Abstract

   Novel treatments for autoimmune hepatitis (AIH) are highly demanded due
   to the limitations of existing therapeutic agents. Costunolide is a
   promising candidate due to its anti-inflammatory and hepatoprotective
   function, but its effect in AIH remains obscure. In this study, we
   integrated network pharmacology and experimental validation to reveal
   the effect and mechanism of costunolide in AIH. A total of 73 common
   targets of costunolide and AIH were obtained from databases. Pathway
   enrichment analysis indicated that PI3K-AKT pathway was the core
   pathway of costunolide in AIH. Protein–protein interaction network
   analysis and molecular docking revealed that SRC and IGF1R might play
   critical roles. In two murine AIH models, costunolide significantly
   attenuated liver injury, inflammation, and fibrosis reflected by the
   liver gross appearance, serum transaminases, necrosis area, spleen
   index, immune cell infiltration, and collagen deposition. Western blot
   and immunohistochemistry confirmed that phosphorylated AKT, SRC, and
   IGF1R were upregulated in AIH models, and costunolide administration
   could inhibit the phosphorylation of these proteins. In summary,
   costunolide significantly ameliorates murine AIH. The therapeutic
   effect might work by suppressing the activation of PI3K-AKT pathway and
   inhibiting the phosphorylation of SRC and IGF1R. Our research reveals
   the potent therapeutic effect of costunolide in AIH and the potential
   role of SRC and IGF1R in AIH for the first time, which may further
   contribute to the novel drug development for AIH and other autoimmune
   diseases.

   Keywords: costunolide, autoimmune hepatitis, network pharmacology,
   immune-mediated liver injury, PI3K-AKT, SRC, IGF1R

1. Introduction

   Autoimmune hepatitis (AIH) is a chronic inflammatory liver disease due
   to immune-mediated destruction of hepatocytes. The etiology of AIH
   involves the interaction of both genetic background and environmental
   triggers, which is not entirely understood. Its pathophysiological
   processes are believed to involve specific genetic traits, molecular
   mimicry, impaired immunoregulatory mechanisms, etc. [[38]1]. AIH is
   characterized biochemically by the elevation of serum transaminase,
   serologically by the presence of autoantibodies and elevated
   immunoglobulin G (IgG) levels, and histologically by interface
   hepatitis [[39]2]. It occurs worldwide in all ethnicities, affecting
   all ages with distinct female preponderance [[40]2]. Although most
   patients respond well to the standard immunosuppressive therapy of
   steroids and azathioprine, insufficient response and intolerable side
   effects occur in 10–20% of patients [[41]3]. Long-term or uncontrolled
   AIH may lead to progressive fibrosis, cirrhosis, liver failure, and
   even hepatocellular carcinoma [[42]4]. Therefore, identification of
   novel therapeutic drugs and clarifying the underlying mechanism is of
   great significance for the treatment of AIH.

   Traditional Chinese medicine (TCM) and natural products may provide
   some new options for the drug discovery of AIH [[43]5]. Costunolide
   (COS) is a well-known sesquiterpene lactone in the germacranolides
   series isolated from Aucklandiae Radix (Mu Xiang in Chinese). Numerous
   preclinical studies have indicated that costunolide possesses
   antioxidative, anti-inflammatory, antiallergic, anticancer, and
   antidiabetic properties [[44]6]. Remarkably, costunolide exerts a
   powerful anti-inflammation effect in multiple immune-related diseases
   such as carrageenan-induced paw edema and lung inflammation,
   ethanol-induced gastric ulcer, lipoteichoic acid-induced acute lung
   injury, and dextran sulfate sodium (DSS)-induced murine colitis
   [[45]7,[46]8,[47]9,[48]10,[49]11,[50]12]. Meanwhile, costunolide
   exhibits a hepatoprotective effect against lipopolysaccharide and
   D-galactosamine-induced acute liver injury and alcoholic liver injury
   [[51]13,[52]14]. It also ameliorates liver fibrosis in vitro and in
   vivo [[53]15]. However, the effect of costunolide in AIH remains
   obscure.

   Network pharmacology was proposed by integrating network biology and
   polypharmacology [[54]16]. It is capable of describing complexities
   among biological systems, drugs, and diseases from a network
   perspective, which has been applied in many studies to explore the
   molecular mechanism of drugs [[55]17,[56]18,[57]19,[58]20,[59]21]. In
   this study, we investigated the effect of costunolide in two murine
   models of AIH: Concanavalin A (ConA)-induced acute immune-mediated
   hepatitis and human Cytochrome P4502D6 (CYP2D6) plasmid
   injection-induced chronic autoimmune hepatitis [[60]22]. In addition,
   we integrated network pharmacology with experimental validation to
   clarify underlying mechanisms. Our research indicates that costunolide
   might be a potential option for the therapy of AIH. The flow chart of
   this research is shown in [61]Figure 1.

Figure 1.

   [62]Figure 1
   [63]Open in a new tab

   Flow chart of research. Detailed procedures of analysis are described
   in Materials and Methods. Briefly, AIH-related targets were collected
   from DisGeNET and GeneCards. The potential targets of costunolide were
   obtained from SwissTargetPrediction and PharmMapper. The potential
   targets of costunolide in AIH were generated by taking intersection of
   targets above. KEGG and GO enrichment analysis were performed using
   DAVID and Metascape. STRING database and Cytoscape were used for the
   construction of PPI network analysis. The docking between costunolide
   and key targets were conducted by PyMol and AutoDock. In vivo
   experiments were performed to verify the effect of costunolide on AIH
   and its underlying mechanism.

2. Results

2.1. Identification of the Potential Targets of Costunolide in AIH

   Costunolide is a lactone compound isolated from Aucklandiae Radix. The
   two-dimensional (2D) structure of costunolide is shown in [64]Figure
   2a. The pharmacological and molecular properties of costunolide are
   shown in [65]Table 1. Costunolide shows good oral bioavailability (OB)
   and high drug likeness (DL) in both the Traditional Chinese Medicine
   Systems Pharmacology Database and Analysis Platform (TCMSP) and
   SwissADME.

Figure 2.

   [66]Figure 2
   [67]Open in a new tab

   Potential targets of costunolide in AIH: (a) 2D structure of
   costunolide (PubChem Identifier: CID 5281437
   [68]https://pubchem.ncbi.nlm.nih.gov/compound/5281437#section=2D-Struct
   ure (accessed on 26 February 2022)), (b) Venn diagram of intersecting
   targets of costunolide and AIH.

Table 1.

   Pharmacological and molecular properties of costunolide.
    MW    AlogP Hdon Hacc OB (%) Caco-2 (nm/s) BBB   DL  FASA- TPSA  RBN
   232.35 4.02   0    2   29.07      1.28      1.42 0.11 0.36  26.30  0
   [69]Open in a new tab

   Abbreviations: MW, molecular weight; AlogP, low lipid/water partition
   coefficient; Hdon, hydrogen bond donors; Hacc, hydrogen bond acceptors;
   OB, oral bioavailability; Caco-2, Caco-2 permeability; BBB, blood brain
   barrier; DL, drug likeness; FASA-, fractional water accessible surface
   area of all atoms with negative partial charge; TPSA, topological polar
   surface area; RBN, rotatable bonds number. The molecular and
   pharmacological properties data of costunolide were obtained from TCMSP
   [[70]23] and SwissADME [[71]24].

   From SwissTargetPrediction, we got 33 potential targets, and we
   retrieved 174 targets from PharmMapper. After merging data, removing
   duplicates, and confirming in UniProt, we got 200 potential targets of
   costunolide. Using the search word “autoimmune hepatitis”, we obtained
   190 related genes in DisGeNET and 1183 genes which score >5 in
   GeneCards. After merging and removing duplicates, we obtained 1252
   genes related to AIH. Via intersecting targets of costunolide and
   AIH-related genes, we obtained 73 potential targets of costunolide in
   AIH ([72]Figure 2b).

2.2. Recognition of Enriched Pathway of the Potential Targets of Costunolide
in AIH

   The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis
   was performed using The Database for Annotation, Visualization and
   Integrated Discovery (DAVID). The top 20 clusters were selected based
   on the p-value and were presented in the bubble chart ([73]Figure 3).
   The potential targets of costunolide in AIH were involved in pathways
   in cancer (hsa05200), PI3K-AKT signaling pathway (hsa04151), lipid and
   atherosclerosis (hsa05417), MAPK signaling pathway (hsa04010),
   proteoglycans in cancer (hsa05205), etc.

Figure 3.

   [74]Figure 3
   [75]Open in a new tab

   KEGG pathway enrichment analysis of intersecting targets of costunolide
   and AIH.

   Meanwhile, based on the p-value of Gene Ontology (GO) enrichment
   analysis, the top 10 significantly enriched terms in Biological Process
   (BP), Cellular Component (CC), and Molecular Function (MF) categories
   were selected ([76]Figure 4). Potential targets of costunolide in AIH
   were mainly involved in biological processes such as cellular response
   to lipid (GO:0071396), response to hormone (GO:0009725), and response
   to lipopolysaccharide (GO:0032496). The main cellular component terms
   were membrane raft (GO:0045121), vesicle lumen (GO:0031983), and
   receptor complex (GO:0043235). The main molecular function terms were
   nuclear receptor activity (GO:0004879), protein tyrosine kinase
   activity (GO:0004713), and protein domain specific binding
   (GO:0019904).

Figure 4.

   [77]Figure 4
   [78]Open in a new tab

   Gene Ontology (GO) enrichment analysis of intersecting targets of
   costunolide and AIH.

2.3. Topological Network Analysis of the Potential Targets of Costunolide in
AIH

   Seventy-three potential targets of costunolide in AIH were imported
   into The Search Tool of Retrieval of Interacting Genes (STRING) to
   construct the protein–protein interaction (PPI) network. After removing
   disconnected nodes, the network contained 61 nodes and 156 edges
   ([79]Figure 5a). The average number of neighbors is 5.115 and the
   clustering coefficient is 0.317. Using plug-in CytoHubba, we found top
   10 genes that might play critical roles in the therapeutic effect of
   costunolide in AIH, namely MAPK1, SRC, GRB2, EGFR, LCK, ESR1, JAK2,
   IGF1, HSP90AA1, and IGF1R, ranked by Maximal Clique Centrality (MCC)
   ([80]Figure 5b).

Figure 5.

   [81]Figure 5
   [82]Open in a new tab

   Protein–protein interaction network analysis. (a) Protein–protein
   interaction network of intersecting targets of costunolide and AIH
   visualized by Cytoscape. (b) The core subnetwork constructed by
   CytoHubba.

2.4. Molecular Docking

   Based on the pathway enrichment analysis and literature review, SRC and
   IGF1R ranked high in the PPI network analysis and were reported to
   regulate the PI3K-AKT pathway in inflammation-related diseases.
   However, their roles in AIH were not elucidated yet. Thus, we chose SRC
   and IGF1R as our interested targets of costunolide in AIH. Using PyMol
   and Autodock, molecular docking visually showed the interaction between
   costunolide and SRC, IGF1R. The diagrams of drug-target binding mode
   were shown left and the details right ([83]Figure 6). The dotted yellow
   line represents the hydrogen bond. The lower binding energy indicates
   higher stability. The binding energies between costunolide and SRC and
   between costunolide and IGF1R were −7.7 and −5.9 kcal/mol, respectively
   ([84]Table 2).

Figure 6.

   [85]Figure 6
   [86]Open in a new tab

   Molecular docking of costunolide with SRC and IGF1R.

Table 2.

   Binding energies of costunolide with SRC and IGF1R.
   Target PDB ID Binding Energy (kcal/mol)
    SRC    6E6E            −7.7
   IGF1R   3NW7            −5.9
   [87]Open in a new tab

2.5. Costunolide Attenuated ConA-Induced Acute Hepatitis

   Firstly, we administered costunolide in the ConA-induced acute
   hepatitis model. The liver gross appearance of ConA+Solvent mice showed
   hepatic congestion and suppuration on the surface, while ConA+COS liver
   showed a milder change ([88]Figure 7a). In hematoxylin-eosin (H&E)
   staining, we observed that ConA induced necrosis of liver parenchyma.
   Costunolide administration significantly restrained the hepatic
   necrosis, while the necrosis score of ConA+Solvent and ConA+COS showed
   a significant difference ([89]Figure 7b,c). Similar to the histological
   results, level of the liver enzyme alanine aminotransferase (ALT) and
   aspartate aminotransferase (AST) in serum surged in the ConA+Solvent
   group compared with Solvent and COS group, while costunolide
   significantly suppressed it ([90]Figure 7d,e). Notably, there was no
   obvious difference of gross appearance, H&E staining, and liver enzyme
   level between the Solvent group and the COS group.

Figure 7.

   [91]Figure 7
   [92]Open in a new tab

   Costunolide attenuated ConA-induced acute hepatitis. (a) Liver gross
   appearance. (b) Representative H&E staining of liver sections. (c)
   Necrosis score of H&E staining of liver sections. (d) Effect of
   costunolide on serum ALT levels. (e) Effect of costunolide on serum AST
   levels. Data shown in c–e are from 6 individual mice per group. * p <
   0.05, ** p < 0.01, *** p < 0.001.

2.6. Costunolide Ameliorated Chronic Murine Autoimmune Hepatitis

   In our chronic murine AIH model, the spleens of AIH mice were enlarged
   significantly, while costunolide suppressed it ([93]Figure 8a). There
   was a significant difference of spleen indexes between AIH+Solvent and
   AIH+COS mice ([94]Figure 8b). Unlike ConA-induced acute hepatitis,
   there was no obvious change of the liver gross appearance in the
   chronic murine AIH model ([95]Figure 8c). H&E staining of liver
   sections showed that costunolide inhibited immune cell infiltration in
   AIH mice, which is the typical characteristic of AIH ([96]Figure 8c).
   There was a significant difference of inflammation score between
   AIH+Solvent and AIH+COS group ([97]Figure 8d). The ALT level showed a
   similar trend as histological results ([98]Figure 8e). There was no
   significant change of AST level in our chronic model ([99]Figure 8f).
   Further, Sirius red staining of liver sections indicated that
   costunolide reduced liver fibrosis in the chronic AIH model
   ([100]Figure 8g).

Figure 8.

   [101]Figure 8
   [102]Open in a new tab

   Costunolide ameliorated chronic murine autoimmune hepatitis. (a) Spleen
   gross appearance. (b) Spleen index (mg/g). (c) Representative liver
   gross appearance and H&E staining of liver sections (black arrows
   indicate immune cell infiltration). (d) Inflammation score of H&E
   staining of liver sections. (e) Effect of costunolide on serum ALT
   levels. (f) Effect of costunolide on serum ALT levels. (g)
   Representative Sirius-red staining of liver sections (Red color
   indicates collagen deposition). Data shown in (b,d,e) are from 5
   individual mice per group. * p < 0.05, ** p < 0.01, **** p < 0.0001.

   To further elucidate the landscape of immune cell infiltration in
   different groups, we conducted immunohistochemistry (IHC) of CD4 and
   F4/80 in liver sections. The IHC result revealed the increased
   infiltration of CD4^+ T cells and macrophages in AIH+Solvent mice,
   while costunolide suppressed this trend substantially ([103]Figure
   9a,b).

Figure 9.

   [104]Figure 9
   [105]Open in a new tab

   Costunolide inhibited immune cell infiltration in chronic murine
   autoimmune hepatitis. (a) Immunohistochemistry of CD4^+ T cells. (b)
   Immunohistochemistry of F4/80^+ macrophages.

2.7. Costunolide Inhibited the Activation of PI3K-AKT Pathway and Suppressed
the Phosphorylation of SRC and IGF1R in AIH

   Based on the pathway enrichment analysis, PPI network analysis and
   literature review, we investigated the protein level change of
   phosphorylated AKT, SRC, and IGF1R in liver tissue of two AIH models.
   The phosphorylation of AKT, SRC, and IGF1R were significantly increased
   in the ConA model, whilst costunolide dramatically inhibited it
   ([106]Figure 10a, [107]Supplementary Figure S1a). Similar results were
   found in the chronic AIH model ([108]Figure 10b, [109]Supplementary
   Figure S1b).

Figure 10.

   [110]Figure 10
   [111]Open in a new tab

   Costunolide inhibited the phosphorylation of AKT, SRC, and IGF1R. (a)
   Protein expression in ConA-induced acute hepatitis. Two samples from
   two different mice in each group are presented. The numbers indicate
   the ratio of pAKT/AKT, pSRC/SRC, and pIGF1Rβ/IGF1Rβ, respectively. (b)
   Protein expression in chronic murine autoimmune hepatitis. (c)
   Immunohistochemistry of phospho-AKT in ConA-induced acute hepatitis.
   (d) Immunohistochemistry of phospho-AKT in chronic murine autoimmune
   hepatitis.

   To further confirm this trend, we conducted IHC of phospho-AKT in liver
   sections. The phosphorylated AKT was mainly expressed in nuclei of
   hepatocytes in Solvent and COS mice, and its expression level was
   higher in two murine AIH models, mostly in hepatocytes nuclei and
   cytoplasm as well as infiltrating immune cells. The phospho-AKT
   expression level in ConA/AIH+COS groups were lower than that in the
   ConA/AIH+Solvent groups ([112]Figure 10c,d).

3. Discussion

   Autoimmune hepatitis (AIH) is a chronic liver disease related to
   immunological tolerance disorders targeting hepatocytes. Although AIH
   has long been considered a rare disease, several studies have indicated
   that the incidence and prevalence of AIH were increasing worldwide
   [[113]25]. Although most patients respond well to traditional
   immunosuppressive therapy, many patients were troubled by the
   intolerance, insufficient response to medication, and recurrence after
   withdrawal of treatment [[114]2]. The discovery of novel therapeutic
   strategies is necessary for the management of the increasing population
   of AIH. Costunolide is a promising candidate due to its
   anti-inflammatory and hepatoprotective function in other disease
   models, whereas there is no study about its effect in AIH yet
   [[115]6,[116]12,[117]14]. In this study, we integrated network
   pharmacology and experimental validation to elucidate the therapeutic
   effect and underlying mechanism of costunolide against AIH.

   In the current study, we identified 73 targets that might participate
   in the effect of costunolide on AIH. Based on the KEGG enrichment
   pathway analysis and literature review, we speculated that costunolide
   might mainly restrain the disease progress of AIH through PI3K-AKT
   pathway. Generally, the PI3K-AKT pathway is an important intracellular
   signaling pathway for cell cycle progression, cell survival, and
   metabolism [[118]26]. It is reported that IHC of liver sections from
   AIH patients showed upregulation of phosphorylated AKT [[119]27]. In
   ConA-induced hepatitis, PI3K-AKT pathway is activated in hepatocytes
   and dendritic cells, promoting inflammation via inducing the
   transcription of IL-6 and activation of CD8^+ T cell responses,
   respectively [[120]28,[121]29]. Similarly, in our experiments, we
   observed a significant elevated protein level of phosphorylated AKT in
   the liver of ConA-induced hepatitis and chronic AIH model in WB, and
   costunolide administration induced a decrease of it. This finding was
   further confirmed by IHC showing a similar trend as in WB, while the
   upregulated phosphorylated AKT in AIH models was mainly expressed in
   hepatocytes and infiltrating immune cells. In hepatocytes, the
   activation of AKT generally plays a protective effect [[122]30].
   However, it was recently reported that AKT activation in hepatocytes
   was also associated with increased inflammatory cytokine expression
   [[123]28]. Activation of the PI3K-AKT-mTOR pathway and PI3K-AKT-FOXO1
   signaling axis regulates the activation, differentiation, and function
   of T cells in different conditions
   [[124]31,[125]32,[126]33,[127]34,[128]35]. Attenuation of PI3K
   signaling could lead to defects in T cell activation and neutrophil
   migration, restraining the inflammation in systemic lupus erythematosus
   (SLE) and rheumatoid arthritis (RA) models [[129]36,[130]37]. Several
   studies have reported the suppressing effect of costunolide on PI3K-AKT
   pathway in cancer cells and in DSS-induced colitis
   [[131]11,[132]26,[133]38]. These findings suggest that the therapeutic
   effect of costunolide against AIH might be related to the inhibition of
   PI3K-AKT pathway, whereas the specific cell types should be studied
   further. Contrary to our results, it was reported that phospho-AKT was
   decreased in liver of ConA-induced hepatitis and in ConA-treated L02
   cells compared with the saline control. This difference might be due to
   the different dosages of ConA and the time between ConA injection and
   sacrifice. Moreover, directly administrating ConA on L02 cells might
   not reflect the disease feature of immune-mediated hepatocyte
   destruction in vivo [[134]39]. Interestingly, the top three signaling
   pathways of KEGG analysis included lipid and atherosclerosis pathway.
   It is reported that pre-existing high-fat diet-induced non-alcoholic
   fatty liver disease (NAFLD) in mice potentiates the severity of AIH
   [[135]40]. Whether the lipid metabolism is involved in the
   pathophysiology of AIH and if costunolide could interfere with it
   awaits further research.

   Based on the PPI network analysis using STRING and CytoHubba, we
   further identified 10 important targets in the therapeutic effect of
   costunolide in AIH, namely MAPK1, SRC, GRB2, EGFR, LCK, ESR1, JAK2,
   IGF1, HSP90AA1, and IGF1R, ranked by MCC. After comprehensive
   consideration of KEGG pathway analysis, GO enrichment analysis,
   molecular docking, and literature review, we selected SRC and IGF1R as
   our candidate targets for further research. SRC belongs to the
   non-receptor protein-tyrosine kinases family, playing key roles in cell
   growth, division, migration, and survival signaling pathways [[136]41].
   Multiple studies indicated that SRC regulated the PI3K-AKT pathway in
   inflammation-related diseases such as multiple sclerosis (MS), Grave’s
   Disease (GD), and in the neutrophil extracellular traps (NET) formation
   [[137]42,[138]43,[139]44]. In our experiments, we found that the
   phosphorylation level of SRC was significantly upregulated in hepatic
   tissue in two AIH models and was drastically decreased in the
   costunolide group. Similar to the present literature, our results
   provided clues that SRC phosphorylation rose in AIH, and it might
   exacerbate AIH via the PI3K-AKT pathway, which demands further rigorous
   research to confirm it. Moreover, another protein in SRC family, LCK,
   also ranked high in our network analysis and it is an important
   regulator in T cell proliferation and function [[140]45]. Whether
   costunolide also interacts with LCK in its effect against AIH is an
   interesting focus for further studies. Insulin-like growth factor-1
   receptor (IGF1R), the primary signaling receptor of insulin-like growth
   factors (IGF) system, signals through the PI3K-AKT-mTOR and
   RAS-RAF-MEK-ERK pathways [[141]46]. Dysregulation of the IGF system has
   been directly related to altered CD4^+ T cell function in RA and GD
   [[142]47,[143]48]. Although there is no report about the direct
   contributory role of IGF1R in AIH yet, recently, its specific
   modulatory role in autoimmunity was confirmed by researchers in EAE
   mice. It augmented AKT-mTOR and STAT3 signaling, favoring Th17 cell
   differentiation over that of Treg cells [[144]46]. Our results showed
   that phosphorylated IGF1R level was increased in the liver of AIH
   models, and its level decreased in the costunolide group, indicating
   that IGF1R phosphorylation level was altered in AIH, and costunolide
   might attenuate AIH via inhibiting the phosphorylation of IGF1R. Since
   the appropriate balance of Th17 and Treg cells maintains immune
   tolerance and impairment of it permits progress of AIH, the variation
   of IGF1R might be related to AIH progression by modulating Th17/Treg
   balance by regulating the PI3K-AKT pathway, which should be studied
   further.

   In this study, we found costunolide possessed a strong therapeutic
   effect in two AIH models. Transaminase is the sensitive indicator of
   the liver injury. In our research, we observed that the administration
   of costunolide could lower the elevated transaminase level in two
   murine AIH models. The change of AST level was not statistically
   significant in our chronic AIH model, which might be due to the
   relatively overall lower degree of liver inflammation. As is shown in
   H&E staining, the necrosis area was also diminished by costunolide in
   the ConA model. The spleen index (spleen weight/ body weight) could
   roughly reflect the degree of chronic inflammation [[145]49]. We found
   that the spleen index in AIH group was significantly higher than the
   Solvent group and COS group, while the spleen index of COS+AIH group
   was significantly lower, indicating that costunolide could suppress the
   chronic inflammation. This finding was further supported by the
   suppressed immune cell infiltration in liver observed in H&E and the
   reduced CD4^+ T cells and macrophage infiltration in liver as is shown
   in IHC. Based on the present literature, costunolide inhibits the
   production of pro-inflammatory factors in macrophages and influences
   the differentiation of CD4^+ T cells in vitro [[146]50,[147]51]. Since
   T cells and macrophages both play pivotal roles in AIH, the specific
   mechanism of costunolide attenuating AIH and the exact role of PI3K-AKT
   pathway inside await further investigation such as flow cytometry. We
   also observed that costunolide suppressed liver fibrosis in chronic AIH
   model. That antifibrotic function might mainly be due to its inhibition
   of inflammation, which removed the etiological factor causing liver
   injury [[148]52]. It also could be related to its direct antifibrotic
   effect via inhibiting hepatic stellate cell (HSC) activation as
   reported before [[149]15]. It is worth noting that methacrylic acid
   copolymer (MAC)-coated PH-responsive mesoporous silica nanoparticles
   (MSNs) carrying costunolide significantly repressed liver fibrogenesis
   at a reduced dose in vitro and in vivo [[150]53]. Adopting similar
   technology might also reduce the dosage needed for the therapeutic
   effect of costunolide against AIH.

   This study has some limitations. The binding affinity of costunolide
   with potential targets awaits further verification using surface
   plasmon resonance (SPR) or bio-layer interferometry (BLI) assays. The
   specific cell types in which those phosphorylated protein level
   increased are not identified. Since in different cell types those
   signaling pathways might exert distinct roles, we plan to perform
   immunofluorescence to reveal it. The specific molecular mechanism is
   not clear due to no rescue experiment being conducted. In addition, the
   CYP2D6 model reflects type 2 AIH. The conclusion of our study remains
   to be confirmed in more AIH models.

4. Materials and Methods

4.1. Chemical Information Collection and Targets Retrieval of Costunolide

   The molecular and pharmacological properties data of costunolide were
   obtained from TCMSP [[151]23] and SwissADME [[152]24]. The canonical
   SMILES (Simplified Molecular-Input Line-Entry System), image, and SDF
   format file of 2D structure of costunolide were obtained from the
   PubChem database (PubChem Identifier: CID 5281437
   [153]https://pubchem.ncbi.nlm.nih.gov/compound/5281437#section=2D-Struc
   ture (accessed on 26 February 2022)) [[154]54]. SwissTargetPrediction
   [[155]55] and PharmMapper [[156]56] were utilized to predict the
   potential targets of costunolide. For outcome of PharmMapper, targets
   which norm fit score > 0.25 were included in subsequent analysis. After
   integrating data and removing duplicates, all proteins were verified in
   the UniProt database [[157]57].

4.2. Prediction of Potential Targets of Costunolide in AIH and Enrichment
Analysis

   AIH-related targets were collected from DisGeNET [[158]58] and
   GeneCards ([159]https://www.genecards.org/ (accessed on 21 June 2022))
   [[160]59]. In GeneCards, genes that Score > 5 were included in
   subsequent analysis. The potential targets of costunolide and
   AIH-related targets obtained in above steps were imported into an
   online Venn diagram drawing tool
   ([161]https://bioinformatics.psb.ugent.be/webtools/Venn/ (accessed on
   13 September 2022)).

   KEGG enrichment analysis was conducted using DAVID [[162]60,[163]61].
   Species was set as Homo sapiens. Threshold and EASE were set as 2 and
   0.1, respectively. GO enrichment analysis was performed using Metascape
   [[164]62]. Species was set as Homo sapiens and all genes in the genome
   were used as the enrichment background. Min overlap, p value cutoff,
   and min enrichment were set as 3, 0.01, and 1.5, respectively. Bubble
   chart and bar graph were plotted by
   [165]http://www.bioinformatics.com.cn (accessed on 14 September 2022),
   an online platform for data analysis and visualization.

4.3. Protein–Protein Interaction Target Network Construction and
Visualization

   STRING database ([166]https://cn.string-db.org/ (accessed on 14
   September 2022)) was utilized to construct a PPI network [[167]63]. The
   organism was set as Homo sapiens. The minimum required interaction
   score was set as 0.900 (highest confidence), and disconnected nodes in
   the network were hidden.

   Network visualization and analysis were performed using Cytoscape
   version 3.9.1 [[168]64]. The TSV-format file was downloaded from the
   STRING database and imported into Cytoscape. The size and color of
   nodes were defined continuously according to the degree (number of
   edges), and the color of stroke was set according to the combined score
   calculated by network analyzer of Cytoscape. The plug-in CytoHubba
   ([169]https://apps.cytoscape.org/apps/cytohubba (accessed on 14
   September 2022)) was used to find the core subnetwork. The node’s
   scores were calculated and top 10 hub genes ranked by MCC were
   selected, then the network graph was constructed.

4.4. Molecular Docking

   The crystal structures of candidate targets were obtained from the PDB
   database ([170]https://www.rcsb.org/ (accessed on 20 September 2022)).
   Open-Source PyMol was used to pre-process the structure, including
   removing the ligand and water molecules, adding hydrogen [[171]65].
   AutoDock 4.2 was used to conduct docking between costunolide and key
   targets [[172]66].

4.5. Establishment of Two Experimental AIH Models and Administration of
Costunolide

   C57BL/6 mice were supplied by the Laboratory Animal Centre of Tongji
   Hospital. All animals were housed in the SPF environment and a 12 h
   light/12 h dark cycle at room temperature of 22 ± 2 °C with 55 ± 2%
   humidity. All animals had free access to food and water. The animal
   study protocol was approved by the Laboratory Animal Welfare and Ethics
   Committee of Tongji Hospital of Tongji Medical College, Huazhong
   University of Science and Technology (IACUC Issue No.: TJH-202207036;
   date of approval: 15 July 2022). The specific animal model
   establishment procedure and treatment scheme was depicted in
   [173]supplementary materials (Supplementary Figure S2).

   Concanavalin A (ConA) was used to establish the acute immune-mediated
   liver injury model. ConA (C2010, CAS: 11028-71-0, LOT: 091610120V) was
   purchased from Sigma-Aldrich. Costunolide (CAS: #553-21-9, LOT:
   J22GB152180) was purchased from Shanghai YuanYe Bio-Technology Co.,
   Ltd. Twenty-four male mice (6–8 weeks old, 20–25 g) were randomly
   divided into solvent group (Solvent, n = 6), costunolide group (COS, n
   = 6), ConA group (ConA+Solvent, n = 6), and ConA+costunolide group
   (ConA+COS, n = 6). A single dose of ConA (15 mg/kg) dissolved in normal
   saline was injected via tail vein of mice, and the same volume of
   normal saline injection was used as control. Based on the existing
   literature, costunolide (10 mg/kg) was injected intraperitoneally once
   daily for three days, the first dose was given two days before giving
   ConA [[174]11,[175]26]. Costunolide was dissolved in DMSO, then mixed
   with 40% PEG300 (CAS: 25322-68-3, MCE), 5% Tween 80 (CAS: 9005-65-6,
   MCE), and normal saline. Mice were sacrificed 24 h after ConA
   injection. After anesthesia, blood samples were collected by removing
   the eyeball from the socket using tissue forceps [[176]67]. Mice were
   then sacrificed and livers were harvested. Part of the liver was fixed
   in 4% paraformaldehyde for 48 h, the rest was collected in EP tubes,
   cooled in liquid nitrogen, and preserved in a −80 °C refrigerator.

   As described in our previous publication, a single dose of empty
   adenovirus and multiple high-pressure tail vein injections of human
   CYP2D6 plasmid were administered to establish the chronic AIH model
   [[177]22]. Empty adenovirus was purchased from Shanghai DesignGene
   Company. Human CYP2D6 plasmid was obtained from our lab. Twenty male
   mice (6–8 weeks old, 20–25 g) were randomly divided into four groups:
   solvent group (Solvent, n = 5), costunolide group (COS, n = 5), AIH
   group (AIH+Solvent, n = 5), and AIH+costunolide group (AIH+COS, n = 5).
   In the COS group and AIH+COS group, costunolide (10 mg/kg) was injected
   intraperitoneally once daily from day 14 after the adenovirus injection
   continuously for 20 days. Mice were sacrificed at day 34 after
   adenovirus injection. The sacrifice and specimen taking procedure was
   similar to the ConA model described above. In addition, spleens were
   harvested to measure the spleen weight, and spleen index was calculated
   using spleen weight (mg) divided by mice weight (g).

4.6. Histopathology, Immunohistochemistry, and Biochemical Analysis

   The fixed liver tissue was embedded in paraffine and sliced at a
   thickness of 4 µm. H&E staining was performed to evaluate liver injury
   and inflammation. Afterward, all sections were graded of whole sections
   blindly under optical microscope (Olympus IX71) and representative
   pictures were imaged. Necrosis scores and inflammation scores were
   calculated as below: 0, no area of necrosis/immune cells infiltration;
   1, very mild, few interspersed necrosis/ infiltration; 2, mild,
   necrosis/infiltration area ≤30%; 3, moderate, 30% <
   necrosis/infiltration area ≤60%; 4, severe, necrosis/infiltration area
   >60%. Plus, Sirius-red staining was performed to evaluate the degree of
   liver fibrosis in chronic AIH model. Immunohistochemistry (IHC) was
   performed to detect the expression of CD4, F4/80, and phospho-AKT
   expression in liver tissue according to protocol using SP9000 IHC kit
   (ZSGB-Bio, Beijing, China). Primary antibodies: anti-CD4 (1:2000;
   ab183685, Abcam (Cambridge, UK)), anti-F4/80 (1:200; #70076S, CST
   (Danvers, MA, USA)), and anti-phospho-AKT (Ser473) (1:250; AF0016,
   Affinity (Cincinnati, OH, USA)) were used.

   Blood samples collected were centrifuged at 3000 rpm for 15 min to
   obtain the serum. The liver enzyme ALT and AST in serum were measured
   using the Rayto automatic biochemistry analyzer Chemray 420 by Wuhan
   ServiceBio Company (Wuhan, China).

4.7. Western Blot

   Liver tissue specimens were preserved in a −80 °C refrigerator before
   use. After being cut to an appropriate size, liver tissue was lysed in
   RIPA lysis buffer (harsh) containing protease inhibitor cocktail and
   phosphatase inhibitor cocktail (Wuhan Servicebio Technology Co., Ltd.
   (Wuhan, China)). Homogenizer was used to grind tissue into homogenate
   and sonication was performed using ultrasonic processor. After 40 min
   lysis, all samples were centrifuged at 12,000 rpm for 10 min. Protein
   concentration was measured using BCA kit (Wuhan Servicebio Technology
   Co., Ltd. (Wuhan, China)). Tissue protein (40 µg) was separated in an
   SDS-PAGE and was transferred to PVDF membranes. The blotted PVDF
   membranes were blocked with TBST containing 5% BSA at room temperature
   for 1 h. Then the primary antibodies were incubated overnight at 4 °C
   including: anti-phospho-AKT (Ser473) (1:1000; #4060T, CST),
   anti-AKT(1:1000; #9272S, CST), anti-phospho-SRC (Tyr416) (1:1000;
   #6943S, CST), anti-SRC (1:1000; #A19119, Abclonal (Wuhan, China)),
   anti-phospho-IGF1Rβ (Tyr1135/1136) (1:1000; #3024S, CST), anti-IGF1Rβ
   (1:800; #9750S, CST), anti-GAPDH (1:2000; 30202ES60, Yeason (Shanghai,
   China)). On the next day, secondary antibodies were incubated for 1 h
   at room temperature. After being washed 3 times in TBST for 10 min,
   membranes were exposed to hypersensitive electrochemiluminescence (ECL)
   kit (NCM Biotech (Suzhou, China)). Protein bands were visualized on the
   Tanon 5200 Multi System (Tanon Science and Technology Co., Ltd.
   (Shanghai, China)).

4.8. Statistical Analysis

   GraphPad Prism software (version 9.0.0) was utilized to analyze data.
   The continuous data was presented as mean ± standard deviation (SD),
   and the Student’s t-test was performed to analyze the comparison
   between two groups. Welch’s correction was used if variances were not
   equal based on the F test. The discontinuous data were analyzed using
   the Mann–Whitney test. p < 0.05 was considered statistically
   significant.

5. Conclusions

   In our present study, network pharmacology and experimental validation
   were integrated to investigate the effect and mechanism of costunolide
   in AIH. According to the results, costunolide could relieve the
   inflammation and fibrosis in two murine AIH models, and the therapeutic
   effect might work by suppressing the activation of PI3K-AKT pathway and
   inhibiting the phosphorylation of SRC and IGF1R. Our research reveals
   the effect and possible mechanism of costunolide in AIH and may further
   contribute to novel drug development for AIH and other autoimmune
   diseases.

Acknowledgments