Integrating network pharmacology, molecular docking, and experimental verification to investigate the mechanism of baicalein in colorectal cancer

Introduction

Colorectal cancer (CRC) is among the most prevalent malignancies worldwide and represents the second leading cause of cancer-related death (1,2). Current clinical treatments mainly involve surgical resection, combined radiotherapy and chemotherapy, and molecular targeted therapies (3). Although chemotherapy may prolong patient survival, it frequently causes serious complications, including myelosuppression and gastrointestinal toxicity. CRC is among the most prevalent malignancies worldwide and represents the second leading cause of cancer-related death (4).

In recent decades, traditional Chinese medicine (TCM) and naturally derived compounds have been extensively used in oncology, given their therapeutic benefits and relatively favorable safety profiles (5). Both classical medical texts and modern studies suggest that formulas, including Gegen Qinlian, Huanglian Jiedu, and Baitouweng decoction, hold therapeutic potential in CRC treatment, all of which contain Scutellaria baicalensis Georgi (6-8). Modern pharmacological research has identified flavonoids as its major bioactive constituents, including baicalein (BA), baicalin, wogonoside, and wogonin. These flavonoids are now understood to exhibit diverse pharmacological properties, including anticancer, anti-inflammatory, antimicrobial, antiviral, and neuroprotective effects (9-12). Among these compounds, BA (Figure 1A) has attracted significant interest for its distinctive role in cancer suppression. Prior research indicated that BA suppresses CRC cell proliferation by blocking the JAK2/STAT3/GPX4 axis, thereby triggering ferroptosis (13). Moreover, BA suppresses Snail-mediated epithelial-mesenchymal transition (EMT), which reduces CRC cell proliferation and invasion (14). However, the exact mechanisms of BA in CRC therapy remain largely unclear.

Figure 1 NP prediction of compound-target interactions of BA for CRC treatment. (A) Chemical structure of BA. (B) Overlap of CRC-related genes identified across five database sources. (C) Intersected genes between BA and CRC. (D) Compound-disease-target network. (E) BA-CRC PPI network diagram. (F) Network diagram depicting protein interactions with pivotal BA-CRC targets. (G) GO annotation of 171 BA-related genes. (H) DO analysis of 171 genes with the top 20 disease column plot. (I) KEGG gene set enrichment and column graph highlighting the top 20 pathways. GO, DO, and KEGG analyses were performed using the R package, including clusterProfiler, org.Hs.eg.db, enrichplot, and ggplot2. BA, baicalein; BP, biological process; CC, cellular component; CRC, colorectal cancer; DO, Disease Ontology; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; NP, network pharmacology; OMIM, Online Mendelian Inheritance in Man; PPI, protein-protein interaction; TTD, Therapeutic Target Database.

BA-modulated targets are highly complex, making it challenging to fully elucidate their pharmacological mechanisms using conventional approaches. Network pharmacology (NP) is an emerging interdisciplinary field that integrates systems biology and bioinformatics, offering distinct advantages in identifying therapeutic targets and clarifying complex pharmacological mechanisms. Moreover, molecular docking is widely applied to model receptor-ligand interactions, enabling the prediction of binding conformations and affinity. The past few years have witnessed a burgeoning interest in molecular dynamics simulations (MDS), grounded in Newtonian mechanics, are used to analyze the temporal evolution of molecular systems. This approach allows researchers to evaluate the structural stability and conformational flexibility while identifying the critical dynamic events underlying drug action (15).

Recent studies have suggested that BA may contribute to combination therapy by modulating EGFR/ERK-mediated metabolic reprogramming and coordinately inhibiting the Wnt/β-catenin and PI3K/AKT/mTOR signaling pathways (16,17). However, whether BA exerts antitumor effects in CRC through EGFR-mediated MAPK and PI3K/AKT signaling pathways remains unclear. Therefore, in the present study, a comprehensive approach involving NP, protein-protein interaction (PPI) network analysis, molecular docking, molecular dynamics simulations, and in vitro validation was used to systematically identify overlapping targets and related signaling pathways associated with BA and CRC. The predicted mechanisms were further validated in HCT-8 and HCT116 cells, selected for their differential EGFR expression, with an emphasis on EGFR-mediated downstream signaling. Importantly, this study provides systematic experimental evidence for elucidating the multi-target antitumor mechanisms of BA in CRC. We present this article in accordance with the MDAR reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0200/rc).

Methods

NP analysis

Screening of potential BA targets in CRC

The Simplified Molecular Input Line Entry System (SMILES) identifier and two-dimensional SDF file of BA were obtained from the PubChem database. The SMILES code was C1=CC=C(C=C1)C2=CC(=O)C3=C(O2)C=C(C(=C3O)O)O (18). The potential drug targets of BA were predicted using SwissTargetPrediction (19), PharmMapper (20), and SuperPred (21). CRC-related targets were retrieved from the Therapeutic Target Database (TTD) (22), GeneCards (23), DisGeNET (24), Online Mendelian Inheritance in Man (OMIM) (25), and PharmGKB databases (26). Venn diagram analysis was performed using Venny to identify the overlapping targets between BA and CRC. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Constructing the PPI network

Targets with intersections were added to the STRING repository (version 12.0) to build a PPI network, which was then depicted in Cytoscape software (version 3.10.1). Topological analysis was performed using the Network Analysis tool to calculate parameters, including degree, betweenness centrality (BC), and closeness centrality (CC). Targets with parameter values above the average were considered significant, and the 20 genes with the highest degree values were identified as key targets.

Gene Ontology (GO) enrichment, Disease Ontology (DO) enrichment, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses

GO, DO, and KEGG enrichment analyses were performed using the R packages (“clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, “ggplot2”, “pathview”). The significance thresholds were set at FDR <0.05 for GO and DO terms and P<0.05 for KEGG pathways. GO analysis covered biological processes (BPs), cellular components (CCs), and molecular functions (MFs), whereas DO analysis focused on disease-related gene collections and their categorizations. KEGG enrichment analysis was conducted to systematically elucidate the potential signaling pathways and molecular mechanisms of BA in CRC, thereby providing a theoretical foundation for its clinical application. The major enriched KEGG pathways were subsequently visualized.

Molecular docking

Molecular docking connected BA with six critical targets: EGFR, AKT1, ERK1, BCL-2, GSK3B, and SRC. Directly, the three-dimensional (3D) architecture of BA was extracted from TCMSP and stored as MOL2 and then converted to PDBQT with AutoDockTools (version 1.5.6). The X-ray crystallographic structures of the six receptors were obtained from the RCSB Protein Data Bank. The protein structures were preprocessed and visualized using PyMOL (version 2.1), followed by molecular docking calculations using Sybyl-X (version 2.1.1) and AutoDock Vina. The docked findings were prioritized based on binding affinity, with the leading conformations selected for subsequent examination.

MDS

MDS was performed using the GROMACS software package (version 2022.3) (27). Small molecule preparation involved the use of AmberTools22 to apply the GAFF force field parameters, while Gaussian 16-W handled hydrogenation and the calculation of RESP atomic charges. The obtained force field parameters were incorporated into the topology file of the molecular dynamics system. Simulations were performed at a constant temperature (300 K) and pressure (1 bar). We used the Amber99sb-ildn force field for our calculations, representing the water molecules using the TIP3P model. To ensure the electrical neutrality of the system, we introduced the requisite quantity of sodium ions to balance the total charge. Energy minimization was first achieved using the steepest descent method, followed by equilibration under the isothermal-isovolumic ensemble (NVT) and isothermal-isobaric ensemble (NPT) conditions. Each equilibration phase consisted of 100,000 simulation steps using a 0.1 ps coupling constant and spanning 100 ps in total. Following the initial setup, the MDS was run for five million steps using a two-femtosecond time increment, yielding a cumulative simulation duration of 100 ns. Upon completion, the trajectory data were processed using the software’s native analysis tools. Essential dynamic parameters, including the root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), and radius of gyration of amino acid residues, were calculated to evaluate system stability. These were supplemented with additional thermodynamic data, including the binding free energy [molecular mechanics/generalized Born surface area (MM/GBSA)] and free energy landscape visualizations.

Experimental validation

Chemicals and reagents

BA (purity >95%) was obtained from Meilunbio Biotechnology (Dalian, China) and dissolved in dimethyl sulfoxide (Solarbio, Beijing, China) to prepare a 400 µM stock solution. Crystal violet staining solution and Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) apoptosis detection kit were obtained from KeyGEN Biotechnology (Nanjing, China). The cell counting kit-8 (CCK-8) kit was bought from ApexBio Biotechnology (Houston, TX, USA). The 5-ethynyl-2'-deoxyuridine (EdU) cell proliferation imaging kit was obtained from Abbkine (Wuhan, China). Recombinant human epidermal growth factor (EGF) and the EGFR tyrosine kinase inhibitor AG1478 were purchased from Sigma-Aldrich (St. Louis, MO, USA). EGF was used at a final concentration of 100 ng/mL, and AG1478 was used at a final concentration of 10 µM. The bicinchoninic acid (BCA) protein assay kit was acquired from Beyotime Biotechnology (Shanghai, China). The following antibodies were used: EGFR (YM8344), MEK (YM8273), ERK (YM8336), PI3K (YM8829), AKT (YM3618), p-EGFR (YM8664), p-MEK (YP0167), p-ERK (YM8452), p-AKT (YM8531), Bax (YM8175), Bcl-2 (YM8319), vimentin (YM8324), E-cadherin (YM0207), N-cadherin (YM8097), β-actin (YM8343), and GAPDH (YM8394) (all 1:1,000, ImmunoWay, Plano, TX, USA). Antibody against p-PI3K (#17336) was acquired by Cell Signaling Technology (Danvers, MA, USA) and used at a dilution of 1:1,000.

Cell culture

Human CRC cell lines HCT-8 and HCT116 were obtained from the Chinese Academy of Sciences (Shanghai, China). High-glucose Dulbecco’s modified Eagle medium (DMEM) and RPMI-1640 medium were obtained from Gibco (Grand Island, NY, USA). Fetal bovine serum (FBS) was obtained from CellMax Cell Technology (Beijing, China). HCT-8 cells were cultured in RPMI-1640 medium containing 10% fetal bovine serum and 1% antibiotic mixture (penicillin, streptomycin, and amphotericin B; Solarbio), whereas HCT116 cells were maintained in high-glucose DMEM with the same additives. Cells were maintained at 37 ℃ in a humidified incubator with 5% CO₂, and only logarithmically growing cells were used in the experiments.

Cell viability assay

Cells were seeded in 96-well plates at a density of 5×10³ cells/well. For routine experiments, cells were treated with BA at various concentrations and conditions, and morphological changes were observed. For inhibitor experiments, cells were pretreated with AG1478 (10 µM) for 2 h before exposure to BA (40 µM). After the indicated treatments, the cells were exposed to the CCK-8 reagent for an additional hour, following the manufacturer’s guidelines. The absorbance values at 450 nm were measured using a microplate reader. The cell viability rate (%) was calculated using the following equation:

Cellviabilityrate=OD(As)−OD(Ab)OD(Ac)−OD(Ab)×100%[1]

Colony formation assay

Cells in the logarithmic phase of growth were diluted to a concentration of 1,000 cells/mL. Each well of a 6-well plate received a 1.5 mL suspension containing exactly 1,000 cells to ensure a consistent spread. After incubation for 24 h at 37 ℃, adherent single cells were observed under a microscope (20×). BA was then administered at concentrations of 0, 20, 40, and 80 µM, with the 0 µM group serving as the control. The culture medium was refreshed thrice weekly, enabling undisturbed cellular proliferation for approximately 10 days until visible, uniform colonies developed in the control group. Following treatment, the cells were washed three times with phosphate-buffered saline (PBS) and then immobilized in 4% paraformaldehyde at ambient temperature for 20 min. The cells were then stained with a 0.1% crystal violet solution for the same duration. Any surplus dye was eliminated by rinsing with water. Finally, only colonies containing upwards of 50 cells were tallied under microscopic examination.

EdU staining assay

A manufacturer-recommended EdU detection kit was used to determine cell proliferation. After attachment, the cells were treated with different concentrations of BA for 24 h. Logarithmically growing cells were seeded into 96-well plates at the appropriate density. EdU was diluted 1:1,000 in complete medium, with 100 µL of the EdU mixture introduced/well, followed by incubation for 2 h. Following two PBS washes, 50 µL of a 4% paraformaldehyde solution was introduced into each well and incubated at ambient temperature for 30 min. The wells were washed with PBS for 5 min before adding 100 µL of 0.5% Triton X-100 for permeabilization. Next, 100 µL of Apollo staining reagent was dispensed into each well, incubated for 30 min at 25 ℃, and removed. After adding Hoechst 33342 staining solution to each well, the samples were incubated for 15 min before being thoroughly rinsed with PBS in three separate 5-min washes. The presence of EdU-positive cells was examined using fluorescence microscopy.

Transwell assay

Cells in six-well plates were exposed to different BA concentrations for 24 h. After centrifugation, a 200 µL aliquot of the cell suspension (3×10⁵ cells/mL) was placed in the upper compartment of the Transwell chamber. For invasion assays, the upper chamber was pre-coated with Matrigel to mimic basement membrane conditions, whereas no coating was applied in the migration assays. The bottom well received 600 µL of 10% FBS-supplemented medium, and the inserts were cultured in a humidified environment for 24–48 h. Meanwhile, the cells that had either moved or breached the barrier were set to dry with 4% paraformaldehyde at room temperature for half an hour. Subsequently, they were treated with a 0.5% crystal violet solution for a brief 15-min session. After incubation, the cells remaining on the upper surface of the membrane were gently removed using a cotton swab. Five randomly selected fields from each group were photographed at 200× magnification, and ImageJ software was used to measure the number of cells that had moved or infiltrated.

Wound healing assay

Cells were plated in six-well plates. When the cells reached approximately 90% confluence, a linear scratch was made across the cell monolayer using a 200 µL pipette tip. The wells were rinsed with sterile PBS to eliminate cellular debris, followed by the addition of fresh 2% FBS-supplemented medium containing different concentrations of BA. The scratching time was set at 0 h, and cell morphology was observed microscopically. Cell migration and wound closure were subsequently monitored at 24 and 48 h using an inverted microscope, and the wound area was quantified with ImageJ software.

Migrationrate=cell-freeareaat(0h)-cell-freeareaat(24hor48h)cell-freeareaat(0h)×100%[2]

Flow cytometry assay

The cells were plated in six-well plates and exposed to different BA concentrations for 24 h. Following harvesting and PBS rinsing, the cells were reintroduced into 500 µL of binding buffer. To prevent light-induced degradation, the samples were incubated at room temperature for 15 min after adding 5 µL of Annexin V-FITC and 10 µL of PI. The stained samples were analyzed using a BD flow cytometer (San Jose, CA, USA), and the resulting data were processed using FlowJo software.

Western blot analysis

Logarithmically growing cells were adjusted to 2.5×10⁵ cells/mL. One milliliter of cell suspension was added to each well of a six-well plate. After confirming cell dispersion, the plates were cultured for 24 h to promote cell adherence. For routine Western blot analysis, cells were exposed to various BA concentrations for 24 h, while untreated cells served as the control group. For EGF stimulation experiments, after cell attachment, HCT-8 and HCT116 cells were serum-starved for 12 h, treated with BA (40 µM) or vehicle for 24 h, and then stimulated with EGF (100 ng/mL) for 30 min before protein extraction. Cellular proteins were isolated using RIPA lysis buffer. After centrifugation at 12,000 ×g for 15 min at 4 ℃, the supernatant was collected, and the protein concentration was determined using a BCA protein assay kit. Equivalent protein quantities were then separated using sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) before being transferred onto PVDF membranes for further analysis. GAPDH or β-actin was used as a loading control. The expression levels of EGFR, MEK, ERK, AKT, and PI3K, and their phosphorylated forms (p-EGFR, p-MEK, p-ERK, p-AKT, and p-PI3K), along with Bax, Bcl-2, caspase-3, cleaved caspase-3, E-cadherin, N-cadherin, and vimentin, were analyzed using the respective primary antibodies. After blocking for 2 h, the membranes were incubated with primary antibodies overnight at 4 ℃ and then washed thrice with 1× Tris-buffered saline with Tween 20 (TBST). The membranes were then incubated with secondary antibodies at room temperature for 2 h. The protein bands were visualized and quantified by measuring the grayscale intensity using ImageJ software. Phosphorylated protein levels were normalized to their corresponding total protein levels, and total protein expression was normalized to GAPDH or β-actin, as appropriate. The normalized value of the control group was set to 1.0, and the values of the treatment groups were expressed relative to the control group. Quantitative data were obtained from three independent experiments and are presented as the mean ± standard deviation (SD). The error bars in the Western blot quantification graphs represent the SD.

Statistical analysis

All data processing was conducted using GraphPad Prism software (version 10), with the results displayed as mean ± SD. All data shown represent biological replicates from at least three independent experiments. For comparing two groups, Student’s t-tests were applied, and for multiple group comparisons, a one-way analysis of variance (ANOVA) test was used. Variability was assessed for each dataset. *, P<0.05; **, P<0.01; ***, P<0.001; ns, not significant (P>0.05).

Results

Results of NP

Predicted targets of BA against CRC

The genetic disease-related targets for CRC were meticulously compiled from various databases, including GeneCards, TTD, OMIM, DisGeNet, and PharmGKB. This screening process resulted in the identification of 2,852 genes, which constituted the CRC disease target set (Figure 1B). In parallel, 279 predicted targets of BA were identified. Venn analysis revealed 171 overlapping targets (Figure 1C). These shared molecules served as prospective therapeutic targets for BA in CRC.

PPI networks

Cytoscape was used to construct the “active compound-disease-target” network (Figure 1D), where the elliptical nodes denote the overlapping targets between the active compounds and the disease. A PPI network was generated using the STRING database (version 12.0) (Figure 1E), and the interaction data were imported into Cytoscape for visualization, producing a PPI network of BA-associated targets (Figure 1F). In this network, protein nodes were depicted as circles, with their diameter and color saturation directly proportional to their degree, organized radially from the core based on degree value. The PPI network contained 171 nodes and 2,433 interaction edges. Based on topological parameters, including degree, betweenness centrality, and closeness centrality, 20 key hub targets were identified (Table 1).

GO, DO, and KEGG enrichment analysis

GO, DO, and KEGG enrichment analyses were performed using the R package. GO enrichment (Figure 1G) included BP, CC, and MF categories, with the top 10 terms selected based on P values. BP was mainly enriched in cellular responses to xenobiotic or chemical stress, oxidative stress, hypoxia, peptide hormones, and oxygen level changes, along with terms associated with reactive oxygen species (ROS) metabolism and regulation. CC was significantly enriched in membrane rafts/microdomains, protein kinase complexes, serine/threonine protein kinase complexes, and phosphorus-containing transferase complexes, indicating that key signaling events mainly occur and converge within these structures. MF was enriched in protein tyrosine kinase activity, transmembrane receptor kinase activity, serine/threonine kinase activity, and transcription factor binding. Collectively, these three domains highlight phosphorylation events centered on protein kinase activity and localized within membrane-associated microstructures and kinase complexes. DO enrichment (Figure 1H) further revealed significant associations with gastrointestinal tumors and stress-related conditions, including ischemia/reperfusion, consistent with ROS-phosphorylation-driven signaling features. KEGG enrichment analysis (Figure 1I) revealed that MAPK and PI3K/AKT signaling pathways were among the top-enriched pathways, along with chemical carcinogenesis-ROS, Ras, and FoxO signaling pathways, all linked to stress responses and cell proliferation. Integrating GO findings, ROS/hormone response (BP), membrane raft/kinase complex localization (CC), and protein kinase activity (MF), suggests that BA-related targets primarily converge on phosphorylation-dependent signaling cascades, with MAPK and PI3K/AKT constituting the key axes. Accordingly, we used R to construct mechanistic diagrams of MAPK and PI3K/AKT pathways and incorporated them into our experimental studies for validation (Figure 2A,2B).

Figure 2 Key signaling pathways and molecular docking analysis of BA in CRC. (A) PI3K-AKT signaling pathway. (B) MAPK signaling pathway. (C) Molecular docking scores of BA with key target proteins, exhibiting Total Score (Sybyl) and Binding Energy (AutoDock Vina). BA, baicalein; CRC, colorectal cancer.

Molecular docking results analysis

The binding interactions between BA and key target proteins were assessed using the TotalScore values from Sybyl-X (version 2.1.1) and the binding energies calculated with AutoDock Vina (Figure 2C). The results indicated that all targets had TotalScore values ≥3.0 and binding energies ≤−6 kcal/mol, suggesting that BA exhibited strong binding affinity for these proteins. Notably, EGFR, AKT1, ERK1, BCL2, GSK3B, and SRC demonstrated particularly strong binding characteristics (TotalScore ≥3.0 and binding energies ≤−6.5 kcal/mol), with molecular docking visualizations revealed in Figure 3. These proteins are key regulatory factors in EGFR/MAPK and EGFR/PI3K/AKT signaling pathways, suggesting that BA may regulate signal transduction by modulating their activity or disrupting PPIs.

Figure 3 3D visualization of the molecular docking conformations between BA and its key protein targets. (A) Binding conformation of BA with EGFR (PDB ID: 1m14). (B) Binding conformation of BA with AKT1 (PDB ID: 1unq). (C) Binding conformation of BA with ERK1 (PDB ID: 2zoq). (D) Binding conformation of BA with BCL2 (PDB ID: 1g5m). (E) Binding conformation of BA with GSK3B (PDB ID: 4j1r). (F) Binding conformation of BA with SRC (PDB ID: 2cud). 3D, three-dimensional; BA, baicalein; PDB, Protein Data Bank.

Evaluation of the effects of BA on the MDS of EGFR

Based on the docking results, MDS was performed using GROMACS to evaluate the dynamic stability of the BA-EGFR complex. The EGFR-BA complex maintained an RMSD under 1 nm during simulation, demonstrating strong structural integrity (Figure 4A). RMSF values remained below 1 nm, indicating that BA binding had a negligible effect on EGFR residue flexibility and preserved the structural integrity of the complex (Figure 4B). The gyration radius (Rg) consistently hovered near 2.0 nm (Figure 4C), while the solvent-accessible surface area (SASA) remained constant between 135 and 160 nm² (Figure 4D). These results confirm that BA binding induces minimal structural changes in EGFR, facilitating stable EGFR-BA complex formation.

Figure 4 MDS analyses of the BA-EGFR complex. (A) RMSD profile illustrating conformational stability during the simulation. (B) RMSF plot depicting residue-level flexibility of the protein. (C) Rg trajectory describing the overall compactness of the complex. (D) SASA variation reflecting changes in protein surface exposure. (E) Hydrogen bond dynamics exhibiting the interaction frequency and persistence between BA and EGFR. (F) Free energy landscape map representing conformational energy distribution and stability basins. (G) Mean binding free energy (MM/GBSA) summarizing the overall binding affinity between BA and EGFR. (H) Per-residue energy decomposition highlighting the key amino acids contributing to ligand stabilization. ns, not significant (P>0.05). BA, baicalein; MDS, molecular dynamics simulations; MM/GBSA, molecular mechanics/generalized Born surface area; Rg, gyration radius; RMSD, root-mean-square deviation; RMSF, root-mean-square fluctuation; SASA, solvent-accessible surface area.

Hydrogen bond analysis revealed that 1–3 hydrogen bonds were maintained between the ligand and EGFR throughout the 100-ns MDS, with transient increases to 4–5 at specific time points (Figure 4E). Although the hydrogen bond number fluctuated slightly, it remained relatively stable throughout the simulation. A detailed examination of the free energy landscape, using RMSD, Rg, and Gibbs free energy as X-, Y-, and Z-axes, respectively, uncovered a solitary pronounced energy basin (Figure 4F), which lends additional credence to complex structural integrity.

Following system equilibration, MM/GBSA analysis determined the EGFR-BA binding free energy to be −31.09 kcal·mol⁻¹, reflecting a high-affinity interaction (Figure 4G). Moreover, BA established robust molecular interactions with specific EGFR residues, namely, GLU-738, LEU-820, and ASP-831, demonstrating binding affinities of −3.065, −1.664, and −1.342 kcal·mol⁻¹, respectively (Figure 4H). These findings highlight the critical role played by these amino acids in facilitating ligand-receptor binding. The observed molecular interactions aligned perfectly with our initial docking predictions, demonstrating that the binding configuration maintained its integrity throughout the entire simulation process and provided additional proof of the remarkable stability of the EGFR-BA complex.

Results of the in vitro experiment

BA inhibited CRC cell growth

To evaluate the effect of BA on CRC cell growth, the CCK-8 test was used to determine cell viability after exposure to different BA concentrations. BA suppressed CRC cell growth in a concentration-dependent manner. The half-maximal inhibitory concentration (IC₅₀) values of BA were 59.63 and 56.81 µM for HCT-8 cells at 24 and 48 h, and 73.46 and 67.41 µM for HCT116 cells, respectively, indicating substantial inhibition of cell growth (Figure 5A-5C). Based on the IC₅₀ values obtained from the CCK-8 assay, representative concentrations below, near, and above the IC₅₀ (20, 40, and 80 µM) were selected for subsequent experiments to evaluate the dose-dependent effects of BA. Moreover, colony formation assays revealed that BA markedly reduced colony number and size (Figure 5D,5E). EdU staining further confirmed that BA treatment significantly inhibited CRC-cell proliferation (Figure 5F,5G). Collectively, these findings demonstrated that BA suppresses CRC cell proliferation concentration-dependently.

Figure 5 BA suppresses CRC cell proliferation and growth. (A,B) Cell viability of HCT-8 and HCT116 cells treated with different BA concentrations for 24 or 48 h, as determined by the CCK-8 assay. (C) IC₅₀ values of BA in HCT-8 and HCT116 cells at 24 and 48 h. (D) HCT-8 and HCT116 were treated with BA for 10 days, and colony formation of cells was detected by crystal violet staining. (E) Quantification of colony numbers from ImageJ. (F) Cell proliferation in HCT-8 and HCT116 cells following BA treatment for 24 h was assessed using an EdU incorporation assay. (G) Histogram of EdU detection results by immunofluorescence staining. Data are expressed as mean ± SD. *, P<0.05; **, P<0.01; ***, P<0.001, compared with the control group (n=3). BA, baicalein; CCK-8, cell counting kit-8; CRC, colorectal cancer; EdU, 5-ethynyl-2'-deoxyuridine; IC₅₀, half-maximal inhibitory concentration; SD, standard deviation.

BA suppressed CRC cell migration and invasion

To investigate the effects of BA on CRC cell migration, invasion, and EMT, wound healing and Transwell migration and invasion assays were conducted using HCT-8 and HCT116 cells. The results (Figure 6A-6E) revealed that the untreated cells exhibited a strong migratory capacity, whereas BA treatment significantly and dose-dependently inhibited migration. Specifically, cells were pretreated with BA for 24 h before the Transwell migration and invasion assays, which were conducted for 24–48 h. For the wound healing assay, cells were treated with BA immediately after scratching, and images were captured at 0, 24, and 48 h. Further analysis suggested that the inhibitory effect of BA on CRC cell migration may be associated with EMT. To verify this association, the expression levels of EMT-related marker proteins were examined after BA treatment for 24 h. Western blotting (Figure 6F-6I) demonstrated that BA treatment significantly upregulated E-cadherin expression while downregulating N-cadherin and vimentin levels. Collectively, these findings demonstrated that BA inhibited CRC cell migration and invasion by impeding EMT.

Figure 6 BA inhibited migration, invasion, and EMT in CRC cells. (A) Transwell migration and invasion assays performed on HCT-8 and HCT116 cells pretreated with varying BA concentrations for 24 h. The Transwell assays were conducted for 24–48 h, and the migrated and invaded cells were stained with 0.5% crystal violet. (B) Quantification of migrated cells after BA treatment. (C) Quantification of invaded cells after BA treatment. (D) Wound healing assay used to evaluate the migration rate of HCT-8 and HCT116 cells following BA treatment. Images were captured at 0, 24, and 48 h after scratching. (E) Quantification of the migration rate at 24 and 48 h relative to 0 h from ImageJ. (F,G) Western blot analysis and quantification of E-cadherin, N-cadherin, and vimentin expression in HCT-8 cells after BA treatment for 24 h. (H,I) Western blot analysis and quantification of E-cadherin, N-cadherin, and vimentin expression in HCT116 cells after BA treatment for 24 h. Data are expressed as mean ± SD. ns, not significant (P>0.05); *, P<0.05; **, P<0.01; ***, P<0.001. For Transwell assays and Western blot analysis, BA-treated groups were compared with the control group (n=3). For the wound-healing assay, BA-treated groups were compared with the corresponding control groups at the same time point (n=3). BA, baicalein; CRC, colorectal cancer; EMT, epithelial-mesenchymal transition; SD, standard deviation.

BA induced apoptosis in CRC cells

Our previous NP analysis indicated that the targets associated with the anti-CRC effects of BA were significantly enriched in the apoptotic pathway. Among these targets, Bcl-2 was identified as a core node that exhibited a strong binding affinity for BA. To further confirm the pro-apoptotic effects of BA, Annexin V/PI co-staining was performed using flow cytometry. The results revealed that BA treatment significantly increased apoptosis in CRC cell lines, with a marked elevation in late apoptotic cell proportion (Figure 7A,7B). Western blot analysis revealed that BA treatment notably increased Bax and cleaved caspase-3 levels and decreased Bcl-2 expression (Figure 7C-7F). Together, these results demonstrated that BA triggered programmed cell death in CRC cells.

Figure 7 BA promotes apoptosis in CRC cells in a concentration-dependent manner. (A) Flow cytometry assessment of apoptotic populations in HCT-8 and HCT116 cells after 24 h of BA treatment. (B) Histogram of apoptosis results detected by flow cytometry. (C,D) Western blot analysis and quantification of Bax, Bcl-2, caspase-3, and cleaved caspase-3 expression in HCT-8 cells after BA treatment for 24 h. (E,F) Western blot analysis and quantification of Bax, Bcl-2, caspase-3, and cleaved caspase-3 expression in HCT116 cells after BA treatment for 24 h. Data are expressed as mean ± SD. ns, not significant (P>0.05); *, P<0.05; **, P<0.01; ***, P<0.001, compared with the control group (n=3). BA, baicalein; CRC, colorectal cancer; FITC, fluorescein isothiocyanate; SD, standard deviation.

BA inhibited EGFR activation and downstream MAPK and PI3K/AKT signaling

Guided by the findings from NP and molecular docking studies, we further aimed to study the EGFR-associated signaling pathways implicated in the activity of BA. Compared with the control group, BA treatment significantly reduced the phosphorylation of EGFR, MEK, ERK, PI3K, and AKT in both CRC cell lines dose-dependently, while the total protein levels of these molecules remained unchanged (Figure 8A-8D). Moreover, exposure to EGF substantially enhanced the phosphorylation levels of EGFR and associated downstream effectors, whereas BA administration significantly reduced EGF-induced activation (Figure 9A-9D). Collectively, these data indicate that BA mediates its anti-CRC activity by suppressing EGFR tyrosine kinase function, thereby inhibiting downstream MAPK and PI3K/AKT signaling cascades.

Figure 8 BA dose-dependently inhibits EGFR and its downstream MAPK and PI3K/AKT signaling pathways in CRC cells. (A,B) Western blot analysis and quantification of phosphorylated and total EGFR, MEK1/2, ERK1/2, PI3K, and AKT expression in HCT-8 cells after BA treatment for 24 h. (C,D) Western blot analysis and quantification of phosphorylated and total EGFR, MEK1/2, ERK1/2, PI3K, and AKT expression in HCT116 cells after BA treatment for 24 h. Data are expressed as mean ± SD. ns, not significant (P>0.05); *, P<0.05; **, P<0.01; ***, P<0.001, compared with the control group (n=3). BA, baicalein; CRC, colorectal cancer; SD, standard deviation.

Figure 9 BA reverses EGF-induced activation of EGFR and its downstream MAPK and PI3K/AKT signaling pathways in CRC cells. (A-D) HCT-8 and HCT116 cells were treated with BA (40 µM) or vehicle for 24 h, then stimulated with EGF (100 ng/mL) for 30 min as indicated. Phosphorylated and total EGFR, MEK1/2, ERK1/2, PI3K, and AKT proteins were analyzed by Western blotting. (E) HCT-8 and HCT116 cells were treated with BA (40 µM) or vehicle for 24 h, with or without pretreatment with AG1478 (10 µM, an EGFR tyrosine kinase inhibitor) for 2 h, as indicated. Cell viability was determined using the CCK-8 assay. Data are expressed as mean ± SD, n=3. ns, not significant (P>0.05); *, P<0.05; **, P<0.01; ***, P<0.001. Brackets indicate the groups being compared. BA, baicalein; CCK-8, cell counting kit-8; CRC, colorectal cancer; EGF, epidermal growth factor; SD, standard deviation.

BA suppressed CRC cell proliferation through EGFR-related mechanisms and potential additional pathways

AG1478 treatment was able to partially counteract the BA-mediated reduction in cell viability. The CCK-8 assay results (Figure 9E) demonstrated that exposure to 40 µM BA alone led to a significant reduction in cell viability. Notably, pretreating cells with 10 µM AG1478 for 2 h partially mitigated the decline in viability caused by BA. We propose that, beyond suppressing EGFR phosphorylation, BA may also interact with additional molecular targets and modulate alternative signaling cascades. These findings imply that although BA is not exclusively reliant on EGFR blockade, EGFR-related mechanisms are importantly involved in its antitumor activity.

Discussion

CRC, one of the most common malignancies, poses a significant threat to human health (28,29). In recent years, a growing body of research has highlighted the therapeutic potential of natural bioactive compounds in cancer prevention and therapeutic intervention (30). BA, a flavonoid derived from the Chinese herbal medicine Scutellaria baicalensis Georgi, has demonstrated potential anti-CRC activity and multifaceted tumor-suppressive effects (31-34). Nonetheless, the exact molecular processes governing its anti-CRC effects remain unclear.

In the present study, an integrated analysis combining NP, molecular docking, and MD simulations identified EGFR, AKT1, ERK1, BCL2, GSK3B, and SRC as central nodes in the BA-CRC PPI network. GO analysis indicated that the anticancer effects of BA were associated with the regulation of oxidative stress responses and protein-tyrosine kinase activity. Moreover, KEGG pathway analysis revealed that MAPK, PI3K/AKT, and Ras signaling pathways were closely associated with its anti-CRC effects. These core targets have been established as essential regulators of cell proliferation and survival. It is now understood that EGFR, a key upstream receptor tyrosine kinase, undergoes ligand-induced dimerization upon EGF binding. This event subsequently induces the self-phosphorylation of tyrosine sites located in the cytoplasmic domain of the receptor. Subsequent phosphorylation events initiate several downstream signaling networks, notably MAPK and PI3K/AKT pathways (35-37). ERK1 and AKT1 function as key nodes in these two signaling pathways. SRC forms a complex with EGFR to synergistically activate ERK and AKT pathways, thereby amplifying pro-survival signaling in tumor cells (38). GSK3B, a downstream effector, is inhibited by AKT-mediated phosphorylation and contributes to the regulation of the cell cycle and cellular metabolism (39). Bcl-2 upregulation inhibits mitochondria-dependent apoptosis, thereby promoting tumor cell survival.

Among the pathways identified in the preceding analysis, the ERK1/2 signaling pathway, which is a central component of the MAPK cascade, emerged as a key regulatory pathway. This pathway is essential for CRC cell survival (40). Previous studies have revealed that gossypin promotes apoptotic cell death in CRC cells by stimulating the ERK/p38 MAPK signaling cascade (41). As a major downstream effector of the PI3K/AKT cascade, AKT regulates CRC cell proliferation, survival, metabolic reprogramming, and metastatic progression (42). Research indicates that this route exhibits abnormal activation in approximately 60% of CRC instances, and is strongly associated with drug resistance and tumor progression (43). Notably, concurrent inhibition of PI3K and ERK pathways has been revealed to produce a more pronounced synergistic antitumor effect (44), supporting the hypothesis that BA inhibits cancer via multi-target action.

MDS further confirmed that BA binds stably to EGFR, suggesting that it may exert antitumor effects by disrupting EGFR-mediated signaling. Based on these observations, in vitro assays were conducted to confirm the proposed mechanisms. BA significantly reduced CRC cell viability, triggered apoptosis, and suppressed cellular migration and invasion. Mechanistically, BA treatment decreased the phosphorylation levels of EGFR, MEK1/2, ERK1/2, PI3K, and AKT. Previous studies have demonstrated that ERK1/2 activation can promote anti-apoptotic responses by upregulating Bcl-2 and downregulating Bax, thereby increasing the resistance of cancer cells to apoptosis (45). BA elevated Bax and cleaved caspase-3 expression levels while reducing Bcl-2 expression, suggesting that BA activates apoptosis-related signaling. Metastasis, a multifaceted, multistage process, involves primary tumor cell detachment, extracellular matrix (ECM) breakdown, vascular invasion, and dissemination to distant sites (46-48). EMT has been established as a key mechanism driving tumor cell migration and invasion (49,50). Importantly, EMT is characterized by the loss of E-cadherin in epithelial cells and a concomitant increase in mesenchymal markers, including N-cadherin and vimentin (50). These findings collectively demonstrated that BA effectively counteracted these changes by substantially upregulating E-cadherin expression while downregulating N-cadherin and vimentin, thereby inhibiting EMT and diminishing the metastatic capabilities of CRC cells.

Given its central roles in tumor stemness, proliferation, invasion, and metastasis, and its high expression in 25–82% of CRC cases (51), EGFR is a promising therapeutic target. However, EGFR inhibitors, including gefitinib, demonstrate limited efficacy when used as monotherapy in CRC. This limited efficacy is largely attributed to primary and acquired resistance resulting from downstream pathway reactivation and signaling redundancy. Natural small molecules can simultaneously modulate multiple biological pathways. Their multi-target activity may generate synergistic or additive effects, contributing to resistance mitigation to EGFR inhibitors. In this study, EGFR activation was transiently inhibited using AG1478, followed by BA treatment. The observed partial restoration of cell viability indicates that BA’s antitumor activity is attributable to the suppression of EGFR activation and may involve the regulation of other signaling cascades. Conversely, AG1478 targets only a single pathway.

Although this study provides valuable insights, several limitations should be acknowledged. First, only two CRC cell lines were used, which may not fully represent the biological heterogeneity of CRC. Second, although BA exerted direct inhibitory effects on CRC cells in vitro, the relatively high IC₅₀ values suggest that its effective dose and antitumor activity in vivo require further validation. Third, the mechanistic validation mainly focused on the EGFR-centered MAPK and PI3K/AKT signaling axis. Therefore, SRC and GSK3B, although identified as potential targets in the integrated analysis, were not directly examined. Further studies are warranted to validate the in vivo antitumor efficacy of BA and clarify the contribution of additional candidate targets to its anti-CRC effects. In addition, formulation optimization, prodrug design, targeted delivery, or combination therapy may help further improve the in vivo activity of BA.

Conclusions

In summary, BA showed significant inhibitory effects on CRC cells by suppressing cell proliferation, migration, invasion, and EMT while promoting apoptosis. Mechanistically, BA exerted its anti-CRC activity by inhibiting EGFR phosphorylation, thereby suppressing downstream MAPK and PI3K/AKT signaling pathways. The proposed molecular mechanism underlying BA activity is illustrated in Figure 10. These findings support BA as a promising therapeutic candidate for CRC.

Figure 10 Proposed molecular mechanism underlying the anticancer effects of BA in CRC cells. BA, baicalein; CRC, colorectal cancer; EGF, epidermal growth factor; EMT, epithelial-mesenchymal transition.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0200/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0200/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0200/prf

Funding: This study was supported by the School Management Project of Fujian University of Traditional Chinese Medicine (No. X2024009).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0200/coif). All authors report funding support from the School Management Project of Fujian University of Traditional Chinese Medicine (No. X2024009). The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Zheng RS, Zhang SW, Sun KX, et al. Cancer statistics in China, 2016. Zhonghua Zhong Liu Za Zhi 2023;45:212-20. [Crossref] [PubMed]
3. Eng C, Yoshino T, Ruíz-García E. Lancet 2024;404:294-310. [Crossref] [PubMed]
4. Liotti A, La Civita E, Cennamo M, et al. Periprostatic adipose tissue promotes prostate cancer resistance to docetaxel by paracrine IGF-1 upregulation of TUBB2B beta-tubulin isoform. Prostate 2021;81:407-17. [Crossref] [PubMed]
5. Liu Y, Fang C, Luo J, et al. Traditional Chinese Medicine for Cancer Treatment. Am J Chin Med 2024;52:583-604. [Crossref] [PubMed]
6. Chan YT, Cheung F, Zhang C, et al. Ancient Chinese Medicine Herbal Formula Huanglian Jiedu Decoction as a Neoadjuvant Treatment of Chemotherapy by Improving Diarrhea and Tumor Response. Front Pharmacol 2020;11:252. [Crossref] [PubMed]
7. Chen D, Cai Y, Gao W, et al. Study on the regulation mechanism of TBX5 gene and Gegen Qinlian decoction on colorectal cancer. Front Oncol 2025;15:1732015. [Crossref] [PubMed]
8. Ji X, Chen Z, Lin W, et al. Esculin induces endoplasmic reticulum stress and drives apoptosis and ferroptosis in colorectal cancer via PERK regulating eIF2α/CHOP and Nrf2/HO-1 cascades. J Ethnopharmacol 2024;328:118139. [Crossref] [PubMed]
9. Tuli HS, Aggarwal V, Kaur J, et al. Baicalein: A metabolite with promising antineoplastic activity. Life Sci 2020;259:118183. [Crossref] [PubMed]
10. Zhao K, Zhang J, Zhou L, et al. Scutellaria baicalensis and its flavonoids in the treatment of digestive system tumors. Front Pharmacol 2024;15:1483785. [Crossref] [PubMed]
11. Zhang J, Tan B, Wu H, et al. Scutellaria baicalensis Extracts Restrict Intestinal Epithelial Cell Ferroptosis by Regulating Lipid Peroxidation and GPX4/ACSL4 in Colitis. Phytomedicine 2025;141:156708. [Crossref] [PubMed]
12. Wen Y, Wang Y, Zhao C, et al. The Pharmacological Efficacy of Baicalin in Inflammatory Diseases. Int J Mol Sci 2023;24:9317. [Crossref] [PubMed]
13. Lai JQ, Zhao LL, Hong C, et al. Baicalein triggers ferroptosis in colorectal cancer cells via blocking the JAK2/STAT3/GPX4 axis. Acta Pharmacol Sin 2024;45:1715-26. [Crossref] [PubMed]
14. Zeng Q, Zhang Y, Zhang W, et al. Baicalein suppresses the proliferation and invasiveness of colorectal cancer cells by inhibiting Snail induced epithelial mesenchymal transition. Mol Med Rep 2020;21:2544-52. [Crossref] [PubMed]
15. Wu X, Xu LY, Li EM, et al. Application of molecular dynamics simulation in biomedicine. Chem Biol Drug Des 2022;99:789-800. [Crossref] [PubMed]
16. Zhou P, Zeng Q, Zhang Q, et al. Beta-sitosterol-baicalein-guanosine synergistically alleviates Warburg effect in colorectal cancer via EGFR/ERK pathway. J Pharm Pharmacol 2026;78:rgag006. [Crossref] [PubMed]
17. Anlei W, Kaihao W, Yazhao G, et al. Baicalein suppresses colorectal cancer progression through dual inhibition of the Wnt/β--catenin and PI3K/AKT/mTOR signaling pathways. Cytotechnology 2026;78:87. [Crossref] [PubMed]
18. Kim S, Chen J, Cheng T, et al. PubChem 2025 update. Nucleic Acids Res 2025;53:D1516-25. [Crossref] [PubMed]
19. Daina A, Michielin O, Zoete V. SwissTargetPrediction: updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res 2019;47:W357-64. [Crossref] [PubMed]
20. Wang X, Shen Y, Wang S, et al. PharmMapper 2017 update: a web server for potential drug target identification with a comprehensive target pharmacophore database. Nucleic Acids Res 2017;45:W356-60. [Crossref] [PubMed]
21. Gallo K, Goede A, Preissner R, et al. SuperPred 3.0: drug classification and target prediction-a machine learning approach. Nucleic Acids Res 2022;50:W726-31. [Crossref] [PubMed]
22. Zhou Y, Zhang Y, Zhao D, et al. TTD: Therapeutic Target Database describing target druggability information. Nucleic Acids Res 2024;52:D1465-77. [Crossref] [PubMed]
23. Safran M, Dalah I, Alexander J, et al. GeneCards Version 3: the human gene integrator. Database (Oxford) 2010;2010:baq020. [Crossref] [PubMed]
24. Piñero J, Bravo À, Queralt-Rosinach N, et al. DisGeNET: a comprehensive platform integrating information on human disease-associated genes and variants. Nucleic Acids Res 2017;45:D833-9. [Crossref] [PubMed]
25. Amberger JS, Hamosh A. Searching Online Mendelian Inheritance in Man (OMIM): A Knowledgebase of Human Genes and Genetic Phenotypes. Curr Protoc Bioinformatics 2017;58:1.2.1-1.2.12.
26. Barbarino JM, Whirl-Carrillo M, Altman RB, et al. PharmGKB: A worldwide resource for pharmacogenomic information. Wiley Interdiscip Rev Syst Biol Med 2018;10:e1417. [Crossref] [PubMed]
27. Van Der Spoel D, Lindahl E, Hess B, et al. GROMACS: fast, flexible, and free. J Comput Chem 2005;26:1701-18. [Crossref] [PubMed]
28. Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024;4:47-53. [Crossref] [PubMed]
29. Xi Y, Xu P. Global colorectal cancer burden in 2020 and projections to 2040. Transl Oncol 2021;14:101174. [Crossref] [PubMed]
30. Agrawal S, Chavan P, Dufossé L. Hidden Treasure: Halophilic Fungi as a Repository of Bioactive Lead Compounds. J Fungi (Basel) 2024;10:290. [Crossref] [PubMed]
31. Morshed AKMH, Paul S, Hossain A, et al. Baicalein as Promising Anticancer Agent: A Comprehensive Analysis on Molecular Mechanisms and Therapeutic Perspectives. Cancers (Basel) 2023;15:2128. [Crossref] [PubMed]
32. Li J, Zhang D, Wang S, et al. Baicalein induces apoptosis by inhibiting the glutamine-mTOR metabolic pathway in lung cancer. J Adv Res 2025;68:341-57. [Crossref] [PubMed]
33. Jin Y, Wu Q, Pan S, et al. Baicalein enhances cisplatin sensitivity in cervical cancer cells by promoting cuproptosis through the Akt pathway. Biomed Pharmacother 2024;179:117415. [Crossref] [PubMed]
34. He G, Huang X, Dong Y, et al. Preliminary investigation on the mechanism of baicalein regulating the effects of Nischarin on invasion and apoptosis of human breast cancer cells MCF-7 through Wnt3α/β-catenin pathway. Int Immunopharmacol 2024;143:113262. [Crossref] [PubMed]
35. Li Q, Li Z, Luo T, et al. Targeting the PI3K/AKT/mTOR and RAF/MEK/ERK pathways for cancer therapy. Mol Biomed 2022;3:47. [Crossref] [PubMed]
36. Qu X, Hamidi H, Johnson RM, et al. Ligand-activated EGFR/MAPK signaling but not PI3K, are key resistance mechanisms to EGFR-therapy in colorectal cancer. Nat Commun 2025;16:4332. [Crossref] [PubMed]
37. Chan XY, Chang KP, Yang CY, et al. Upregulation of ENAH by a PI3K/AKT/β-catenin cascade promotes oral cancer cell migration and growth via an ITGB5/Src axis. Cell Mol Biol Lett 2024;29:136. [Crossref] [PubMed]
38. Liang Y, Wang Y, Zhang X, et al. Melatonin alleviates valproic acid-induced neural tube defects by modulating Src/PI3K/ERK signaling and oxidative stress. Acta Biochim Biophys Sin (Shanghai) 2024;56:23-33. [Crossref] [PubMed]
39. Zheng Y, Yang Y, Zhu W, et al. GSK3B inhibition reduced cervical cancer cell proliferation and migration by modulating the PI3K/Akt signaling pathway and epithelial-to-mesenchymal transition. Braz J Med Biol Res 2024;57:e13796. [Crossref] [PubMed]
40. Zhou L, Zhang J, Zhao K, et al. Natural products modulating MAPK for CRC treatment: a promising strategy. Front Pharmacol 2025;16:1514486. [Crossref] [PubMed]
41. Moon JM, Lee SW, Jang YS, et al. Gossypin induces apoptosis and autophagy via the MAPK/JNK pathway in HT 29 human colorectal cancer cells. Int J Mol Med 2025;56:107. [Crossref] [PubMed]
42. Narayanankutty A. PI3K/ Akt/ mTOR Pathway as a Therapeutic Target for Colorectal Cancer: A Review of Preclinical and Clinical Evidence. Curr Drug Targets 2019;20:1217-26. [Crossref] [PubMed]
43. He Y, Sun MM, Zhang GG, et al. Targeting PI3K/Akt signal transduction for cancer therapy. Signal Transduct Target Ther 2021;6:425. [Crossref] [PubMed]
44. Wright TD, Raybuck C, Bhatt A, et al. Pharmacological inhibition of the MEK5/ERK5 and PI3K/Akt signaling pathways synergistically reduces viability in triple-negative breast cancer. J Cell Biochem 2020;121:1156-68. [Crossref] [PubMed]
45. Cook SJ, Stuart K, Gilley R, et al. Control of cell death and mitochondrial fission by ERK1/2 MAP kinase signalling. FEBS J 2017;284:4177-95. [Crossref] [PubMed]
46. Li Y, Liu F, Cai Q, et al. Invasion and metastasis in cancer: molecular insights and therapeutic targets. Signal Transduct Target Ther 2025;10:57. [Crossref] [PubMed]
47. Zhang L, Zhou J, Kong W. Extracellular matrix in vascular homeostasis and disease. Nat Rev Cardiol 2025;22:333-53. [Crossref] [PubMed]
48. Gerstberger S, Jiang Q, Ganesh K. Metastasis. Cell 2023;186:1564-79. [Crossref] [PubMed]
49. Hu JL, Wang W, Lan XL, et al. CAFs secreted exosomes promote metastasis and chemotherapy resistance by enhancing cell stemness and epithelial-mesenchymal transition in colorectal cancer. Mol Cancer 2019;18:91. [Crossref] [PubMed]
50. Xue W, Yang L, Chen C, et al. Wnt/β-catenin-driven EMT regulation in human cancers. Cell Mol Life Sci 2024;81:79. [Crossref] [PubMed]
51. Ríos-Hoyo A, Monzonís X, Vidal J, et al. Unveiling acquired resistance to anti-EGFR therapies in colorectal cancer: a long and winding road. Front Pharmacol 2024;15:1398419. [Crossref] [PubMed]