Resveratrol (purity ≥98.0%) was purchased from NatureStandard (Shanghai Standard Technology Co., Ltd.; cat. no. ST00670120), prepared as a 100-mM stock solution with DMSO (Beyotime Institute of Biotechnology; cat. no. ST038) and diluted to different concentrations according to the experimental requirements. Osimertinib, purchased from Selleck Chemicals (cat. no. S7297), was reconstituted in DMSO. The solution was prepared to a concentration of 200 mM and stored at −20°C to preserve its stability.

The potential targets of resveratrol were identified via searches of the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database (https://www.tcmspw.com/tcmsp.php) (23). The disease-related genes were obtained from the GeneCards (http://www.genecards.org) database (24), with the search term ‘lung cancer osimertinib resistance’. The gene names were standardized via the UniProtKB database (https://www.uniprot.org/) (25). Subsequently, the targets associated with both resveratrol and osimertinib-resistant lung cancer were obtained.

The PPI network of the overlapping targets was established via the STRING database (https://string–db.org/) (26). With ‘Homo sapiens’ as the chosen species and a confidence score of ≥0.4, the PPI network was visualized via Cytoscape (version 3.7.2) (27). The PPI network was subsequently analyzed and the hub targets with degrees above the median were filtered.

To further determine the mechanism by which resveratrol delays osimertinib resistance in lung cancer, the biological functions and potential pathways involved were investigated. The Metascape database (www.metascape.org/) (28) was employed to perform GO and KEGG analyses on the hub targets identified from the PPI network. The results were visualized in the form of a bubble chart via the ggplot2 package (version 3.5.1) (29). The network of resveratrol, the common targets and the pathways were subsequently mapped via Cytoscape.

The 3D model of resveratrol was retrieved from the PubChem database (30) and Chem3D software (version 20.1; library.bath.ac.uk/chemistry-software/chem3d) was utilized to minimize the energy of the structure. The protein structures of TP53 (4AGQ) (31), STAT3 (6NJS) (32), IGF1R (1P4O) (33), MCL1 (6OQC) (34) and BCL2L11 (1PQ1) (35) were subsequently extracted from the Protein Data Bank repository (www.rcsb.org) (36) and converted into the pdbqt format. Discovery Studio 2016 (Dassault Systèmes S.E.) was then employed for the processes of dehydration and hydrogenation, the modification of amino acid residues and energy minimization. The CDOCKER protocol within the software was applied to estimate the binding affinity between resveratrol and the target proteins. Finally, PyMOL (version 2.2.0; pymol.org/2/) was selected to visualize the molecular interactions.

PC9 cells, initially procured from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences, were cultivated in RPMI-1640 medium (HyClone; Cytiva; cat. no. SH30809.01) enriched with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.; cat. no. 10091148) and 1% penicillin-streptomycin (HyClone; Cytiva; cat. no. SV30010) at 37°C in a humidified incubator with 5% CO₂ (Thermo Fisher Scientific, Inc.). Subsequently, osimertinib-resistant PC9 cells (PC9OR) were developed by gradually increasing the osimertinib concentration from 5 to 3,000 nM over a period of 5 months, as previously described (37). PC9OR cells were treated with osimertinib at a concentration of 3,000 nM to sustain the drug resistance.

The Cell Counting Kit-8 (CCK-8) assay (cat. no. C6005; NCM Biotech) was utilized to evaluate the viability of PC9OR cells after treatment with resveratrol. PC9OR cells were seeded in 96-well plates and exposed to different concentrations of resveratrol (0, 10, 20, 40, 80 and 160 µM) for 48 h. A CCK-8 solution (10 µl/well) was subsequently added, and the cells were incubated for 2 h, after which the absorbance at 450 nm was measured with a spectrophotometer.

A total of 1,000 PC9OR cells were cultured in 6-well plates and then treated with 60 µM resveratrol for 2 weeks. Once visible colonies had formed (>50 cells per colony) (38), the plates were fixed with 4% paraformaldehyde (cat. no. BL539A; Biosharp Life Sciences) at room temperature for 30 min and stained with a 0.1% crystal violet solution (cat. no. C0121; Beyotime Institute of Biotechnology) for 15 min at room temperature. After staining, the plates were washed with PBS, dried and imaged, and then cells were counted using ImageJ (v1.8.0; National Institutes of Health).

PC9OR cells were plated in 6-well plates at a density of 5×10³ cells per well. After the cells had attached, 60 µM resveratrol was added for treatment for 48 h. Total RNA was extracted from PC9OR cells with the EZ-press RNA Purification Kit (EZBioscience; cat. no. B0004DP). cDNA was reverse transcribed using the PrimeScript^™ RT reagent Kit (Takara Bio, Inc.; cat. no. RR037A). The reverse transcription conditions were as follows: 37°C for 15 min, then 85°C for 5 sec and finally cooling to 4°C. For qPCR, PerfectStart^® Green qPCR SuperMix (TransGen Biotech Co., Ltd.; cat. no. AQ601-02) was utilized. The qPCR thermal cycling parameters were set as follows: Preliminary denaturation at 95°C for 30 sec, followed by 40 cycles of 95°C for 15 sec and 60°C for 60 sec. The quantification of mRNA levels was performed using the 2^−ΔΔCq method (39), and the results were further normalized against the expression levels of GAPDH. The specific sequences of the primers used were as follows: BCL2L11 forward, 5′-CTTGTGACCCCAGCCAT-3′ and reverse, 5′-CTTCCTTCCACTGCACTAA-3′; Bcl-2 forward, 5′-GATTGTGGCCTTCTTTGAG-3′ and reverse, 5′-GTTCCACAAAGGCATCC-3′; Bax forward, 5′-CCTGTGCACCAAGGTGCCGGAACT-3′ and reverse, 5′-CCACCCTGGTCTTGGATCCAGCCC-3′; and GAPDH forward, 5′-GGAAGCTTGTCATCAATGGAAATC-3′ and reverse, 5′-TGATGACCCTTTTGGCTCCC-3′. All primers were synthesized by Sangon Biotech Co., Ltd.

PC9OR cells were treated with 60 µM resveratrol for 48 h and then lysed in ice-cold RIPA lysis buffer (Beyotime Institute of Biotechnology; cat. no. P0013B). The protein concentrations within the lysates were measured with a BCA Protein Assay Kit (Epizyme; Ipsen Pharma; cat. no. ZJ101). Proteins (20 µg/lane) were resolved through 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then electrophoretically transferred onto polyvinylidene fluoride membranes. The membranes were blocked with a 5% milk solution at room temperature for 1 h, then incubated with primary antibodies (1:5,000) at 4°C overnight. Afterwards, the membranes were incubated with a 1:5,000 dilution of HRP-conjugated goat anti-rabbit IgG secondary antibody (Proteintech Group, Inc.; cat. no. SA00001-2) for 1 h at 37°C and visualized using an enhanced chemiluminescence solution (New Cell Molecular Biotech; cat. no. P10200). The analysis of the bands was conducted with ImageJ (v1.8.0; National Institutes of Health). The BCL2L11 antibody (cat. no. ET1608-14) was obtained from HUABIO, and the β-actin antibody (cat. no. 20536-1-AP) was purchased from Proteintech Group, Inc.

All data are displayed as the mean ± SD derived from a minimum of three separate experimental trials. GraphPad Prism 8.0 software (Dotmatics) was utilized for both plotting graphs and statistical analysis. All the results were calculated using the unpaired t-test. P<0.05 was considered to indicate a statistically significant difference.

A flowchart illustrating the design of the present study is depicted in Fig. 1. A total of 151 target genes were initially obtained from the TCMSP database and 150 genes were retained from the UniProt database with standardized gene symbols. Specifically, the standardized gene symbols for the target genes PKA catalytic subunit C-α and protein kinase C α type were consistent, both being PRKCA. Simultaneously, 131 disease-related genes were identified from the GeneCards database. A total of 13 overlapping targets (TP53, BRCA2, IGF1R, STAT3, ABCG2, HGF, MCL1, CFLAR, BCL2L11, TNFRSF10A, TGFB2, SREBF1 and CDK7) were obtained when the resveratrol targets were compared with the disease-related targets. A PPI network was then built to identify the key targets (Fig. 2). The 5 targets with degree values >5 were screened as the key therapeutic targets, from 1 to 5 in the order of TP53, STAT3, IGF1R, MCL1 and BCL2L11.

To explore the potential mechanisms by which resveratrol delays osimertinib resistance, GO and KEGG analyses of the 13 genes derived from the PPI network were conducted (Fig. 3). The identified biological process terms included ‘apoptotic signaling pathway’ and ‘cellular response to peptide’. The molecular function terms included ‘protease binding’, ‘protein kinase binding’ and ‘kinase binding’. The cellular component terms included ‘membrane raft’ and ‘membrane microdomain’. Moreover, the top 20 pathways were related to KEGG pathways including ‘Pathways in cancer’, ‘Apoptosis’ and ‘EGFR tyrosine kinase inhibitor resistance’ (Fig. 3B). Consequently, the apoptosis pathway was selected as a pivotal focus for subsequent investigation on the basis of this pathway enrichment.

To visualize and elucidate the pharmacological potential of resveratrol for modulating osimertinib resistance, Cytoscape was utilized to construct a network of resveratrol-common target genes and KEGG pathways with 35 nodes and 93 edges (Fig. 4). Among the core target genes, TP53 (19 edges), STAT3 (13 edges), IGF1R (10 edges), MCL1 (4 edges), BCL2L11 (9 edges) and the apoptosis pathway (5 edges) had a high degree of connectivity, suggesting their involvement in the development of osimertinib resistance.

Molecular docking simulations were performed with the 5 identified targets (TP53, STAT3, IGF1R, MCL1 and BCL2L11) and resveratrol. It was found that resveratrol was predicted to bind to TP53, STAT3, IGF1R, MCL1 and BCL2L11 with energies of −5.9, −6.0, −6.5, −7.9 and −7.3 kJ/mol, respectively (Fig. 5). The smaller the binding energy is, the better the binding activity between the ligand and receptor. When the binding energy is <-7 kcal/mol, this signifies a strong binding affinity (40). Therefore, resveratrol may have strong binding affinities with MCL1 and BCL2L11, indicating that the mechanism by which resveratrol delays osimertinib resistance may be closely linked to binding to these targets. Furthermore, the distribution of the targets of resveratrol in the apoptosis pathway was mapped, and it was found that resveratrol can regulate the apoptosis pathway by targeting MCL1 and BCL2L11 (Fig. 6).

To verify the results of the network pharmacological analysis, in vitro experiments were conducted. Various concentrations of resveratrol (0–160 µM) were used to treat PC9OR cells for 48 h and cell viability was determined via a CCK-8 assay. The results showed that resveratrol significantly reduced the cell viability of PC9OR cells, with an IC₅₀ of ~60 µM (Fig. 7A). Furthermore, resveratrol inhibited the long-term proliferation of PC9OR cells (Fig. 7B) and led to a notable alteration in the morphology of PC9OR cells, resulting in a flattened appearance (Fig. 7C).

RT-qPCR was performed to confirm whether resveratrol delays osimertinib resistance through the targets identified via network pharmacology (MCL1 and BCL2L11). Resveratrol increased the mRNA expression levels of BCL2L11 in PC9OR cells (P<0.001) (Fig. 7D). Notably, resveratrol decreased the expression levels of MCL1, but there was no significant difference (P>0.05). Concurrently, resveratrol reduced the mRNA expression levels of Bcl-2 and increased the expression levels of Bax (P<0.001) (Fig. S1). Furthermore, resveratrol increased the protein expression levels of BCL2L11 in PC9OR cells (Fig. 7E). These findings demonstrated that resveratrol might promote apoptosis to delay osimertinib resistance.