Network Pharmacology and Bioinformatics Approach Reveals the Multi-Target Pharmacological Mechanism of Fumaria indica in the Treatment of Liver Cancer

Liver cancer (LC), a frequently occurring cancer, has become the fourth leading cause of cancer mortality. The small number of reported data and diverse spectra of pathophysiological mechanisms of liver cancer make it a challenging task and a serious economic burden in health care management. Fumaria indica is a herbaceous annual plant used in various regions of Asia to treat a variety of ailments, including liver cancer. Several in vitro investigations have revealed the effectiveness of F. indica in the treatment of liver cancer; however, the exact molecular mechanism is still unrevealed. In this study, the network pharmacology technique was utilized to characterize the mechanism of F. indica on liver cancer. Furthermore, we analyzed the active ingredient-target-pathway network and uncovered that Fumaridine, Lastourvilline, N-feruloyl tyramine, and Cryptopine conclusively contributed to the development of liver cancer by affecting the MTOR, MAPK3, PIK3R1, and EGFR gene. Afterward, molecular docking was used to verify the effective activity of the active ingredients against the prospective targets. The results of molecular docking predicted that several key targets of liver cancer (along with MTOR, EGFR, MAPK3, and PIK3R1) bind stably with the corresponding active ingredient of F. indica. We concluded through network pharmacology methods that multiple biological processes and signaling pathways involved in F. indica exerted a preventing effect in the treatment of liver cancer. The molecular docking results also provide us with sound direction for further experiments. In the framework of this study, network pharmacology integrated with docking analysis revealed that F. indica exerted a promising preventive effect on liver cancer by acting on liver cancer-associated signaling pathways. This enables us to understand the biological mechanism of the anti liver cancer activity of F. indica.


Introduction
Liver cancer is the fourth major cause of cancer-related deaths globally. According to a report from Cancer Research UK, liver cancer will be one of the fastest-growing malignancies by 2035 [1]. It is a fatal condition that affects people with hepatitis A or C, fatty liver disease, and diabetes, and it has been associated with excessive alcohol consumption, smoking, and dietary toxins [2]. Some hereditary variables may also play a role in the development of this disease. Abnormalities in the cell cycle, metabolic networks, in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets. found 1500 genes that could be the target of the 15 active ingredients (Table 1). After finding the most promising targets for drugs, a total of 38,477 genes linked to liver cancer were found in the GeneCards and DisGeNET databases. Later, a Venn diagram was used to figure out the common targets of both liver cancer and compound linked genes. A total of 557 genes from F. indica that could fight against liver cancer were considered as key targets.

Compounds-Target Network Construction
A total of 15 active compounds from F. indica were found to be satisfactory. To construct the 'active compound-targeted genes-connected pathway' network diagram, 15 active compounds, 1500 key targets, and their associated pathways with high gene count were chosen. Each of these active compounds have multiple targets, which shows that many targets induce a synergistic effect when F. indica is used as an anti-hepatic cancer agent. The degree of these 12 compounds in the compound-targeted genes-connected pathways network was then assessed ( Table 2). As indicated in Table 2, alkaloids together with tyramines have comparably maximum degree; however, the degree of both alkylamides and steroids are comparatively low as compared to alkaloids and tyramines. Furthermore, from these 12 compounds, 4 compounds were selected for docking analysis: two alkaloids with a higher degree of connectivity, especially Fumaridine and Cryptopine, one tyramine compound, namely N-Feruloyltyramine, and one alkylamides compound, namely Lastourvilline.

PPI Network Construction
The 557 genes that overlapped were uploaded to the STRING database to build a PPI network. A PPI network shows how different targets work together during the development of a disease. The nodes and their connections show how these targets work together ( Figure 1A). Later, Cytoscape was utilized at the PPI network of genes  Figure 1D). This means that the highest degree genes are greatly linked to each other; thus, all of these genes might be hub targets.   Comparing these findings with those provided by enrichment analysis (Table 3) four genes in particular, EGFR, MAPK3, MTOR, and PIK3R1, were identified as the main anti-liver cancer targets of F. indica and were chosen for molecular docking experiments.

GO and KEGG Pathway Analysis
The functional annotation and enrichment analysis unveiled the potential biological roles of F. indica targets. The targets of F. indica, according to GO functional analysis, were related to protein phosphorylation, inflammatory response, integral part of plasma membrane, and so on. (Figure 2). The KEGG pathway analysis was performed to identify the significant signaling pathways linked to the anti-liver cancer effect of F. indica. It is noteworthy that most of the genes were involved in following pathways. These include the NF-κB, the stress-responsive mitogen-activated protein kinase (MAPK), and the STAT pathways [19], including the PI3K-AKT-mTOR, AMPK-mTOR, EGF, MAPK, Wnt/β-catenin, p53, and NF-κB pathways [20]. KEGG pathway analysis revealed that EGFR, MAPK3, MTOR, and PIK3R1 were significantly enriched genes ( Figure 3).

Molecular Docking
The top four targets EGFR, MTOR, MAPK3, and PIK3R1 were chosen for molecular docking after a thorough analysis of the PPI network. PDB structure was used to find the 3D structure of the target protein (EGFR (PDB id: 1IVO), MTOR (PDB id: 1FAP), MAPK3 (PDB id: 2ZOQ), and PIK3R1 (PDB id: 1HPO)). All the PDB structures were selected on the basis of their resolution. Low numeric values in Å mean the resolution of the structure is good and can be considered for further analysis. In the framework of the current study, EGFR had a resolution of 3.30 Å, while other targets MTOR, MAPK3, and PIK3R1 had resolution of 2.70 Å, 2.39 Å, and 2.50 Å, respectively. Furthermore, the CPort tool was used for active site prediction of the selected proteins. The compounds interacted with the active site of the EGFR receptor via forming a bond with the following amino acid residues: Tyr B251, Gln A8, Leu A38, Ala A62, Asn A86, Thr B249, Pro B248, and Lys A407.
In the case of MTOR, the active compounds bound with the Ser B2035, Glu A54, Tyr A26, Phe A46, Asp A37, Arg A42, Thr B2098, Asp B2102, Lys B2095, Trp B2101, Phe B2039, Tyr B2105, Phe B2108, and Leu B2031 residues. On the other hand, MAPK3 bound with active compounds via Arg A41, Arg A64, Thr B347, Glu B194, Pro B193, Asn B161, Phe A371, Arg A370, Pro A373, and Asp A105. Lastly, the selected compounds bound with PIK3R1 by forming a bond with Arg A8,Ala A28,Gly B27,Asp A25,Ile A84,Asp B25,Gly B48,Ile B47,Asp B29,Ile A50,Gly A4, and Ile B50 (Figure 4). The drug candidates showed hydrogen bond, Pi-pi-stacked, and van der Waals interactions with the receptor proteins, indicated with dotted lines mentioned in the additional file (Supplementary Material: Table S1). All those binding pockets were selected by the site finder tool present in the molecular operating environment. The top four inhibitors, Fumaridine, N-feruloyl tyramine, Cryptopine, and Lastourvilline, were screened out of the 15 showing a good docking score along with RMSD values for all targets as shown in Table 4.     The red nodes represent the hub genes, the orange nodes represent active compounds, and the blue nodes are the pathways associated with the core targets.

Molecular Docking
The top four targets EGFR, MTOR, MAPK3, and PIK3R1 were chosen for molecular docking after a thorough analysis of the PPI network. PDB structure was used to find the 3D structure of the target protein (EGFR (PDB id: 1IVO), MTOR (PDB id: 1FAP), MAPK3 (PDB id: 2ZOQ), and PIK3R1 (PDB id: 1HPO)). All the PDB structures were selected on the basis of their resolution. Low numeric values in Å mean the resolution of the structure is good and can be considered for further analysis. In the framework of the current study, EGFR had a resolution of 3.30 Å , while other targets MTOR, MAPK3, and PIK3R1 had resolution of 2.70 Å , 2.39 Å , and 2.50 Å , respectively. Furthermore, the CPort tool was used for active site prediction of the selected proteins. The compounds interacted with the active site of the EGFR receptor via forming a bond with the following amino acid residues: Tyr B251, Gln A8, Leu A38, Ala A62, Asn A86, Thr B249, Pro B248, and Lys A407. In the case of MTOR, the active compounds bound with the Ser B2035, Glu A54, Tyr A26, Phe A46, Asp A37, Arg A42, Thr B2098, Asp B2102, Lys B2095, Trp B2101, Phe B2039, Tyr B2105, Phe B2108, and Leu B2031 residues. On the other hand, MAPK3 bound with active compounds via Arg A41, Arg A64, Thr B347, Glu B194, Pro B193, Asn B161, Phe A371, Arg A370, Pro A373, and Asp A105. Lastly, the selected compounds bound with PIK3R1 by forming a bond with Arg A8,Ala A28,Gly B27,Asp A25,Ile A84,Asp B25,Gly B48,Ile B47,Asp B29,Ile A50,Gly A4, and Ile B50 (Figure 4). The drug candidates showed hydrogen bond, Pi-pi-stacked, and van der Waals interactions with the receptor proteins, indicated with dotted lines mentioned in the additional file (Supplementary Material: Table  S1). All those binding pockets were selected by the site finder tool present in the molecular operating environment. The top four inhibitors, Fumaridine, N-feruloyl tyramine, Cryptopine, and Lastourvilline, were screened out of the 15 showing a good docking score along with RMSD values for all targets as shown in Table 4.   Fumaridine and N-feruloyl tyramine showed good binding affinity with all the targets, except EGFR, having docking scores between −13.86 kcal/mol and −10.69 kcal/mol. However, EGFR showed top binding with the compounds Lastourvilline and Cryptopine, having docking scores of −12.69 kcal/mol and −10.29 kcal/mol, respectively. Similarly, all the compounds also showed strong hydrogen bond interactions with interacting residues of MTOR. Thus, docking analysis strengthened our findings that predicted stable target bonds with active compounds of F. indica. Figure 4 represents the sketch map of target proteins together with their strongest binding components.

ADMET Profiling
ADMET analysis is a challenging process in drug discovery. The SwissADME tool was applied to forecast different types of pharmacokinetic properties. The pharmacokinetic factor may be used to predict the absorption, distribution, metabolism and elimination (ADME), and toxicity of the top therapeutic novel compounds. ADMET profiling of all the top selected drug candidates showed that there are no negative consequences of the pharmacokinetic properties, first and foremost, of the potential compounds ( Table 5). The ADMET characteristics of possible drugs for various models such as P-glycoprotein substrates, BBB penetration, and CYP2C19 inhibitors, CYP2C9 inhibitors, CYP2D6 inhibitors, and CYP3A4 inhibitors produced promising results that substantially confirm the compound ability to function as a drug candidate. Furthermore, the skin permeation lop Kp values describe that, depending on its size and chemophysical qualities, a chemical can permeate the stratum corneum via intercellular, transcellular, or appendageal channels. It is noteworthy that all the compounds showed non-toxic behavior, although different types of toxicity were measured for all compounds, and none of the compounds showed toxic behavior.

Discussion
Natural product research has received a lot of interest in recent years [21]. The network pharmacology method aids in the understanding of the complicated interactions that exist between medicines and their targets, as well as the probable mechanisms of action [22,23]. Moreover, the diversity of developing new drugs from plant sources provides methodological challenges [24]. Because of the lack of ADME qualities in the newly discovered medication, and because of the high-budget nature of research, drug discovery methodologies face additional challenges [25]. As a result, in the creation of medications, pharmaceutical specialists place a high value on ADME-based screening [26]. Liver fibrosis, viral hepatitis, fatty liver, cirrhosis, and liver cancer are all serious disorders that endanger human health and are the top causes of mortality globally [27]. Despite significant advances in the treatment of liver disorders over the last several centuries, the majority of medications still fail to provide satisfying results in patients [28]. Hepatocellular carcinoma is significantly linked to chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, aflatoxin-contaminated food consumption, and high alcohol usage [29]. Multiple drug resistance (MDR), a high clearance rate, severe side effects, undesirable drug distribution to the specific site of liver cancer, and a low concentration of medication that reaches liver cancer cells are just a few of the drawbacks of traditional liver cancer chemotherapy. As a result, new techniques and network pharmacology must be developed to convey the medication molecules specific to the malignant hepatocytes in enough of an amount and for a sufficient period of time within the therapeutic window [30,31]. As a consequence, the search for novel drugs is becoming targeted. A high-potency origin of phytochemicals with medical advantage would have a potential liver cancer therapy option in this scenario.
Fumaria indica is a medicinal plant of the fumitory family that is rich with phytochemical constituents, which have huge medicinal value [32,33]. F. indica has antipyretic, antidiarrheal, and hypoglycemic effects, according to pharmacological investigations [34,35]. Various in vitro studies revealed the therapeutic significance of F. indica against liver diseases. However, the exact molecular mechanism remains unclarified [8]. This study provides a foundation for the initial screening of F. indica bioactive compounds, as well as a novel therapeutic concept for future research into F. indica processes for liver cancer therapy. The hallmark of this age will be the identification of potential bioactive ingredients that stop the pathophysiology of disorders and disease.
In the current study we uncover several target genes that are revealed to be involved in various pathways in cancer. The pathogenesis of disease can be avoided by targeting the genes that cause disruption in pathways in cancer. A slew of studies strengthened our findings, such as most people that are suffering from liver cancer include chronic infections with HBV or HCV, or cirrhosis, and certain people inherit liver diseases and diabetes [36]. It is important to note that two of our major targets, MTOR and MAPK3, are primarily implicated in liver cancer resistance pathways [37]. Our research proposed that MTOR, MAPK3, and EGFR are directly involved in hepatitis B pathways. As a result, changes in these genes may disrupt the pathways that interconnect them, leading to disorder. Beyond this, the targeted genes of active constituents are also enriched in various inflammatory conditions such as arthritis and so forth, which seems to indicate that they can act on various antiinflammatory cytokines and exert an effect on liver cancer. It is worth noting that our key gene, namely EGFR, has been shown to play a key role during liver regeneration following acute and chronic liver damage, as well as in cirrhosis and hepatocellular carcinoma, highlighting the importance of EGFR in the development of liver diseases [38]. Hence blocking the EGFR genes might help in the treatment of liver cancer. Furthermore, the mammalian target of the rapamycin (mTOR) signaling system is involved in many aspects of cancer such as cell growth, the inhibition of apoptosis, and metabolic reprogramming proliferation [39]. This demonstrates conclusively that the dysregulation of mTOR is emphasized in the pathogenesis of liver cancer.
The biological information of target genes was obtained using GO enrichment analysis. According to GO functional analysis, anti-liver cancer targets of F. indica were mainly involved in protein phosphorylation, peptidyl-tyrosine phosphorylation phosphatidylinositol 3-kinase complex, class IA, and GO protein serine/threonine/tyrosine kinase activity. KEGG pathway studies revealed that targets were involved in liver cancer-related pathways. The KEGG pathway enrichment results revealed that the putative targets were significantly enriched in hepatitis B and viral carcinogenesis, and the cAMP signaling pathway, PI3K-Akt signaling route, MAPK signaling pathway, estrogen signaling pathway, p53 signaling pathway, and cell cycle signaling pathway were all found to be enriched in cancer pathways.
It is noteworthy that our core genes are mainly enriched in the cyclic adenosine monophosphate (cAMP) signaling pathway. Previous studies demonstrated that the cAMP signaling pathway controls a variety of cellular activities such as lipid, metabolism, inflammation, cell differentiation, and injury and regulates gene-protein expression and function [40]. Hence, disturbance in the cAMP signaling pathway might be associated with liver cancer. Furthermore, through the KEGG pathways it has been revealed that MAPK3 genes are directly involved in the mitogen-activated protein kinase (MAPK) signaling pathway. MAPK inhibitors are effective at reducing pro-inflammatory cytokinesis and increasing anticancer activity, especially in human pancreatic cancer cells [31]. This gives clear evidence that dysregulation of the MAPK3 gene causes disturbance in the MAPK signaling pathway, which ultimately leads to liver cancer. Therefore, by targeting MAPK3, the pathophysiology of liver cancer can be halted. Beyond this, the targeted genes of active constituents are also enriched in various other cancer-related signaling pathways, which seems to indicate that these compounds and their associated target genes exert a strong effect on liver cancer.
According to topological parameters of the compound-genes-pathway network, four major targets named EGFR, MAPK3, MTOR, and PIK3R1 were identified as the core targets. Furthermore, these core targets were validated using molecular docking, which revealed that Fumaridine, N-feruloyl tyramine, Lastourvilline, and Cryptopine bound stably with these core targets. The findings of docking analysis indicated that these four compounds can be used for treatment of liver cancer because of their ability to bind stably with core targets. In the light of current network pharmacology, this research predicted the active compounds, their prospective targets, and associated pathways for the treatment of liver cancer, thereby providing a theoretical foundation for future experimental research. Given the limitation of network pharmacology, the basic pharmacological mechanism for liver cancer treatment can only be discovered by data mining. The mining of active compounds is particularly based on different databases. Although the information in databases are curated, a lot of inconsistencies may, however, occur due to variety of information resources and experimental data. In this regard, modern high throughput techniques including chromatography can be used; liquid chromatography and mass spectrometry can help to solve this problem. Even though we have given some intriguing evidence, more research and clinical trials are required to fully investigate the potential of F. indica and to validate its medicinal applications.

Virtual Screening of Active Constituents
The information on active phytocompounds of Fumaria indica was collected from the literature using different databases such as PubMed and Google scholar and KNApSAcK. PubChem Explore Chemistry [41] was used to obtain the Canonical Simplified Molecular-Input Line-Entry System (SMILES) of each active ingredient, while PubMed [42] and ChemSpider [43] were used to obtain the chemical structures of active compounds. All constituents of F. indica were virtually screened by applying bioavailability (OB) and drug likeness (DL) parameters, which are crucial in the characteristics of absorption, distribution, metabolism, and excretion (ADME) characteristics of drugs. Compounds were only retained if DL ≥ 0.18 and OB ≥ 30% to satisfy ADME criteria. Biologically active compounds that did not match these conditions were discarded and were not investigated further. In this regard, DL and OB of all active constituents were calculated using Molsoft [44] and SwissADME [45].

Target Genes Screening
The potential target genes of screened active constituents were retrieved by entering their Canonical SMILES to SwissTarget Prediction tool [46]. Therefore, the target with probability ≥ 0.7 were selected. Prediction of disease-related genes is the next step to uncover the molecular mechanism of medicinal herb to treat multiple diseases. Two databases, GeneCards and DisGeNET, were searched using keywords 'primary liver cancer' and 'Hepatic cancer' to retrieve disease-related genes. DisGeNET is a multipurpose data system that provides information related to genes, disorders, and their related empirical studies [47]. GeneCards database contains information related to the genome, proteome, and transcriptomes of an organism [48]. The Venn online tool was used to identify the overlap genes between predicted target genes of screened compounds and disease-related targets [49]. Therefore, the common targets of active constituents and disease were obtained for subsequent analysis.

Pathway and Functional Enrichment Analysis
To perform gene enrichment analysis and KEGG pathway analysis, Database for annotation, visualization, and integrated discovery (DAVID) [50] was used. List of key genes were subjected to DAVID to perform functional annotation at three levels: cellular component (CC), molecular function (MF), and biological process (BP). DAVID is a webbased functional enrichment database that helps researchers to comprehend the bioactivity of a huge number of genes. In current study, p-value ≤ 0.01 was selected, and top 10 GO enrichments and top 10 KEGG pathways were chosen for subsequent analysis.

Network Construction
The mechanism of F. indica in liver cancer was performed by network analysis. The software Cytoscape 3.8.0, which is a freely available, graphical user interface for importing, visually exploring, and analyzing bimolecular interaction networks, was used to construct and visualize the network [51]. Active constituents and the target genes in the network were represented by nodes, while the edges were used to represent the interaction between active constituents and their target genes. Network analyzer tool was used to calculate degree, a topological property that reveals the importance of compound-target gene-pathways in network diagram. Moreover, target genes with the highest degree of connectivity were considered as 'key target'.

PPI Network Construction and Molecular Docking Analysis
Protein-protein interaction (PPI) data were obtained from the Search Tool for the retrieval of Interacting Genes (STRING) database with a confidence score of >0.7 to construct PPI network by uploading common genes on a database [52]. The PPI network obtained from STRING was subjected to the cytoHubba plugin of Cytoscape, which was used to analyze the core regulatory genes of the PPI web and the identification of key targets. The observed co-expression of predicted key targets was also obtained through STRING database. Moreover, key targets were validated through molecular docking approach. The RCSB PDB [53] was used to obtain the X-ray crystal structure of candidate target; it was used to obtain crystal structures of potential targets. Moreover, refinement of structure was performed using Chimera. After that, they were brought into molecular operating environment (MOE) [54], which was used to extract ligands from proteins, adjust their structure, and remove water from them. The best docked score with the lowest RMSD and binding energy were selected for further analysis. Furthermore, Chimera X and discovery studio was used for visualization of interaction among active compounds and predicted target. The workflow of the present study is displayed in Figure 5.

Conclusions
This research establishes a scientific foundation for determining the efficacy of multicomponent, multi-target drug treatment as well as finding novel anti liver cancer thera-

Conclusions
This research establishes a scientific foundation for determining the efficacy of multicomponent, multi-target drug treatment as well as finding novel anti liver cancer therapeutic targets. In this study, network pharmacology along with molecular docking was employed to explore the underlying mechanism for the treatment of liver cancer. According to network analysis, F. indica contains multi-targeting compounds that function on numerous disease-related pathways; hence, they might be considered as novel therapeutic options against liver cancer. Furthermore, our studies revealed that the EGFR, MAPK3, MTOR, and PIK3R1 genes are effective and potential therapeutic agents for lowering the incidence of liver cancer and potentially exhibiting therapeutic effects on liver cancer. However, the current study has significant limitations since further phytochemical and pharmacological research is needed to verify these findings.