Cancer remains one of the most formidable health challenges of the modern era, claiming millions of lives each year despite decades of therapeutic advances. According to GLOBOCAN 2022 estimates, roughly 20 million new cancer cases and 9.7 million deaths were recorded worldwide in a single year, and the burden continues to climb in developing nations such as Bangladesh, where population-based studies report more than 109 new cases per 100,000 people. Now, a team of researchers from Bangladesh has turned to the vast chemical library of traditional medicinal plants to find new weapons against the disease, and their computational investigation has singled out a familiar steroidal saponin, diosgenin, as a remarkably potent candidate against one of cancer’s most stubborn survival mechanisms.
The target of the study is a family of proteins that cancer cells exploit with devastating efficiency: the antiapoptotic Bcl-2 family. Apoptosis, or programmed cell death, is the body’s built-in quality control system, removing damaged, aged, or unwanted cells through two main routes, the mitochondrial intrinsic pathway and the extrinsic death receptor pathway. The intrinsic route is governed by the Bcl-2 protein family, which includes antiapoptotic members such as Bcl-2, Bcl-xL, Bcl-w, and Mcl-1, and proapoptotic players like Bax, Bak, and a suite of BH3-only proteins. When proapoptotic proteins oligomerize on the mitochondrial outer membrane, they trigger mitochondrial outer membrane permeabilization, releasing cytochrome c, assembling the apoptosome with Apaf-1, and activating the caspase enzymes that dismantle the cell. Antiapoptotic members act as guardians of the mitochondrial membrane, sequestering their proapoptotic relatives and preventing this point of no return.
Cancer cells frequently overexpress these antiapoptotic guardians, blocking the release of caspase-activating factors and thereby enhancing survival, proliferation, tumor progression, metastasis, angiogenesis, and drug resistance. This makes them attractive drug targets, particularly because they sit at the final, largely irreversible step of cell death, unlike upstream signaling pathways such as EGFR, PI3K, ALK, and BRAF, which tumors can bypass through compensatory mutations. Drugs like ABT-737, ABT-263, and venetoclax have demonstrated the therapeutic promise of this strategy, but resistance often emerges when other family members such as Mcl-1 or Bcl-xL are overexpressed. For this reason, the researchers selected obatoclax, a pan-Bcl-2 family inhibitor capable of targeting multiple antiapoptotic proteins simultaneously, as their reference compound rather than the more selective venetoclax.
The study, published in Results in Chemistry, cast a wide net across the plant kingdom. The team assembled a library of 698 bioactive phytocompounds drawn from eight medicinal plants with long histories of therapeutic use: Lycium barbarum, Asparagus racemosus, Curcuma longa, Acorus calamus, Moringa oleifera, Aristolochia indica, Taxus baccata, and Taxus brevifolia. These plants are rich sources of carotenoids, flavonols, polyphenols, flavonoids, sterols, alkaloids, and terpenoids, with documented antioxidant, antineoplastic, anti-inflammatory, and cytotoxic activities. Compound structures were retrieved from the NPASS, IMPPAT, and KNApSAcK databases, prepared with OpenBabel and Discovery Studio, and docked against crystal structures of Bcl-2, Bcl-xL, Bcl-w, and Mcl-1 obtained from the Protein Data Bank using AutoDock Vina within the PyRx platform.
The docking results were striking. Among the 698 compounds screened, only three, diosgenin, friedelin, and roridin E, outperformed obatoclax across all four antiapoptotic proteins, and diosgenin emerged as the clear leader. It achieved binding energies of −9.0 kcal/mol against Bcl-2, −9.3 kcal/mol against Bcl-xL, −8.5 kcal/mol against Bcl-w, and −8.2 kcal/mol against Mcl-1, consistently surpassing the control compound. Ligand efficiency analysis, which normalizes binding affinity by molecular size, confirmed that diosgenin offered a favorable balance between potency and molecular complexity, whereas larger compounds like violaxanthin achieved respectable scores only through sheer size. Receptor preference margin analysis further revealed that diosgenin favors multiple homologous Bcl-2 family proteins rather than a single receptor, with a preference order of Bcl-xL, followed by Bcl-2, Bcl-w, and Mcl-1, echoing the broad-spectrum profile that makes obatoclax clinically interesting.
Structural analysis showed why diosgenin binds so effectively. The compound occupies the hydrophobic groove formed by the BH1, BH2, and BH3 domains, the very pocket that antiapoptotic proteins use to grip the BH3 regions of their proapoptotic targets. With Bcl-2, diosgenin formed eight hydrophobic contacts involving residues such as Met-115, Val-156, Ala-149, Leu-137, Phe-104, and Phe-112. With Bcl-w, it generated thirteen hydrophobic interactions plus a hydrogen bond through Arg-95, engaging aromatic residues Phe-102 and Trp-137. Residue contact fingerprint analysis showed that phenylalanine, tyrosine, and leucine residues were conserved contact points across 75 percent of the target proteins, indicating that van der Waals and hydrophobic forces drive the interaction, and that diosgenin partially mimics the binding pattern of obatoclax while also forging unique contacts of its own. Redocking of co-crystallized ligands produced root-mean-square deviations below 2 angstroms for all targets, validating the docking protocol.
To confirm that these interactions would hold up in a dynamic, physiological environment, the team ran 200-nanosecond molecular dynamics simulations with the AMBER14 force field in YASARA under near-physiological conditions. The diosgenin-Bcl-xL complex proved more stable than the obatoclax control, with lower root-mean-square deviation, while the other complexes remained comparably stable. Root-mean-square fluctuation, solvent-accessible surface area, and radius of gyration analyses all pointed to compact, stable complexes, with diosgenin generally producing tighter protein conformations than the control. Hydrogen bond analysis revealed that diosgenin maintained roughly three hydrogen bonds with Bcl-2, more than double the control’s average, and secondary structure analysis confirmed that all proteins retained their predominantly alpha-helical folds throughout the simulations. MM-PBSA binding free energy calculations reinforced the picture, showing diosgenin complexes with Bcl-2 and Bcl-xL achieving more favorable energies than the control.
Perhaps most compelling are the drug-likeness and safety predictions. Diosgenin satisfied both Lipinski’s Rule of Five and Veber’s rules, showed high gastrointestinal absorption, crossed the blood-brain barrier, and, critically, did not inhibit any of the key cytochrome P450 enzymes, whereas obatoclax was predicted to interfere with CYP1A2, CYP3A4, and CYP2C19, raising the specter of drug-drug interactions. Toxicity profiling with ProTox-3 found diosgenin inactive for neurotoxicity, mutagenicity, hepatotoxicity, nephrotoxicity, carcinogenicity, and cardiotoxicity, with only mild predicted immunotoxicity, and a predicted median lethal dose of 8000 mg/kg compared with obatoclax’s 3066 mg/kg. Density functional theory calculations added a quantum mechanical dimension, showing a HOMO-LUMO energy gap of 5.279 eV for diosgenin versus 3.224 eV for obatoclax and a molecular electrostatic potential map indicating greater electron density and nucleophilicity for the plant compound.
These computational findings align intriguingly with a growing body of experimental literature. Previous studies have shown that diosgenin suppresses proliferation and triggers caspase-driven apoptosis in skin squamous carcinoma cells through AKT and JNK signaling, overcomes TRAIL resistance in colon cancer cells by downregulating the p38 MAPK pathway and elevating death receptor DR5, and curbs the invasion of triple-negative breast cancer cells by suppressing Vav2 phosphorylation and Cdc42 activation. In animal models, intra-tumoral diosgenin treatment markedly suppressed breast cancer xenograft growth in mice, and the compound reduced azoxymethane-induced aberrant crypt foci in rat models of colorectal cancer. The new study adds a mechanistic explanation for these observations: by wedging into the hydrophobic groove of antiapoptotic Bcl-2 family proteins, diosgenin could liberate proapoptotic effectors, drive mitochondrial outer membrane permeabilization, and push cancer cells irreversibly toward apoptosis.
The authors are careful to note the limits of their work. Every result, from the docking scores to the ADMET predictions, rests on computational models that require laboratory confirmation, and further in vitro and in vivo studies will be essential to establish diosgenin’s efficacy, bioavailability, and off-target profile in living systems. Still, the convergence of strong binding across all four antiapoptotic targets, stable simulated complexes, clean pharmacokinetics, and an exceptional predicted safety margin positions this humble plant saponin, found in fenugreek, wild yam, and other traditional remedies, as a genuinely promising scaffold for the next generation of Bcl-2 family inhibitors. If experimental validation keeps pace, a compound known to ancient healers may yet find a place in the modern oncology clinic.
Subject of Research: Computational identification of the natural plant compound diosgenin as an inhibitor of antiapoptotic Bcl-2 family proteins for cancer therapy
Article Title: Computational discovery of natural product-based anti-cancer agents targeting antiapoptotic family proteins through bioinformatics approach
Article References: Bulbul, M. I. A., Khan, T., Sharid, M. A. G., Shuvo, M. N., Debnath, A. C., Ali, M. A., & Haque, M. A. (2026). Computational discovery of natural product-based anti-cancer agents targeting antiapoptotic family proteins through bioinformatics approach. Results in Chemistry, 30, Article 103804. https://doi.org/10.1016/j.rechem.2026.103804
Image Credits: AI Generated
DOI: 10.1016/j.rechem.2026.103804
Keywords: diosgenin, Bcl-2 family proteins, apoptosis, molecular docking, molecular dynamics simulation, natural products, cancer drug discovery, MM-PBSA, ADMET prediction, obatoclax, phytocompounds, in silico drug design
Cite Scienmag News
APA
MLA
Chicago
Nathaniel Bowman. (September 12, 2026). Plant Compound Diosgenin Emerges as Powerful Computational Anti-Cancer Candidate. Scienmag. https://scienmag.com/plant-compound-diosgenin-emerges-as-powerful-computational-anti-cancer-candidate/
Nathaniel Bowman. “Plant Compound Diosgenin Emerges as Powerful Computational Anti-Cancer Candidate.” Scienmag, 12 September 2026, https://scienmag.com/plant-compound-diosgenin-emerges-as-powerful-computational-anti-cancer-candidate/. Accessed 12 September 2026.
Nathaniel Bowman. “Plant Compound Diosgenin Emerges as Powerful Computational Anti-Cancer Candidate.” Scienmag. September 12, 2026. https://scienmag.com/plant-compound-diosgenin-emerges-as-powerful-computational-anti-cancer-candidate/
Copy citation
Download RIS
Tags: ADMET predictionAdvances in cancer therapeutics using plant moleculesapoptosisApoptosis regulation in cancer cellsBangladesh-based cancer research studiesBCL-2 family proteinsBcl-2 family proteins as cancer targetscancer drug discoveryComputational anti-cancer drug discoverydiosgeninDiosgenin as natural anti-cancer compoundEmerging natural antiin silico drug designIn silico screening of plant compounds for cancerMM/PBSAmolecular dockingmolecular dynamics simulationNatural compounds targeting antiapoptotic proteinsnatural productsobatoclaxphytocompoundsPlant-derived steroidal saponins in cancer therapyRole of mitochondria in programmed cell deathTraditional medicinal plants in cancer research
