The study, led by Tshwarelo E. Mohale and colleagues, focused on the interplay between docetaxel, a widely used semisynthetic chemotherapy derived from the needles of the European yew, and kaempferol, a natural flavonoid with a growing reputation as an anticancer agent. The team’s central question was whether these two molecules could each bind to and disrupt Bcl-2, an anti-apoptotic protein that acts as a molecular bodyguard for cancer cells. Bcl-2 works by forming heterodimers with pro-apoptotic proteins such as Bax and BAK, blocking mitochondrial outer membrane permeabilization and preventing the release of cytochrome C that would otherwise activate the caspase cascade leading to programmed cell death. Because Bcl-2 is overexpressed in many breast cancer subtypes and is strongly linked to therapy resistance, it has long been considered an attractive drug target.

Docetaxel’s problem is well documented. It kills cancer cells by inducing mitotic arrest followed by apoptosis, but tumours frequently counter by overproducing Bcl-2, which blunts the death signal. Docetaxel is also a poor drug by the standards of oral bioavailability: it violates two of Lipinski’s rule of five criteria, with a molecular weight of 807.88 g/mol and fourteen hydrogen bond acceptors, and it carries a topological polar surface area of 224.45 square angstroms, which explains its low gastrointestinal absorption. Worse, it is a substrate of P-glycoprotein, an efflux pump expressed in the liver and gut lining that actively pumps the drug back out of cells, and it is metabolically inactivated by the CYP3A4 enzyme. Kaempferol, by contrast, is a model of drug-likeness: a molecular weight of just 286.24 g/mol, a single rotatable bond, zero Lipinski violations, a bioavailability score of 0.55 compared with docetaxel’s 0.17, and high predicted gastrointestinal absorption. Crucially, it is not a P-glycoprotein substrate.

To test whether both molecules could engage Bcl-2, the researchers first prepared the human Bcl-2 crystal structure from the Protein Data Bank, removing the co-crystallized ligand to leave an empty receptor. They then performed blind molecular docking using AutoDock Vina, enclosing the entire protein in the search grid so that the algorithm could find the most favourable binding site without bias. The results were striking: docetaxel docked with a score of −7.3 kcal/mol and kaempferol with −7.7 kcal/mol, both outperforming the FDA-approved Bcl-2 inhibitor venetoclax, which scores −6.793 kcal/mol in the same framework, and the potent inhibitor paclitaxel at −7.2 kcal/mol. These scores suggested that both compounds deserve serious consideration as Bcl-2 modulators.

Docking scores alone can mislead, so the team subjected each complex to 150-nanosecond molecular dynamics simulations in explicit water using the Amber20 software suite with the ff14SB protein force field and the General Amber Force Field for the ligands. The simulations tracked how the protein-ligand complexes behaved at body temperature over time. Root mean square deviation analysis showed that the unbound Bcl-2 protein averaged 2.89 angstroms of deviation from its starting structure, while the docetaxel-bound complex averaged 3.44 angstroms and the kaempferol-bound complex 2.73 angstroms. In other words, docetaxel loosened the protein’s architecture, increasing its flexibility by 0.55 angstroms relative to the unbound state, whereas kaempferol actually produced a slightly more compact conformation than the protein achieved on its own.

Root mean square fluctuation analysis told a complementary story at the level of individual amino acids. The docetaxel complex showed elevated fluctuations across many regions, indicating that the drug destabilizes the residues that maintain Bcl-2’s function, particularly within the conserved hydrophobic cleft where the protein normally sequesters BH3-only pro-apoptotic proteins. Disrupting this cleft could weaken Bcl-2’s grip on its targets and restore apoptotic signalling. The kaempferol complex, in contrast, remained comparatively rigid. Radius of gyration measurements, which ranged between 14.49 and 14.63 angstroms across all systems, confirmed modest structural rearrangements upon ligand binding, and dynamic cross-correlation matrix analysis revealed that docetaxel increased positively correlated, synchronized motions in the BH3 region spanning residues 40 to 110, while kaempferil restricted movement in ways that preserved conformational stability.

To quantify binding strength more rigorously than docking allows, the researchers applied the Molecular Mechanics Poisson-Boltzmann Surface Area method to the simulation trajectories. The docetaxel-Bcl-2 complex yielded a binding free energy of −28.90 ± 11.93 kcal/mol, stronger than the kaempferol complex at −23.63 ± 4.87 kcal/mol. In both cases van der Waals interactions dominated the binding energy, with docetaxel contributing −37.44 kcal/mol and kaempferol −29.01 kcal/mol, confirming that hydrophobic contacts within the BH3-binding cleft are the principal driver of recognition. Per-residue energy decomposition identified Tyr114, Ala65 and Glu69 as the biggest contributors in the docetaxel complex, while Arg80, Gly79 and Asp37 anchored kaempferol. Notably, the highest-contributing residues in each complex also formed direct hydrogen bonds with their ligands, stabilizing both molecules inside the same binding pocket.

The pharmacokinetic analysis added a provocative twist. Kaempferol was predicted to inhibit CYP3A4 and CYP1A2, the very enzymes responsible for metabolizing docetaxel. In theory, this could raise systemic docetaxel exposure and enhance its anticancer effect, but it also raises the spectre of increased toxicity, a consideration the authors flag as a significant translational caveat. The combination has precedent: kaempferol and docetaxel have previously been reported to act synergistically against pancreatic and prostate cancer cells, and kaempferol alone has been shown to suppress breast cancer proliferation and metastasis by downregulating Bcl-2 and modulating the PI3K/Akt pathway. The new study provides a structural rationale for why the two molecules might cooperate at the level of a single protein target.

The authors are careful to frame these findings as hypothesis-generating rather than clinical. The study relies entirely on computational methods: a single 150-nanosecond trajectory per complex, docking, MM-PBSA energetics and in silico ADME predictions. Key open questions remain, including whether docetaxel and kaempferol can bind Bcl-2 simultaneously and whether a ternary complex would show cooperative effects. The team calls for in vitro studies testing the pair individually and in combination against breast cancer cells, alongside experimental CYP450 inhibition assays to verify the predicted metabolic interaction. If those experiments bear out the simulations, a common dietary flavonoid could become a genuine partner for one of oncology’s workhorse drugs, offering a cheaper and more tolerable route to disarming the survival machinery that keeps breast cancer cells alive.

Subject of Research: Computational study of kaempferol and docetaxel binding to the Bcl-2 anti-apoptotic protein in breast cancer therapy

Article Title: Enhancing therapeutic efficacy through Kaempferol and docetaxel combination: integrative binding and molecular dynamics study targeting Bcl-2 protein in breast cancer

Article References: Mohale, T. E., Machaba, K. E., Hlengwa, N., & Mbazima, V. (2026). Enhancing therapeutic efficacy through Kaempferol and docetaxel combination: integrative binding and molecular dynamics study targeting Bcl-2 protein in breast cancer. Results in Chemistry, 31, Article 103906. https://doi.org/10.1016/j.rechem.2026.103906

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.103906

Keywords: breast cancer, Bcl-2, kaempferol, docetaxel, molecular docking, molecular dynamics, MM-PBSA, apoptosis, drug resistance, flavonoid, chemotherapy, drug-likeness