Warfarin and bevacizumab suppress tumor progression in pancreatic ductal adenocarcinoma by targeting EGFR-PI3K-Akt signaling: inhibition of proliferation/migration and apoptosis induction

Introduction

Pancreatic ductal adenocarcinoma (PDAC), accounting for approximately 85% of pancreatic malignancies, is characterized by aggressive biological behavior and dismal prognosis, with rising incidence in younger populations (1,2). According to the most recent Cancer Statistics report, the 5-year relative survival rate for pancreatic cancer is only 13%, the lowest among all major cancers (3). PDAC exhibits a high tumor mutational burden, chemoradiotherapy resistance, and a densely desmoplastic stromal microenvironment that critically compromise drugs delivery (4-6). Only 15–20% of patients are eligible for surgical resection. For those with locally advanced or metastatic disease, first-line treatments include gemcitabine combined with albumin-bound paclitaxel or FOLFIRINOX regimens (7,8). Despite these advances, overall survival has improved only marginally over the past decade (9,10).

Angiogenesis, a hallmark of tumor progression, plays a critical role in promoting tumor growth and metastatic dissemination through vascular nutrient supply (11). Vascular endothelial growth factor (VEGF) has emerged as a dominant angiogenic driver in PDAC, wherein VEGF binding triggers VEGF receptor (VEGFR) autophosphorylation and subsequent activation of proliferation and survival signaling in endothelial cells (12,13). Bevacizumab, a VEGF-A-targeting monoclonal antibody, demonstrated promising results in preclinical and phase II trials (14-16), but failed to produce significant survival benefits in later trials. Recent evidence indicates that resistance to VEGF-targeted therapy may arise through compensatory activation of alternative angiogenic pathways, including fibroblast growth factor and platelet-derived growth factor signaling (17). Furthermore, metabolic reprogramming of endothelial cells and remodeling of the tumor microenvironment further contribute to diminished responsiveness to anti-angiogenic agents (18).

Hypercoagulability, a hallmark of PDAC pathophysiology, contributes to both thromboembolic complications and tumor perfusion deficits through microvascular dysfunction (19). Anticoagulants such as warfarin may improve microvascular perfusion and has been shown to possess antitumor properties beyond their anticoagulant effects. Warfarin targets vitamin K epoxide reductase complex subunit 1 (VKORC1), inhibiting the synthesis of vitamin K-dependent proteins, including growth arrest-specific protein 6 (Gas6), a ligand of the Axl receptor tyrosine kinase (Axl). Inhibition of the Gas6/Axl axis has been associated with reduced epithelial plasticity and metastasis in PDAC (20,21). Previous studies demonstrated that Gas6-induced Axl activation is a critical driver of PDAC progression, and that low-dose warfarin can block Gas6/Axl signaling to inhibit tumor migration, invasion, and proliferation while enhancing apoptosis (20). Our previous work further confirmed that warfarin blocks the Gas6/Axl-PI3K-Akt-nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) cascade in lung cancer models, supporting its broader relevance in oncogenic signaling beyond coagulation (22).

In PDAC, epidermal growth factor receptor (EGFR) also functions as a key upstream regulator of the PI3K-Akt pathway. Integrative bioinformatics analysis identified EGFR as a central node connecting angiogenic and proliferative signaling. Recent studies have highlighted that the immunosuppressive microenvironment of PDAC, characterized by dense desmoplastic stroma, activated fibroblasts, and infiltration of immunosuppressive myeloid cells, represents a major obstacle to both immune-mediated and pharmacologic interventions (23). Collectively, these observations support the hypothesis that dual targeting with warfarin and bevacizumab may synergistically suppress PDAC progression by modulating the EGFR-PI3K-Akt signaling axis, thereby overcoming oncogenic pathway plasticity and enhancing translational potential. To test this hypothesis, we combined integrative bioinformatics, molecular docking, and in vitro functional assays to elucidate the mechanistic basis and therapeutic implications of this strategy. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1190/rc).

Methods

Potential target identification and molecular docking

Pancreatic cancer-related genes were identified through the OMIM (24) and GeneCards (25) databases (score ≥10), while potential warfarin targets were retrieved from DrugBank (26) and SwissTargetPrediction (27). The intersection of these targets was used to construct a protein-protein interaction (PPI) network via the STRING (28), and the top 10 hub genes were identified using Cytoscape (29) (v3.10). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using R (v4.3.3) with the clusterProfiler, org.Hs.eg.db, enrichplot, and pathview packages.

Key molecular targets were prioritized through integration of PPI network core genes with significant GO/KEGG pathway enrichment (30). Warfarin’s 3D structure (PubChem SDF) was converted to mol2 format using OpenBabel (31). Receptor structures were acquired from AlphaFold (32) and RCSB PDB (33), with subsequent structural optimization performed in PyMOL (34) and AutoDockTools (35). Molecular docking simulations implemented in AutoDockTools quantified binding affinities through calculated free energy values (ΔG), where more negative values indicated stronger ligand-receptor interactions. Spatial configurations of top-ranked complexes were visualized using PyMOL.

Cell culture and passage

The PANC-1 PDAC cell line was obtained from Wuhan PriCells Life Technologies Co., Ltd. (Wuhan, China; Catalog No. CL-0184). Pancreatic Cancer Cell Line-1 (Panc-1 cells), which harbors a G12D mutation in codon 12 of the Kirsten rat sarcoma viral oncogene homolog (KRAS) gene, was cultured in Roswell Park Memorial Institute (RPMI)-1640 complete medium supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin (Meilune, Dalian, China, MA0215). The cells were maintained at 37 ℃ in a humidified incubator with 5% CO₂ (Thermo, Waltham, MA, USA, Model 311). For subculturing, cells were incubated with 1 mL of 0.25% trypsin (Meilune, MA0233) for 3–5 minutes to facilitate detachment. Enzymatic digestion was neutralized by adding 2 mL of complete medium, followed by centrifugation at 1,000 rpm for 3 minutes. The supernatant was discarded, and the cell pellet was resuspended in fresh complete medium. The resulting cell suspension was then seeded at the desired ratio for further culture or experimentation. All experiments were conducted using cells from the same validated batch to minimize biological variability.

Cell growth and proliferation assay

To establish suitable working concentrations, preliminary dose-response experiments were conducted using the Cell Counting Kit-8 (CCK-8) assay. Panc-1 cells were exposed to warfarin (0.2–1.6 mmol/L) and bevacizumab (100–500 mg/L) across multiple time points. From these exploratory assays, 0.8 mmol/L warfarin and 500 mg/L (≈3.36 µM, based on a molecular weight of ~149 kDa) were chosen as the working concentrations for subsequent combination and functional studies. Bevacizumab was tested at 100, 250, and 500 mg/L in monotherapy.

Panc-1 cells (logarithmic phase) were seeded in 96-well plates (5×10³ cells/mL, 6-8 replicates/group). After stabilization, cells were treated with: (I) blank control group (blank, no treatment); (II) warfarin group (WA, 0.8 mmol/L warfarin); (III) bevacizumab group (BE, 500 mg/L bevacizumab); (IV) warfarin + bevacizumab group (WB, 0.8 mmol/L warfarin + 500 mg/L bevacizumab).

Following 48-hour incubation, cell viability was assessed at 12-hour intervals via CCK-8 assay: 10 µL reagent added per well, incubated (37 ℃, 1 h), and absorbance (450 nm) measured. Survival rate was calculated as:

Cell viability (%)=(D(450)treatment groupD(450)control group)×100%[1]

Drug synergy was quantified using the Bliss Independence Model (36):

EABBliss=EA+EB−EA×EB[2]

Bliss Score=EABobs−EABBliss[3]

Where EABobs= observed inhibition (combination); EABBliss= theoretical inhibition rate (the same combination). E_A/E_B = inhibition (monotherapy). Bliss score >1: synergistic interaction; Bliss score ≈1 (±0.05): additive effect; and Bliss score <1: antagonistic interaction.

Synergy scores were calculated using R (v4.3.3) with the synergyfinder, dplyr, and ggplot2 packages.

Wound healing assay

Under sterile conditions, culture inserts (Ibidi, Gräfelfing, Germany, 80209) were carefully placed into 6-well plates (NEST, Wuxi, China, 705001) and gently pressed to ensure complete adhesion. Panc-1 cells were adjusted to a concentration of 3×10⁵ cells/mL, and a volume of 70 µL of the cell suspension was added into each insert well. Cells were incubated until achieving a confluent monolayer. Subsequently, the culture inserts were carefully removed, and cell monolayers were gently washed with phosphate-buffered saline (PBS) to eliminate floating or unattached cells. Thereafter, 2 mL of culture medium containing specified drug concentrations was added to each well according to the designated experimental groups. Images of the wound area were captured at 0 and 48 hours using an inverted fluorescence microscope (Nikon, Tokyo, Japan, Ts2-FL) to evaluate cell migration capability. All experiments were performed in triplicate using independent biological replicates.

Cell cycle analysis and apoptosis-related protein expression

Panc-1 cells in the logarithmic growth phase were seeded into 6-well plates at a density of 1×10⁵ cells/mL and allowed to adhere. Once cell growth and confluency reached approximately 80%, cells were treated with different drug conditions, consistent with the grouping scheme described in preceding sections. After treatment, cell cycle distribution was analyzed using flow cytometry, and the expression levels of apoptosis-related proteins, including B-cell lymphoma 2 (Bcl-2), Bcl-2-associated X protein (Bax), and cysteine-aspartic acid protease-3 (caspase-3), were assessed via Western blot analysis. All experiments were performed in triplicate, and the mean values were calculated for statistical analysis.

Detection of EGFR, PI3K, and AKT1 protein expression in cells

Cells cultured in 6-well plates were collected and washed 2–3 times with PBS. Radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitors was added, and cells were incubated for 3–5 minutes. The cells were then scraped off using a sterile cell scraper and transferred into 1.5 mL centrifuge tubes. To ensure complete lysis, the samples were incubated on ice for 30 minutes with intermittent pipetting. The lysates were then centrifuged at 12,000 rpm for 10 minutes at 4 ℃, and the supernatant was collected for total protein extraction. Protein concentration was determined using the bicinchoninic acid (BCA) assay. Following protein denaturation, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed, and proteins were transferred onto a nitrocellulose (NC) membrane. The membrane was blocked and sequentially incubated with primary and secondary antibodies. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection system, and quantitative analysis was conducted using Image Lab software. Each group was tested in triplicate using identical cell preparations, and the mean values were calculated for statistical analysis.

Quantification of messenger RNA (mRNA) expression levels of EGFR, PI3K, and AKT1 by quantitative real-time polymerase chain reaction (qRT-PCR)

Panc-1 cells from 6-well plates were collected, washed twice with PBS, and lysed with RNAiso Plus (1 mL per well). After incubation at room temperature for 5 minutes, 200 µL of chloroform was added, mixed, and centrifuged at 12,000 rpm for 12 minutes at 4 ℃. The aqueous phase (400 µL) was transferred, mixed with 400 µL of isopropanol, and incubated at −20 ℃ for 15 minutes, followed by centrifugation at 12,000 rpm for 10 minutes at 4 ℃. The RNA pellet was washed with 75% ethanol, air-dried, and dissolved in 20 µL of nuclease-free water. RNA concentration and purity were determined, and samples were adjusted to 200 ng/µL. For complementary DNA (cDNA) synthesis, 2 µg of RNA was mixed with oligo(dT)18 primers, incubated at 65 ℃ for 5 minutes, and cooled on ice. Reverse transcription was performed using a standard reaction buffer, deoxynucleotide triphosphates (dNTPs), RNase inhibitor, and reverse transcriptase, incubated at 42 ℃ for 60 minutes, followed by inactivation at 70 ℃ for 5 minutes. qPCR reactions were performed in using 2× qPCR mix, gene-specific primers (Table 1) and cDNA template. Relative gene expression was analyzed using the 2^−ΔΔCT method. Reactions were conducted in triplicate using RNA from the same batch per group.

Statistical analysis

All statistical analyses were rigorously performed in accordance with biostatistical guidelines. Data visualization and preliminary analyses were conducted using GraphPad Prism (version 10.1.2) and R (version 4.3.3). Continuous variables were presented as mean ± standard deviation (SD). Intergroup differences were evaluated using two-way analysis of variance (ANOVA), with post hoc Tukey tests for multiple comparisons. The statistical significance threshold was set at α=0.05 (two-tailed). A P<0.05 was considered statistically significant.

Results

Integrated bioinformatics and molecular docking analysis

A total of 2,923 pancreatic cancer-related genes were identified from the OMIM and GeneCards databases (score ≥10). Concurrently, 111 potential warfarin targets were obtained from the DrugBank and Swiss Target Prediction databases. The intersection of these two datasets yielded 70 common targets between warfarin and pancreatic cancer. A PPI network was constructed for these intersecting targets using the STRING database with a confidence score threshold >0.7 (Figure 1A).

Figure 1 Warfarin and pancreatic cancer integrated bioinformatics analysis. (A) Intersecting targets; (B) the hub genes; (C) GO enrichment analysis; (D) KEGG pathway analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

The resulting PPI network was imported into Cytoscape (v3.10) and analyzed using the CytoHubba plugin, through which the top 10 core targets were identified: PIK3CA, PIK3CD, PIK3R1, PIK3CB, EGFR, JAK2, PLCG2, SRC, KDR, and AKT1 (Figure 1B). These targets have been shown to play central roles in tumor-related signaling networks, particularly in pathways associated with cell proliferation, angiogenesis, and immune response.

To explore the functional relevance of these targets, GO and KEGG enrichment analyses were performed using R software (v4.3.3). GO enrichment analysis revealed statistically significant involvement in various biological processes (BPs), including responses to lipopolysaccharide, cellular responses to biotic stimuli, and smooth muscle cell proliferation, indicating roles in inflammation, host defense, and tumor progression. In terms of cellular components (CCs), target genes were enriched in PI3K complexes, interleukin receptor complexes, and membrane-associated structures, suggesting localization to key signaling platforms. Within the molecular function (MF) category, significant enrichment was observed in serine/threonine kinase activity, histone modification, oxidoreductase activity, and tetrapyrrole and heme binding, which highlighted their involvement in signal transduction, epigenetic regulation, and oxidative stress responses. (Figure 1C).

KEGG pathway analysis identified 50 significantly enriched pathways (adjusted P<0.05). These included the PI3K-Akt signaling pathway, VEGF signaling, Erb-B receptor signaling, EGFR tyrosine kinase inhibitor resistance, forkhead box O (FoxO) signaling, hypoxia-inducible factor-1 (HIF-1) pathway, and C-type lectin receptor signaling, many of which were involved in key processes such as inflammation, immune modulation, tumor angiogenesis, and cell cycle regulation (Figure 1D).

Based on the GO and KEGG enrichment results, and the identification of core targets in the PPI network, six targets (EGFR, PIK3CA/CB/CD/R1, AKT1) were prioritized for molecular docking analysis. Molecular docking analysis revealed that warfarin exhibited favorable binding affinities with all selected targets. Warfarin exhibited strong binding to EGFR (ΔG=−7.6 kcal/mol), PIK3CA (−8.2), PIK3CB (−7.2), PIK3CD (−8.6), and AKT1 (−9.1), with moderate affinity for PIK3R1 (ΔG =−6.1) (Table 2). Representative docking conformations, including key hydrogen bond interactions, are illustrated in Figure 2, further supporting the predicted binding stability and interaction specificity between warfarin and these targets.

Figure 2 Simulation of molecular docking of warfarin to the core target. (A) Warfarin and EGFR; (B) warfarin and PIK3CA; (C) warfarin and PIK3CB; (D) warfarin and PIK3CD; (E) warfarin and PIK3R1; (F) warfarin and AKT1.

Among the 70 overlapping targets, EGFR, PI3K isoforms, and AKT1 were identified as central hubs in the PPI network and showed strong binding affinities with warfarin. These nodes were therefore prioritized for experimental validation, ensuring that subsequent functional assays focused on the EGFR-PI3K-Akt axis as a mechanistically central pathway.

Bevacizumab-warfarin combinatorial treatment inhibits Panc-1 cell growth: dose-response and Bliss model evaluation

A concentration-dependent inhibitory effect of warfarin on Panc-1 cell proliferation was observed. At the highest tested concentration (1.6 mmol/L), cell viability decreased to approximately 15% at 72 h, whereas 100 mg/L bevacizumab showed minimal inhibition, with cell viability exceeding 80%. Intermediate concentrations, particularly 0.8 mmol/L warfarin and 500 mg/L bevacizumab, produced moderate yet consistent growth suppression and were selected for subsequent combination studies (Table 3).

Bliss independence analysis revealed that the combination of 0.8 mmol/L warfarin and 500 mg/L bevacizumab exhibited a synergistic inhibitory effect at 24 h (Bliss score =0.335), resulting in an inhibition rate of 57.44%, compared with 18.61% for warfarin and 6.53% for bevacizumab monotherapy. At 48 h, the interaction approached additivity (Bliss score ≈0), and the combination group achieved an inhibition rate of 50.51%, higher than either monotherapy (37.76% and 22.84%). The distribution of Bliss scores and interaction patterns are visualized in Figure 3. The inhibitory effect remained stable at approximately 50% at 72 h, indicating sustained antiproliferative activity over time (Table 3). In comparison, the combination of 0.8 mmol/L warfarin and 250 mg/L bevacizumab showed a slightly higher Bliss score (0.009) but lower overall inhibition, indicating that higher bevacizumab concentration enhanced growth suppression.

Figure 3 The inhibitory effects of warfarin and bevacizumab combinations on Panc-1 cells at 24 and 48 h were evaluated using the Bliss independence model. Data points are categorized by interaction type: synergistic effect (blue; Bliss score >+0.05), additive effects (red; −0.05≤ Bliss score ≤+0.05), and antagonism (green; Bliss score <−0.05). The diagonal dashed line (y = x) represents theoretical agreement between predicted and observed inhibition rates.

To further elucidate the biological consequences of the observed growth inhibition, subsequent experiments were conducted to assess the effects of the combined treatment on apoptosis and cell migration.

Bevacizumab-warfarin combinatorial treatment attenuates Panc-1 cell migration: quantitative analysis via the wound healing assay

The wound healing assay demonstrated notable differences in the migratory capacity of Panc-1 cells among the treatment groups (Figure 4). In the blank control group, extensive wound closure was observed after 48 hours, indicating strong intrinsic migratory activity. Treatment with WA significantly inhibited cell migration compared to the blank group (P<0.001). In contrast, BE alone did not significantly reduce migration relative to the control (P>0.05), suggesting a limited effect on motility in this assay setting. Importantly, the WB did not further enhance the anti-migratory effect compared to WA monotherapy, with no statistically significant difference observed between the two groups.

Figure 4 Inhibitory effects of warfarin and bevacizumab on the migratory capacity of Panc-1 cells. Cells were treated with 0.8 mmol/L WA, 500 mg/L BE, or WB, with untreated cells serving as the blank control. Cell migration was assessed following drug exposure. Statistical significance was determined via hypothesis testing, with annotations as follows: ns, not significant, P≥0.05; ****, P<0.0001. Images were captured at an original magnification of ×200 (10× objective). BE, bevacizumab; WA, warfarin; WB, warfarin + bevacizumab.

Bevacizumab-warfarin combinatorial treatment promotes Panc-1 cell apoptosis: mechanism verified by flow cytometry and Western blot assay

Flow cytometry analysis revealed significantly increased apoptosis rates in the WA, BE, and WB groups compared to the blank control (8.46%±0.48%). The apoptosis rate reached 19.35%±1.63% in the WA group and 23.14%±2.29% in the WB group, both showing a statistically significant elevation versus blank (P<0.001). In contrast, the BE group exhibited a modest increase (10.92%±1.11%) that did not reach statistical significance (P>0.05), indicating a limited pro-apoptotic effect. Although the WB group showed a higher apoptosis rate than WA alone, although the difference was not statistically significant (Figure 5, upper).

Figure 5 Effects of warfarin and bevacizumab of Panc-1 cell apoptosis. Upper: Panc-1 cell apoptosis detected by flow cytometry. Below: apoptosis-related proteins of Panc-1 cell detected by Western blot. Cells were treated with 0.8 mmol/L WA, 500 mg/L BE, or WB, with untreated cells serving as the blank control. Statistical significance was determined via hypothesis testing, with annotations as follows: ns, not significant, P≥0.05; *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. BE, bevacizumab; WA, warfarin; WB, warfarin + bevacizumab.

To delineate the molecular mechanisms of apoptosis induction, Western blot analysis was performed to evaluate key apoptotic regulators (Figure 5, lower). Both WA and WB treatments reduced the expression of the anti-apoptotic protein Bcl-2 (WA: 0.74±0.04-fold; WB: 0.41±0.08-fold vs. blank: 1.20±0.09-fold; P<0.001) and increased levels of pro-apoptotic Bax (WA: 0.93±0.06-fold; WB: 1.12±0.08-fold vs. blank: 0.41±0.05-fold; P<0.01). A concomitant upregulation of caspase-3 was observed in both treatment groups (WA: 0.73±0.04-fold; WB: 0.93±0.09-fold vs. blank 0.13±0.01; P<0.001), though no significant interventional difference was detected (P=0.21). Quantification of Bcl-2/Bax ratios (Table 4) demonstrated the WA group exhibiting a 72.60% decrease (0.80±0.07 vs. blank 2.92±0.42; P<0.001) and the WB group showing further suppression (0.36±0.08 vs. WA group; P<0.001). These coordinated molecular alterations, which include Bcl-2 downregulation, Bax and caspase-3 upregulation, collectively supported the activation of the intrinsic mitochondrial apoptotic pathway.

Modulation of the EGFR-PI3K-AKT signaling pathway by warfarin/bevacizumab combination therapy

Western blot analysis was conducted to evaluate the effects of WA, BE, and WB on the EGFR-PI3K-Akt signaling pathway in Panc-1 cells (Figure 6). Both warfarin and bevacizumab monotherapy significantly reduced the expression levels of EGFR, PI3K, and AKT1 proteins compared to the blank group (P<0.05 to P<0.001). The combination therapy resulted in a more pronounced suppression of EGFR and PI3K expression than either monotherapy (P<0.001, P<0.001), indicating an inhibitory effect. Although AKT1 expression was also reduced in the WB group compared to control, the difference between the WB and WA groups was not statistically significant (P>0.05).

Figure 6 The expression of EGFR-PI3K-Akt pathway protein in Panc-1 cells. Cells were treated with 0.8 mmol/L WA, 500 mg/L BE, or WB, with untreated cells serving as the blank control. Statistical significance was determined via hypothesis testing, with annotations as follows: ns, not significant, P≥0.05; *, P<0.05; ***, P<0.001; ****, P<0.0001. BE, bevacizumab; WA, warfarin; WB, warfarin + bevacizumab.

Transcriptional modulation of EGFR-PI3K-AKT signaling axis by warfarin-bevacizumab combination therapy: evidence from qRT-PCR profiling

qRT-PCR analysis was performed to assess the effects of WA, BE, and WB on the transcriptional expression of EGFR, PI3K, and AKT1 in Panc-1 cells (Figure 7). All treatment groups showed significantly reduced mRNA levels of these genes compared to the blank control (P<0.001), with the combination group exhibiting the most pronounced downregulation. While no significant differences were observed between WA and BE groups (P>0.05), the WB group consistently demonstrated lower mRNA expression levels than either monotherapy.

Figure 7 The expression of EGFR-PI3K-Akt pathway mRNA in Panc-1 cell. Cells were treated with 0.8 mmol/L WA, 500 mg/L BE, or WB, with untreated cells serving as the blank control. Statistical significance was determined via hypothesis testing, with annotations as follows: ns, not significant, P≥0.05; ****, P<0.0001. BE, bevacizumab; WA, warfarin; WB, warfarin + bevacizumab.

Discussion

Recent studies have increasingly highlighted the critical role of dual-targeting strategies involving oncogenic signaling pathways and tumor microenvironmental remodeling to improve therapeutic outcomes in solid tumors (37-39). Accumulating evidence demonstrates that monotherapies frequently yield suboptimal clinical responses, a phenomenon largely attributed to compensatory crosstalk between neoplastic cells and stromal components (40). Consequently, mechanistically rational combination regimens incorporating repurposed pharmacological agents have emerged as a prioritized research focus (41,42). Of particular translational significance is the development of multi-targeted synergistic approaches, with combined anti-angiogenic therapy and blockade of core survival signaling pathways representing a strategic paradigm currently under intensive investigation for precision intervention in PDAC.

Previous investigations established that warfarin exhibited oncostatic properties through dual mechanisms: suppression of the AXL/Gas6 signaling axis and consequent disruption of PI3K-Akt-mediated survival pathways, ultimately triggering both cell cycle arrest and apoptosis across multiple malignancies (20,22). In parallel studies, pharmacological VEGFR inhibition was demonstrated to effectively attenuate tumor-associated angiogenesis and impede disease progression in PDAC (43). However, the mechanistic interplay between these therapeutic modalities and the PI3K-Akt cascade remained incompletely characterized. Furthermore, clinical-pathological analyses revealed that elevated VEGF expression exhibited a robust association with advanced tumor-node-metastasis (TNM) staging, lymph node metastasis, and hepatic dissemination in PDAC cohorts (44,45). In this study, the antitumor effects of warfarin and VEGFR-targeting agents in PDAC were systematically investigated through an integrated approach combining network pharmacology, molecular docking, and functional validation.

Target prediction and KEGG enrichment analysis jointly identified 70 intersecting genes enriched in the EGFR-PI3K-Akt axis (adjust P<0.05), with the hub proteins including EGFR, PIK3CA/CB/CD/3R1, and AKT1, all previously implicated in PDAC progression through multi-omics studies. Functionally, these hub proteins were mechanistically linked to apoptosis regulation, as EGFR (the upstream orchestrator of PI3K-Akt signaling) had been shown to sustain tumor growth by transcriptional upregulation of anti-apoptotic Bcl-2 and suppression of Bax-dependent mitochondrial permeability transition (46). Furthermore, constitutive activation of the PI3K-Akt axis had been associated with enhanced metastatic potential and acquired chemoresistance in PDAC (47). Taken together, these findings justify the focus on the EGFR-PI3K-Akt pathway as it is supported by prior biological evidence and further corroborated by our experimental validation.

Molecular docking analysis demonstrated that warfarin exhibited stable binding interactions with critical components of the EGFR-PI3K-Akt signaling axis, including EGFR, PI3K catalytic subunits (PIK3CA/CB/CD), and AKT1, with binding energies consistently below −7.0 kcal/mol. While the regulatory subunit PIK3R1 showed a moderately weaker interaction (−6.1 kcal/mol), this value remained within the range predictive of potential biological activity.

Although bioinformatics analysis identified 70 potential targets, experimental validation was focused on EGFR, PI3K, and AKT1, which were prioritized as hub nodes in the PPI network and are well recognized for their roles in PDAC progression. Other predicted targets, mainly related to cell cycle, metabolism, and immune pathways, were beyond the scope of this study but remain important for future research. By combining computational prediction with experimental assays, we validated key nodes in the EGFR-PI3K-Akt pathway, thereby substantiating the network pharmacology predictions and reinforcing the mechanistic basis of our study. Consistently, molecular docking supported that warfarin directly engages key nodes of the EGFR-PI3K-Akt signaling cascade, providing a mechanistic basis for its modulation of PDAC phenotypes.

In vitro experiments demonstrated that the combination of warfarin and bevacizumab exerted significant antitumor effects on PDAC cells. Dose-response profiling identified 0.8 mmol/L warfarin and 500 mg/L bevacizumab as the optimal regimen for combination studies. Bliss independence analysis revealed a time-dependent interaction pattern. At 24 h, the combination produced clear synergism (Bliss score =0.335), whereas at 48 h the interaction approximated additivity (Bliss score ≈0) while still achieving the sustained overall growth suppression. Notably, this temporal transition coincided with progressive modulation of the EGFR-PI3K-Akt pathway, which may reflect adaptive feedback regulation, a phenomenon also observed in other targeted therapies (48,49). Collectively, these pharmacodynamic features supported the selection of this regimen for functional validation, balancing maximal antiproliferative efficacy with biological relevance.

Similarly, wound healing assays demonstrated that while warfarin effectively impaired cell migration, bevacizumab alone had a limited effect. However, their combination further attenuated migratory capacity, although no statistical difference was observed compared to warfarin alone, indicating that the anti-migratory effect may be primarily driven by warfarin. The anti-migratory activity of warfarin was associated with alterations in Akt-mediated cytoskeletal regulation (50-52). Notably, prior studies have established that Akt inactivation suppresses Rac1-driven lamellipodia formation (53) and cofilin-mediated actin depolymerization (54). These coordinated mechanisms may collectively underlie the observed impairment of PDAC cell motility, suggesting that warfarin exerts multi-tiered interference with cytoskeletal remodeling networks.

The combined flow cytometry and western blotting analyses revealed a multi-layered pro-apoptotic effect of warfarin-based regimens. While both warfarin monotherapy and its combination with bevacizumab significantly increased total apoptosis rates compared to blank (WA: 19.35%±1.63%; WB: 23.14%±2.29% vs. 8.46%±0.48%, P<0.001), the interventional difference between WA and WB groups remained statistically non-significant (P>0.05). This was further supported by molecular synergy in the WB group, which showed amplified suppression of the Bcl-2/Bax ratio (0.36±0.08 vs. WA: 0.80±0.07, P<0.001) and caspase-3 activation (0.93±0.09-fold vs. WA: 0.73±0.04-fold). The dissociation between equivalent terminal apoptosis rates and superior mitochondrial priming in WB group might reflect threshold saturation of caspase-3 activation downstream of mitochondrial permeabilization, as previously observed in anticoagulant-potentiated apoptosis models (55). These findings indicated that bevacizumab augmented warfarin’s capacity to destabilize the Bcl-2/Bax axis.

Molecular validation via Western blotting and qRT-PCR revealed that warfarin and bevacizumab, both monotherapy and in combination, synergistically suppressed EGFR, PI3K and AKT1 expression at the transcriptional and translational levels. The combination regimen (WB) exerted the most pronounced inhibition, significantly outperforming monotherapies in suppressing EGFR (P<0.001 vs. WA/BE) and PI3K (P<0.001), consistent with the enhanced efficacy of dual receptor tyrosine kinase/intracellular kinase targeting observed in oncogenic pathway modulation (56).

Notably, while WB reduced AKT1 mRNA to the lowest numerally levels among all groups, its transcriptional suppression relative to WA alone was non-significant (P>0.05). This WA-comparable mRNA reduction, coupled with unaltered AKT1 protein levels in WB versus WA (P>0.05), suggests compensatory mechanisms preserving AKT1 signaling, potentially through enhanced mechanistic target of rapamycin (mTOR)-mediated mRNA translation efficiency or HSP90-dependent protein stabilization, as observed in kinase inhibitor-adapted malignancies (57,58).

Collectively, the findings demonstrated that the combination of warfarin and bevacizumab had a significant impact on the EGFR-PI3K-Akt pathway, primarily through transcriptional suppression and post-translational modulation. Given the critical role of this pathway in PDAC progression and treatment resistance (59,60), its multilevel suppression is likely to have underpinned the efficacy of the regimen.

Several limitations of this study should be acknowledged. First, the findings relied mainly on in vitro models and computational predictions, and lack of in vivo validation restricts the conclusions to the cellular level. Future animal experiments are planned to confirm the antitumor efficacy and to assess potential dose-related safety concerns such as bleeding. Second, the network pharmacology analysis primarily served as a hypothesis-generating approach rather than direct mechanistic evidence; although key nodes of the EGFR-PI3K-Akt axis were experimentally validated, other predicted targets were not examined and warrant further study. Third, although suppression of EGFR and PI3K was confirmed, AKT1 expression remained largely unchanged, suggesting possible compensatory mechanisms. Downstream molecules, including phosphorylated Akt (p-Akt), were not assessed in this study and will be examined in future experiments to clarify the detailed signaling response. Finally, the immunological implications of the combination therapy remain unexplored, despite the established immunomodulatory roles of VEGF in the PDAC microenvironment. Further in vivo and pharmacodynamic studies will be required to bridge these gaps and support clinical translation. In addition, future dose-optimization and exploratory preclinical studies are needed to identify a therapeutic window that balances antitumor efficacy with bleeding risk, providing a foundation for potential clinical application.

Conclusions

This study demonstrated that the combination of warfarin and bevacizumab exerted antitumor effects against PDAC by targeting multiple nodes of the EGFR-PI3K-Akt signaling axis. The dual regimen suppressed tumor cell proliferation and migration, while enhancing mitochondrial apoptosis through destabilization of the Bcl-2/Bax axis. These coordinated effects provide mechanistic support for the development of warfarin-based combination strategies aimed at overcoming oncogenic signaling plasticity in PDAC therapy.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1190/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1190/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1190/prf

Funding: The study was supported by Medical Quality (Evidence-Based) Management Research Project of the National Institute of Hospital Administration, National Health Commission of the People’s Republic of China (grant No. YLZLXZ24G112), and Natural Science Foundation of Fujian, China (grant No. 2022J011011).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1190/coif). The authors have no 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.

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. Alese OB, Jiang R, Shaib W, et al. Young Adults With Pancreatic Cancer: National Trends in Treatment and Outcomes. Pancreas 2020;49:341-54. [Crossref] [PubMed]
2. Mehra S, Deshpande N, Nagathihalli N. Targeting PI3K Pathway in Pancreatic Ductal Adenocarcinoma: Rationale and Progress. Cancers (Basel) 2021;13:4434. [Crossref] [PubMed]
3. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12-49. [Crossref] [PubMed]
4. Rishi A, Goggins M, Wood LD, et al. Pathological and molecular evaluation of pancreatic neoplasms. Semin Oncol 2015;42:28-39. [Crossref] [PubMed]
5. Apte MV, Xu Z, Pothula S, et al. Pancreatic cancer: The microenvironment needs attention too! Pancreatology 2015;15:S32-8. [Crossref] [PubMed]
6. Biankin AV, Waddell N, Kassahn KS, et al. Pancreatic cancer genomes reveal aberrations in axon guidance pathway genes. Nature 2012;491:399-405. [Crossref] [PubMed]
7. Ryan DP, Hong TS, Bardeesy N. Pancreatic adenocarcinoma. N Engl J Med 2014;371:2140-1. [Crossref] [PubMed]
8. Garrido-Laguna I, Hidalgo M. Pancreatic cancer: from state-of-the-art treatments to promising novel therapies. Nat Rev Clin Oncol 2015;12:319-34. [Crossref] [PubMed]
9. Zheng R, Liu X, Zhang Y, et al. Frontiers and future of immunotherapy for pancreatic cancer: from molecular mechanisms to clinical application. Front Immunol 2024;15:1383978. [Crossref] [PubMed]
10. Farhangnia P, Khorramdelazad H, Nickho H, et al. Current and future immunotherapeutic approaches in pancreatic cancer treatment. J Hematol Oncol 2024;17:40. [Crossref] [PubMed]
11. Wang J, Zhu C. Anticoagulation in combination with antiangiogenesis and chemotherapy for cancer patients: evidence and hypothesis. Onco Targets Ther 2016;9:4737-46. [Crossref] [PubMed]
12. Guryanov I, Tennikova T, Urtti A. Peptide Inhibitors of Vascular Endothelial Growth Factor A: Current Situation and Perspectives. Pharmaceutics 2021;13:1337. [Crossref] [PubMed]
13. Balsano R, Tommasi C, Garajova I. State of the Art for Metastatic Pancreatic Cancer Treatment: Where Are We Now? Anticancer Res 2019;39:3405-12. [Crossref] [PubMed]
14. Luo J, Guo P, Matsuda K, et al. Pancreatic cancer cell-derived vascular endothelial growth factor is biologically active in vitro and enhances tumorigenicity in vivo. Int J Cancer 2001;92:361-9. [Crossref] [PubMed]
15. Kindler HL, Friberg G, Singh DA, et al. Phase II trial of bevacizumab plus gemcitabine in patients with advanced pancreatic cancer. J Clin Oncol 2005;23:8033-40. [Crossref] [PubMed]
16. Kindler HL, Niedzwiecki D, Hollis D, et al. Gemcitabine plus bevacizumab compared with gemcitabine plus placebo in patients with advanced pancreatic cancer: phase III trial of the Cancer and Leukemia Group B (CALGB 80303). J Clin Oncol 2010;28:3617-22. [Crossref] [PubMed]
17. Liu X, Zhang J, Yi T, et al. Decoding tumor angiogenesis: pathways, mechanisms, and future directions in anti-cancer strategies. Biomark Res 2025;13:62. [Crossref] [PubMed]
18. Liu ZL, Chen HH, Zheng LL, et al. Angiogenic signaling pathways and anti-angiogenic therapy for cancer. Signal Transduct Target Ther 2023;8:198. [Crossref] [PubMed]
19. Cohen O, Caiano LM, Tufano A, et al. Cancer-Associated Splanchnic Vein Thrombosis. Semin Thromb Hemost 2021;47:931-41. [Crossref] [PubMed]
20. Kirane A, Ludwig KF, Sorrelle N, et al. Warfarin Blocks Gas6-Mediated Axl Activation Required for Pancreatic Cancer Epithelial Plasticity and Metastasis. Cancer Res 2015;75:3699-705. [Crossref] [PubMed]
21. Paolino M, Choidas A, Wallner S, et al. The E3 ligase Cbl-b and TAM receptors regulate cancer metastasis via natural killer cells. Nature 2014;507:508-12. [Crossref] [PubMed]
22. Yang L, Wang L, Huang X, et al. Warfarin affects the proliferation and apoptosis of lung cancer cells by regulating the Gas6/Axl/PI3K/Akt/NF-κB pathway. Journal of Chinese Pharmaceutical Sciences 2023;32:190-9. [Crossref]
23. Ju Y, Xu D, Liao MM, et al. Barriers and opportunities in pancreatic cancer immunotherapy. NPJ Precis Oncol 2024;8:199. [Crossref] [PubMed]
24. Amberger JS, Bocchini CA, Schiettecatte F, et al. OMIM.org: Online Mendelian Inheritance in Man (OMIM®), an online catalog of human genes and genetic disorders. Nucleic Acids Res 2015;43:D789-98. [Crossref] [PubMed]
25. Stelzer G, Rosen N, Plaschkes I, et al. The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses. Curr Protoc Bioinformatics 2016;54:1.30.1-1.30.33.
26. Wishart DS, Feunang YD, Guo AC, et al. DrugBank 5.0: a major update to the DrugBank database for 2018. Nucleic Acids Res 2018;46:D1074-82. [Crossref] [PubMed]
27. 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]
28. Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 2019;47:D607-13. [Crossref] [PubMed]
29. Shannon P, Markiel A, Ozier O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 2003;13:2498-504. [Crossref] [PubMed]
30. Kim S, Chen J, Cheng T, et al. PubChem in 2021: new data content and improved web interfaces. Nucleic Acids Res 2021;49:D1388-95. [Crossref] [PubMed]
31. O'Boyle NM, Banck M, James CA, et al. Open Babel: An open chemical toolbox. J Cheminform 2011;3:33. [Crossref] [PubMed]
32. Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature 2021;596:583-9. [Crossref] [PubMed]
33. Burley SK, Bhikadiya C, Bi C, et al. RCSB Protein Data Bank: powerful new tools for exploring 3D structures of biological macromolecules for basic and applied research and education in fundamental biology, biomedicine, biotechnology, bioengineering and energy sciences. Nucleic Acids Res 2021;49:D437-51. [Crossref] [PubMed]
34. DeLano WL. The PyMOL molecular graphics system. 2002. Available online: https://www.pymol.org/
35. Morris GM, Huey R, Lindstrom W, et al. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J Comput Chem 2009;30:2785-91. [Crossref] [PubMed]
36. BLISS CI. The calculation of microbial assays. Bacteriol Rev 1956;20:243-58. [Crossref] [PubMed]
37. Liu D, Che X, Wu G. Deciphering the role of neddylation in tumor microenvironment modulation: common outcome of multiple signaling pathways. Biomark Res 2024;12:5. [Crossref] [PubMed]
38. Gunes M, Rosen ST, Shachar I, et al. Signaling lymphocytic activation molecule family receptors as potential immune therapeutic targets in solid tumors. Front Immunol 2024;15:1297473. [Crossref] [PubMed]
39. Xu Y, Miller CP, Tykodi SS, et al. Signaling crosstalk between tumor endothelial cells and immune cells in the microenvironment of solid tumors. Front Cell Dev Biol 2024;12:1387198. [Crossref] [PubMed]
40. Bina D, Tatiana M, Daria M, et al. Abstract 552: Stromal facilitated multifactorial resistance to tumor cells against targeted therapies in ALK+ NSCLC. Cancer Res 2023;83:552. [Crossref]
41. Vafaei S, Zekiy AO, Khanamir RA, et al. Combination therapy with immune checkpoint inhibitors (ICIs); a new frontier. Cancer Cell Int 2022;22:2. [Crossref] [PubMed]
42. Ravensbergen CJ, Polack M, Roelands J, et al. Combined Assessment of the Tumor-Stroma Ratio and Tumor Immune Cell Infiltrate for Immune Checkpoint Inhibitor Therapy Response Prediction in Colon Cancer. Cells 2021;10:2935. [Crossref] [PubMed]
43. Nakashio A, Fujita N, Tsuruo T. Topotecan inhibits VEGF- and bFGF-induced vascular endothelial cell migration via downregulation of the PI3K-Akt signaling pathway. Int J Cancer 2002;98:36-41. [Crossref] [PubMed]
44. Seo Y, Baba H, Fukuda T, et al. High expression of vascular endothelial growth factor is associated with liver metastasis and a poor prognosis for patients with ductal pancreatic adenocarcinoma. Cancer 2000;88:2239-45. [Crossref] [PubMed]
45. Liang QL, Wang BR, Chen GQ, et al. Clinical significance of vascular endothelial growth factor and connexin43 for predicting pancreatic cancer clinicopathologic parameters. Med Oncol 2010;27:1164-70. [Crossref] [PubMed]
46. Suh DS, Park SE, Jin H, et al. LRIG2 is a growth suppressor of Hec-1A and Ishikawa endometrial adenocarcinoma cells by regulating PI3K/AKT- and EGFR-mediated apoptosis and cell-cycle. Oncogenesis 2018;7:3. [Crossref] [PubMed]
47. Mohite R, Doshi G. Elucidation of the Role of the Epigenetic Regulatory Mechanisms of PI3K/Akt/mTOR Signaling Pathway in Human Malignancies. Curr Cancer Drug Targets 2024;24:231-44. [Crossref] [PubMed]
48. Niessner H, Forschner A, Klumpp B, et al. Targeting hyperactivation of the AKT survival pathway to overcome therapy resistance of melanoma brain metastases. Cancer Med 2013;2:76-85. [Crossref] [PubMed]
49. Yu L, Wei J, Liu P. Attacking the PI3K/Akt/mTOR signaling pathway for targeted therapeutic treatment in human cancer. Semin Cancer Biol 2022;85:69-94. [Crossref] [PubMed]
50. Wang X, Huang M, Xie W, et al. Eupafolin regulates non-small-cell lung cancer cell proliferation,migration, and invasion by suppressing MMP9 and RhoA via FAK/PI3K/AKT signaling pathway. J Biosci 2023;48:1. [Crossref] [PubMed]
51. Huang CY, Fong YC, Lee CY, et al. CCL5 increases lung cancer migration via PI3K, Akt and NF-kappaB pathways. Biochem Pharmacol 2009;77:794-803. [Crossref] [PubMed]
52. Wei R, Penso NEC, Hackman RM, et al. Epigallocatechin-3-Gallate (EGCG) Suppresses Pancreatic Cancer Cell Growth, Invasion, and Migration partly through the Inhibition of Akt Pathway and Epithelial-Mesenchymal Transition: Enhanced Efficacy when Combined with Gemcitabine. Nutrients 2019;11:1856. [Crossref] [PubMed]
53. Gautam S, Ishrat N, Singh R, et al. Aegeline from Aegle marmelos stimulates glucose transport via Akt and Rac1 signaling, and contributes to a cytoskeletal rearrangement through PI3K/Rac1. Eur J Pharmacol 2015;762:419-29. [Crossref] [PubMed]
54. Xie H, Zhang C, Liu D, et al. Erythropoietin protects the inner blood-retinal barrier by inhibiting microglia phagocytosis via Src/Akt/cofilin signalling in experimental diabetic retinopathy. Diabetologia 2021;64:211-25. [Crossref] [PubMed]
55. Liu MD, Xiong SJ, Tan F, et al. Physcion 8-O-β-glucopyranoside induces mitochondria-dependent apoptosis of human oral squamous cell carcinoma cells via suppressing survivin expression. Acta Pharmacol Sin 2016;37:687-97. [Crossref] [PubMed]
56. Yu Y, Xiao Z, Lei C, et al. BYL719 reverses gefitinib-resistance induced by PI3K/AKT activation in non-small cell lung cancer cells. BMC Cancer 2023;23:732. [Crossref] [PubMed]
57. Codenotti S, Zizioli D, Mignani L, et al. Hyperactive Akt1 Signaling Increases Tumor Progression and DNA Repair in Embryonal Rhabdomyosarcoma RD Line and Confers Susceptibility to Glycolysis and Mevalonate Pathway Inhibitors. Cells 2022;11:2859. [Crossref] [PubMed]
58. Tsimberidou AM, Skliris A, Valentine A, et al. AKT inhibition in the central nervous system induces signaling defects resulting in psychiatric symptomatology. Cell Biosci 2022;12:56. [Crossref] [PubMed]
59. Zhou Q, Xu CR, Cheng Y, et al. Bevacizumab plus erlotinib in Chinese patients with untreated, EGFR-mutated, advanced NSCLC (ARTEMIS-CTONG1509): A multicenter phase 3 study. Cancer Cell 2021;39:1279-1291.e3. [Crossref] [PubMed]
60. Wu Y, Zhang F, Xu P, et al. Brucine Inhibits Proliferation of Pancreatic Ductal Adenocarcinoma through PI3K/AKT Pathway-induced Mitochondrial Apoptosis. Curr Cancer Drug Targets 2024;24:749-59. [Crossref] [PubMed]