Integrating Network Pharmacology, Molecular Docking, and Experimental Validation to Investigate the Mechanism of (-)-Guaiol Against Lung Adenocarcinoma.

BACKGROUND Lung adenocarcinoma (LUAD) is the most common type of lung cancer, which poses a serious threat to human life and health. -(-)Guaiol, an effective ingredient of many medicinal herbs, has been shown to have a high potential for tumor interference and suppression. However, knowledge of pharmacological mechanisms is still lacking adequate identification or interpretation. MATERIAL AND METHODS The genes of LUAD patients collected from TCGA were analyzed using limma and WGCNA. In addition, targets of (-)-Guaiol treating LUAD were selected through a prediction network. Venn analysis was then used to visualize the overlapping genes, which were further condensed using the PPI network. GO and KEGG analyses were performed sequentially, and the essential targets were evaluated and validated using molecular docking. In addition, cell-based verification, including the CCK-8 assay, cell death assessment, apoptosis analysis, and western blot, was performed to determine the mechanism of action of (-)-Guaiol. RESULTS The genes included 959 differentially-expressed genes, 6075 highly-correlated genes, and 480 drug-target genes. Through multivariate analysis, 23 hub genes were identified and functional enrichment analyses revealed that the PI3K/Akt signaling pathway was the most significant. Experiment results showed that -(-)Guaiol can inhibit LUAD cell growth and induce apoptosis. Additional evidence suggested that the PI3K/Akt signaling pathway established an inseparable role in the antitumor processes of -(-)Guaiol, which is consistent with network pharmacology results. CONCLUSIONS Our results show that the effect of (-)-Guaiol in LUAD treatment involves the PI3K/Akt signaling pathway, providing a useful reference and medicinal value in the treatment of LUAD.

Background

According to the latest survey of American Cancer Society [1], lung cancer is still the leading cause of cancer-related death for both men and women (22% for men and 22% for women), with a high rate of metastases. This malignant disease, which has high morbidity and mortality rates, is classified into 2 subtypes: small-cell lung carcinoma (SCLC) and non-small-cell lung carcinoma (NSCLC). Lung adenocarcinoma (LUAD), a subtype of NSCLC, is the most common subtype of lung cancer, accounting for 40% of all cases [2]. LUAD treatment differs depending on the stage of the disease. Surgical resection or radiotherapy is considered at the early stage [3]. Patients with advanced-stage disease are treated with chemotherapeutic medications such as platinum agents, taxanes, and therapies that target oncogenetic mutation drivers [4]. Unfortunately, the clinical benefit of conventional therapy is limited by toxicity and acquired resistance, causing recurrence and distant metastasis. Patients are eventually susceptible to disease progression, with a poor prognosis. From 2010 to 2016, the 5-year survival rate for metastatic LUAD was only 6% [1]. It is essential to develop new drugs for combination therapy, which is considered a more effective and advanced treatment for overcoming resistance rather than monotherapies.

-(−)Guaiol, which had distinct antibacterial activity, is a natural compound found in medicinal plants. The formula of -(−)Guaiol is C15H26O, and the molecular weight is 222.37g/mol [5]. This sesquiterpene alcohol with the guaiane skeleton has been used for thousands of years as a traditional Chinese medicine and is the main active ingredient found in Alisma orientale [6], Ferula ferulaeoides [7], Aloysia gratissima [8], and other traditional medicinal plants. -(−)Guaiol is a naturally occurring insecticidal product because of its mosquito larvicidal activity, antimicrobial activity, and anti-leishmania activity [9-11]. However, recent research indicates that -(−)Guaiol has a high potential for antitumor activity [12-14], although the molecular mechanisms are still unclear. In our previous study, the pharmacological investigation of antitumor mechanisms was focused on the treatment of lung cancer. -(−)Guaiol impairs mTOR signaling to induce autophagy in lung cancer cells [15,16], which should encourage more LUAD-related research. PI3K/Akt is one of the important signaling pathways in intracellular processes; it is known to inhibit apoptosis and to mediate cell survival. The literature has indicated that inhibiting the PI3K/Akt signaling pathway can reverse resistance caused by cisplatin in lung cancer [17], and it can activate downstream signaling molecules when activated inappropriately, affecting the occurrence and progression of lung cancer [18]. It is intriguing that (−)-Guaiol inhibits the proliferation of LUAD cells by specifically targeting mTOR, which is one of the downstream signaling molecules of the PI3K/Akt signaling pathway. Therefore, we conducted a series of investigations to further explore this topic.

In this study, a network was built using network pharmacology, weighted gene co-expression network analysis (WGCNA), and a system bioinformatics approach, which was then validated using molecular docking and experiments. The main goal of the current study was to determine the anticancer effect and molecular mechanism of -(−)Guaiol against LUAD, which complements those of earlier studies. A general study flow chart is shown in Figure 1.

Figure 1

In the flow chart, the systems pharmacology approach was used to determine the mechanism of (−)-Guaiol in the treatment of LUAD by integrating target identification, network construction, enrichment analysis, molecular docking and experimental validations.

Material and Methods

Identification of Differentially-Expressed Genes

The level 3 RNA sequencing (RNA-seq) profiles were downloaded from the The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/) by searching the following keywords “bronchus and lung”, “adenomas and adenocarcinomas”, and “TCGA-LUAD”, consisting of 375 LUAD tissues and 32 adjacent normal tissues transcriptome profiles. After processing the available data with the limma [19] (v: 3.42.2) R package, differentially-expressed genes (DEGs) between tumor and normal tissues were specifically identified in the TCGA cohort, with a false-discovery rate of 0.05.

WGCNA Module Construction

Additionally, WGCNA, a systems biology modality for distinguishing the correlation patterns among genes across microarray samples [20], was used to screen the highly-correlated genes from RNA-seq profiles for summarizing such clusters and correlating modules to external sample traits. To create a scale-free network, the soft-thresholding parameter was defined as=6 (scale-free R2=0.85) using the soft function pickSoftThreshold. Integrating 2 systems biology approaches increased the discrimination ability of candidate genes selection.

Prediction of the Targets of (−)-Guaiol

The study used PubChem [21] (https://pubchem.ncbi.nlm.nih.gov/), the world’s largest chemical information collection website, to search for the SMILES structural formula of (−)-Guaiol. The obtained structure was then fed into ChemMapper [22] (http://lilab.ecust.edu.cn/chemmapper/) to predict corresponding targets. The names of the corresponding targets were normalized in the Uniprot [23] database (https://www.uniprot.org/). Finally, using Cytoscape (v: 3.7.2) software, all attribute data were combined and the “disease-drug-target” network was visualized.

Co-Expression Network Construction and Screening of Hub Genes

After thoroughly preparing for data collection, a co-expression network was created, which connected 3 major gene cohorts (DEGs, WGCNA module genes, and (−)-Guaiol target genes). The network was displayed by a Venn diagram created with the R package VennDiagram [24] (v: 1.6.20). Following that, the co-expression genes were entered into Cytoscape (v: 3.7.2) software for protein–protein interactions (PPI) network construction. They were processed using “Biogenet” [25] (v: 3.0.0), an application for visualizing the relationships between multiple biomolecules, and CytoNCA [26] (v: 2.1.6) for topological analysis. In the network, only eligible genes can be selected. The selection criteria were set as: Interaction Degree (DC) >50 and Betweenness (BC) >100. After double standard screening, the genes that met the selection criteria were ultimately enrolled as key biological targets in the subsequent research.

Enrichment Analysis

System functional enrichment analyses were performed to elucidate the biological functions and signaling pathways associated with (−)-Guaiol against LUAD to gain more insights into the regulation of selected genes. Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) are knowledge-base resources. GO provides information on biological processes (BPs), cellular components (CCs), and molecular functions (MFs), while KEGG connects genomic information with higher-order functional intelligence [27,28]. GO and KEGG pathway analyses with a P value of 0.05 were performed using 3 R packages: clusterProfiler [29] (v: 3.14.3), org.Hs.eg.db (v: 3.10.0), and enrichplot (v: 1.6.1). The R package ggplot2 (v: 3.2.1). were used to create the bubble diagrams for the GO enrichment and KEGG pathway analyses.

Molecular Docking

The current study constructed a KEGG pathway-target relation network based on the enriched pathways with related target proteins, with only the top-ranked pathways enrolled. In addition, Cytoscape (v: 3.7.2) was used for topological analysis with the “Network Analyzer” tools. The analysis revealed that targets with a high degree score were of interest and may participate in the molecular docking analysis. Molecular docking, an in silico structure-based approach that allows the identification of active ingredients of therapeutic interest [30], was used to estimate the molecular interactions between targets and (−)-Guaiol. In addition, the chemical formulas and 3D structures of the samples, CDK2 (PDB ID: 1B39: 2.1Å), FN1 (PDB ID: 2HAZ: 1.7Å), HSP90AA1 (PDB ID: 5NJX: 2.49Å), and HSP90AB1 (PDB ID: 1UYM: 2.45Å), were obtained from the PDB (https://www.rcsb.org/) and PubChem [31] (https://pubchem.ncbi.nlm.nih.gov/) databases. To improve the quality of molecular docking, AutoDockTools (1.5.6) was used to add charge and display rotatable keys to the compound structures, and Pymol was used to modify the protein crystal structures corresponding to the core target genes by removing water molecules and hetero-molecules. Moreover, the AutoDockTools (1.5.6) were applied to add hydrogen atoms and charge operations. It is necessary to transform the protein format from “pdb” to “pdbqt” format and search for active pockets via AutoDock software. After completing the preparation of the ligand (the compound) and receptors (the proteins corresponding to the core target genes), docking operations with AutoDock Vina could finally be initiated. The binding affinity assisted in measuring the conformational stability as a reference.

Chemicals and Reagents

(−)-Guaiol was commercially obtained from Sigma (448575, St. Louis, MO, USA). DMEM medium was purchased from KeyGEN BioTECH (KGM12800-500, Jiangsu, China). Fetal bovine serum (FBS) was provided by Gibco Life Technologies (10099-141, Grand Island, NY, USA). The reagent Z-VAD-FMK (Z-VAD) was obtained from Selleck (S7023, Shanghai, China). Annexin V-FITC/PI Apoptosis Detection Kits were purchased from KeyGEN BioTECH (KGA105-KGA108, Jiangsu, China). Cell Counting Kit-8 was purchased from YEASEN (40203ES76, Shanghai, China). LY294002 was purchased from MedChemExpress (HY-10108, Monmouth Junction, New Jersey, USA). SYTOX Green was purchased from Invitrogen (S7020, Carlsbad, California, USA). The following antibodies – Bcl-2(15071, 1: 1000), BAX (5023, 1: 1000), β-actin (4970, 1: 1000), p-Akt (4060, 1: 1000), Akt (2920, 1: 1000), PI3K (4255, 1: 1000), p-PI3K (4249, 1: 1000), and IgG secondary antibody (10700, 1: 2000) – were purchased from Cell Signaling Technology (Boston, Massachusetts, USA).

Cell Culture and Drug Administration

The LUAD cells A549 and Calu-1 and normal lung cells BEAS-2B were obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin streptomycin in a humidified atmosphere with 5% CO2 at 37°C. (−)-Guaiol was diluted in methanol at a primary stock concentration of 10 mM. The control group was DMEM with 10% FBS and methanol.

Cell Survival Assay

Cell survival was detected with a CCK-8 assay kit. Approximately 5000 cells were seeded into a 96-well culture plate, and after 24 h, the cells were exposed to various doses of (−)-Guaiol (60 μM or150 μM) or Z-VAD (20 μM). The solution was then removed, and all wells were washed twice with PBS before adding 10 μl CCK-8 solution to the wells for incubation. After waiting 1h, a Multiplate Reader (Bio Tek, United States) was used to measure the OD value at 450 nm. The cell survival rate (%) was calculated as ([experimental group-blank group]/[control group-blank group]) 100% using the average OD value of 3 wells.

SYTOX Green staining assay was performed according to the manufacturer’s protocol, visualizing cell death through imaging applications. The working concentration was 100 nM and experiments were carried out 3 times each.

Apoptosis Analysis

The cells were inoculated into 6-well plates and each well contained 2 mL of media with 10% FBS and approximately 2×105 cells. They were cultured for 24 h at 37°C and 5% CO2, then treated by (−)-Guaiol (60 μM or 120 μM) for 24 h. The cells were collected after drug treatment. We added binding buffer (500 μL), Annexin V-FITC (5 μL), and propidium iodide (5 μL) to each tube, followed by incubation at room temperature in the dark for 10 min. Finally, a fluorescence microscope was used to observe the cell morphology and flow cytometry was used to detect the apoptosis rate. The experiments were carried out 3 times each.

Western Blot Analysis

After treatment, the cells were collected and washed twice with PBS, then added to cell lysis buffer with protease and phosphatase inhibitors for 30 min at 4°C. The whole protein was quantified with a bicinchoninic acid kit (Beyotime, China). Samples were obtained after boiling within the 1% SDS loading buffer at 100°C for 8 min. Afterward, 20 μg protein was subjected to SDS-PAGE and subsequently transferred to a PVDF membrane. After immersing the membrane in BCA for blocking, the samples were incubated overnight with the following primary antibodies – β-actin, p-Akt (Ser473), Akt, PI3K, p-PI3K (110α), BAX, and Bcl-2 – diluted 1: 1000. Next, we used an anti-rabbit IgG secondary antibody (1: 2000) to conjugate each primary antibody. Finally, the samples mixed with ECL (Tanon, Shanghai, China) solution were exposed and recorded by chemiluminescence reagent. The gray values of the bands, which were calculated automatically using Image J software, were normalized to β-actin as the internal control. The relative protein levels were calculated by normalizing the densitometry results of the experimental group to those of the control group. One-way ANOVA was used to statistically analyze the significance. The experiments were carried out 3 times each.

Data Analysis

All results were obtained from experiments independently repeated 3 times. The continuous variables of normal distribution or nearly normal distribution were displayed as average value±standard deviation. After homogeneity of variance test, one-way ANOVA was used for those meeting data requirements in SPSS 25.0. or GraphPad Prism 8.0. P<0.05 was considered to indicate a statistically significant difference.

Results

Construction of DEGs and WGCNA Module in TCGA Cohort

There were 15 208 RNA-seq genes of LUAD downloaded from the TCGA cohort, with 959 genes pertaining to DEGs and 6075 genes pertaining to WGCNA key modules. The 959 DEGs were described by the volcano plot and heat map, using the R package ggplot2 [32] (v: 3.2.1) for visualization (Figure 2A, 2B). The 6075 highly-related genes were classified into 15 color modules and are presented in Figure 2C, 2D.

Figure 2

(A) Volcano plot of DEGs in TCGA. (B) Heat map of DEGs in TCGA. The green parts present down-regulated genes, red parts present up-regulated genes, and black parts presents non-differential genes. (C) The cluster dendrogram with different colors of TCGA. (D) Module-trait relationships. Each row corresponds to a color, while each column corresponds to the normal or tumor group.

Target Prediction and Analysis of (−)-Guaiol

A prediction network integrating targets of (−)-Guaiol against LUAD was obtained during this process. Figure 3 depicts the “disease-drug-target” network. As a result, 480 targets were identified and are represented by the green ovals in the plot; thus far, all of the target genes have been collected under the current conditions. The file containing total gene (DEGs, WGCNA module genes and drug target genes) information was included as Supplementary Table 1.

Figure 3

Disease-drug-target network. Purple triangular node presents the drug, orange rhombic node presents disease, and green oval nodes present 480 targets.

Identification of Hub Genes

The genes were then optimized and reduced to a few hub genes in the following step of the workflow. The selected co-expression genes among DEGs, WGCNA module genes, and (−)-Guaiol target genes were presented in a Venn diagram (Figure 4A), and only 17 genes were filtered out in the plot. The 17 co-expression genes were then analyzed by PPI network-to-topology analysis. The PPI network was established for topology feature analysis with the Cytoscape software platform based on degree and betweenness parameters for further investigation. There were 23 genes that met the degree >50 and betweenness >100 thresholds (Figure 4B), which meant these genes played an essential role in (−)-Guaiol treatment of LUAD. The specific gene information on 17 identified genes and 23 hub genes is shown in Supplementary Tables 2 and 3.

Figure 4

(A) Venn diagram was used to summarize the co-expression genes of DEGs, WGCNA module genes, and (−)-Guaiol target genes. (B) The 23 hub genes were predicted by topological screening of the protein–protein interaction (PPI) network with screening 2 times. DC – degree; BC – betweenness.

GO and KEGG Enrichment Analysis

The 23 genes that emerged were used in subsequent functional enrichment analyses. The GO and KEGG enrichment analyses identified a number of related biological activities that were statistically significant (Figure 5). In total, 883 significant GO terms and 27 significant pathways were enriched (P<0.05) (Supplementary Table 4). The BPs aspect changes were clearly associated with 732 functional categories, including DNA metabolic process regulation, developmental cell growth regulation, protein binding regulation, and positive cell cycle regulation. Furthermore, there were 81 terms in the CCs analysis aspect, primarily vesicle lumen, secretory granule lumen, apical plasma membrane, and membrane microdomain. There were also 70 terms in MFs, such as ubiquitin protein ligase binding, cell adhesion molecule binding, MHC class II protein complex binding, and ATPase activity. The KEGG pathway enrichments revealed that the identified targets were mainly enriched in 27 pathways, including the hosphatidylinositol3-kinase/ProteinKinase-B (PI3K/Akt) signaling pathway, protein processing in the endoplasmic reticulum, cell cycle, and antigen processing and presentation. Of note, the PI3K-Akt signaling pathway was significant, which is listed at the top of the chart.

Figure 5

(A) Representative results (P<0.05) of GO enrichment analysis. (B) Representative results (p<0.05) of KEGG enrichment analyses. Bubble size indicates gene count and the color indicates P value.

KEGG Pathway-Target Network and Molecular Docking

The pathway-target network for (−)-Guaiol and LUAD, which shows 19 targets and 20 pathways, was built to reveal potential mechanisms of action. Based on the calculated degree, each node pertaining to the pathway or target was displayed in an ordered arrangement (Figure 6). According to the network, the top 5 candidate targets (CDK2, EGFR, FN1, HSP90AA1, HSP90AB1) were investigated further for their interaction with (−)-Guaiol; notably, these genes were also involved in the PI3K/Akt signaling pathway in KEGG analysis. Each target was docked with the (−)-Guaiol 3D structure after a series of modifications. As shown in Figure 7, CDK2, FN1, HSP90AA1 and HSP90AB1 displayed a favorable binding affinity, with ideal conjunction in general. The docking score in binding affinity represents the cohesiveness of the models. It is clear that the lower binding affinity score was preferred.

Figure 6

KEGG pathway-target network. The orange round nodes represent the top 19 related targets and the green round nodes represent the top 20 pathways.

Figure 7

Molecular docking diagrams of (−)-Guaiol with (A) CDK2, (B) EGFR, (C) HSP90AA1, (D) HSP90AB1, and (E) FN1. (F) The binding affinity for the 5 targets docked into the (−)-Guaiol crystal structure.

(−)-Guaiol Can Suppress LUAD Cells and Induce Apoptosis

The CCK-8 assay kit was used to assess the cytotoxicity of (−)-Guaiol in A549 cells and Calu-1 cells, taking normal lung cells BEAS-2B as a control. The survival rates of A549 and Calu-1 were concentration-dependently decreased when treated with different concentrations of (−)-Guaiol, as suggested by the CCK-8 assay (Figure 8A). According to the results of GraphPad Prism, (−)-Guaiol dramatically inhibited growth of A549 cells and Calu-1 cells, with IC50 values of 75.7 μM and 92.1 μM, respectively, whereas its IC50 value for BEAS-2B cells was 322.3 μM. To investigate the apoptosis action of (−)-Guaiol, the appropriate concentration (75 μM) of (−)-Guaiol and the apoptosis inhibitor Z-VAD were used in the cell death assays. When compared to the (−)-Guaiol single group, Z-VAD significantly reversed cell viability in A549 (p=0.048) and Calu-1 cells (P=0.033) (Figure 8B). Similarly, the SYTOX Green staining assay showed that (−)-Guaiol damaged LUAD cells. Under fluorescence microscopy, the exceptional brightness of the signal produced by the (−)-Guaiol group indicated more dead cells, whereas Z-VAD weakened the fluorescence intensity, counteracting the effect of (−)-Guaiol, which was consistent with the CCK-8 assay results (Figure 8C).

Figure 8

(A) Cell survival rate was detected by performing a CCK-8 assay. The LUAD cells (A549 and Calu-1) and normal lung cells BEAS-2B cells treated with different concentrations of (−)-Guaiol. IC50 values were 75.7 μM for A549 cells, 92.1 μM for H1299 cells, and 322.3 μM for BEAS-2B cells. Data from 3 independent experiments were represented as means±SD. (B) A549 and Calu-1 were treated with (−)-Guaiol (75 μM) or (−)-Guaiol (75 μM) companied Z-VAD (20μM) for 24 h. ** P<0.01 vs control in A549. # P<0.05 vs (−)-Guaiol in Calu-1. Data from 3 independent experiments were represented as means±SD. (C) A549 and Calu-1 were treated with or without (−)-Guaiol (75 μM) or (−)-Guaiol (75 μM) and Z-VAD (20 μM), and stained with SYTOX Green (100 nM), then observed under a fluorescence microscope.

(−)-Guaiol Increased Early Apoptosis and Altered Expression of Related Proteins

The Annexin V-FITC/PI dual-staining assay was used for quantitative analysis to gain more insight into the apoptosis induced by (−)-Guaiol. As shown in Figure 9A, the early apoptosis of the control group, 60 μM (−)-Guaiol group, and 120 μM (−)-Guaiol group for A549 was 17.75%, 25.9%, and 35.7%, respectively, and for Calu-1 it was 1.07%, 10.9%, and 13.9%, respectively. (−)-Guaiol induced apoptosis by increasing early apoptosis, as demonstrated by the detected green fluorescence indicating PS in apoptotic cells with low red fluorescence in the image, and the proportion of cells with a higher level of green and lower level of red fluorescence was higher in the flow cytometry analysis (Figure 9B). According to western blot analysis, (−)-Guaiol inhibited the expression of the anti-apoptotic protein Bcl-2 (p<0.001) while increasing the expression of the pro-apoptotic protein BAX (P=0.028) (Figure 9C).

Figure 9

(A) Flow cytometry to detect the apoptosis rate in A549 and Calu-1 treated with different concentrations of (−)-Guaiol. (B) Fluorescence microscope was used to observe apoptosis in A549 and Calu-1 treated with different concentrations of (−)-Guaiol. (C) Western blot analysis was used to detect the protein expression of BAX and Bcl-2 in A549 treated with (−)-Guaiol (60 μM or 120 μM). The relative BAX and Bcl-2 levels from 3 independent experiments were statistically analyzed using ANOV A test in the quantitative analysis. * P<0.05, *** P<0.001 vs control.

(−)-Guaiol Regulates the Expression of Related Proteins in PI3K/Akt Signaling Pathway

The PI3K/Akt signaling pathway, which includes classical proteins PI3K, phosphorylated PI3K, Akt, and phosphorylated Akt, was found to be highly likely to be involved in the mechanisms of (−)-Guaiol’s anti-LUAD activity. To validate the conclusion, western blotting was used to examine the protein expression of PI3K/Akt signaling in A549 cells treated with (−)-Guaiol (75 μM) and LY294002 (15 μM), a PI3K/Akt signaling pathway inhibitor. The expression of total PI3K and Akt was little changed (P>0.05). Treatment with (−)-Guaiol decreased p-PI3K (P=0.001) and p-Akt (P<0.001) levels and furthered the LY294002-induced decrease in p-PI3K (P<0.001) and p-Akt levels (P=0.004), as expected (Figure 10). These findings suggest that (−)-Guaiol inhibits the PI3K/Akt signaling pathway, which regulates apoptosis in A549 cells, thereby validating the network pharmacology result.

Figure 10

(−)-Guaiol regulates the expression of related proteins in PI3K/Akt Signaling Pathway. Western blot analysis was used to detect the protein expression of PI3K, p-PI3K, Akt, and p-Akt in A549 treated with or without (−)-Guaiol (75 μM) or LY294002 (15 μM). The relative protein levels from 3 independent experiments were statistically analyzed using ANOVA in the quantitative analysis. * P<0.05, ** P<0.01, *** P<0.001 vs control. ## P<0.01, ### P<0.001 vs LY294002.

Discussion

Lung cancer is the most prevalent human malignancy worldwide; historically, LUAD became the most common subtype of lung cancer, outranking squamous cell carcinoma (SCC) by 1988 to 2002 [33]. As a result, numerous studies on LUAD treatment have been conducted. Medical research combined with network pharmacology is an appealing approach that is becoming increasingly popular in scientific research fields [34]. In any case, a better understanding of the underlying mechanism is essential for guiding effective and precise therapy.

(−)- Guaiol has shown potent antineoplastic activity, but the molecular mechanism has yet to be fully understood. To clarify the underlying mechanism, a network was constructed based on the significant genes, and through a series of network analyses, pivot targets and pathways were discovered. First, the gene information of LUAD patients was collected using the publicly available database TCGA, which was then processed using computational methods to generate 959 DEGs between tumor and non-tumor tissues, as well as 6075 genes highly correlated with WGCNA. At the same time, network pharmacology identified another 480 (−)-Guaiol target genes. These were extracted and shrunk to 17 co-expression genes, and they eventually yielded 23 hub genes through topological data analysis of PPI. Based on the findings, a GO biological process and KEGG pathway enrichment analyses were performed to investigate the potential mechanisms. However, there are some limitations to our study. As we know, bioinformatics has opened up a new field of pharmacological research that develops computing technology and networks public databases for analysis of biological data. Researchers use network databases for data collection, but some related genes may still have been missed, which is unavoidable. Only strongly correlated and significant genes can be used for subsequent analysis. Bioinformatics analysis provides significant options, although it may not be comprehensive. Based on the result of the enrichment analysis, we mainly focused on the PI3K/Akt signaling pathway, which was the most significant pathway in enrichment analysis.

The study was then completed with validation, which included molecular docking and experiments. The molecular docking study confirmed that the top 5 candidate targets from the KEGG pathway-target network (CDK2, EGFR, FN1, HSP90AA1, HSP90AB1), which also were found to be involved in the PI3K/Akt signaling pathway, had a high affinity for (−)-Guaiol, enhancing the results of network pharmacology. Specifically, Cyclin-dependent kinase 2 (CDK2) plays a vital role in cycle regulation and DNA damage response. In addition, CDK2 induces apoptosis by targeting FOXO1, a pivotal protein in triggering DNA-damage-induced apoptosis following dsDNA breaks [35], but CDK2 acts has opposite effects in cell apoptosis. In KRAS-mutant lung cancer models, inhibiting CDK2 kinase activity causes anaphase catastrophe and apoptosis [36]. The epidermal growth factor receptor (EGFR) is another cell growth regulator that belongs to the receptor tyrosine kinase (RTK) family, stimulating specific downstream signaling pathways such as PI3K/Akt, RAS/RAF/MAPK, and JAK/ATAT [37]. EGFR, which has been identified as an oncogenic driver of NSCLC, is associated with advanced disease stage and poor prognosis [38]. Coincidentally, fibronectin 1(FN1), heat shock protein 90AA1 (HSP90AA1), and heat shock protein 90AB1 (HSP90AB1) are overexpressed in cancer, predicting a poor prognosis [39-41].

Furthermore, a series of experiments were carried out to investigate the PI3K/Akt pathway and effect of (−)-Guaiol in LUAD. The PI3K/Akt pathway, which had the lowest P value among the 27 candidate pathways in functional analysis, has a close relationship with malignant tumors because it regulates cell proliferation, differentiation, and metabolism [42]. One of the consequences of this process is apoptotic inhibition, which can lead to cell dysfunction and mutation. As a result, a number of PI3K and Akt inhibitors for lung cancer are being studied [43]. In LUAD cells, (−)-Guaiol also inhibits the PI3K/Akt pathway and regulates apoptosis.

RTK stimulation and somatic mutations in specific elements pathway elements are the 2 most common mechanisms of PI3K/Akt activation in human cancers [44]. However, somatic mutations and amplifications of the PIK3CA gene, which encodes the class I PI3K p110α, frequently occur in patients with NSCLC. Specifically, somatic mutations in PI3KCA are found in 3% to 10% of squamous cell carcinomas and 0% to 2.7% of adenocarcinomas, while PI3KCA amplification is found in 35% of squamous cell carcinomas and 7% of adenocarcinomas [45,46]. Moreover, laboratory research investigated various lung cancer call lines and tumor tissues, including 86 NSCLC cell lines, 43 SCLC cell lines, 3 extrapulmonary small-cell cancer (ExPuSC) cell lines, and 691 resected NSCLC tumors, showing that 12.8% of cell lines and 19.1% of tumors were associated with PIK3CA mutations or amplification [47]. Furthermore, high phosphorylated Akt expression was found in a significant proportion of NSCLC patients, which influenced PIK3CA mutations or amplification [43,47].

The aberrantly activated PI3K/Akt signaling pathway triggers a variety of molecular mechanisms, resulting in a series of changes in cell metabolism and cell growth. Active Akt has been shown to promote cell cycle progression and inhibit apoptosis via a variety of biological mechanisms, including phosphorylation of intermediate targets (GSK3, FOXO, BAD, and MDM2), inhibition of NFB, and activation of mTOR [43]. The Bcl-2 family, a related downstream pathway component, is well known as an apoptosis regulator. The Bcl-2 family plays 2 roles in apoptotic regulation: anti-apoptotic executor and pro-apoptotic executor [48]. Bcl-2 and Bcl-xl are anti-apoptotic proteins that sequester caspases, a unique set of cysteine proteases that cleave critical cellular proteins to cause apoptotic changes or prevent mitochondrial apoptogenic factors from release into the cytoplasm. Other members of the Bcl-2 family, such as BAX and Bak, are responsible for pro-apoptotic functions such as facilitating caspase release and activation depending on mitochondrial outer-membrane permeabilization (MOMP) [49]. According to the findings, (−)-Guaiol inhibited Bcl-2 expression while increasing BAX expression, resulting in lung cancer cell apoptosis.

(−)-Guaiol was found to kill LUAD cells by inhibiting the PI3K/Akt signaling pathway and inducing cell apoptosis. However, (−)-Guaiol has little influence on normal lung cells. The apoptosis inhibitor Z-VAD was used to investigate the apoptosis effect of (−)-Guaiol in cell death assays. The CCK8 and SYTOX Green staining assay illustrated that the Z-VAD rescued the LUAD cells in (−)-Guaiol treatment, suggesting (−)-Guaiol causes apoptosis in LUAD. The apoptotic cells were mainly early apoptotic cells in the apoptosis analysis. Interestingly, previous research found that (−)-Guaiol activated double-strand breaks (DSBs)-induced cell apoptosis via caspase signaling and targeting RAD51, a key factor in homologous recombination repair [15]. Another study looked into the mechanism of autophagy and discovered that (−)-Guaiol targeted the mTORC1 and mTORC2 signaling pathways to cause autophagic cell death. Notably, the mTORC2-AKT signaling was blocked due to mTOR phosphorylation failure, and the mTORC1 signaling was impaired due to downstream factor inhibition [16]. In light of this, PI3K/Akt is regarded as a common upstream signaling pivot, initiating multiple downstream activities leading to cell death.

Conclusions

The purpose of this study was to assess the effects and mechanism of (−)-Guaiol against LUAD using network pharmacology and bioinformatics. As part of the validation process, molecular docking and experiments were carried out. The results showed that (−)-Guaiol targeted the PI3K/Akt pathway to exert its effects via the apoptotic pathway. These findings support the results of previous research on (−)-Guaiol, which indicate that (−)-Guaiol is a potential candidate for clinical application and promotion. Additional research on (−)-Guaiol, clinical trials, and more elaborate experiments are needed.

Supplementary Table 1

The identified genes of DEGs, WGCNA module genes, and drug target genes.

DEGs	ADAMTS1	AKAP12
A2M	ADAMTS8	AKR1B10
AADAC	ADAMTSL4	AKR1C2
AATK	ADGRE5	ALDH18A1
ABCA3	ADGRF5	ALDH1A1
ABCA8	ADGRL2	ALDH2
ABCC3	ADH1B	ALDH3B1
ABCG1	ADPRH	ALDOA
ABHD11	ADRB1	ALG1L
ABI3BP	ADRB2	ALOX15
AC007906.2	AFAP1L1	ALOX15B
ACADL	AGER	ALOX5
ACE	AGR2	ALOX5AP
ACKR1	AGR3	ALPL
ACP5	AGRP	AMOTL1
ACVRL1	AGTR2	ANGPT1
ADA2	AHNAK	ANKRD1
ADAM8	AK1	ANKRD22
ANKRD29	BZW2	CCL23
ANLN	C11orf96	CCL24
ANOS1	C14orf132	CCM2L
ANXA3	C15orf48	CCN1
AOC1	C16orf89	CCN2
AOC3	C1orf115	CCN5
APLN	C1orf116	CCNA2
APOBR	C1orf162	CCNB1
APOC1	C1orf194	CCNB2
APOL3	C1orf198	CCND2
APOLD1	C1QA	CCT3
AQP1	C1QB	CCT5
AQP4	C1QC	CD101
AQP9	C2	CD163
ARHGAP18	C20orf85	CD24
ARHGAP31	C4BPA	CD300C
ARHGAP44	C5AR1	CD300LF
ARHGEF15	C5orf38	CD300LG
ARHGEF26	C7	CD34
ARHGEF6	C8B	CD36
ARRB1	C9orf24	CD44
ASF1B	CA2	CD52
ATF3	CA4	CD53
ATIC	CA9	CD74
ATOH8	CACNA2D2	CD79A
ATP13A4	CADM1	CD83
AURKA	CALCOCO1	CD93
AURKB	CALCRL	CDC20
B3GNT3	CAMK2N1	CDCA5
B3GNT6	CASKIN2	CDCA7
BASP1	CAT	CDCA8
BCHE	CAV1	CDH3
BCL2A1	CAV2	CDH5
BCL6B	CAVIN1	CDK1
BEX4	CAVIN2	CDKN2B
BIRC5	CBLC	CDO1
BMP2	CBX7	CEACAM1
BMPR2	CCBE1	CEACAM5
BOP1	CCDC167	CELF2
BPIFA1	CCDC69	CENPF
BTG2	CCL18	CEP55
BTNL9	CCL21	CES1
CFD	CTHRC1	ECRG4
CGNL1	CTNNAL1	ECSCR
CHI3L2	CTSH	ECT2
CHRDL1	CTSS	EDN1
CITED2	CX3CL1	EDNRB
CKS1B	CXCL12	EEF1A2
CLDN18	CXCL13	EFCC1
CLDN3	CXCL14	EFEMP1
CLDN5	CXCL16	EFNA4
CLEC14A	CXCL2	EFNB2
CLEC3B	CYB5A	EGFL7
CLIC3	CYBB	EGLN3
CLIC5	CYBRD1	EGR1
CNN1	CYP24A1	EGR2
COL10A1	CYP27A1	EHD2
COL11A1	CYP4B1	EIF4EBP1
COL17A1	CYYR1	EMCN
COL1A1	DAPK1	EMP1
COL1A2	DCN	EMP2
COL3A1	DENND3	EMP3
COL5A1	DEPP1	ENG
COL5A2	DERL3	ENO1
COL6A6	DES	EPAS1
COLEC12	DHCR24	EPCAM
COMP	DKK3	EPHX3
COX4I2	DLC1	ERG
COX7A1	DLGAP5	ERO1A
CP	DMBT1	ESAM
CPA3	DNAJC12	ETS1
CPAMD8	DNASE1L3	ETV4
CPB2	DNTTIP1	EVI2B
CRABP2	DOK2	FABP4
CRIM1	DPEP2	FABP5
CRLF1	DPT	FAM107A
CRTAC1	DPYSL2	FAM162B
CRYAB	DRAM1	FAM167A
CSF3	DSG2	FAM183A
CSF3R	DSP	FAM189A2
CSRNP1	DUOX1	FAM216B
CST1	DUOXA1	FAM83A
CSTA	DUSP1	FBLN1
CT83	DUSP4	FBLN5
FBP1	GATA2	HBEGF
FBXO32	GATA6	HCAR2
FCER1G	GBP4	HCK
FCGR3A	GCNT3	HEG1
FCN1	GDF10	HES6
FCN3	GGCT	HHIP
FEN1	GGTLC1	HIGD1B
FERMT1	GIMAP1	HJURP
FERMT2	GIMAP4	HK3
FGB	GIMAP6	HLA-DOA
FGD5	GIMAP7	HLA-DPA1
FGFBP2	GIMAP8	HLA-DPB1
FGFR2	GJA1	HLA-DQA1
FGFR4	GJA4	HLA-DQB1
FGL1	GJA5	HLA-DRA
FGR	GJB2	HLA-DRB1
FHL1	GKN2	HLA-DRB5
FHL2	GLIPR2	HLA-E
FHL5	GMFG	HLF
FIBIN	GNAQ	HMGA1
FKBP10	GNG11	HMGB3
FKBP11	GOLM1	HPGD
FLI1	GPA33	HS6ST2
FLRT3	GPC3	HSD17B6
FMO2	GPD1	HSPA12B
FOLR1	GPIHBP1	HSPB6
FOS	GPM6A	HSPB8
FOSB	GPM6B	HSPD1
FOXF1	GPRC5A	HSPE1
FOXM1	GPRIN2	HYAL1
FPR1	GPT2	HYAL2
FPR2	GPX2	ID1
FUT2	GPX3	ID3
FUT3	GRAMD2A	ID4
FXYD6	GRASP	IFT57
FZD4	GRK5	IGFBP3
GABARAPL1	GYPC	IGLL5
GADD45B	H1-0	IL1RL1
GALNT18	H1-2	IL33
GALNT7	H2BC5	IL3RA
GAPDH	HBA2	IL6
GAS6	HBB	IL7R
INAVA	KLF6	LTBP4
INMT	KLF9	LY86
IQANK1	KPNA2	LYVE1
IQGAP3	KRT4	LYZ
IRAK1	KRT6A	MAL
IRX2	KRT8	MAMDC2
ITGA8	KRT80	MANF
ITGB4	KRTCAP3	MAOA
ITIH5	LAD1	MAOB
ITLN1	LAMA3	MARCKSL1
ITLN2	LAMP3	MARCO
ITM2A	LAPTM4B	MATN3
ITPKA	LAPTM5	MCEMP1
ITPRID2	LARGE2	MCM2
ITPRIP	LBH	MCM4
JAM2	LCN2	MDK
JAML	LCP1	MELK
JCAD	LDB2	METTL7A
JPT1	LDHA	METTL7B
JUN	LDLR	MEX3A
JUNB	LEPR	MFAP4
JUND	LGALSL	MFNG
KANK2	LGI3	MFSD2A
KANK3	LGR4	MGAT3
KAT2A	LHFPL6	MGP
KCNJ15	LIFR	MID1IP1
KCNJ8	LILRA5	MKI67
KCNK3	LIMCH1	MME
KCNN4	LIMS2	MMP1
KCTD12	LIPA	MMP11
KDELR3	LMCD1	MMP12
KIAA0040	LMNB1	MMP13
KIAA1324L	LMO2	MMP19
KIF11	LMO7	MMP28
KIF1C	LMOD1	MMP9
KIF20A	LPL	MMRN1
KIF2C	LRRC32	MMRN2
KIF4A	LRRC36	MNDA
KIFC1	LRRK2	MRC1
KLF13	LRRN4	MROH6
KLF2	LSR	MS4A15
KLF4	LST1	MS4A4A
MS4A7	NOTCH4	PGC
MS4A8	NPM3	PGGHG
MSMO1	NPNT	PHACTR1
MSR1	NPR1	PHLDA2
MSRB3	NQO1	PI16
MT1A	NR4A1	PID1
MT1M	NR4A3	PIGR
MTARC2	NRGN	PILRA
MTHFD2	NRN1	PIP5K1B
MT-ND6	NTN4	PKIG
MTURN	NUSAP1	PKP3
MUC13	OCIAD2	PLA2G1B
MUC20	OGN	PLA2G4A
MUC21	OLFML1	PLA2G4F
MUC5B	OLR1	PLAC8
MYADM	OSCAR	PLAC9
MYBL2	OTUD1	PLAU
MYCT1	OTULINL	PLBD1
MYH10	P3H2	PLEK2
MYH11	P3H4	PLEKHO2
MYL9	P4HB	PLK1
MYRF	P53	PLLP
MYZAP	PAEP	PLOD2
MZB1	PAFAH1B3	PLPP2
NAPSA	PAICS	PLPP3
NCF2	PAPSS2	PLSCR4
NCKAP5	PCDH12	PMP22
NDNF	PCOLCE2	PNPLA6
NDRG2	PCP4	PODXL2
NDST1	PCSK9	POSTN
NEBL	PCYOX1	PPARG
NECAB1	PDGFB	PPBP
NECTIN4	PDIA4	PPP1R14A
NEDD9	PDK4	PPP1R14B
NEK2	PDLIM1	PPP1R14D
NES	PDLIM2	PPP1R15A
NET1	PDPN	PRAM1
NFIX	PDZD2	PRC1
NHSL1	PEBP4	PRDX4
NKG7	PECAM1	PRELP
NME1	PER1	PRF1
NME4	PFKP	PRG4
PROM2	ROBO4	SERPING1
PROS1	RPL39L	SERTM1
PRX	RRAD	SESN1
PSAT1	RRAS	SEZ6L2
PSMG3	RRM2	SFN
PTGDS	RTKN2	SFTPA1
PTGER4	S100A2	SFTPA2
PTGES	S100A3	SFTPB
PTGIS	S100A4	SFTPC
PTPN21	S100A8	SFTPD
PTPRB	S100P	SGCA
PTPRM	S1PR1	SGCE
PTTG1	S1PR4	SGK1
PXDC1	SAMHD1	SGPP2
PYCR1	SAPCD2	SH2D3C
QKI	SASH1	SHMT2
QPCT	SBK1	SHROOM4
RAB11FIP1	SCARA5	SIRPA
RAC3	SCD	SIRPB1
RAI2	SCD5	SLC11A1
RAMP1	SCEL	SLC15A2
RAMP2	SCGB1A1	SLC15A3
RAMP3	SCGB3A1	SLC19A3
RARRES2	SCGB3A2	SLC1A1
RASGRF1	SCN4B	SLC25A10
RASIP1	SCN7A	SLC25A25
RASL12	SCNN1B	SLC25A39
RBMS2	SCUBE1	SLC27A3
RBP4	SDC2	SLC2A1
RCAN1	SEC14L6	SLC2A3
RCC1	SECISBP2L	SLC34A2
RCN3	SELENBP1	SLC35F2
RECQL4	SELENOP	SLC39A8
RETN	SELP	SLC46A2
RGCC	SELPLG	SLC50A1
RGS5	SEMA3B	SLC52A2
RHOJ	SEMA3G	SLC6A4
RHOV	SEMA4B	SLC7A5
RIPOR1	SEMA5A	SLC7A7
RMI2	SERINC1	SLCO2A1
RNASE1	SERINC2	SLCO2B1
RNF144B	SERPINA1	SLIT2
SLIT3	STXBP6	TMPRSS4
SLPI	SULF1	TNFRSF21
SMAD6	SUSD2	TNFSF12
SMIM22	SVEP1	TNFSF13
SMPDL3B	SYNPO	TNNC1
SNX25	SYT7	TNNT1
SOCS2	TACC1	TNS1
SOCS3	TBC1D2	TNS2
SOD3	TBX2	TNS3
SORBS3	TBX3	TNXB
SOSTDC1	TBX4	TOM1L2
SOX17	TCF21	TOP2A
SOX4	TCN1	TPPP
SOX7	TEK	TPPP3
SPAG4	TENT5B	TPSAB1
SPAG5	TESC	TPSB2
SPARCL1	TFF1	TPX2
SPDEF	TFPI2	TRAF4
SPI1	TFRC	TREM1
SPINK1	TGFBR2	TRIP13
SPN	TGFBR3	TRPV2
SPOCK2	THBD	TSC22D3
SPP1	THBS2	TSKU
SPTBN1	THSD1	TSPAN12
SPTBN2	THY1	TSPAN18
SRD5A3	TIE1	TSPAN7
SRGN	TIMP1	TSTA3
SRPK1	TIMP3	TTYH3
SRPX	TK1	TUBA1A
SRPX2	TLCD1	TUBB6
SSR4	TLR4	TXNDC17
ST14	TMEM100	TXNIP
ST6GALNAC1	TMEM125	TYMS
ST6GALNAC5	TMEM132A	TYROBP
ST6GALNAC6	TMEM139	UBE2C
STAC	TMEM184A	UBE2T
STARD13	TMEM190	UBL3
STARD8	TMEM204	UCHL1
STEAP1	TMEM45B	UGDH
STK39	TMEM47	UNC5CL
STOM	TMEM88	UPK3B
STX11	TMPRSS11E	UTRN
VAMP2	VSIG4	WNT3A
VCAN	VSIR	WWC2
VEGFD	VSTM2L	XPR1
VEPH1	VWF	ZBED2
VIM	WASF3	ZDHHC9
VIPR1	WFDC2	ZFP36
VMP1	WFS1	ZNF385B
VSIG2	WIF1	ZWINT

Supplementary Table 2

The 17 co-expression genes.

Co-expression genes	Co-expression genes
BCHE	ADRB1
CA2	ADRB2
CA4	ALDH1A1
FABP4	CES1
FABP5	COX7A1
MMP13	HBA2
PLA2G1B	KCNJ8
SCD	SGK1
SLC6A4

Supplementary Table 4.

The result of GO and KEGG enrichment analysis.

ID	Description			Gene ratio	Bg ratio
hsa04151	PI3K-Akt signaling pathway			20.07.2021	354/8102
hsa04141	Protein processing in endoplasmic reticulum			20.05.2021	171/8102
hsa05215	Prostate cancer			20.04.2021	97/8102
hsa04110	Cell cycle			20.04.2021	124/8102
hsa04612	Antigen processing and presentation			20.03.2021	78/8102
hsa04914	Progesterone-mediated oocyte maturation			20.03.2021	100/8102
hsa04114	Oocyte meiosis			20.03.2021	129/8102
hsa04915	Estrogen signaling pathway			20.03.2021	138/8102
hsa05418	Fluid shear stress and atherosclerosis			20.03.2021	139/8102
hsa04120	Ubiquitin mediated proteolysis			20.03.2021	140/8102
hsa05022	Pathways of neurodegeneration - multiple diseases			20.05.2021	475/8102
hsa03040	Spliceosome			20.03.2021	151/8102
hsa05160	Hepatitis C			20.03.2021	157/8102
hsa05165	Human papillomavirus infection			20.04.2021	331/8102
hsa05134	Legionellosis			20.02.2021	57/8102
hsa04510	Focal adhesion			20.03.2021	201/8102
hsa05140	Leishmaniasis			20.02.2021	77/8102
hsa04810	Regulation of actin cytoskeleton			20.03.2021	218/8102
hsa04512	ECM-receptor interaction			20.02.2021	88/8102
hsa05222	Small cell lung cancer			20.02.2021	92/8102
hsa05131	Shigellosis			20.03.2021	246/8102
hsa04657	IL-17 signaling pathway			20.02.2021	94/8102
hsa04933	AGE-RAGE signaling pathway in diabetic complications			20.02.2021	100/8102
hsa04659	Th17 cell differentiation			20.02.2021	107/8102
hsa04670	Leukocyte transendothelial migration			20.02.2021	114/8102
hsa04068	FoxO signaling pathway			20.02.2021	131/8102
hsa05135	Yersinia infection			20.02.2021	137/8102
ID	p value	p. adjust	q value	Gene ID	Count
hsa04151	1.36E-05	0.001519047	0.001214	FN1/EGFR/HSP90AA1/YWHAZ/ITGA4/CDK2/HSP90AB1	7
hsa04141	4.74E-05	0.002652306	0.002119	VCP/HSP90AA1/HSPA5/HSP90AB1/CUL1	5
hsa05215	8.08E-05	0.003014678	0.002408	EGFR/HSP90AA1/CDK2/HSP90AB1	4
hsa04110	0.000209	0.005864916	0.004685	MCM2/YWHAZ/CDK2/CUL1	4
hsa04612	0.00087	0.019482004	0.015564	HSP90AA1/HSPA5/HSP90AB1	3
hsa04914	0.001786	0.033332938	0.026629	HSP90AA1/CDK2/HSP90AB1	3
hsa04114	0.003687	0.051336929	0.041012	YWHAZ/CDK2/CUL1	3
hsa04915	0.004458	0.051336929	0.041012	EGFR/HSP90AA1/HSP90AB1	3
hsa05418	0.004549	0.051336929	0.041012	HSP90AA1/VCAM1/HSP90AB1	3
hsa04120	0.004641	0.051336929	0.041012	UBC/CUL1/CUL7	3
hsa05022	0.005042	0.051336929	0.041012	VCP/UBC/FUS/APP/HSPA5	5
hsa03040	0.005733	0.053504025	0.042743	FUS/HNRNPU/U2AF2	3
hsa05160	0.006388	0.055033564	0.043965	EGFR/YWHAZ/CDK2	3
hsa05165	0.007885	0.063080921	0.050394	FN1/EGFR/ITGA4/CDK2	4
hsa05134	0.008519	0.063604833	0.050812	VCP/EEF1A1	2
hsa04510	0.012561	0.087929521	0.070244	FN1/EGFR/ITGA4	3
hsa05140	0.015164	0.097201735	0.077652	ITGA4/EEF1A1	2
hsa04810	0.015622	0.097201735	0.077652	FN1/EGFR/ITGA4	3
hsa04512	0.01952	0.112474118	0.089852	FN1/ITGA4	2
hsa05222	0.02122	0.112474118	0.089852	FN1/CDK2	2
hsa05131	0.021516	0.112474118	0.089852	EGFR/UBC/CUL1	3
hsa04657	0.022093	0.112474118	0.089852	HSP90AA1/HSP90AB1	2
hsa04933	0.0248	0.120765971	0.096477	FN1/VCAM1	2
hsa04659	0.028122	0.131236173	0.104841	HSP90AA1/HSP90AB1	2
hsa04670	0.031614	0.141631452	0.113145	ITGA4/VCAM1	2
hsa04068	0.040766	0.175607639	0.140288	EGFR/CDK2	2
hsa05135	0.044211	0.183394284	0.146509	FN1/ITGA4	2