Abstract

Objective

   Jinbei oral liquid (JBOL), which is derived from a traditional hospital
   preparation, is frequently utilized to treat idiopathic pulmonary
   fibrosis (IPF) and has shown efficacy in clinical therapy. However,
   there are now several obstacles facing the mechanism inquiry, including
   target proteins, active components, and the binding affinity between
   crucial compounds and target proteins. To gain additional insight into
   the mechanisms underlying JBOL in anti-IPF, this study used
   bioinformation technologies, including network pharmacology, molecular
   docking, and molecular dynamic simulation, with a substantial amount of
   data based on realistic constituents.

Methods

   Using network pharmacology, we loaded 118 realistic compounds into the
   SwissTargetPrediction and SwissADME databases and screened the active
   compounds and target proteins. IPF-related targets were collected from
   the OMIM, DisGeNET, and GeneCards databases, and the network of
   IPF-active constituents was built with Cytoscape 3.10.1. The GO and
   KEGG pathway enrichment analyses were carried out using Metascape, and
   the protein-protein interaction (PPI) network was constructed to screen
   the key targets with the STRING database. Finally, the reciprocal
   affinity between the active molecules and the crucial targets was
   assessed through the use of molecular docking and molecular dynamics
   simulation.

Results

   A total of 122 targets and 34 tested active compounds were summarized
   in this investigation. Among these, kaempferol, apigenin, baicalein
   were present in high degree. PPI networks topological analysis
   identified eight key target proteins. AGE-RAGE, EGFR, and PI3K-Akt
   signaling pathways were found to be regulated during the phases of cell
   senescence, inflammatory response, autophagy, and immunological
   response in anti-IPF of JBOL. It was verified by molecular docking and
   molecular dynamics simulation that the combining way and binding energy
   between active ingredients and selected targets.

Conclusions

   This work forecasts the prospective core ingredients, targets, and
   signal pathways of JBOL in anti-IPF, which has confirmed the multiple
   targets and pathways of JBOL in anti-IPF and provided the first
   comprehensive assessment with bioinformatic approaches. With empirical
   backing and an innovative approach to the molecular mechanism, JBOL is
   being considered as a potential new medication.

   Keywords: Jinbei oral liquid, Idiopathic pulmonary fibrosis, Network
   pharmacology, Molecular docking, Molecular dynamics simulation,
   Realistic ingredients

Graphical abstract

   Image 1
   [39]Open in a new tab

Highlights

     * •
       Multi-bio information technology was used to analyze first in
       Jinbei oral liquid.
     * •
       Bioinformatics analysis based on realistic components has reduced
       data deviation.
     * •
       Multiple components developed anti-idiopathic pulmonary fibrosis
       via multi-pathway.
     * •
       The latest evidence was provided for Jinbei's anti-idiopathic
       pulmonary fibrosis.
     * •
       We provide a reference for the further new drug development of
       Jinbei oral liquid.

1. Introduction

   Idiopathic pulmonary fibrosis (IPF) is a chronic intractable
   progressive fibrotic interstitial pulmonary disease, characterized by
   irreversible destruction of the lung parenchyma, excessive
   extracellular matrix (ECM) deposition, and an aberrant wound healing
   response brought on by damage to alveolar epithelial cells (AECs)
   [[40]1]. Due to its uncertain etiology and poorer prognosis-especially
   with a median survival of 3.8 years (95 % CI 3.5–3.8) [[41]2] since the
   time of diagnosis in adults aged 65 years or older-it is a serious
   issue in global public health. End-stage respiratory failure, pulmonary
   hypertension, and even death may occur following IPF and result in a
   significant financial burden [[42]3]. Comorbidities including cardiac
   disease and gastroesophageal reflux [[43]4] could further raise the
   risk of dying sooner.

   According to epidemiological research, the fatality rate from fibrotic
   pulmonary dysfunction has reached 60%–70 % [[44]5] since COVID-19.
   Glucocorticoids and cytotoxic medications are currently the mainstays
   of conventional IPF therapy, albeit their efficacy appears to be
   limited when used as an anti-IPF measure. Tyrosine kinase inhibitors,
   such as pirfenidone and nintedanib, which have received FDA approval,
   are being utilized as first-line treatments for IPF [[45]6]. According
   to the data survey, the average expenses of nintedanib and pirfenidone
   for clinical treatment in the United States from 2014 to 2019 was USD
   397.51 and USD 394.49, respectively. This indicates a low clinical
   utilization ratio, which is primarily attributable to the higher
   monthly amounts of self-paid patients. However, pirfenidone and
   nintedanib's potential medication side effects-gastrointestinal
   distress, skin eruption [[46]7], diarrhea, and elevated liver enzymes
   [[47]8]- avoid their widespread usage. Currently, lung transplantation
   [[48]9] is the unique treatment that is known to improve the prognosis
   and increase survival for IPF patients, however, strategies that
   involve a limited number of appropriate IPF patients and a scarcity of
   donor organs have intrinsic advantages and disadvantages.

   Recent studies have demonstrated that the intricate pathophysiology of
   IPF exhibits intertwined connection of cell types and signaling
   pathways, whereby the exploration of certain crucial signaling pathways
   serves as a catalyst for elucidating the IPF pathogenesis [[49]10].
   More medications, particularly Chinese medicinal formulae that target
   the anti-IPF signal pathways, have been verified in recent years.
   Herbal formulations have been demonstrated via numerous clinical
   studies to operate on the body through multi-component and
   multiple-target/pathway in order to combat respiratory disorders and
   improve the prognosis. To a greater extent, Traditional Chinese
   Medicine (TCM) is unable to realize modernization and globalization due
   to the mechanism of action and security are not sufficiently validated.
   A relatively new field, network pharmacology is intertwined with
   systems biology, genomics, polypharmacology, computational biology, and
   cutting-edge technologies integrated with massive data analysis and
   artificial intelligence. It has articulated the mechanism of disease
   and medicine function from the biological network [[50]11], built a
   bridge between empirical and evidence-based medicine by constructing
   the “herb-compound-target-disease” relationship that was proposed by
   Professor Shao [[51]12], and is thus appropriate for investigating
   pharmacological mechanisms in herbal formulas. Massive data sources
   would, however, induce a great deal of variation in the data algorithm.
   Other causes including acquisition from various databases and
   inconsistent standardization in data application [[52]13] would
   generally result in deviations between predicted and the realistic
   constituents in holistic analysis as well.

   Currently, there is no newly approved Chinese medicine for IPF
   treatment, leaving a significant gap in the field of treatment of this
   disease in China. With its origins in 40 years of clinical experience,
   Jinbei Oral Liquid (JBOL) is one of the few Chinese herbal medicine
   compounds specifically designed for the therapy of IPF. Multi-center
   clinical study findings have demonstrated that JBOL undoubtedly
   provides an advantage when it comes to considerably improving blood gas
   analysis, pulmonary function index, and blood oxygen saturation. In
   recent years, JBOL with unquestionable clinical efficacy and safety is
   frequently utilized at healthcare facilities as an approach to treating
   IPF [[53]14]. JBOL consists of twelve herbs including Angelica sinensis
   (Oliv.) Diels, Astragalus membranaceus (Fisch.) Bge, Codonopsis
   pilosula (Franch.) Nannf, Glycyrrhiza uralensis Fisch, Lonicera
   japonica Thunb, Glehnia littoralis Fr. Schmidt ex Miq, Pinellia ternata
   (Thunb.) Breit, Salvia miltiorrhiza Bge, Ligusticum striatum DC,
   Scutellaria baicalensis Georgi, Forsythia suspensa (Thunb.) Vahl, and
   Fritillaria cirrhosa D. Don [[54]15]. In order to precisely position
   the key components and targets, we have created a target network by
   utilizing realistic constituents rather than those that were predicted.
   We then imported the realistic constituents into the
   SwissTargetPrediction database and traditional Chinese medicine systems
   pharmacology (TCMSP) to investigate the hub genes in the biological
   information of JBOL that target IPF.

   The goal of this study was to promote data mining of multi-component
   and multi-target via network pharmacology by analyzing realistic
   constituents as a benchmark. In order to confirm the inference of
   network pharmacology, we further investigated the crosstalk between
   ligand and protein of core components and targets using molecular
   docking and molecular dynamics simulation. In a nutshell, this research
   aims to offer a reference for methodically delving into the JBOL
   mechanism in anti-IPF using realistic ingredients.

2. Materials and methods

2.1. Screening the active ingredients and targets of JBOL

   A total of 118 peaks were detected and identified [[55]16] using
   ultra-high-performance liquid chromatography in conjunction with
   quadruple time-of-flight mass spectrometry (UPLC-Q-TOF-MS). This
   mixture contained 43 flavonoids, 26 phenylpropanoids, 14 glycosides, 9
   phthalides, 8 alkaloids, and other substances. The substances listed
   above that were assessed served as the research object. The databases
   SwissTargetPrediction ([56]http://www.swisstargetprediction.ch/) and
   SwissADME ([57]http://www.swissadme.ch/) [[58]16] were utilized to
   uncover potential active ingredients and targets in JBOL, while the
   TCMSP database ([59]http://tcmspw.com/index.php) [[60]17,[61]18] and
   literature mining were utilized to fill in the component targets that
   were left out of SwissTargetPrediction. The screening criteria for this
   study mainly include two indicators: OB ≥ 30 % and DL ≥ 0.1; GI
   absorption is “High” and DL index greater than or equal to 3 Yes;
   Probability≥0.1. Subsequently, the final targets set was uniformly
   encoded using UniProt ([62]https://www.uniprot.org/) genes [[63]19],
   which eliminated duplicates, using the species setting “Homo sapiens”,
   and the structures of the active compounds were summarized in [64]Table
   1.

Table 1.

   Prediction of the component of JBOL based on SwissADME and TCMSP.

   [65]2.1.

   [66]2.1.

   [67]2.1.

   [68]2.1.
   [69]Open in a new tab

2.2. Collection of the active ingredients-disease intersection targets

   The target genes of IPF-related were retrieved with the keyword
   “Idiopathic pulmonary fibrosis”, “pulmonary fibrosis” in OMIM
   ([70]https://omim.org/) [[71]20], DisGeNET
   ([72]https://www.disgenet.org/) and GeneCards
   ([73]https://www.genecards.org/) databases [[74]21,[75]22]. The overlap
   between the targets of the active ingredient and the IPF-related
   targets was obtained using a Venn diagram, which was utilized to
   determine the potential therapeutic targets of IPF-related.

2.3. Construction of herb-active ingredient-target network

   The active components of JBOL were matched to pertinent targets using
   the network biology visualization tool Cytoscape (3.10.1). To establish
   the herb-active ingredient-target network, the XLS files containing the
   node attribute information and node interaction connection were loaded
   into Cytoscape. The node size and (or) depth of color directly
   reflected the degree value of every node.

2.4. Construction of a PPI network diagram

   Targets of ingredient-IPF intersection were loaded into the STRING
   database [[76]23] ([77]https://string-db.org/) to gather data on the
   protein interaction network. “Homo sapiens” was the criterion for
   screening the “organisms” and “medium confidence (0.4)" was the minimal
   required interaction score. The generated data was then extracted and
   loaded into Cytoscape (3.10.0) for the purpose to establish a PPI
   network. In the PPI network, the degree centrality (DC), closeness
   centrality (CC) [[78]24], and betweenness centrality (BC) of the node
   are calculated, while the node size was visually represented according
   to DC. The key targets were extracted based on the degree value in the
   PPI network [[79]25].

2.5. Genomic functional enrichment

   Gene Ontology (GO) [[80]26] and the Kyoto Encyclopedia of Genes and
   Genomes (KEGG) pathway enrichment analysis [[81]27] were conducted out
   to obtain insight into the biological function of possible targets in
   IPF therapy using Metascape ([82]https://www.metascape.org) database
   [[83]28]. Molecular functions (MF), cellular components (CC), and
   biological processes (BP) were screened using GO analysis. Important
   signaling pathways involved in biological activities were identified by
   means of KEGG [[84]29] enrichment analysis. During the analytic
   procedure, p < 0.01, minimum overlap of 3, and enrichment factor >1.5
   were collected and subsequently categorized based on how similar their
   members were. (last update date: 2024-02-19). For visual analysis, GO
   and KEGG data were uploaded to the Bioinformatics
   ([85]http://www.bioinformatics.com.cn/).

2.6. Construction of a ingredients-disease-target-pathway network

   To further investigate the functional mechanism of JBOL's active
   ingredients against IPF, the “ingredient-disease-target-pathway”
   network was constructed using the Cytoscape (3.10.1) software package
   [[86]30] (Boston, MA, USA) to visualize and illuminate the intricate
   relationships between ingredients, targets, and pathways. The degree of
   every node in the network would be directly indicated by node size
   and/or color depth.

2.7. Molecular docking

   To verify the binding ability among the active ingredients and core
   target proteins, nine core ingredients were selected for molecular
   docking with nine target proteins [[87]31]. We used AutoDock Vina for
   molecular docking, which adopted a semi-flexible docking mode [[88]32].
   The crystal structure of the core target proteins was made reference to
   RSCB PDB ([89]http://www.rcsb.org/) [[90]33]. The options of 3D
   structure need to coincide as follows: (1) the species type is “Homo
   sapiens”, (2) the pH of crystallization should approach the normal
   range of the human body, (3) Crystal proteins need to contain the
   information of original small molecule ligand, (4) High resolution
   (≤2.5A). The ligands and water molecules of each protein were removed
   by PyMOL software. The structure of the 3D files containing the active
   components was acquired using the PubChem database. The receptors and
   ligands were pretreated with hydrogenation using Auto Dock Tools 1.5.7
   before docking, and the Grid Box was set with the original ligand as
   the center. The scope of the grid box should cover the entire protein
   receptor structure and ensure that any possible docking pockets are
   included. In order to assess free binding energies, AutoDock Vina was
   used for the docking process. Finally, we used the Discovery Studio
   (2024) and PyMOL 2.5.8 ([91]https://www.pymol.org/) to visualize and
   evaluate the interactions of typical docking data. The minimum binding
   energy among each pair of complexes was calculated to evaluate the
   binding stability between JBOL core components and target proteins. The
   binding energy value is inversely proportional to the binding ability,
   with binding energy ≤ -5 kcal/mol signifying strong binding activity
   [[92]34].

2.8. Molecular dynamics simulation

   Molecular dynamics simulations were run on the Amber 22 (San Francisco,
   CA, USA) to determine the binding mode and binding energy of the
   receptor protein with the lowest binding energy to the ligand during
   the molecular docking process. The ff19SB force field [[93]35] was
   trained using TIP3P water, and counterions were added to neutralize the
   system. The system was heated up to 300 K within 500 ps when energy was
   minimized, and then the system achieved pre-equilibrated at 300 K using
   the regular system (NVT), lastly, molecular dynamics simulations were
   run for 100 ns at the isothermal isobaric system (NPT). The SHAKE
   limited all covalent bonds of hydrogen, and the results were analyzed
   with AmberTools23.

3. Results

3.1. Analysis of active components screening and target prediction

3.1.1. Prediction of nuclear ingredients combined with the realistic
constituents

   A total of 118 chemical ingredients of JBOL were identified through
   UPLC-Q-TOF-MS. In this investigation, 34 active compounds were screened
   according to the standards of SwissADME database systems, of which 14
   components were shared by the 12 herbs. Including 8, 2, 6, 10, 4, 2, 5,
   7, 6, 3, 4, and 6 compounds were obtained from HQ, DangS, BSS, DanS,
   DG, CX, JYH, LQ, HQIN, CBM, QBX, and GC, respectively. In addition,
   these substances contained two phenolic acids, five alkaloids, four
   tanshinones, seven phenylpropanols, and eleven flavonoids. The
   remaining compounds appertained to isoflavone, glycoside,
   phenylpropionic acid, and its derivatives, phthalide, and psoralen,
   respectively ([94]Table 1). The UpSet plot acquired the information of
   intersecting components in 12 herbs and showed the amounts of active
   ingredients and common components distribution in each herb of JBOL
   ([95]Fig. 1).

Fig. 1.

   [96]Fig. 1
   [97]Open in a new tab

   Distribution of the predicted active ingredients in 12 herbs. The
   number of active compounds found in every herb is indicated in the
   strip at the bottom left. The intersection situations of the individual
   sets are displayed by the dots and lines at the bottom right. A
   histogram illustrates the quantity of relevant components exist in each
   subset.

3.1.2. Construction of JBOL herb-active ingredient-target network

   The herbs, active ingredients, and related putative targets of JBOL
   were entered into Cytoscape 3.10.1 to produce the herb-active
   ingredient-target network diagram ([98]Fig. 2), which includes 549
   nodes and 1550 edges. According to the network analysis tool, the
   network heterogeneity is 2.43 and the network centralization is 0.187.
   The top 10 active components with degree values were taken as the core
   nodes that including Kaempferol (MOL000422, degree:108), Apigenin
   (MOL000008, degree:107), Methylnissolin (14077830, degree:106),
   Baicalein (MOL002714, degree:105), Liquiritigenin (MOL001792,
   degree:102), Phellopterin (MOL002644, degree:80), Peimine (MOL004445,
   degree:77), Cryptotanshinone (MOL007088, degree:71), Isoliquiritigenin
   (MOL001789, degree:70), 7-hydroxy coumarin (MOL003875, degree:68). The
   network revealed that the most significant active ingredients
   interacting with the targets were kaempferol, apigenin, methylnissolin,
   and baicalein, among others.

Fig. 2.

   [99]Fig. 2
   [100]Open in a new tab

   Herb-active ingredient-target network of JBOL. (Red arrows represent
   the herbs in JBOL, the octagon node represents components of JBOL, 2
   blue octagon nodes are BSS compounds, 3 crimson octagon nodes are CBM
   compounds, 1 purple octagon node is CX compound, 5 rose red octagon
   nodes are DanS compounds, 2 navy blue octagon nodes are GC compounds, 1
   yellow octagon node is DG compound, 3 Red octagon nodes are HQ
   compounds, 2 blue octagon nodes are CX compounds, 1 bottle-green
   octagon node is LQ compound, and the 14 light green circular nodes are
   common components in the twelve herbs of JBOL. Purple rhombus node
   represent predict targets. The edges depict the interaction between
   ingredients and targets, and the node size and depth of color are
   proportional to the degree value of interaction.

3.2. Prediction of potential targets of JBOL in the treatment of IPF

   The retrieval terms were set to “Idiopathic Pulmonary Fibrosis”,
   “Pulmonary Fibrosis”, and the condition was set to “Homo sapiens”. The
   targets associated to IPF were found using the DisGeNET, GeneCards, and
   OMIM databases. The findings indicated that 1708, 1671, and 356 were
   the confirmed or potential IPF targets, respectively. Uniprot was
   applied to the screen and standardized all disease targets with
   corresponding gene symbols. A total of 1169 targets were extracted
   after deduplication. According to the Venn diagram data, a total of 122
   intersection targets between JBOL component targets and IPF disease The
   Venn diagram data indicates that there are 122 intersection targets
   between IPF targets and JBOL component targets ([101]Fig. 3).

Fig. 3.

   [102]Fig. 3
   [103]Open in a new tab

   Venn diagram of the target of JBOL and the target of IPF.

3.3. PPI network construction and core target selection

   We entered 122 common targets to the STRING database into create a PPI
   network with the aim to investigate the mechanism potential therapeutic
   effects of JBOL against IPF ([104]Fig. 4A). Proteins were represented
   by nodes, while protein associations were displayed via lines. The
   network displayed 120 nodes and 1793 edges, with median DC values of
   26. Based on topology analysis, the core PPI network was identified,
   with DC ≥ 2 times the median and DC, BC, and CC ≥ median as the
   screening criterion. Topological analysis indicated that 8 core
   objectives stand out distinctly. ([105]Table 2 and [106]Fig. 4A),
   namely RAC-alpha serine/threonine-protein kinase (AKT1), Epidermal
   growth factor receptor (EGFR), Proto-oncogene tyrosine-protein kinase
   Src (SRC), Signal transducer and activator of transcription 3 (STAT3),
   Matrix metalloproteinase-9 (MMP9), mitogen-activated protein kinase 3
   (MAPK3), Caspase-3 (CASP3) and Heat shock protein HSP 90-alpha
   (HSP90AA1) ([107]Table 2). The network with essential and non-essential
   targets was constructed simultaneously ([108]Fig. 4B), with node sizes
   in this network proportional to the target degree.

Fig. 4.

   [109]Fig. 4
   [110]Open in a new tab

   The PPI network of JBOL and IPF predict key targets. (A) The PPI
   network screening process. The 8 core targets were screened by
   screening from 122 common targets via the value of DC, BC, and CC. (B)
   Construction of core target and non-core target network. Nodes
   represent proteins (The important target is indicated by the red
   circular node; the node size and color depth are graphically depicted
   based on the DC value.), edge represents protein-protein association.

Table 2.

   Core targets information of PPI network.
   UniProt ID Gene symbol Protein name Degree Betweenness Centrality
   Closeness Centrality
   [111]P31749 AKT1 RAC-alpha serine/threonine-protein kinase 87 0.0628
   0.7778
   [112]P00533 EGFR Epidermal growth factor receptor 86 0.0599 0.7727
   [113]P12931 SRC Proto-oncogene tyrosine-protein kinase Src 80 0.0761
   0.7438
   [114]P40763 STAT3 Signal transducer and activator of transcription 3 79
   0.0284 0.7391
   [115]P14780 MMP9 Matrix metalloproteinase-9 74 0.0309 0.7169
   [116]P27361 MAPK3 Mitogen-activated protein kinase 3 72 0.0261 0.7083
   [117]P42574 CASP3 Caspase-3 71 0.0222 0.7041
   [118]P07900 HSP90AA1 Heat shock protein HSP 90-alpha 66 0.0669 0.6879
   [119]Open in a new tab

3.4. GO enrichment and KEGG pathway analysis

   The core targets were imported into the Metascape database for
   enrichment analysis. The top 10 items of biological function (BP), cell
   function (CC), and molecular function (MF) (P < 0.01) in GO analysis
   were shown in [120]Fig. 5. Among them, BP involved positive regulation
   of phosphorus and phosphate, transmembrane receptor protein tyrosine
   kinase signaling pathway, positive regulation of protein
   phosphorylation, transferase, and kinase activity. There were vesicle
   lumen, caveola, membrane raft, membrane microdomain, and focal
   adhesion. In CC. MF was related to protein kinase activity and
   phosphotransferase activity with the alcohol group as acceptor, kinase
   activity, protein tyrosine kinase activity, and transmembrane receptor
   protein tyrosine kinase activity.

Fig. 5.

   [121]Fig. 5
   [122]Open in a new tab

   GO analysis (Top 10 terms of biological process (BP), molecular
   function (MF) and cellular component (CC)).

   The first 20 signal pathways involved pathways in cancer, AGE-RAGE
   signaling pathway, lipid and atherosclerosis, EGFR tyrosine kinase
   inhibitor resistance PI3K-Akt signaling pathway, and so on ([123]Table
   3 and [124]Fig. 6A and B). Using the Bioinformatics platform
   ([125]http://www.bioinformatics.com.cn/), a Sankey diagram was
   established on the basis of the relationship between the target genes
   and the enrichment pathways ([126]Fig. 6C).

Table 3.

   Analysis of the KEGG pathway of phillygenin for weight loss (Top 20).
   ID Description Count Gene ID P-Value
   hsa05200 Pathways in cancer 44
   ABL1,AKT1,ALK,AR,CCND1,CASP3,EDNRA,EGFR,ESR1,F2,FGFR1,FLT4,MTOR,HDAC2,H
   SP90AA1,IGF1R,IL2,JAK2,KIT,MET,MMP1,MMP2,MMP9,NFKB1,NOS2,PDGFRB,PGF,PIK
   3CA,PIK3CB,PIK3CD,PPARG,PRKCA,MAPK1,MAPK3,MAPK8,MAP2K1,PTGS2,PTK2,RAF1,
   RET,SLC2A1,STAT3,TERT,VEGFA 2.96E-43
   hsa04151 PI3K-Akt signaling pathway 30
   AKT1,CCND1,EGFR,FGFR1,FLT4,MTOR,HSP90AA1,IGF1R,IL2,ITGB3,JAK2,KDR,KIT,M
   ET,NFKB1,PDGFRB,PGF,PIK3CA,PIK3CB,PIK3CD,PIK3CG,PRKCA,MAPK1,MAPK3,MAP2K
   1,PTK2,RAF1,RET,TEK,VEGFA 2.73E-29
   hsa05205 Proteoglycans in cancer 29
   AKT1,CCND1,CASP3,MAPK14,EGFR,ESR1,FGFR1,MTOR,IGF1R,ITGB3,KDR,MET,MMP2,M
   MP9,PIK3CA,PIK3CB,PIK3CD,PLAU,PRKCA,MAPK1,MAPK3,MAP2K1,PTK2,PTPN11,RAF1
   ,SRC,STAT3,VEGFA,EZR 1.55E-34
   hsa05417 Lipid and atherosclerosis 26
   AKT1,CASP3,MAPK14,CYP1A1,CYP2C9,HSPA1A,HSP90AA1,ICAM1,JAK2,LYN,MMP1,MMP
   3,MMP9,NFKB1,PIK3CA,PIK3CB,PIK3CD,PPARG,PRKCA,MAPK1,MAPK3,MAPK8,PTK2,SE
   LE,SRC,STAT3 2.73E-29
   hsa04014 Ras signaling pathway 25
   ABL1,AKT1,EGFR,FGFR1,FLT4,IGF1R,KDR,KIT,MET,NFKB1,PDGFRB,PGF,PIK3CA,PIK
   3CB,PIK3CD,PLA2G2A,PRKCA,MAPK1,MAPK3,MAPK8,MAP2K1,PTPN11,RAF1,TEK,VEGFA
   6.44E-27
   hsa04010 MAPK signaling pathway 25
   AKT1,CASP3,MAP3K8,MAPK14,EGFR,FGFR1,FLT4,HSPA1A,IGF1R,KDR,KIT,MAPT,MET,
   NFKB1,PDGFRB,PGF,PRKCA,MAPK1,MAPK3,MAPK8,MAP2K1,RAF1,RET,TEK,VEGFA
   1.82E-24
   hsa05167 Kaposi sarcoma-associated herpesvirus infection 24
   AKT1,CCND1,CASP3,CCR3,MAPK14,MTOR,HCK,ICAM1,JAK2,LYN,NFKB1,PIK3CA,PIK3C
   B,PIK3CD,PIK3CG,MAPK1,MAPK3,MAPK8,MAP2K1,PTGS2,RAF1,SRC,STAT3,VEGFA
   2.03E-27
   hsa05207 Chemical carcinogenesis-receptor activation 24
   ADRB2,AKT1,AR,CCND1,CHRNA7,CYP1A1,CYP1A2,EGFR,ESR1,MTOR,HSP90AA1,JAK2,N
   FKB1,PIK3CA,PIK3CB,PIK3CD,PRKCA,MAPK1,MAPK3,MAP2K1,RAF1,SRC,STAT3,VEGFA
   1.51E-26
   hsa04510 Focal adhesion 23
   AKT1,CCND1,EGFR,FLT4,IGF1R,ITGB3,KDR,MET,MYLK,PDGFRB,PGF,PIK3CA,PIK3CB,
   PIK3CD,PRKCA,MAPK1,MAPK3,MAPK8,MAP2K1,PTK2,RAF1,SRC,VEGFA 1.76E-25
   hsa04015 Rap1 signaling pathway 23
   AKT1,MAPK14,EGFR,FGFR1,FLT4,IGF1R,ITGB3,KDR,KIT,MET,PDGFRB,PGF,PIK3CA,P
   IK3CB,PIK3CD,PRKCA,MAPK1,MAPK3,MAP2K1,RAF1,SRC,TEK,VEGFA 3.64E-25
   hsa05163 Human cytomegalovirus infection 23
   AKT1,CCND1,CASP3,CCR3,MAPK14,EGFR,MTOR,CXCR2,ITGB3,NFKB1,PIK3CA,PIK3CB,
   PIK3CD,PRKCA,MAPK1,MAPK3,MAP2K1,PTGS2,PTK2,RAF1,SRC,STAT3,VEGFA
   1.62E-24
   hsa04933 AGE-RAGE signaling pathway in diabetic complications 22
   AKT1,CCND1,CASP3,MAPK14,F3,ICAM1,JAK2,MMP2,NFKB1,SERPINE1,PIK3CA,PIK3CB
   ,PIK3CD,PRKCA,PRKCD,MAPK1,MAPK3,MAPK8,SELE,STAT3,VEGFA,NOX4 1.39E-30
   hsa05208 Chemical carcinogenesis-reactive oxygen species 22
   ABL1,AKT1,MAPK14,CYP1A1,CYP1A2,EGFR,MET,NFKB1,PIK3CA,PIK3CB,PIK3CD,PRKC
   D,MAPK1,MAPK3,MAPK8,MAP2K1,PTK2,PTPN11,RAF1,SRC,VEGFA,NOX4 4.01E-23
   hsa05215 Prostate cancer 21
   AKT1,AR,CCND1,EGFR,FGFR1,MTOR,HSP90AA1,IGF1R,MMP3,MMP9,NFKB1,PDGFRB,PIK
   3CA,PIK3CB,PIK3CD,PLAT,PLAU,MAPK1,MAPK3,MAP2K1,RAF1 2.86E-29
   hsa04062 Chemokine signaling pathway 21
   AKT1,CCR3,HCK,CXCR1,CXCR2,JAK2,LYN,NFKB1,PIK3CA,PIK3CB,PIK3CD,PIK3CG,PR
   KCD,MAPK1,MAPK3,MAP2K1,PTK2,RAF1,SRC,STAT3,CCR2 4.73E-23
   hsa01521 EGFR tyrosine kinase inhibitor resistance 20
   AKT1,AXL,EGFR,MTOR,IGF1R,JAK2,KDR,MET,PDGFRB,PIK3CA,PIK3CB,PIK3CD,PRKCA
   ,MAPK1,MAPK3,MAP2K1,RAF1,SRC,STAT3,VEGFA 2.73E-29
   hsa04926 Relaxin signaling pathway 20
   AKT1,MAPK14,EGFR,MMP1,MMP2,MMP9,MMP13,NFKB1,NOS2,PIK3CA,PIK3CB,PIK3CD,P
   RKCA,MAPK1,MAPK3,MAPK8,MAP2K1,RAF1,SRC,VEGFA 5.76E-25
   hsa01522 Endocrine resistance 19
   AKT1,CCND1,MAPK14,EGFR,ESR1,MTOR,IGF1R,MMP2,MMP9,PIK3CA,PIK3CB,PIK3CD,M
   APK1,MAPK3,MAPK8,MAP2K1,PTK2,RAF1,SRC 1.42E-25
   hsa04066 HIF-1 signaling pathway 18
   AKT1,EGFR,MTOR,IGF1R,NFKB1,NOS2,SERPINE1,PIK3CA,PIK3CB,PIK3CD,PRKCA,MAP
   K1,MAPK3,MAP2K1,SLC2A1,STAT3,TEK,VEGFA 4.73E-23
   hsa04917 Prolactin signaling pathway 16
   AKT1,CCND1,MAPK14,ESR1,JAK2,NFKB1,PIK3CA,PIK3CB,PIK3CD,MAPK1,MAPK3,MAPK
   8,MAP2K1,RAF1,SRC,STAT3 5.56E-23
   [127]Open in a new tab

Fig. 6.

   [128]Fig. 6

   [129]Fig. 6
   [130]Open in a new tab

   Results of KEGG pathway enrichment analysis. (A) The bubble diagram for
   the top 20 pathways in the KEGG enrichment analysis. (B) The KEGG type
   of the top 20 pathways. (C) The KEGG enrichment analysis sankey diagram
   for the JBOL anticipated targets in the anti-IPF. The KEGG pathways are
   symbolized via the right rectangle nodes in the sankey diagram, the
   anticipated targets via the left rectangle nodes, and the interaction
   between the target and the pathway with the lines. The right side is a
   conventional bubble diagram. The bubble size represents the number of
   genes to which the pathway belongs, and the bubble color represents the
   p-value.

3.5. Construction of JBOL ingredients-disease-target-pathway network

   Finally, we constructed a ingredients-disease-target-pathway network
   and visualized using Cytoscape 3.10.1 ([131]Fig. 7), containing 185
   nodes and 905 edges. The degree value indicated that kaempferol,
   apigenin, baicalein, vanillic acid, and isoimperatorin were the top 5
   active components. The top five targets included EGFR, AKT1, PIK3CA,
   PIK3CD and MAP2K1.

Fig. 7.

   [132]Fig. 7
   [133]Open in a new tab

   The ingredients-disease-target-pathway network of JBOL in ant-IPF. The
   green rhombus nodes represent the targets, the lavender hexagon nodes
   represent the herb, the pink circle nodes represent the ingredients,
   orange circle nodes represent the common ingredients, the purple
   octagon nodes represent the pathways.

3.6. Analysis of molecular docking

   Molecular docking was used to evaluate the interaction of screening
   active ingredients and core targets to validate the findings from
   network pharmacology. Based on synthesis analysis of the degree values
   from the PPI network and KEGG, 9 key protein targets were chosen. 9
   active components that achieved high degree values through the
   “herb-active ingredient-targets” network was used for molecular
   docking. Realistic ingredients with high contents from previous
   experiments were also taken into account. Based on the aforementioned
   outcome, we docked 9 receptor proteins-AKT1 (PDB ID: [134]4EJN), EGFR
   (PDB ID: [135]4LQM), SRC (PDB ID: [136]1T65), STAT3 (PDB ID:
   [137]5AX3), MMP9 (PDB ID: [138]4WZV), MAPK3 (PDB ID: [139]2ZOP), CASP3
   (PDB ID: [140]3KJF), HSP90AA1 (PDB ID: [141]5VYY), MTOR (PDB ID:
   [142]4DRI)-with 9 ingredients as ligands, including kaempferol,
   apigenin, methylnissolin, baicalein, liquiritigenin, phellopterin,
   peimine, cryptotanshinone, and wogonin. An average free binding energy
   value of −8.27 kcal·mol-1 was found in the docking results of 81 groups
   of active ingredient-target protein, indicating tight binding activity
   in screening active components with core targets ([143]Fig. 8). The top
   eight compound-target interactions and the binding mode with the
   highest free binding energy score were visualized using PyMoL2.5.8 and
   Discovery Studio 2024 ([144]Fig. 9 and [145]Table 4).

Fig. 8.

   [146]Fig. 8
   [147]Open in a new tab

   Heat map of Molecular docking binding energy.

Fig. 9.

   [148]Fig. 9

   [149]Fig. 9

   [150]Fig. 9

   [151]Fig. 9
   [152]Open in a new tab

   Molecular docking mode of core targets and screened active ingredients:
   Apigenin-MMP9 (A), baicalein-MMP9 (B), cryptotanshinone-AKT1 (C),
   cryptotanshinone-MTOR (D), kaempferol-MMP9 (E), liquiritigenin-MMP9
   (F), peimine-MTOR (G), wogonin-MMP9 (H).

Table 4.

   Molecular docking results and interaction analysis for identified
   compounds.
   Ligands Compounds Receptor Protein PDB ID Hydrogenbonds Hydrophobic
   interaction E. kcal·mol^−1
   peimine MTOR 4DRI TYR-57, ARG-73 TRP-2102, TRY-2105, PHE-77, ILE-87,
   VAL-86, PHE-2039, PHE-2108 −11.1
   cryptotanshinone AKT1 4EJN SER-205 LYS-268, LEU-264, TYR-272, VAL-270
   −11.0
   cryptotanshinone MTOR 4DRI SER-2035 PHE-77, PHE-2039 −10.4
   apigenin MMP9 4WZV ALA-189 VAL-223 −10.4
   kaempferol ALA-189, GLU-227 VAL-223 −10.2
   baicalein LEU-188, ALA-189, MET-247, PRO-246 VAL-223 −10.1
   liquiritigenin ALA-189, TYR-248 VAL-223, LEU-188 −10.0
   wogonin ALA-189 VAL-223 −9.7
   [153]Open in a new tab

   The smaller the value of the binding free energy, the lower the binding
   free energy and the stronger the binding. A ligand would demonstrate a
   certain binding activity, a satisfactory binding activity, and a tight
   binding activity with the receptor if its binding energy was less than
   −4.25 kcal·mol-1, -5.0 kcal·mol-1, and -7.0 kcal·mol-1, respectively
   [[154]36]. Out of all of them, the free binding energy of MTOR with
   peimine exhibited the highest affinity, it was −11.1 kcal mol^−1. The
   interaction was mainly reflected in the formation of hydrogen bonds
   with TYR-57, and ARG-73 residues and hydrophobic interactions with
   TRP-2102, TRY-2105, PHE-77, ILE-87, VAL-86, PHE-2039, and PHE-2108
   residues of MTOR. Other groups of target protein-active ingredient,
   like apigenin-MMP9 and cryptotanshinone-AKT1, also exhibited high free
   binding energies, all of which were less than −9.4 kcal·mol-1.
   Additionally, this investigation discovered that MMP9 could interact
   with apigenin, kaempferol, baicalein, liquiritigenin, and wogonin,
   among additional compounds. The interaction was concerned with the
   formation of hydrogen bonds and hydrophobic interactions, whose
   residues included ALA-189, VAL-223, etc. We predicted that screening
   compounds and targets would be important in the management of IPF based
   on the aforementioned findings. It should be mentioned, nonetheless,
   that more experimental verification is required for the potential
   active components.

3.7. Molecular dynamic simulation

   The molecular dynamic simulation was run to investigate the binding
   abilities of AKT1 to cryptotanshinone, MTOR to peimine, and MMP9 to
   apigenin. Root Mean Square Fluctuation (RMSD) curves of protein
   indicated position variability between the conformation and initial
   conformation during the MD simulation. As shown in [155]Fig. 10. The
   protein structures of peimine-MTOR and apigenin-MMP9 were stable after
   40 ns, cryptotanshinone- AKT1 remained more stable after 60 ns, which
   signified the protein structure of three complexes showed a slight
   modification, but the bindings were mostly stable. The fact that the
   RMSD curve was not disrupted indicated that the core chemicals might
   remain tightly linked to screened proteins throughout the simulation
   without separating from the protein pocket.

Fig. 10.

   [156]Fig. 10
   [157]Open in a new tab

   Molecular dynamics simulation of peimine-MTOR, cryptotanshinone-AKT1,
   and apigenin-MMP9. (A) RMSD values extracted from the protein-ligand
   docked complexes. (B) H bond formation among peimine-MTOR,
   cryptotanshinone-AKT1, and apigenin-MMP9.

   The quantity of hydrogen bonding in the protein-compound interactions
   assessed the binding strength, the strongest and most dense hydrogen
   bonds were found in peimine-MTOR among them. According to MM/PBSA
   calculations, binding free energy represents the stability and
   variation of ligand and binding protein properties. The total binding
   free energies of peimine-MTOR, cryptotanshinone-AKT1, and apigenin-MMP9
   were −37.78, −33.36, and −21.44 kJ/mol, respectively. Peimine-MTOR
   exhibited the lowest binding free energy and the maximum binding
   intensity, in agreement with the outcomes of molecular docking and
   dynamic simulation.

4. Discussion

   The development of pulmonary fibrosis is associated with the
   stimulation of inflammatory cytokines and the promotion of the
   epithelial-mesenchymal transition (EMT), which are both aided by the
   significant upregulation of alveolar epithelial cells and vascular
   endothelial cells in hyperplastic during the course of IPF [[158]37].
   Moreover, it involves the following: angiogenesis, oxidative stress,
   extracellular matrix deposition, insufficient autophagy, immune
   response misaligned, imbalance of the coagulation and fibrinolysis
   systems, etc. [[159]38]. In some instances, the above-mentioned
   provides future directions of treatment, prevention in clinical
   practice of anti-fibrosis. Modern medicine has not by any means
   clarified the pathophysiology of IPF. Patients still have unmet medical
   demands despite the wide range of conventional medication alternatives
   available for the management of IPF because they don't feel satisfied
   with their treatment or stick with it. The application of TCM,
   particularly with regard to the use of herbs or extracts from them, has
   garnered significant attention in the wake of its exceptional
   efficiency in treating COVID-19. TCM classifies the IPF into two
   categories: “pulmonary flaccidity” and “pulmonary arthralgia”
   [[160]39]. The pathophysiology of “pulmonary flaccidity” and “pulmonary
   arthralgia” is described in ancient medical literature as being
   characterized by an embodied deficiency of vital qi, blood and qi
   stagnation, and phlegm stasis blocking [[161]40]. The principles of TCM
   include replenishing qi and nourishing yin, activating blood and
   dredging collaterals, and resolving phlegm and masses as the entry
   point to resist IPF.

   On the other hand, there are rarely any reports on the range of TCM
   preparations that have obvious advantages in treating IPF. JBOL is a
   hospital prescription that was developed based on clinical experience.
   It possessed the advantage in therapy of pulmonary interstitial
   fibrosis and acute lung injury, some of which benefits include
   replenishing qi and nourishing yin, clearing heat and detoxifying and
   relieving cough, and resolving phlegm. Prior multi-center clinical
   trials have demonstrated that JBOL provides unique clinical treatment
   advantages with a notable improvement in lung function indicators and
   blood gas analysis, suggesting that JBOL might be an additional
   therapeutic scheme option for the treatment of IPF. A prior study
   revealed that JBOL prevented the production of inflammatory factors,
   which in turn helped to ameliorate bleomycin-induced IPF in rats.
   Anti-IPF was also validated by JBOL inhibiting EMT in vitro experiment
   [[162]41] and downregulating the expression of ERK1/2, JNK, and p38 by
   decreasing phosphorylation [[163]42]. All these results indicate that
   JBOL plays a critical role in anti-IPF with multiple components and
   pathways. Nevertheless, data mining and experimental validation are
   still required for active components and mechanisms. It is critical for
   the determination of the key indicators to ensure their safety and
   efficacy in clinical use. In order to lay the groundwork for more
   investigation into the function mechanism, this research aims to
   tentatively forecast the potential active components and mechanisms
   based on realistic constituents of JBOL in anti-IPF using biological
   methodologies and analysis. In the meanwhile, it also contributes to
   offering a point of reference for the determination of the quality
   control index and formulation's optimization.

   Based on 118 components found in earlier studies, 34 active components
   were selected for this research, and the key targets were taken via the
   SwissTargetPrediction and TCMSP databases. 1169 IPF-related targets
   were obtained from OMIM, DisGeNET, and GeneCards databases, and 122
   component-IPF targets were determined with intersection. We screened
   out the top 8 potential active ingredients by analyzing rank of degree
   value in the herb-compound-target network, including kaempferol,
   apigenin, methylnissolin, baicalein, liquiritigenin, phellopterin,
   peimine, cryptotanshinone.

   Kaempferol is one of the core ingredients existed in 5 herbs of JBOL,
   which exhibited a wide range of inhibition on the expression of
   fibrosis-related genes in transforming growth factor beta 1 (TGF-β1)
   and triggered the process of EMT by reversing the expression of α-SMA
   and E-cadherin [[164]43,[165]44]. Previous in vitro experiments have
   shown that kaempferol inhibits TGF-β1-induced fibroblast proliferation
   and myofibroblast differentiation by promoting the AMPK/PPAR-γ pathway,
   which in turn reduces excessive collagen production and the formation
   of scar tissue in fibrotic conditions [[166]45]. Additional diverse
   pathways, like the TGF-β/Smad and Wnt/β-catenin pathways, were
   implicated in the anti-fibrosis process [[167]46]. Moreover, kaempferol
   could prevent the formation of inflammatory cytokines and regulate
   autophagy to have an anti-fibrosis effect [[168]47]. Recent studies
   have shown that kaempferol influences the progression of organ fibrosis
   through the bone morphogenic protein 7-Smad1/5, Notch pathways, and
   autophagy [[169]48], respectively.

   Apigenin is a significant member of the flavonoid family with
   anti-fibrosis and anti-inflammatory characteristics. At the same time,
   the exact mechanism of apigenin in inhibiting lung fibrosis and
   inflammation has not yet been articulated. It is involved in Salvia
   miltiorrhiza Bge; Lonicera japonica Thunb; Scutellaria baicalensis
   Georgi; Codonopsis pilosula (Franch.) Nannf in JBOL. Phosphorylated p65
   levels were downregulated by apigenin, which also reduced the
   expression of pro-inflammatory cytokines such as caspase-3,
   interleukin-1β (IL-1β), TNF-α and TGF-β that were motivated by
   lipopolysaccharide (LPS) [[170]49]. Numerous investigations have
   demonstrated that oxidative stress plays a major role in the
   progression of lung fibrosis. TGF-β, a powerful fibrosis promoter, was
   produced in a condition of high oxidation. Apigenin can decrease TGF-β
   expression in a concentration-dependent way and lessen the accumulation
   of collagen [[171]50]. In short, it can prevent lung cells from
   experiencing oxidative stress and fibrosis cytokines [[172]51]. Yet,
   further experimental validation is required to confirm the active
   mechanism of apigenin due to its low levels of solubility in vitro and
   bioavailability in vivo. Zhang J. et al. [[173]52] looked examined the
   utilization of apigenin loaded in a biodegradable carrier (apigenin-NP)
   to treat pulmonary fibrosis and found that it could dose-dependently
   arrest cells in the G0/G1 phases.

   Baicalein is a common plant flavonoid with a variety of advantageous
   pharmacological characteristics, including anti-inflammatory,
   antioxidative, anti-apoptotic, pro-immunoregulatory functions. Sun
   [[174]53] investigated that baicalein inhibited the differentiation of
   lung fibroblasts into myofibroblasts via downregulating the levels of
   Col1a1 and CTGF. In addition, baicalein significantly improved the
   proliferation and activation of NIH/3T3 by alleviating TGF β1-induced
   type I collagen [[175]54]. EGFR and AKT1 are indispensable targets for
   anti-IPF, while peimine could suppress the phosphorylation of EGFR,
   PI3K, and AKT1 to inhibit fibroblast activation [[176]55]. Yuan et al.
   [[177]56] demonstrated that baicalein arrests the development of
   pulmonary fibrosis by suppressing the TGF-β1/Smad signaling pathway by
   downregulating the BLM-induced of lung Sirt3 expression. In addition,
   it also attenuated markedly mRNA levels of profibrotic
   senescence-associated secretory phenotype (SASP) factors, including
   MCP-1, PAI-1, TNF-a, MMP-10 and MMP-12 to relieve the senescence of
   lung fibroblasts.

   Major tanshinone derived from Salvia miltiorrhiza Bge is identified as
   cryptotanshinone. According to recent research, cryptotanshinone
   inhibits the expression of COL1A1 and α-SMA, which provides the effects
   of anti-inflammatory, anti-angiogenesis, and anti-lymphangiogenesis in
   addition to preventing atherosclerosis and Alzheimer's disease
   [[178]57,[179]58], but also exerts multiple effects on
   anti-inflammatory, anti-angiogenesis, and anti-lymphangiogenesis
   through inhibiting the expression of α-SMA, collagen-I And the
   phosphorylation level of Stat3 also participate in anti-IPF [[180]59].
   To put it briefly, the aforementioned components are speculated to be
   the material basis of IPF.

   The above-mentioned studies stated that screened ingredients are
   closely linked with JBOL in anti-IPF. Additionally, three core
   components were found in Salvia miltiorrhiza Bge and Astragalus
   membranaceus (Fisch.) Bge, and two key components involved in
   Scutellaria baicalensis Georgi, Lonicera japonica Thunb, and
   Glycyrrhiza uralensis Fisch, respectively. The herb-component-target
   network predicts that Astragalus membranaceus (Fisch.) Bge and Salvia
   miltiorrhiza Bge are essential groups in JBOL and correspond to the IPF
   therapeutics in TCM, which followed the principles of tonifying the
   spleen to nourish qi and promoting blood circulation to remove blood
   stasis. Thus, we anticipate that the network conclusion will serve as a
   guide for JBOL compatibility and dosage optimization.

   We performed GO and KEGG enrichment analysis to investigate the
   mechanism of JBOL in the treatment of IPF. Numerous biological
   processes, including as protein phosphorylation, signal transduction,
   drug reaction, inflammation, and apoptosis, were identified by GO
   functional analysis as being connected to the effects of JBOL in
   anti-IPF. Cell senescence, inflammatory response, autophagy, and
   immune-related pathways, which focused on the AGE-RAGE, PI3K-Akt
   signaling pathway, and EGFR tyrosine kinase inhibitor resistance were
   among the enriched targets.

   Based on accumulating data, the EGFR-ligand system was crucial in the
   irreversible lung tissue scarring. In the meantime, the EGFR receptors
   and ligands were found to be overexpressed in IPF lungs [[181]60]. It
   has already been demonstrated that the stimulation of fibroblasts with
   TGF-β1 elevates the expression of various EGFR ligands, such as AREG,
   EREG, and HB-EGF, which in turn increases cell proliferation and the
   expression of profibrotic genes. These effects were reversed by the
   treatment of TGF-β1-stimulated fibroblasts with AREG siRNA or EGFR
   inhibitors [[182]61]. Moreover, several reviews indicate that the EGFR
   signaling cascade is upregulated in lung fibroblasts from patients with
   IPF [[183]62]. The EGFR pathway promotes fibrosis by phosphorylating
   EGFR, which in turn stimulates a number of downstream signaling
   pathways, such as PIK3/Akt pathway, PLC pathway, STAT pathway, and MAPK
   pathway [[184]63]. Previous research has demonstrated that the activity
   of EGFR in pulmonary fibrosis can be restricted through decreasing the
   expression of TGF and VEGF/PDGF/FGF1 receptors [[185]64], which in turn
   decreases collagen synthesis and ECM deposition [[186]65], and
   inhibiting fibroblast proliferation.

   Abnormal activation of EMT in IPF is associated with accelerated aging,
   which is the key to changes in wound healing [[187]66]. The advanced
   glycation end-products (AGEs) have been proposed as markers of
   oxidative stress and aging. AGEs were proven to postpone the wound
   healing of epithelial cells and TGF-β1 synthesis in vitro studies, and
   their ligand RAGEs decrease led to AGEs accumulation. All of these
   findings indicated that the formation and progression of IPF is caused
   by an imbalance between AGEs and RAGEs [[188]67], which are correlating
   with the upward regulation of the AGEs-RAGEs pathway.

   The PI3K/AKT signaling pathway has been significant in the
   intracellular regulation of cell growth, proliferation, autophagy,
   metabolism, and apoptosis, and it has been identified as a hotspot
   route in IPF [[189]68,[190]69], which play an important role of
   cellular proliferative and migration [[191]70], cell adhesion
   [[192]71]. The activation of the PI3K/AKT pathway accelerates the
   overexpression of α-SMA and TGF-β leading to the formation of pulmonary
   fibrosis [[193]72]. In addition, it promotes pulmonary fibrosis by
   upregulating its downstream cytokines including mTOR and HIF-1alpha, to
   regulate inflammatory immune response, induces EMT and ECM deposition
   [[194]73,[195]74]. In addition to regulating the biological processes
   associated with IPF alone, the pathophysiology of IPF involves a range
   of interaction and crosstalk with signaling pathways, such as TGF,
   VEGF, WNT, mTOR, Jun N-terminal kinase, CTGF, Hedgehog, and Notch
   pathway [[196]75]. Recent studies suggest that the suppression of
   PI3K/Akt signaling pathway induces the down-regulation of HIF-1α and
   VEGF expression in lung tissue, which in turn interfere with
   angiogenesis and the IPF process [[197]76]. It is necessary to confirm
   our hypothesis that these pathways play an essential part in the JBOL
   therapy mechanism in IPF.

   To explore the potential molecular mechanism of JBOL in the treatment
   of IPF and verify the prediction results of network pharmacology,
   molecular docking studies, and molecular dynamic simulation were
   performed with core bioactive components and targets based on the
   outcome of the herb-active ingredient-target network and PPI network.
   The results of docking suggested that the components exhibited a
   satisfactory affinity with the key target proteins. These compounds and
   targets might contribute to the therapeutic effects of JBOL in IPF.
   Peimine-MTOR is among the compounds with the lowest binding affinities
   groups.

   To identify the active components and putative signaling pathways in
   this study, we first performed data mining analysis based on the actual
   measurement of 118 components and merged it with the quantitative
   analysis of chemical components in JBOL, and the outcomes of the
   bioinformation analysis were more rational. However, it should be
   pointed out that there are still some limitations in this research. As
   we all know, there are large discrimination of effects for the content
   and efficacy of active ingredients with different processing
   specifications and combinations [[198]77]. For example, Pinellia
   ternata (Thunb.) Breit has some other artillery products but is not
   distinguished in SwissTargetPrediction and other traditional Chinese
   medicine information databases. Consequently, it is essential to screen
   accurately for each artillery drug via scientific verification. On the
   other hand, the accuracy and timeliness of the chemical library need to
   be updated, such as the amount of evidence for toxicological mechanisms
   is insufficient, which may lead to biases in the analysis results. Our
   research findings will be further validated in pharmacodynamic and
   molecular biology experiment in the following time.

5. Conclusion

   In this study, we firstly forecasted an investigation of main active
   ingredients and prospective mechanisms in JBOL for the therapy of IPF,
   which utilizing network pharmacology, molecular docking, and molecular
   dynamic modeling, etc. Effect mechanism of JBOL in anti-IPF was
   concerned with decreasing autophagy, inflammatory, immunological
   response, and cell senescence through AGE-RAGE signaling pathway, EGFR
   signaling pathway, PI3K-Ak signaling pathway, and etc. In summary, we
   reviewed how JBOL exerted its therapeutic effects on anti-IPF through
   multi-target and multi-pathway in accordance with the core ingredients
   and signaling pathways predicted. Bioinformatics techniques were used
   to investigate the potential mechanism, providing the theoretical
   foundation for experimental research. The study has indicated where
   additional research should be performed in order to confirm the quality
   indicators and optimize the preparation technology. Numerous studies
   indicate that machine learning has been increasingly popular in recent
   years and has applications in a variety of fields, including biology,
   medicine, and healthcare [[199]78,[200]79]. This inspires us to combine
   this strategy with experiments to obtain additional verification. From
   a clinical application standpoint, only few research has verified that
   TCM compound preparation is feasible to enhance the clinically
   therapeutic effect of IPF, JBOL has enhanced the efficacy of IPF
   treatment and raised patient acceptance by decreasing adverse events.
   Based on extensive studies investigating, the future research should
   focus on the clinical applications, including determining optimal
   dosages, conducting clinical trials to explore the absorption,
   distribution, metabolism and elimination in the human body, as well as
   the safety and efficacy in the treatment of IPF. With the advantages of
   IPF treatment in multi-target and multi-pathway, JBOL, a novel TCM
   preparation, present a promising frontier in IPF therapy, and we
   supposed that conventional therapy assisted the application of
   phytopharmaceuticals would provide an option for reducing IPF symptoms,
   which could be a supplementary and alternative therapy.

CRediT authorship contribution statement

   Xinru Han: Data curation, Writing – original draft. Aijun Zhang:
   Writing – review & editing. Zhaoqing Meng: Resources, Supervision. Qian
   Wang: Software, Validation. Song Liu: Data curation, Visualization.
   Yunjia Wang: Software. Jiaxin Tan: Validation. Lubo Guo:
   Conceptualization, Methodology, Software. Feng Li: Methodology, Project
   administration, Supervision.

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgments