Western medicines, TCMs are commonly prescribed as herbal formulas that contain a mixture of herbs. Each herb may contain many active ingredients that have single or multiple targets; thus, it is difficult to pinpoint the mechanisms of action of TCM. Network pharmacology shares the core concepts of the holistic philosophy of TCM and meets the requirements to treat complex diseases systematically [35]. However, in the traditional research of network pharmacology, the compounds are mostly collected from databases. Some compounds cannot be detected in the decoction, which may yield false positive results. In this study, LC-MS/MS identification of compounds in the decoction followed by network pharmacology analyses provided insights into the mechanism of MFXD in the treatment of AR. In total, 24 bioactive compounds were identified in MFXD and 37 common targets were obtained and analyzed. e results indicated that MFXD was effective in the treatment of AR by regulating key pathways, including the Fc epsilon RI, B-cell receptor, Toll-like receptor, NF-κB, T-cell receptor, PI3K-AKT, cAMP, and AMPK signaling pathways. Ten compounds in the decoction prepared from MFXD were identified as candidates that could target AR. ese results reduce the prediction range, increase the accuracy of the prediction results, and provide important information for further pharmacological investigations on MFXD. is method, LC-MS/MS for bioactive compound identification followed by network pharmacology analyses, can increase the understanding of the mechanisms of Chinese herbal formulas and promote drug research and development.

Data Availability
e data used to support the findings of this study are included within the supplementary materials.

Conflicts of Interest
e authors report no conflicts of interest in this study.  Table S1: information on 37 therapeutic targets in MFXD acting directly on AR. (Supplementary Materials)