Abstract
Introduction
Chronic obstructive pulmonary disease (COPD) is a heterogeneous complaint associated with abnormal inflammatory responses of the lungs to harmful gases or particles, and is a common disease of the respiratory system, eventually leading to irreversible airflow limitation. 1 The prevalence of COPD will continue to rise over the next 40 years, with more than 5.4 million deaths per year predicted to occur in 2060, causing an intense economic burden. 2 The management and prevention of COPD is the key mission in Healthy China 2030 Action Plan. 2 The therapeutic effect of Traditional Chinese Medicine (TCM) on COPD has been known for numerous years. However, the mechanism of action of the beneficial effects of TCM remains to be elucidated.3,4
COPD belongs to the category of “syndrome characterized by dyspnea” and “lung distention” in TCM. 5 Vital qi deficiency plays the key in COPD, and the positive deficiency accumulation damage is the main pathogenesis. 6 The theory of Plain Questions (Suwen) said that “if vital qi is kept inside, the pathogens would not disturb.” In TCM, vital energy is a concept of the body's ability to self-regulate, resistance to pathogens, and self-recovery. “Vital qi dispels evil” means to drive away pathogens, which is similar to the defense function of the immune system to remove pathogenic antigens. The role of vital qi in regulating the balance of Yin and Yang is similar to self-stabilizing function of the immune system. Vital qi coordinates the organs, regulates qi and blood, and harmonizes Yin and Yang without phlegm and stagnation of blood stasis, which is similar to the surveillance function of the immune system. 7 Deficiency of vital energy is a syndrome of TCM. The clinical manifestations of this disease include physical weakness, paleness, shortness of breath, limb weakness, dizziness, sweating, and low voice. It is very similar to the concepts of “low immunity” and “sub healthy” in modern medicine. 8 Patients mainly have lung disease in TCM at the early stage of COPD. As qi was consumed in the long course of the disease, deficiency of the spleen, kidney and other organs arose. It is difficult for the body to remove the pathogens and recover vital energy, which leads to a vicious circle and lung distension characterized by panting, large amount of phlegm, palpitations, and edema. TCM pathogenesis of stable COPD is characterized by the deficiency of lung, spleen, kidney, phlegm-turbid retention, and deficiency complicated with excessiveness, but mainly deficiency. Lung in TCM is in charge of the vital qi and respiratory function. Invasion of exogenous pathogenic factors leads to the failure of the lung to disperse and descend qi. Rebellion of lung qi can cause cough, and qi deficiency also can cause short breath. Kidney in TCM governs reception of qi. Therefore, breathlessness will appear if kidney fails to receive qi. In TCM theory, lung regulates water while spleen transports and transforms liquid. Body fluid may be assembled and turn into stasis and phlegm because of the dysfunction of lung and spleen. 7 In addition, the spleen dominates the muscles and limbs, as well as their functions. When spleen qi deficiency occurs, systemic muscles and limbs become undernourished. Then their movements and functions are reduced. 9 Therefore, the typical clinical symptoms of COPD are dyspnea, cough, sputum production and decreased exercise capacity and endurance. Usually, qi deficiency can be evaluated by symptoms, lung function, exercise endurance and dyspnea index. 10 Treatment is mainly tonifying, including tonifying lung qi, spleen and lung, tonifying lung and kidney, and eliminating pathogenic factors such as resolving phlegm and promoting blood circulation. According to TCM theory, herbs with Qi-tonifying character are involved in improving the defense capacity of the immune system and the enhancement of ATP generation capacity.11,12 Many studies have confirmed that the nourishment of the lung and kidney has dramatic and long-term effects on boosting a weak immune system, decreasing inflammatory responses, reducing pulmonary pathological impairment and airway remodeling.13,14 Based on this theory, Bufei Nashen Pill (BFNSP, original name: Bufei Jianshen Naqi pill) was developed, as in-hospital preparations of Ningxia Chinese Medicine Research Center in the course of research. Previous studies have confirmed that BFNSP can improve the wheezing and shortness of breath in patients with pulmonary and kidney Qi deficiency syndrome in a COPD-stabilized period and interstitial lung disease. It can improve the degree of airflow limitation in COPD patients and reduce the level of interleukin 8 (IL-8) and tumor necrosis factor (TNF)-α in peripheral blood to reduce airway inflammation. BFNSP greatly improves a patient's prognosis and quality of life.15,16 Tentatively, it appears that BFNSP can be an effective treatment for COPD. Confirmatory evidence, however, is needed from large, and ideally, multicentered trials.
Based on the characteristics of “multi-component, multi-target and multi-way,” Chinese medicine has obvious advantages in the prevention and treatment of complex diseases. Furthermore, with the rapid development of bioinformatics, systems biology and multi-pharmacology, “network pharmacology” has proposed new methods and strategies for drug discovery to reveal the potential interactions between a Chinese compound formula and cytokines. It detects the effects of their interactions on the function of human life systems. 17 This is also in line with the holistic theory of TCM. The molecular docking technique which plays a pivotal role in the study of Chinese herbal medicinal products can simulate interactions between receptors and ligands to probe the active sites of drugs. BFNSP has a definite therapeutic effect on COPD, but the mechanism is still unclear. This study aimed to explore the effect and mechanism of BFNSP on COPD through network pharmacological analysis and molecular docking verification.
Materials and Methods
Screening of Active Ingredients and Potential Targets of BFNSP
The active components of BFNSP were identified from literature and the TCM Systems Pharmacology Database and Analysis Platform (TCMSP, https://old.tcmsp-e.com/tcmsp.php). 18 Pharmacokinetic ADME evaluation represents the ability of a compound to circulate in the body after oral administration. OB indicates whether the active compounds in a formula can be delivered throughout the body and produce a physicochemical effect. DL is an indicator for determining the similarity or likeness of a compound and its physicochemical properties with conventional drugs. 18 The active ingredients of BFNSP were acquired based on the criterion of OB ≥30% and DL ≥0.18. 19 Moreover, some compounds of scorpion and earthworm, which did not satisfy the criteria of DL values but were reported to possess extensive pharmacological activities, 20 were also enrolled into the potential compounds of BFNSP. All compounds had three-dimensional structures, which could be downloaded directly. Furthermore, the relevant targets were collected. The Uniprot database (https://www.uniprot.org/) was used to standardize the gene name. 21
Screening of COPD Disease Targets
The keywords “chronic obstructive pulmonary disease” were used to search the databases GeneCard (https://www.genecards.org/ 22 ), Online Mendelian Inheritance in Man (OMIM) (https://omim.org/), 23 and Therapeutic Target Database (TTD) (http://db.idrblab.net/ttd/ 24 ) to obtain COPD-related targets. Targets with reference scores greater than 30 were selected as candidates, and duplicates were eliminated.
Acquisition of Intersection Targets and Construction of “Drug-Ingredients-Potential Targets” Network
The intersection targets between the COPD-related targets and active compounds were retained by using a Venn tool (https://bioinfogp.cnb.csic.es/tools/venny/). 25 These targets were considered as potential therapeutic targets of BFNSP for COPD patients. A “Drug-Ingredients-Potential Targets” network was constructed by Cytoscape 3.3.0 software (https://cytoscape.org/download.html). 24
Construction of Protein–Protein Interaction Network and Analysis
The potential targets were imported into the STRING database (https://string-db.org/)
24
and the selected species was
GO Function Enrichment and Kyoto Encyclopedia of Genes and Genomes Pathway Enrichment Analysis
The potential targets of BFNSP for COPD were submitted to the Database for Annotation, Visualization, and Integrated Discovery database (https://david.ncifcrf.gov/), and the parameters were set as follows: “Select Identifier” was set to “Official Gene Symbol,” “List Type” was set to “Gene List,” and the species was defined as “
Molecular Docking
Complete structural information of ligand molecules and receptors was calculated by molecular docking. Then, according to the principles of energy matching, geometric matching and chemical environment matching, the binding ability of inhibitors to protein was evaluated, and the best binding mode between them was predicted. SYBYL 2.2.1 is a molecular docking software for flexible docking. 24 The steps are as follows: (1) Preparation of ligand molecules: importing the “mol2 file” of the ingredients from the TCMSP database, adding all hydrogens and modifying the charge to Gasteiger-Huckel, setting 1000 max iterations. (2) Preparation of receptor molecules: the “PDB file” of core targets was obtained from the RCSB PDB database (http://www.rcsb.org/). 26 (3) Molecular docking: Firstly, ligand substructures were extracted, then water and other unwanted ligands were removed. Then we analyzed the structure and added hydrogens. The binding site was generated by the ligand-based mode. Finally, surflex docking results between ingredients and proteins were analyzed and observed by Pymol software.
Results
Active Ingredients and Targets of BFNSP
A total of 1822 chemical constituents were retrieved from the TCMSP database. Using specific criteria (OB ≥ 30% and DL ≥ 0.18), and eliminating duplicates and components without any targets, 172 active ingredients were identified, including 17 from

Venn diagram of coincidence components. The horizontal histogram shows the number of components of each botanical drug, while the vertical histogram shows the size of different intersections of these 17 botanical drug sets.
Basic Information of Repeat Active Components in Bufei Nashen Pill (BFNSP).
Abbreviations: HQ, Huang Qi; HJ, Huang Jing; RCR, Rou Cong Rong; FF, Fang Feng; XX, Xi Xin; CT, Chan Tui; BZ, Bai Zhu; TLZ, Ting Li Zi; JJS, Jing Jie Sui; ZSY, Zi Su Ye; ZSZ, Zi Su Zi; DS, Dan Shen; XD, Xu Duan; YZR, Yi Zhi Ren; GL, Gua Lou; DL, Di Long; QX, Quan Xie.
COPD Disease Targets
GeneCards, TTD, and OMIM databases were applied to identify 6843 COPD-related targets, and merging and eliminating repetitive values. A total of 1219 disease targets were obtained based on a relevance score ≥ 30.
Acquisition of Potential Targets and Construction of “Drug-Ingredients-Potential Targets” Network
One hundred eighty three coincidence targets of BFNSP and COPD were collected by the Veen tool (Figure 2), and Cytoscape software was used to set a Drug-Ingredients-Potential Targets network (Figure 3).

Venn diagram of coincidence targets. In total, 183 targets were common to Bufei Nashen Pill (compound genes) and chronic obstructive pulmonary (COPD) (disease genes).

Drug-ingredients-potential targets network. Green represents the potential active ingredient of Bufei Nashen Pill, purple the potential targets, and red the drugs of Bufei Nashen Pill (BFNSP).
Construction and Analysis of Potential Target PPI Network
The potential targets were submitted to STRING database to obtain the protein interaction relationship. The PPI network was constructed by Cytoscape software, as shown in Figure 4. From this, 177 nodes (6 free targets were cut) and 1625 edges were seen. The greater the degree value, the larger is the node. According to the function of network analysis, we also could acquire the top 20 targets of DC, BC and CC. The repeated targets were selected as the key ones (Figure 5). The main functions of the core targets are shown in Table 2.

Protein–protein interaction (PPI) network. Network nodes represent proteins and edges represent protein-protein associations. The node's size is proportional to the degree value. The larger the degree value, the more is the interaction with the nodes of the gene.

Venn diagram of key targets. Key targets were identified by intersection of 20 genes from 3 algorithms including Degree Centrality, Closeness Centrality, and Betweenness Centrality.
Description of the 14 key Genes.
GO and KEGG Enrichment Analysis
Enrichment analysis of the 183 potential targets was calculated by omicshare tools. BFNSP includes 3726 GO terms (

Go enrichment analysis of potential targets of COPD treated by BFNSP,

Bubble chart of the results of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment. The size of the dot represents the number of genes involved in the process. The larger the point, the greater the number of genes. P indicates the statistical significance of the

Herb-compound-key target-pathway of the first 30 network. The red node represents the herbs in Bufei Nashen Pill (BFNSP); the orange triangular nodes represent the herbs’ ingredients, the purple squares the key target, and the green circles the pathways enriched with the targets.
Molecular Docking
164 components of BFNSP were docked with 15 key receptors: AKT1 (PDB ID: 2UZS), TP53 (1GZH), IL6 (4O9H), TNF (1FT4), MAPK1 (6G54), Vascular endothelial growth factor A (VEGFA) (5HHD), STAT3 (6NJS), JUN (1JNM), SRC (6E6E), epidermal growth factor receptor (EGFR) (4LQM), IL1B (3LTQ), CXCL8 (1QE6), RELA (2O61), matrix metalloproteinase-9 (MMP9) (1GKD), and MAPK8 (4G1W). Total score (TS) was used as the screening condition. The score took into account the role of hydrogen bonds and tensile forces in the ligand molecule, as well as the contribution of hydrogen bonds and Van der Waals’ forces between ligand and receptor. 26 The total docking results are shown in Figure 9. It was found that small molecular ligands generally combined well with IL-6, MAPK1, MAPK8, SRC, EGFR, and MMP9. Figure 10 indicates part of the molecular docking structures in detail.

Total docking results of 164 active components. Total score (TS) ≥ 4.0 as cutoff value for calculating selectivity score. Ligands with a total score ≥4 were divided into 3 docking groups, group 1 (4 ≤ TS<5), group 2 (5 ≤ TS<7), and group 3 (7 ≤ TS).

Molecular docking of some components.
Discussion
Due to chronic and persistent hypoxia, patients with COPD often develop serious diseases such as pulmonary heart disease and respiratory failure. Cigaret smoking is a major risk factor for the development and progression. COPD has a complex pathogenesis, which is currently generally considered to be related to inflammation, oxidative stress, protease/antiprotease imbalance, decreased immunity and other factors. 27 In recent years, TCM has become a hot research topic due to its significant efficacy in the prevention and treatment of COPD. BFNSP contains 17 Chinese natural medicines of which HQ has been demonstrated to possess tonic, hepatoprotective, diuretic, and expectorant properties and exhibit antihyperglycemic, immunomodulating, anti-inflammatory, antiviral and antioxidant activities. Nowadays, HQ is used extensively in clinical therapy, such as in Huangqi injections for treating renal diseases. 28 HJ has a long history with many chemical components with anticancer, antiaging, antidiabetes, and antifatigue properties. 29 RCR contains various bioactive compounds, such as phenylethanoid glycosides, flavonoids, lignin, and alkaloids, with potent antioxidant properties with the capability to improve mitochondrial function,30,31 and displaying a range of functions, including antioxidant, antinociceptive, anti-inflammatory, and immunity. 32 Ting Li Zi can decrease lung qi to relieve dyspnea. Dan Shen aims to remove blood stasis and promote blood circulation, and is widely used in clinical practice for the treatment of cardiovascular diseases, and severe pulmonary bloating. 33 The whole prescription can reduce symptoms and improve pulmonary function, and lower the risk of exacerbations for patients with COPD. 9
According to the results of network pharmacology, there are 172 components in BFNSP. After the intersection of targets of BFNSP and COPD-related genes, 164 effective components were identified with potential targets through the “Drug-Ingredients-Potential Targets” network. Quercetin, kaempferol, luteolin, and wogonin were ranked in the top 4 targets, which may play a wide role in treating COPD. Quercetin, kaempferol, luteolin and wogonin are all flavonoids, which have different therapeutic activities, and can act as cardioprotective, antiviral, antidiabetic, anti-inflammatory, antibacterial, and anticancer agents. 34 , 35 Research has showed that quercetin can reduce the production of IL-6, IL-8, IL-1β, and TNF-α, and suppress the activation of NF-κB to achieve an anti-inflammatory effect. 36 Moreover, quercetin possesses anticarcinogenic, antioxidant, free-radical scavenging, antifibrotic, and antiproliferative properties, inhibiting apoptosis and enhancing early innate immunity that can be expressed in different cell types and animal models.37,38,39 Kaempferol has the same anti-inflammatory effect as quercetin.40,41 Liu et al 42 conducted a study that showed that kaempferol can inhibit the release of inflammatory mediators, and reduce inflammatory response and oxidative stress by down-regulating reactive oxygen species (ROS)-dependent MAPKs-NF-кB and pyrophosphate signaling pathways. Luteolin and wogonin alleviated lung injury and suppressed inflammation eruption, which may have significant effect on the control of COPD with acute exacerbation.43,44 Therefore, BFNSP may protect against COPD by decreasing oxidative stress, repressing inflammatory cascade, and enhancing the immune system.
From the PPI network and topological analysis, AKT1, TP53, IL6, MAPK1, TNF, MMP9, VEGFA, MAPK8, JUN, STAT3, CXCL8, and EGFR are the intersection targets of the first 20 targets in the 3 topological analysis algorithms. In addition, by retrieving relevant articles,45,46 SRC, RELA and IL1B among the top 20 intersection targets of DC and CC were incorporated into the key targets of BFNSP in the treatment of COPD. 164 active compounds of BFNSP were docked with 15 key receptors by SYBYL2.2.1, with a docking score ≥4.0 as the cutoff value for calculating selectivity score. Ligands with a TS ≥4 were divided into 3 docking groups, group 1 (4 ≤ TS<5), indicating that it had certain binding capacity, group 2 (5 ≤ TS<7), indicating that the molecule had better binding activity to the target, and group 3 (7 ≤ TS), indicating a stronger binding affinity.26,47 As shown in figure 9, receptors with greater than 50% binding to ligands are mainly MAPK8 (89%), MAPK1 (88%), SRC (85%), IL6 (83%), EGFR (81%), TP53 (75%), MMP9 (71%), AKT1 (64%), and VEGFA (51%). Among them, monomers had good binding to AKT1, MAPK8, MAPK1, EGFR, IL6, and MMP9. Among the components of BFNSP, 12 have binding ability with 15 receptors, 60 with more than 10 receptors, and 99 with more than 8 receptors. This reflects the multi-component, multi-target and strong functionality characteristics of TCM. AKT1 is involved in cell metabolism, survival migration, and gene expression, and blocks the migration of leukocytes to damaged tissues to reduce inflammation. 48 The PI3K/Akt1-dependent pathway may contribute to the inflammatory response in the airway system. 49 Several recent studies have shown that autophagy has an anti-inflammatory mechanism. Autophagy prevents endothelial damage triggered by various endogenous or infectious agents and prevents unnecessary or excessive inflammation. The down-regulation of expression genes of PI3K and AKT1 mRNA can restore impaired autophagy in vivo and in vitro to reduce the injury of lung tissue and block tumor necrosis factor (TNF)-stimulated gene 8-mediated acute airway inflammation. It has a protective effect on lung injury.49,50 This is consistent with the previous studies that BFNSP can reduce the level of IL-8 and TNF-α in peripheral blood. Differently, a study showed that AKT1 activation contributes to proliferation and hypertrophy of airway myocytes, which are accompanied by proportional contractile protein accumulation, resulting in airway stenosis. 51 The control of cell size by the PI3K/Akt pathway has been in part explained by the ability of this pathway to regulate protein synthesis via the downstream targets of rapamycin (mTOR), ribosomal protein p70 S6 kinase, and eukaryotic initiation factor (eIF)4E-binding protein-1 (4E-BP1)/PHAS-I (Figure 11). 52 Thus, AKT1 may repair respiratory muscles and alleviate clinical symptoms of patients via increasing muscle protein synthesis. However, the hypothesis of BFNSP's ability to activate, inhibit or biphasic regulate AkT1 needs to be verified in normal cells and bodies. After all, many studies have shown that both inhibitors and activators of AKT1 may be useful for treating different clinical subpopulations of COPD patients. 53 IL-6 is a hypersensitive inflammatory factor that predicts the progression of COPD in smokers and regulates the expression of immunomodulatory and inflammatory mediators and oxidant/antioxidant imbalance.54,55 Moreover, IL-6 affects vascular remodeling. 56 MAPK1 functions as a binding point for numerous biochemical signals. The MAPK1 pathway can alter gene expression by phosphorylating transcription factors to promote the production of inflammatory factors (for example, TNF-α, IL-1, IL-6) in macrophages. 57 Known as c-Jun N-terminal kinase 1 (JNK1), MAPK8 can phosphorylate and activate activator protein-1 (AP-1) and then activate the expression of a series of downstream genes, playing a vital role in the regulation and control of cell proliferation, cell differentiation, cell survival, cell death, inflammation and other pathological processes. 58 MMP9 plays a significant role in the MMP-mediated degradation of pulmonary extracellular matrix and tissue remodeling in emphysema, which may lead to the loss of normal small airway traction after the reduction of the elastic retraction force of the lung tissue, further leading to the increase of airway resistance and lung function deterioration.59,60,61,62 The EGFR pathways activate one or more downstream PI3K/AKT and mTOR pathways through receptor autophosphorylation and cytoplasmic protein binding, which in turn drive a variety of respiratory pathologies, such as excessive mucus production to alleviate airway obstruction, lung function deterioration, increased mortality and frequent exacerbations.63,64 These clues inspire a hypothesis that BFNSP may improve airway mucus hypersecretion by inhibiting the EGFR-mediated pathway. To recap, BFNSP may treat COPD from inflammatory reaction, oxidative stress, airway remodeling and airway mucus hypersecretion.

PI3K-Akt signaling pathway.
Through GO-BP bio enrichment analysis, we found that BFNSP in the treatment of COPD is mainly related to antioxidation, such as response to oxygen containing compounds, cellular response to oxygen containing compounds, and response to chemical stress. Oxidants can not only damage DNA, lipids, and proteins, but also mediate a variety of processes that could foster the development of COPD, which can further aggravate the inflammatory reactions in the lung parenchyma and airways, and promote fibrosis in the lung and small airways.65,66 Through KEGG functional analysis of potential targets, the pathways corresponding to key targets were found from the first 30 key pathways. By mapping the drug-component-key target-pathway, we have found that the key targets with good molecular docking mainly involve the MAPK, TNF, C-type lectin receptor, Relaxin, IL-17, PI3K-Akt, Apoptosis, HIF-1 signaling pathways, and EGFR tyrosine kinase inhibitor resistance. The MAPK signaling pathway is closely related to inflammatory reaction, oxidative stress and other pathogenesis. Moreover, inhibitors can effectively improve lung function of COPD patients and reduce the level of inflammatory mediators and become therapeutic targets for COPD. 67 The TNF signal pathway is closely related to host defense, inflammation, apoptosis, autoimmunity, and organogenesis. In COPD, it is associated with bronchial hyper-responsiveness. The HIF-1 signaling pathway mediates adaptive responses to reduced O2 availability, including cell proliferation, metabolism, angiogenesis and vascular remodeling.68,69 BFNSP may improve the body's tolerance to an hypoxic environment by regulating the HIF-1 signaling pathway, and thus alleviate chest tightness and shortness of breath and other related symptoms of COPD. The C-type lectin receptor signaling pathway mediates phagocytosis and ROS release, and boosts the immune system to ensure effective pathogen control and tissue repair. 70 It has been proved that the production and recombination of collagen are restricted by relaxin, which can stimulate large-scale degradation of collagen and resist pulmonary fibrosis. 71 IL-17 is the main inflammation-related factor secreted by the Th17 subset of CD4 + T cells. The IL-17 signaling pathway participates in the development of autoimmune diseases by inducing the expression of proinflammatory factors, chemokines, and other inflammation-related factors. 72
Network pharmacology is indeed a new method for studying the relationship between drugs and diseases. The above results have suggested that BFNSP might play a role in the treatment of COPD through the regulation of inflammation, immunity, hypoxia tolerance, and airway remodeling. It embodies the characteristics of multi-pathway, multi-target and multi-pathway synergism of TCM. However, the actual effect and more accurate conclusions still need to be verified by follow-up animal experimental studies.
