Investigation on the antipyretic mechanism of Chaiqin Qingning capsule for the treatment of fever based on network pharmacology, molecular docking, and in vitro experimental validation

Chaiqin Qingning Capsule (CQQNC), a traditional Chinese patent medicine, can effectively shorten the duration of fever and significantly improve fever symptoms. However, the mechanism of its antipyretic effect needs to be further elucidated. Therefore, we aimed to investigate the molecular mechanism of CQQNC in the treatment of fever. We used the network pharmacology method to analyze the mechanism of action of CQQNC in the treatment of fever and validated our study primarily by molecular docking. Finally, the predictive results were verified by IL‐1β‐induced bEnd.3 cells. The results showed that quercetin, kaempferol, cubebin, chenodeoxycholic acid, isorhamnetin, bilirubin, cholic acid, and baicalin were the major components of CQQNC against fever. A total of 381 common targets have been crossed by CQQNC for the treatment of fever. Furthermore, we found that CQQNC targets several deregulated genes in fever such as AKT1, COX2, AVP, cAMP, IL6, IL1B, TNF, mPGES1, and PI3K, biological functions such as endopeptidase, cytokine receptor binding, and phosphatase activity, and signaling pathways such as the PI3K‐Akt pathway and the AGE‐RAGE signaling pathway in diabetic complications. The docking study revealed that the core components of CQQNC both had high affinity for hub targets, especially the targets of COX‐2, cAMP, mPGES1, and PI3K proteins. To further investigate the mechanism of CQQNC, an Elisa assay and Western blot detection were performed as part of an in vitro study. Elisa's result showed that CQQNC can significantly decrease the expression levels of cPLA2, sPLA2, PGE2, cAMP, and 15‐PGDH after stimulating IL‐1β to bEnd.3 cells in a dose‐dependent manner (p < .01, p < .001, p < .0001). In addition, detection of the PGE2/ COX /cMAP pathway via immunoblotting showed that CQQNC can significantly downregulate the protein expression of COX‐1, COX‐2, EP3, cAMP, and mPGES1 in bEnd.3 (p < .0001, p < .001, p < .01). In conclusion, our study confirmed that the antipyretic mechanism of CQQNC affects the synthesis and secretion processes of PGE2 via the PGE2/ COX/cAMP pathway, providing insight into the antipyretic mechanism of CQQNC in clinical application.


| INTRODUCTION
Fever is one of the most common clinical symptoms, which can be seen in many diseases, such as acute upper respiratory tract infection (URTI), common cold, and influenza.It was generally agreed that fever is caused by the invasion of exogenous pathogens into the body, causing immune cells such as macrophages, neutrophils, and endothelial cells to synthesize and release a large number of endogenous pyrogens such as IL-1β, IL-6, TNF-α, and the like.Then, these pyrogens will further act on the hypothalamic thermoregulatory center, causing an upregulation of the temperature setting point, ultimately leading to an increase in body temperature.Although fever is common, it can easily cause physical discomfort to patients and seriously affect the quality of work and life.Therefore, active treatment measures still need to be taken.Clinically, western medicines such as antipyretic analgesics like acetaminophen, aspirin, and ibuprofen can be used to treat fever due to their rapid effect (Moore et al., 2019;Simmons et al., 2000).However, adverse reactions to antipyretic analgesics can occur, such as gastrointestinal discomfort, liver, and renal toxicity, anaphylaxis, dizziness, or headache (Kanabar, 2017).In contrast, under the guidance of traditional Chinese medicine (TCM) theory, TCM has the advantages of stable efficacy and high safety in the treatment of diseases.TCM compounds have been widely used to treat various diseases and their symptoms (Chi et al., 2019;Rong et al., 2017).Numerous studies have confirmed that TCM compounds have antipyretic, antiinflammatory, and analgesic effects (Ma et al., 2021).
Chaiqin Qingning Capsule (CQQNC), a TCM preparation, is composed of Radix Bupleuri (named ChaiHu (CH) in Chinese), baicalin (extracted from Scutellaria baicalensis Georgi, namely Huangqin (HQ) in Chinese), and Bovis Calculus Artifactus (Named Rengong Niuhuang (RGNH) in Chinese).In China, CQQNC has been approved to treat URTI and its typical symptoms such as fever and sore throat (Wang et al., 2022;Xu et al., 2022;Zhao et al., 2021).In the clinic, CQQNC can shorten the time of fever relief, improve disease symptoms, improve disease efficacy, and is effective and safe when used in the treatment of URTI (Xu et al., 2022) and influenza B (Zhao et al., 2021), deserving clinical application.However, the underlying antipyretic mechanisms and potential bioactive materials of CQQNC in the treatment of fever remain unknown.In recent decades, network pharmacology has been widely used to study TCM compounds' active components and underlying mechanisms (Cai et al., 2021;Lyu et al., 2018).It can help us to understand the complex relationship between disease, targets, and agents from the aspects of the network.
In this research, we first used network pharmacology to explore the potential bioactive ingredients, core targets, and GO and KEGG pathways of the potential mechanism of CQQNC in treating fever.Then, molecular docking was performed to study the affinity effects of the key ingredients docking with the core targets.Finally, IL-1β-stimulated bEnd.3 cell models were established, and ELISA assay and Western blot methods were performed to investigate the effects of CQQNC on the PGE2/COX/ cAMP signaling pathway.In conclusion, our study confirmed that the antipyretic mechanism of CQQNC affects the synthesis and secretion processes of PGE2 via the PGE2/ COX/cAMP pathway, providing insight into the antipyretic mechanism of CQQNC in clinical application.

| Cell culture and treatment
Mouse bEnd.3 cells were given by Professor Song Chen, College of Life Science and Technology, China Pharmaceutical University.It was cultured in DMEM supplemented with 10% FBS at 37°C with 5% CO 2 .The cells were randomly divided into five groups: control group, IL-1β group, CQQNC (low, median, high) + IL-1β groups.Cells undergoing different treatments were incubated for 24 h and collected for subsequent experiments.

| Cell viability detection
Bend.3 cells at logarithmic growth phase were collected to prepare cell suspension and were seeded into 96-well culture plates (5 × 10 3 cells/well).To detect cytotoxicity, the cells were incubated in a DMEM medium containing 300, 250, 200, 150, 100, 50, 25, and 12.5 μg/mL CQQNC for 24 h, respectively.Then the cell viability was examined by using MTT method.

| Network pharmacological
predictions and molecular docking of CQQNC in the treatment of fever 2.4.1 | CQQNC bioactive ingredients identification and database establishment TCM System Pharmacology and Analysis Platform (TCMSP, http:// tcmspw.com/ tcmsp.php) (Ru et al., 2014) was searched using the keyword "ChaiHu."TCMSP contains information on chemicals, targets, drug targets, blood-brain barrier (BBB) penetrability, intestinal epithelial permeability, water solubility, and pharmacokinetic properties such as oral bioavailability (OB) and drug-likeness (DL).The absorption, distribution, metabolism, and excretion (ADME) criteria of OB ≥ 30% and DL ≥ 0.18 were applied to identify the CQQNC bioactive ingredients.OB refers to the percentage of unmodified drugs that enter the circulatory system after oral administration, an essential indicator for an objective evaluation of the internal quality of drugs (Alam et al., 2015).The higher the OB of an ingredient, the higher the likelihood of using it clinically.DL is a vague concept that refers to the similarity between components and known drugs (Jia et al., 2020).In drug discovery, ingredients with DL properties mean that they can become drug targets.

| Prediction of targets corresponding to active ingredients
The targets of bioactive ingredients were collected from the TCMSP, Swiss TargetPrediction, and PharmMapper databases.Swiss TargetPrediction (www.swiss targe tpred iction.ch) (Daina et al., 2019) allows us to estimate the most probable macromolecular targets of small molecules by entering 2D or 3D structure files.PharmMapper (http:// lilab.ecust.edu.cn/ pharm mapper) (Wang et al., 2017) contains 23,236 proteins, covering 16,159 pharmacophore models and 51,431 linkable pharmacophore models.The sdf structures of the active ingredients were imported into these two databases to acquire predictive targets.The Uniprot database (https:// www.unipr ot.org) (UniProt, 2021) was used to standardize CQQNC-related target proteins.
2.4.4 | CQQNC-PPI network of fever intersection genes The construction of a protein-protein interaction (PPI) network can help us understand the coexpression, fusion, neighborhood, and colocalization relationships of potential target genes that interact with predicted genes and the overall links between these genes (Bai et al., 2021).
Intersection targets were entered into the STRING database (https:// strin g-db.org) (Szklarczyk et al., 2021), and 'Homo Sapiens' was selected.Each node represents a target protein, and each edge represents a functional relationship between potential target genes.The PPI results were downloaded and imported into Cytoscape (version 3.8.2) (Otasek et al., 2019) to construct visualization and network analysis.The CytoNCA (Tang et al., 2015) and Cytohubba (Chin et al., 2014) plugins in Cytoscape provided the parameters of betweenness centrality (BC), closeness centrality (CC), and degree centrality (DC).Topology information for each node was calculated from BC, CC, and DC parameters and the MCC algorithm in Cytohubba. 2.4.5 | Construction and analysis of drug-

ingredients-targets-disease network
With the development of high-throughput technology and the advent of the big data era, network prediction algorithms or procedures, visualization, and analysis software or tools have been widely used in bioinformatics research (Su et al., 2022;Wong et al., 2023;Zheng et al., 2023) including Cytoscape software (Shannon et al., 2003), which helps to present the results of the analysis directly in a variety of forms such as pictures.
To reveal complex relationships between CQQNC and the fever intersection genes, the visualization and analysis of drug-ingredients-targets-disease networks were performed by Cytoscape software.Each node represents a TCM, ingredient, target gene, or disease.Each edge represents the relationship between TCM, ingredient, target gene, or disease.The network topology parameters were analyzed to select key components and core targets.
2.4.6 | GO function and KEGG pathway enrichment analysis The intersection target genes were uploaded to the DAVID database (https:// david -d.ncifc rf.gov) (Sherman et al., 2022).Further functional annotation analysis was performed by selecting the official gene symbol, Homo sapiens, and Gene list and converting them to Entrez ID.Then the DAVID database (p < .05)was used for GO and KEGG biological function enrichment analysis.The Bioinformatics platform (http:// www.bioin forma tics.com.cn) was used to visualize the top 20 items of the biological process (BP), the cellular components (CC), the molecular function (MF), and the GO and KEGG pathways.
2.4.7 | Molecular docking verification of key active ingredient-core targets Using Autodock vina (version 1.2.0) (Eberhardt et al., 2021), the key components selected from the drug-ingredients-targets-disease network were docked with our core genes of interest (AKT1, JNK, COX2, MAPK, AVP, ERK, cAMP, IL10, IL6, IL1B, TNF, mPGES1, TLR4, NF-KB, and PI3K) to verify the accuracy of key components and predicted targets.First, the three-dimensional structures of the target proteins were searched in the Protein Data Bank (PDB, https:// www.rcsb.org/ ) (Burley et al., 2017), and the files were downloaded in PDB format.The target protein was pretreated with Autodock (Li et al., 2022), including the removal of water molecules, hydrogenation, and designation of the protein as a receptor.The structures were saved as PDBQT protein receptor files.The prepared active ingredient file was then obtained by identifying the CQQNC active ingredients and establishing a data set.Since the file format was sdf, it was first imported into Chem3D (version 19.0, https:// www.chemd raw.com.cn/ ) to find the most dominant conformation, calculate the minimum binding energy, and convert it into mol2 format.The concrete operation steps are as follows: first import the SDF file of small molecular compound structure, and then select calculations-MM2-Minimize Energy-use the default parameters, run, and save the results as a mol2 file.Finally, use the default parameter, run, for calculations-MM2-Molecular Dynamics.
The lower the binding energy, the higher the affinity between receptors and ligands and the more stable conformation (Li et al., 2022).Binding energy less than −5 kcal/mol indicates good binding activity between the ligand and the receptor (Shi et al., 2021;Tao et al., 2020).
Similarly, drug molecules were preprocessed to remove water molecules, add hydrogens, and set the drug as ligands.Drug molecules were exported to PDBQT format ligand files.Finally, molecular docking was performed by Autodock and visualized using PyMol (version 2.3.0,https:// pymol.org/2/ ).The docking method we adopted is to dock the whole domain of the protein receptor, so the pockets and parameters of each component docking with the corresponding target are not consistent.The specific docking steps are that select Grid-Grid box in Auto Dock software-set the box parameters to wrap the whole protein receptor as the next ligand small molecule docking.Bend.3 cells (5 × 10 3 cells/well) were seeded into 96-well culture plates and incubated for 24 h.Then replace the culture medium with serum-free medium and continue to cultivate for 24 h.Subsequently, except for the control group, CQQNC (100, 50, and 25 μg/mL) was added to each treatment group.After further cultivation for 4 h, IL-1β cytokine was added (final concentration 30 ng/ mL) and continued to cultivate for 10 h in according to the previous study (Guo et al., 2006).Collect cell supernatant and centrifuge at 3000 rpm/min for 15 min in a 4°C centrifuge (Thermo Electron LED GmbH).Collect the supernatant and use an ELISA kit to strictly follow the instructions to detect the levels of cPLA2, sPLA2, PGE2, cAMP, and 15-PGDH in the supernatant of each group.

| Western blotting
Cells at logarithmic growth phase were collected to prepare cell suspension and seeded into a T25 cell culture bottle.After 24 h, the culture medium was discarded and washed with PBS two times, then replaced with FBS-free medium to cultivate for another 24 h.The protein was exposed using RIPA buffer (KeyGEN, Nanjing, China) and the concentration was detected with a BCA protein detection kit (Elabscience, Wuhan, China).We used an SDS-PAGE kit to isolate the proteins from samples and the isolated proteins were transferred into the PVDF membrane (Merck Millipore, Germany), and the PVDF membrane was blocked with 10% defatted milk powder for 2 h.The diluted primary antibodies cAMP (1:5000), EP3 (1:1000), COX-1 (1:1000), COX-2 (1:2000), and mPGES1 (1:250) were reacted with PVDF membranes overnight at 4°C, respectively.Next, the PVDF membrane was incubated at room temperature for 1 h with the corresponding secondary antibody (anti-rabbit or anti-mouse).Finally, proteins on PVDF membranes were imaged using an ultrasensitive ECL kit (Tanon, Shanghai, China) and protein expression was quantified using ImageJ software.

| Statistic analysis
Statistical analysis and visualization were performed by GraphPad Prism 8.0.2 software.The data were presented as mean ± standard deviation (SD).The Student's t-test was used to compare the data between two independent groups, and the one-way analysis of variance (ANOVA) with the Dunnett test was used for multiple comparisons.
A p value <.05 was considered statistically significant.

| CQQNC bioactive ingredients identification and database establishment
A total of 357 CQQNC ingredients were obtained, including 349 CH, 1 HQG, and 7 RGNH.In this study, 30 CQQNC active ingredients (22 in CH, 1 in baicalin, and 7 in RGNH) were selected for the database establishment (Table 1).

| Predicting putative targets of corresponding ingredients and fever
The potential targets of the active components in CQQNC were predicted in TCMSP, SwissTargetPrediction, and PharmMapper database, yielding 391, 1563, and 2206 targets, respectively.SwissTargetPrediction and PharmMapper screening criteria were probability >0 and Norm Fit ≥0.7, respectively.After removing duplicate values, 736 genes were selected as potential CQQNC targets (Figure 1a).Similarly, fever-related targets were retrieved from GeneCards, OMIM, and DisGeNET databases that yielded 3081 target genes after removing duplicates (Figure 1b).

| CQQNC-PPI network of fever intersection genes
A total of 381 intersection targets were obtained (Figure 2a, Table S1).A PPI network with 380 nodes and 9281 edges was obtained.Figure 2b-e shows the top 30 target genes and the core targets in the PPI network.The  2).

| Construction and analysis of drug-ingredient-target-disease network
The herb-ingredients-targets-disease network is shown in Figure 3 with 413 nodes and 2182 edges.The top 10 ingredients had high degree values, indicating their importance in the network (degree value >77).With the combination of baicalin (a monomer composition), 11 compounds were the key ingredients of CQQNC (Table 3).

| GO function and KEGG pathway enrichment analysis
GO enrichment analysis obtained 1276 biological functions.Figure 4a-c shows the bubble diagram of GO enrichment analysis.The GO functions were also visualized by R software (version 4.1.1)(Figure 4d).KEGG pathway enrichment analysis was performed, and 190 signaling pathways were obtained (p < .05)(Table S2).
Bubble plots (Figure 4e), including the top 20 crucial signaling pathways, were visualized by R (version 4.1.1)(Table S3).1alu), IL1B (PDB ID: 6Y8I), TNF (PDB ID: 5uui), mPGES1 (PDB ID: 5bqg), TLR4 (PDB ID: 2z62), NF-KB (PDB ID: 1mdi), and PI3K (PDB ID: 3apf).Heat map shows the docking results (Figure 5a).Except for linoleyl acetate, the binding energies of the other key active ingredients were all less than −5.0 kcal/mol, suggesting that these active ingredients had high binding affinity scores with the core target proteins listed above.The partly optimal docking results of receptors and ligands after visualization are depicted in Figure 5b, and in the current molecular docking analysis, we designated hydrogen bond as the main force.

| Cell viability assay
The MTT method was used to identify the optimal concentration of CQQNC for bEnd.3 cells by detecting the differences in cell viability between different concentrations (Figure 6a).When concentration between 150 μg/ mL and 300 μg/mL, CQQNC showed significant effects on bEnd.3'scell viability (p < .01,p < .0001).Therefore, we selected 25, 50, and 100 μg/mL of CQQNC for further study.

| DISCUSSION
Network pharmacology is a research method that aims to elucidate TCM compounds' active components and potential mechanisms of action.In this study, we identified the The drug-ingredient-target-disease network.The shape of the triangle, diamond, ellipse, and round rectangle represent herbs, RGNH, CH, and gene symbols, respectively.The red represent the herbs of CQQNC, the blue represents the compounds of CH, the light green represents the compounds of RGNH, and the sky blue represent the gene symbols.The larger the size of the nodes represent the higher the degree of the corresponding herbs, compounds, or ingredients.

T A B L E 3
The 11 core ingredients of CQQNC in the network.key chemical components and the core targets for the treatment of fever in CQQNC.Baicalin, quercetin, kaempferol, chenodeoxycholic acid, cholic acid, isorhamnetin, and bilirubin are the key ingredients of CQQNC in the treatment of fever (Table 3).Research has shown that baicalin has a significant antipyretic effect and can significantly reduce IL-6, IL-1β, and TNF-α in both serum, cerebrospinal fluid (CSF), and hypothalamus of febrile rats (Li & Ge, 2010).Its antipyretic mechanism is achieved by inhibiting the TLR4 signaling pathway (Ye et al., 2015), N-methyl-D-aspartate receptor-dependent hydroxyl radical pathway, and circulating TNF-α accumulation in the hypothalamus (Tsai et al., 2006).Quercetin and kaempferol have potent antiinflammatory, immunomodulation, analgesic, anticancer, and anti-oxidant effects (Liao & Lin, 2015;Rajendran et al., 2014).Kaempferol can suppress the expression of iNOS, COX-2, and MMP-3 and blocks TLR4 activation, suggesting that it can reduce LPS-induced inflammatory Note: DC, degree values, from the analyzed results of the drug-ingredient-target-disease network.

PubChem ID
T A B L E 3 (Continued) mediators through downregulation of TLR4, NF-κB, p38 MAPK, JNK, and AKT (Park et al., 2011).The molecular docking results showed that the active ingredients were well docked with target proteins except linoleic acid acetate (Figure 5).Among them, we found that the major ingredients both have higher docking affinity effects with COX-2, AVP, cAMP, and mPGES1 of which have caught our attention.The pathogenesis of fever is that exogenous pyrogens (EP) invade the body, activate macrophages, monocytes, and lymphocytes, and produce and release EP.
EP then acts directly or indirectly on the thermoregulatory center through the BBB, causing a febrile response (Dinarello, 1996;Ogoina, 2011).Research has confirmed F I G U R E 5 Molecular docking results and visualizations of main chemical ingredients of CQQNC with hub genes.(a) The docking affinity of AKT1, JNK, COX2, MAPK, AVP, ERK, cAMP, IL10, IL6, IL1B, TNF, mPGES1, TLR4, NF-KB, and PI3K with the major ingredients, respectively.(b) The visualization of the representative ingredients docking with the hub genes, respectively.that EP such as IL-1β and IL-6 in peripheral blood during fever can produce and release other important thermoregulatory mediators such as PGE2, cAMP, and AVP, which act on the BBB and sensory area of the organum vasculosum of the lamina (OVLT) in the hypothalamus, ultimately leading to a range of pathological and physiological responses (Prajitha et al., 2018).Currently, it is widely believed that PGE2 is the most important thermogenic medium.The molecular docking results showed that the main active components of CQQNC had high docking binding with COX2, cAMP, AVP, and mPGES1.Therefore, we further investigated how CQQNC regulates PGE2 to exert its central antipyretic effect through IL-1β-induced bEnd.3 cells.It was known to all, during fever, levels of pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α are significantly increased in both serum and CSF (Prajitha et al., 2018).PGE2 is considered the most critical central thermoregulatory mediator (Blatteis et al., 2005).Research showed that after IL-1β binding to the IL-1 receptor, enzymes related to phospholipid metabolism in the cell membrane or cytoplasm can be activated through different pathways, such as phospholipase A2 (PLA2), which includes intracellular phospholipase A2 (cPLA2) and secretory phospholipase A2 (sPLA2), further promoting the release of arachidonic acid (AA) from phospholipase, and by inducing the activation of COX, PGE2 was further synthesized and released (Park et al., 2006), which can bind to prostaglandin E2 receptors (EP1-EP4) to further activate adenylate cyclase, promote the synthesis and release of cAMP, and cause a large amount of synthesis and release of PGE2 and cAMP in the brain (Kawahara et al., 2015), ultimately leading to body fever.The synthesis process of PGE2 has three key ratelimiting processes including the activation of PLA2, the conversion of AA to unstable prostaglandin H2 (PGH2) by COX, and the conversion of PGH2 to PGE2 by prostaglandin synthase (PGES), indicating that there are multiple rate-limiting stages in the synthesis and production process of PGE2 (Park et al., 2006).The primary molecular mechanism of acetaminophen and ibuprofen is to inhibit the synthesis of COX-2, thus reducing the secretion of PGE2 (Blatteis et al., 2005;Simmons et al., 2000).In our study, after the stimulation of IL-1β on bEnd.3 cells, the expression of cPLA2, sPLA2, PGE2, cAMP, and 15-PGDH have both increased while the intervention of CQQNC can significantly decreased the levels of cPLA2, sPLA2, PGE2, cAMP, and 15-PGDH in a dose relationship (Figure 6).It was suggested that CQQNC can through inhibit the secretion of cPLA2 and sPLA2 then to reduce the levels of PGE2 and cAMP.Studies have shown that Bupleurum and Scutellaria have sound antipyretic effects (Zhao et al., 2019).In addition, CQQNC inhibited the proliferation of different influenza virus strains and downregulated the gene expressions of IL-6, TNF-α, CXCL8, CXCL10, CCL5, and COX in A549 cells in a dose-dependent manner.It also inhibited the expression of NF-κB protein in the signaling pathway (Zhao et al., 2021).
GO analysis shows that target genes are mainly enriched for biological functions.KEGG pathway enrichment analysis indicates that CQQNC is involved in the AGE-RAGE signaling pathway in diabetic complications, blood lipids, and atherosclerosis, the PI3K-Akt signaling pathway.No results demonstrate the direct connections between these signaling pathways and fever treatment.We found that 61 genes were enriched in the PI3K-Akt signaling pathway, and this pathway could be the vital signaling pathway for the treatment of fever.The PI3K-Akt signaling pathway is involved in transducing signals or BP such as cell development, differentiation, cell survival, protein synthesis, and metabolism (Xie et al., 2019).Furthermore, research has shown that the unbalance of the PI3K-Akt signaling pathway has an intimate relationship with tumor drug resistance, cancer, obesity, type 2 diabetes mellitus, and cardiovascular disease (Huang et al., 2018;Li et al., 2018).Many herbal extracts have anti-inflammatory effects on macrophages by regulating NF-κB, MAPK, and KLF4 signaling pathways, especially PI3K/AKT (Hu et al., 2021;Merecz-Sadowska et al., 2020).NF-κB is located downstream of the PI3K-Akt pathway.Therefore, activating the PI3K-Akt signaling pathway will lead to the activation of NF-κB and, ultimately, the release of various inflammatory factors such as IL-1β, IL-6, TNF-α, and PGE2 (Hayden & Ghosh, 2008).To further investigate the mechanism of CQQNC in regulating the PGE2/COX/cAMP signaling pathway, we performed a Western blot to examine the proteins' expression of COX-1, COX-2, mPGES1, and EP3.The results showed that CQQNC has remarkably inhibited the expression of COX-1, COX-2, mPGES1, and EP3 (Figure 7), suggesting CQQNC has the function of inhibiting COX-1, COX-2, mPGES1, and EP3 and to regulate PGE2 of thermoregulatory center in the hypothalamus.
In this study, there still has shortcomings, we found that saikosaponins did not make the top 10 active ingredients, although saikosaponin is one of the main components of saikosaponin.In addition, the pharmacological effects of artificial bezoar are still lacking in research.We have only constructed an in vitro model to investigate the antipyretic mechanism of CQQNC for the treatment of fever, it is necessary for us to conduct in-depth research on animal models in the near future.In summary, our current in vitro and in silico studies provide a basis for elucidating the antipyretic effect of CQQNC, to confirm the potential mechanism of CQQNC in the treatment of fever.

| CONCLUSIONS
Our network pharmacology and molecular docking studies identified the potential main active compounds, drug targets, and signaling pathways of CQQNC against fever.In vitro study indicated that CQQNC can significantly inhibit the expression of COX-1, COX-2, cAMP, mPGES1, and EP3, and thus downregulated the production of PGE2 and cAMP in bEnd.3 cells.The results of this study suggest that CQQNC has central antipyretic effects and its potential mechanisms may possibly inhibit the PGE2/COX/cAMP signaling pathway, providing a theoretical foundation of CQQNCs' clinical use in the treatment of fever.In conclusion, the integrative approaches of network pharmacology and experimental verification were performed to reveal and elucidate the underlying mechanisms of CQQNC in treating fever which provide a new insight into understanding the antipyretic mechanism of CQQNC in the treatment of fever; and our findings are of crucial meanings on CQQNC's clinical evidence-based medication for TCM prescriptions.

F
Potential gene targets of CQQNC and fever from different databases.(a) The Venn diagram of CQQNC with 226, 215, and 482 from PharmMapper, TCMSP, and SwissTargetPrediction database, respectively.(b) The Venn diagram of fever with 2634, 1021, and 16 from GeneCards, DisGeNET, and OMIM database, respectively.

F I G U R E 2
Venn diagram of CQQNC against fever, PPI network map, the top 30 genes in PPI network and the core targets.(a) Venn diagram of CQQNC and fever with 381 overlapped targets.(b) The top 30 genes in PPI network map according to the rank of degree value by R package.(c) The PPI network of 381 target genes from STRING database and were visualized in Cytoscape.(d) The top 30 targets were screened by CytoNCA.(e) The top 30 targets were selected from Cytohubba.(f) The 22 intersected targets Venn diagram from CytoNCA and Cytohubba.The redder the color and the larger the size of the nodes in Figure.C/D/E represent the higher degree or MCC values of the corresponding targets.CQQNC in the treatment of fever (Figure 2f; Table

F
Top 20 GO and KEGG enrichment analyses of the 381 intersected genes of CQQNC against fever.Including BP, CC, MF, and KEGG pathway analysis.(a) The top 20 BP of GO enrichment biological function.(b) The top 20 CC of GO enrichment biological function.(c) The top 20 MF of GO enrichment biological function.(d) The top 20 GO enrichment was analyzed and visualized by R package.(e) The top 20 KEGG pathway enrichment analysis was analyzed and visualized by R package.The deeper the color in the picture, the more significant it was.
Characteristics of the core active ingredients in CQQNC.

Molecular docking verification of key active ingredient-core targets
The top 22 genes in the PPI network.