Computational Study Maps Resveratrol’s Potential Protein Targets in Lung Squamous Cell Carcinoma

Resveratrol, the polyphenol found in grapes, mulberries and peanuts, has long attracted attention for its anticancer properties, but its mechanisms against lung squamous cell carcinoma (LUSC), the subtype that accounts for a large share of non-small cell lung cancer cases, remained poorly mapped. A new computational study published in the Egyptian Journal of Medical Human Genetics has now used network bioinformatics and molecular docking to identify which proteins resveratrol might plausibly engage in this cancer type, singling out TP53, AKT1, IL6 and CCND1 as targets of particular clinical interest.

The research team, led by Dini Maharani together with Jonathan García-Román and Adam Hermawan, combined data mining from several public databases rather than conducting laboratory experiments. They queried SwissTargetPrediction, TargetNet and PubChem with resveratrol’s chemical structure to compile predicted protein targets, gathering 100 genes from the first database, 618 from the second and 5,861 from the third. After removing duplicates, this yielded 6,212 candidate resveratrol targets. These were intersected with 1,070 genes associated with lung squamous carcinoma drawn from the GeneCards database, producing a set of 429 overlapping genes the authors named Resveratrol and Lung Squamous Cell Carcinoma Related Genes, or RLRG.

Functional enrichment analysis of these 429 genes, performed with the DAVID tool at a significance cutoff of p < 0.05, revealed distinct patterns across the three gene ontology categories. Positive regulation of transcription by RNA polymerase II dominated the biological process results, the cytoplasmic side of the plasma membrane was the most enriched cellular component, and MDM2/MDM4 family protein binding topped the molecular function list. Kyoto Encyclopedia of Genes and Genomes pathway analysis showed that the overlapping genes also appear in several other cancers, including hepatocellular, pancreatic, chronic myeloid leukemia, colorectal and prostate cancer, and highlighted EGFR inhibitor pathways among the enriched results.

To find the most connected proteins within this network, the researchers built a protein-protein interaction network using the STRING database at a high confidence threshold above 0.7, then visualized and ranked it in Cytoscape with the CytoHubba plugin using two algorithms: degree centrality and maximal clique centrality. The network showed an average local clustering coefficient of 0.482, meaning roughly 48.2 percent of the proteins interacted with one another on average, and was statistically significant at an enrichment p-value below 0.05. Six proteins appeared in both ranking algorithms: TP53, CTNNB1, STAT3, AKT1, EGFR and IL6. Four others, MMP9, JUN, CASP3 and EGF, emerged uniquely from the maximal clique centrality analysis. The team chose to advance the ten highest-ranked proteins by degree for downstream analyses, citing the direct interpretability of degree centrality as a measure of a protein’s immediate connections.

Expression analysis added a clinical dimension. Using TCGA data through UALCAN and protein staining data from the Human Protein Atlas, the authors found that six genes, TP53, EGFR, SRC, CTNNB1, AKT1 and CCND1, showed concordant upregulation at both the mRNA and protein levels in tumors compared with normal tissue. PTEN, a tumor suppressor, was concordantly downregulated, consistent with its protective function. IL6 and STAT3 displayed a mismatch: their mRNA levels were reduced in tumors while protein staining intensity appeared increased, a discrepancy the authors attribute to possible post-transcriptional or post-translational regulation. EP300 showed no meaningful mRNA difference between normal and tumor groups yet higher protein levels in the Human Protein Atlas.

Genetic alteration analysis through cBioPortal, drawing on six LUSC studies encompassing 1,849 samples from 1,848 patients, showed TP53 as by far the most frequently altered gene, mutated in 69 percent of cases in the CPTAC Cell 2021 cohort, with missense and truncating mutations dominating. The authors note this aligns with prior reports placing TP53 mutation frequency at about 82.3 percent in LUSC. CCND1 was altered in roughly 13 percent of the population, mainly through amplification, while AKT1 mutations occurred in 4 percent and IL6 alterations in 2.8 percent. Mutual exclusivity analysis revealed tendencies toward co-alteration among EGFR, IL6 and STAT3, suggesting synergistic effects, whereas PTEN and CCND1 alterations were mutually exclusive, implying that alteration of either alone may suffice to drive the same cellular process.

Prognosis analysis used the Kaplan-Meier Plotter database, restricted to the squamous cell cancer histology cohort, with 527 patients in the mRNA gene chip analysis. Of the ten hub genes examined, high expression of TP53, AKT1, CCND1 and IL6 showed a trend toward worse overall survival, with hazard ratios above 1, but only CCND1 reached statistical significance. Conversely, four genes showed statistically significant associations in the opposite direction: high expression of CTNNB1 (HR = 0.82; 95% CI: 0.68–0.98), STAT3 (HR = 0.77; 95% CI: 0.61–0.96), SRC (HR = 0.82; 95% CI: 0.68–0.99) and EP300 (HR = 0.81; 95% CI: 0.67–0.99) was associated with better prognosis. The authors emphasize that the hazard ratios for TP53, AKT1 and IL6 were not statistically significant and require experimental confirmation.

Molecular docking formed the study’s predictive centerpiece. The team docked resveratrol against the four proteins with hazard ratios above 1, using crystal structures from the Protein Data Bank and a protocol validated by redocking each protein’s native ligand with a root mean square deviation below 2 angstroms. For AKT1, resveratrol achieved a docking score of −12.57 kcal/mol versus −7.39 kcal/mol for the native ligand, L(+)-tartaric acid, though the interaction involved only a single amino acid contact at AspA439. For CCND1, resveratrol scored −9.54 versus −6.17 kcal/mol for glycerol, the bound native molecule, with two main amino acid interactions; the authors caution that because glycerol is not a known inhibitor, the inhibitory potency against CCND1 remains unclear. For IL6, resveratrol outscored the native ligand but interacted only at the protein surface, suggesting a weak interaction. For TP53, docking was performed against 53BP1, an interacting partner that binds TP53’s Tudor domain, where resveratrol did not outperform the native ligand, 4-(4-methylpiperazine-1-sulfonyl)benzamide, though both engaged similar residues, Asp1521 and Tyr1552.

Taken together, the results suggest that resveratrol could theoretically modulate LUSC through the PI3K/AKT pathway, cell proliferation regulation via CCND1, and inflammatory signaling via IL6. The findings complement earlier work indicating that resveratrol acts on LUSC through ferroptosis, inhibiting HMMR expression while activating ferroptotic signaling. Prior experimental studies cited by the authors support plausibility: resveratrol’s inhibitory effect disappeared in AKT1-knockdown cells, the compound has been reported to downregulate CCND1 and cyclin D1, and it counteracted IL-6-induced cell migration in ovarian cancer cells. The authors stress, however, that docking scores do not account for entropic contributions, solvation effects or protein flexibility, and that the docking results are projections of binding affinity rather than demonstrated inhibition.

The study’s limitations are considerable and openly acknowledged. It relied entirely on in silico methods without in vitro or in vivo validation, used a single database source for disease-related genes, and employed static crystallized protein structures that ignore conformational flexibility in physiological conditions. Resveratrol’s well-known bioavailability limitations and the therapeutic concentrations needed in patients were not modeled. The authors call for in vitro binding assays, cellular functional assays, molecular dynamics simulations and eventually in vivo studies using commercially available resveratrol within a therapeutic range to confirm whether any of the predicted interactions translate into real anticancer activity against lung squamous cell carcinoma. Until then, the four hub proteins stand as computational leads for future experimental validation rather than established therapeutic targets.

Subject of Research: Bioinformatics and molecular docking analysis of resveratrol's potential protein targets in lung squamous cell carcinoma

Article Title: Network-based discovery of resveratrol’s key regulators in lung squamous cell carcinoma

Article References: Maharani, D., García-Román, J., & Hermawan, A. (2026). Network-based discovery of resveratrol’s key regulators in lung squamous cell carcinoma. Egyptian Journal of Medical Human Genetics, 27(1), Article 80. https://doi.org/10.1186/s43042-026-00906-9

Image Credits: AI Generated

DOI: 10.1186/s43042-026-00906-9

Keywords: resveratrol, lung squamous cell carcinoma, molecular docking, bioinformatics, hub genes, TP53, AKT1, IL6, CCND1, protein-protein interaction network, TCGA, drug discovery