Anticancer

Parthenolide induces apoptosis in primary human AML and blast-crisis CML stem and progenitor cells while sparing normal haematopoietic cells, and outperformed cytarabine for specificity in a NOD/SCID xenograft, through combined NF-κB inhibition, p53 activation and a rise in reactive oxygen species ^{91Reference 91Guzman ML et al. · 2005In vitroThe sesquiterpene lactone parthenolide induces apoptosis of human acute myelogenous leukemia stem and progenitor cells — in vitro / animal modelView study →}.

Reviews describe it as the first small molecule found to be selective against cancer stem cells ^{92Reference 92Ghantous A et al. · 2013ReviewParthenolide: from plant shoots to cancer roots — reviewView study →}.

The most economical explanation is that a promiscuous thiol alkylator reaches its threshold first in cells whose redox buffer is already strained.

Drug-likeness is the limit: native parthenolide’s solubility and bioavailability were poor enough that the field moved to the analogue DMAPT (LC-1), a dimethylamine Michael adduct with about 70% oral bioavailability, and it was DMAPT that carried the finding into animal models, canine leukaemias and a first-in-man study ^{93,94,95Reference 93Guzman ML et al. · 2007In vitroAn orally bioavailable parthenolide analog selectively eradicates acute myelogenous leukemia stem and progenitor cells — in vitro / animal modelView study →Reference 94Neelakantan S et al. · 2009In vitroAminoparthenolides as novel anti-leukemic agents: discovery of the NF-κB inhibitor, DMAPT (LC-1) — in vitroView study →Reference 95Hewamana S et al. · 2008In vitroThe novel nuclear factor-κB inhibitor LC-1 is equipotent in poor prognostic subsets of chronic lymphocytic leukemia and shows strong synergy with fludarabine — in vitroView study →}.

The two molecules are distinct, and no human efficacy result has been published for either.

Feverfew is the herb that carries it.

Aurora A is the target with the clearest causal test.

In breast cancer cells the tanshinones inhibited growth dose-dependently, with tanshinone I the most potent and the only member active against triple-negative MDA-MB-231, and siRNA knockdown of Aurora A largely removed its growth-inhibition and apoptosis activity.

The regulation is epigenetic: Aurora A overexpression in those cells is driven by histone acetylation rather than promoter methylation, tanshinone I reduced acetylated histone H3 at the Aurora A gene, and normal mammary epithelium was much less affected ^{126Reference 126Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →}.

The result was replicated in lung cancer, where tanshinone I was again the most potent of three tanshinones and Aurora A siRNA again abolished the effect; in an H1299 xenograft 200 mg/kg cut tumor weight 34%, proliferation 54%, angiogenesis 72% and Aurora A expression 67% without changing food intake or body weight ^{127Reference 127Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}.

In colorectal cancer the effect is p53-dependent, running through an Aurora A → p53 axis in p53-wild-type cells and through survivin downregulation in p53-mutant cells, with much less effect on normal colon epithelium ^{128Reference 128Lu M et al. · 2016Tanshinone I induces human colorectal cancer cell apoptosis: the potential roles of Aurora A-p53 and survivin-mediated signaling pathwaysView study →}.

The second thread is anti-metastatic rather than cytotoxic.

Unlike tanshinone IIA, tanshinone I had no direct cytotoxicity in the invasive lung line CL1-5; it suppressed migration, invasion and gelatinase activity, reduced IL-8 transcription by attenuating AP-1 and NF-κB DNA binding (Ras-MAPK and Rac1 implicated), and reduced tumorigenesis and metastasis in SCID mice ^{129Reference 129Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}.

On the endothelial side it suppressed ICAM-1/VCAM-1, cut TNF-α-induced VEGF and VEGF-mediated tube formation, reduced MDA-MB-231 adhesion and matrix migration, and lowered both tumor mass and metastasis incidence in vivo ^{130Reference 130Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →}.

Later work is consistent: osteosarcoma growth was suppressed in subcutaneous and orthotopic xenografts at 10–20 mg/kg and metastasis inhibited in an intracardiac inoculation model, with IL-6-driven JAK/STAT3 abolished and Bax raised against Bcl-2 ^{131Reference 131Wang W et al. · 2019Tanshinone I inhibits the growth and metastasis of osteosarcoma via suppressing JAK/STAT3 signalling pathwayView study →}, and cervical cancer invasion and migration fell through BNIP3/NIX-mediated mitophagy with mitochondrial metabolic reprogramming, reversible by siRNA ^{132Reference 132Cui S et al. · 2022Tanshinone I inhibits metastasis of cervical cancer cells by inducing BNIP3/NIX-mediated mitophagy and reprogramming mitochondrial metabolismView study →}.

The two Aurora A papers come from one laboratory with no independent replication, no direct kinase-binding or kinome-selectivity data exist, and there is no human data.

Agrimoniin inhibited PANC-1 pancreatic carcinoma proliferation by inducing apoptosis and cell-cycle arrest, acting as a dual AKT and ERK inhibitor (confirmed against SC79, LY294002 and U0126) and potentiating both inhibitors, with in-vivo support ^{1Reference 1Zhang X et al. · 2023In vitroAgrimoniin is a dual inhibitor of AKT and ERK pathways that inhibit pancreatic cancer cell proliferation — in vitro and in vivoView study →}.

A second group attributed its pancreatic-cancer effect to suppressing the Nrf2-dependent ROS-scavenging system, disrupting mitochondrial membrane potential and lowering mTOR/HIF-1α, in a nude-mouse model ^{2Reference 2Zhu X et al. · 2022In vitroAgrimoniin sensitizes pancreatic cancer to apoptosis through ROS-mediated energy metabolism dysfunction — in vitro and nude mouse studyView study →}.

On isolated rat liver mitochondria, 10 µM opened the permeability transition pore, blocked by ADP and cyclosporine A, and the compound chelated iron ^{3Reference 3Fedotcheva TA et al. · 2021In vitroNew properties and mitochondrial targets of polyphenol agrimoniin as a natural anticancer and preventive agent — in vitroView study →}.

Orally, Agrimonia pilosa root extract attenuated AOM/DSS colorectal tumors in mice, shifted the tumor microenvironment toward anti-tumor immune cells and remodelled the gut microbiota, with agrimoniin among the two most active constituents in vivo ^{4Reference 4Wang L et al. · 2024AnimalThe anti-colorectal cancer effect and metabolites of Agrimonia pilosa Ledeb. — AOM/DSS mouse modelView study →}.

The proposed mechanisms do not yet converge, the cytotoxicity is quenched by serum ^{5Reference 5Miyamoto K et al. · 1987AnimalAntitumor effect of agrimoniin, a tannin of Agrimonia pilosa Ledeb., on transplantable rodent tumors — mouse studyView study →}, and no selectivity index against normal cells has been measured for the isolate.

Carnosol has two threads, and only the first is specific to it.

Molecular modelling placed carnosol in the ligand-binding domain of both the androgen receptor and ER-α; TR-FRET confirmed binding to both, and follow-up work established it as a pure antagonist with no agonist activity at either, a profile unlike any antiandrogen or antioestrogen then in development.

LNCaP, 22Rv1 and MCF7 cells at 20–40 µmol/L lost AR and ER-α protein, and in 22Rv1 xenograft mice oral carnosol at 30 mg/kg five times a week for 28 days suppressed tumor growth 36% (P = 0.028) and serum PSA 26% (P = 0.0042) ^{10Reference 10Johnson JJ et al. · 2010Disruption of androgen and estrogen receptor activity in prostate cancer by a novel dietary diterpene carnosol — implications for chemopreventionView study →}.

In androgen-independent PC3 cells it caused G2 arrest across 10–70 µM and acted on AMPK and PI3K/Akt ^{11Reference 11Johnson JJ et al. · 2008Carnosol, a dietary diterpene, displays growth inhibitory effects in human prostate cancer PC3 cells leading to G2-phase cell cycle arrest and targets the 5’-AMP-activated protein kinase (AMPK) pathwayView study →}.

The second thread is ROS-driven apoptosis across cell lines: triple-negative breast (G2 arrest, DNA damage, simultaneous beclin1-independent autophagy and apoptosis, all abolished by the ROS scavenger tiron ^{12Reference 12Al Dhaheri Y et al. · 2014Carnosol induces ROS-mediated beclin1-independent autophagy and apoptosis in triple negative breast cancerView study →}, traced to p38-driven ER stress and proteasomal degradation of mTOR, PCAF and STAT3 ^{13Reference 13Alsamri H et al. · 2022Carnosol induces p38-mediated ER stress response and autophagy in human breast cancer cellsView study →}); HCT116 colon (ROS, p53 up, Mdm2 down, STAT3 inactivated via Jak2/Src ^{14Reference 14Park KW et al. · 2014Carnosol induces apoptosis through generation of ROS and inactivation of STAT3 signaling in human colon cancer HCT116 cellsView study →}); pre-B ALL lines (34–53% fall in Bcl-2 ^{15Reference 15Dörrie J et al. · 2001Carnosol-induced apoptosis and downregulation of Bcl-2 in B-lineage leukemia cellsView study →}); and gingiva carcinoma, one of the few studies to test selectivity ^{16Reference 16Gassib N et al. · 2024Cellular mechanisms mediating the anti-cancer effects of carnosol on gingiva carcinomaView study →}.

Carnosol and carnosic acid are also anti-angiogenic, with greater inhibition of tumor than non-transformed endothelium, confirmed in the chick chorioallantoic membrane assay ^{17Reference 17López-Jiménez A et al. · 2013Anti-angiogenic properties of carnosol and carnosic acid, two major dietary compounds from rosemaryView study →}.

Effective concentrations are 20–90 times the achievable mouse plasma C_max ^{18Reference 18Vemu B et al. · 2021AnimalPharmacokinetic characterization of carnosol from rosemary (Salvia rosmarinus) in male C57BL/6 mice and inhibition profile in human cytochrome P450 enzymesView study →}, and there is no in-vivo tumor model outside the prostate xenograft.

Carnosol also delayed cytarabine-, methotrexate- and vincristine-induced apoptosis in leukemia lines, and the authors warned that increased dietary intake “may potentially decrease the effectiveness of some standard chemotherapy treatments used for leukemia” ^{19Reference 19Zunino SJ · 2009Carnosol delays chemotherapy-induced DNA fragmentation and morphological changes associated with apoptosis in leukemic cellsView study →}.

Cryptotanshinone rapidly suppresses STAT3 Tyr705 phosphorylation (within 30 minutes, JAK2-independently), collapsing cyclin D1, survivin and Bcl-xL, with growth inhibition in DU145 prostate cells ^{37Reference 37Shin DS et al. · 2009In vitroCryptotanshinone inhibits constitutive signal transducer and activator of transcription 3 function through blocking the dimerization in DU145 prostate cancer cells — in vitro studyView study →}.

That much replicates across many laboratories, cell types and several xenografts ^{38,39,40Reference 38Li H et al. · 2021ReviewA review of the biological activity and pharmacology of cryptotanshinone, an important active constituent in Danshen — reviewView study →Reference 39Wu YH et al. · 2020ReviewCryptotanshinone: a review of its pharmacology activities and molecular mechanisms — reviewView study →Reference 40Lu L et al. · 2017In vitroCryptotanshinone inhibits human glioma cell proliferation in vitro and in vivo through SHP-2-dependent inhibition of STAT3 activation — in vitro and mouse xenograft studyView study →}.

Whether it works by binding STAT3 is not established.

The direct-binding claim rests on colocalisation, a dimerisation readout and computational docking of the SH2 domain, and the authors’ own wording is hedged (“might directly bind”) ^{37Reference 37Shin DS et al. · 2009In vitroCryptotanshinone inhibits constitutive signal transducer and activator of transcription 3 function through blocking the dimerization in DU145 prostate cancer cells — in vitro studyView study →}.

No SPR, ITC or BLI binding constant for the natural product has been published, and STAT1/STAT5 selectivity was never tested for it.

Both gaps were later filled for a synthetic derivative, LYW-6, a structure-based remodelling of cryptotanshinone into a more effective STAT3 inhibitor with a measured Kd of 6.6 µM and STAT1/STAT5 controls ^{41Reference 41Wang H et al. · 2020AnimalLYW-6, a novel cryptotanshinone derived STAT3 targeting inhibitor, suppresses colorectal cancer growth and metastasis — medicinal chemistry and in vivo studyView study →}.

A second laboratory assigns the same phenotype to an indirect route: cryptotanshinone raises SHP-2 phosphatase activity, and SHP-2 inhibitor or siRNA reverses the Tyr705 effect ^{40Reference 40Lu L et al. · 2017In vitroCryptotanshinone inhibits human glioma cell proliferation in vitro and in vivo through SHP-2-dependent inhibition of STAT3 activation — in vitro and mouse xenograft studyView study →}.

At least six mutually independent primary mechanisms are in print for the same cells: SH2 binding ^{37Reference 37Shin DS et al. · 2009In vitroCryptotanshinone inhibits constitutive signal transducer and activator of transcription 3 function through blocking the dimerization in DU145 prostate cancer cells — in vitro studyView study →}, SHP-2 ^{40Reference 40Lu L et al. · 2017In vitroCryptotanshinone inhibits human glioma cell proliferation in vitro and in vivo through SHP-2-dependent inhibition of STAT3 activation — in vitro and mouse xenograft studyView study →}, mTOR/cyclin D1/Rb ^{42Reference 42Chen W et al. · 2010In vitroCryptotanshinone inhibits cancer cell proliferation by suppressing mammalian target of rapamycin-mediated cyclin D1 expression and Rb phosphorylation — in vitro studyView study →}, mTORC1–S6K1 complex disruption ^{43Reference 43Jeoung NH et al. · 2021In vitroCryptotanshinone prevents the binding of S6K1 to mTOR/Raptor leading to the suppression of mTORC1-S6K1 signaling activity and neoplastic cell transformation — in vitro studyView study →}, AMPK–TSC2 ^{44Reference 44Chen W et al. · 2017In vitroCryptotanshinone activates AMPK-TSC2 axis leading to inhibition of mTORC1 signaling in cancer cells — in vitro studyView study →}, ERα-dependence ^{45Reference 45Pan Y et al. · 2017In vitroCryptotanshinone inhibition of mammalian target of rapamycin pathway is dependent on oestrogen receptor alpha in breast cancer — in vitro studyView study →} and NQO1-driven necrosis ^{46Reference 46Hou Y et al. · 2025In vitroA small molecule cryptotanshinone induces non-enzymatic NQO1-dependent necrosis in cancer cells through the JNK1/2/iron/PARP/calcium pathway — in vitro and in vivo studyView study →}.

For an ortho-quinone that generates superoxide ^{47Reference 47Lee DS et al. · 1999In vitroAntibacterial activities of cryptotanshinone and dihydrotanshinone I from a medicinal herb, Salvia miltiorrhiza Bunge — in vitro studyView study →} and kills human erythrocytes at 1–10 µM ^{48Reference 48Bissinger R et al. · 2014In vitroStimulation of eryptosis by cryptotanshinone — human erythrocyte in vitro studyView study →}, promiscuity is the simplest explanation.

Dan shen is the herb that carries it.

Purified cycleanine kills four ovarian cancer lines (Ovcar-8, A2780, Ovcar-4, Igrov-1) at IC₅₀ 7–14 µM through a defined apoptotic route of caspase-3/7 activation, PARP cleavage and subG1 accumulation ^{49Reference 49Uche FI et al. · 2016Cytotoxicity effects and apoptosis induction by bisbenzylisoquinoline alkaloids from Triclisia subcordataView study →}, and two-step (aminoalkyl) analogues are more potent, with uptake confirmed by click chemistry ^{50Reference 50Uche FI et al. · 2018Synthesis of (aminoalkyl)cycleanine analogues: cytotoxicity, cellular uptake, and apoptosis induction in ovarian cancer cellsView study →}.

Docking and molecular dynamics predict a stable cycleanine–PARP1 complex, which would place the inhibition upstream of the cleavage rather than only downstream of it ^{51Reference 51Nwaefulu ON et al. · 2022The impact of cycleanine in cancer research: a computational studyView study →}.

Separately it reverses acquired doxorubicin and vincristine resistance in MCF-7/Adr and KBv200 cells, raising intracellular doxorubicin more effectively than verapamil ^{52Reference 52Tian H · 1997Modulation of multidrug resistance by three bisbenzyl-isoquinolines in comparison with verapamilView study →}.

The selectivity index over normal ovarian epithelium is only about 2–5 (35 µM) ^{49Reference 49Uche FI et al. · 2016Cytotoxicity effects and apoptosis induction by bisbenzylisoquinoline alkaloids from Triclisia subcordataView study →}, there is no in-vivo tumor model of any kind, the PARP1 target rests on docking and has never been assayed, and no transporter was identified in the resistance work.

Abuta is the herb that carries it.

Dihydrotanshinone I is the reference natural-product HuR inhibitor.

It competes with target RNA for HuR’s RRM1/RRM2 surface at Ki 3.74 nM, locking the protein in a closed conformation ^{53,54Reference 53D’Agostino VG et al. · 2015In vitroDihydrotanshinone-I interferes with the RNA-binding activity of HuR affecting its post-transcriptional function — in vitro and cell studyView study →Reference 54Lal P et al. · 2017Regulation of HuR structure and function by dihydrotanshinone-I — NMR, molecular dynamics, RIP-chip and xenograft studyView study →}, strips HuR preferentially from its weaker mRNA targets transcriptome-wide, and inhibited xenograft growth in a HuR-dependent model without systemic toxicity ^{54Reference 54Lal P et al. · 2017Regulation of HuR structure and function by dihydrotanshinone-I — NMR, molecular dynamics, RIP-chip and xenograft studyView study →}.

A designed “tanshinone mimic” ortho-quinone series was built from its scaffold and produced analogues more potent than the natural product ^{55Reference 55Manzoni L et al. · 2018Interfering with HuR-RNA interaction: design, synthesis and biological characterization of tanshinone mimics as novel, effective HuR inhibitors — medicinal chemistry studyView study →}, and independent groups now benchmark against it ^{56Reference 56Dev AJR et al. · 2025In vitroBiophysical and molecular characterization of HuR inhibitors, CMLD-2 and dihydrotanshinone I (DHTS), in triple negative breast cancer — in vitro studyView study →}.

Its HIF-1 suppression is plausibly the same mechanism, acting post-transcriptionally rather than on the protein ^{57Reference 57Duan P et al. · 2021In vitroDuan P, Huang Y, Chen K, Cheng C, Wu Z, Wu Y (2021). 15,16-dihydrotanshinone I inhibits EOMA cells proliferation by interfering in posttranscriptional processing of hypoxia-inducible factor 1 — in vitro study. International Journal of Medical Sciences. https://pubmed.ncbi.nlm.nih.gov/34400891/View study →}.

Beyond HuR, it engages ESR1 to produce DNA double-strand breaks in hepatocellular carcinoma ^{58Reference 58Nie Y et al. · 2024In vitroDihydrotanshinone I targets ESR1 to induce DNA double-strand breaks and proliferation inhibition in hepatocellular carcinoma — in vitro and animal studyView study →}, suppresses neutrophil-extracellular-trap formation to reduce breast-cancer lung metastasis in mice ^{59Reference 59Zhao H et al. · 2022AnimalDihydrotanshinone I inhibits the lung metastasis of breast cancer by suppressing neutrophil extracellular traps formation — mouse studyView study →}, drives caspase-dependent apoptosis and autophagy in colon cancer ^{60Reference 60Wang L et al. · 2015In vitroDihydrotanshinone I induced apoptosis and autophagy through caspase dependent pathway in colon cancer — in vitro studyView study →}, glioma ^{61Reference 61Cao Y et al. · 2017In vitroSalvia miltiorrhiza extract dihydrotanshinone induces apoptosis and inhibits proliferation of glioma cells — in vitro studyView study →} and gastric lines ^{62Reference 62Cheng R et al. · 2016In vitroDihydrotanshinone induces apoptosis of SGC7901 and MGC803 cells via activation of JNK and p38 signalling pathways — in vitro studyView study →}, potentiates cisplatin in anaplastic thyroid cancer ^{63Reference 63Allegri L et al. · 2021In vitroDihydrotanshinone exerts antitumor effects and improves the effects of cisplatin in anaplastic thyroid cancer cells — in vitro studyView study →} and synergises with temozolomide in MGMT-positive glioblastoma ^{64Reference 64Kumar V et al. · 2019In vitroStudies examining the synergy between dihydrotanshinone and temozolomide against MGMT+ glioblastoma cells in vitro — in vitro studyView study →}.

In the only screen with a healthy control line, 63 Salvia metabolites tested across MCF-7, HT-29 and a healthy fibroblast line, dihydrotanshinone I was grouped with the compounds showing high toxicity to the healthy cells, and cryptotanshinone and several royleanones had the better cancer-versus-healthy ratio ^{65Reference 65Irtegun Kandemir S et al. · 2022In vitroInvestigation of cytotoxic and apoptotic effects of 63 compounds obtained from Salvia species — in vitro screen including a healthy cell lineView study →}.

It is potent rather than selective.

Dan shen is the herb that carries it.

One 1990 measurement exists: galdosol was the most active of thirteen Canarian Salvia and Sideritis isolates, with a cytostatic ID₅₀ of 0.5 µg/mL (about 1.5 µM) ^{71Reference 71Darias V et al. · 1990In vitroCytostatic and antibacterial activity of some compounds isolated from several Lamiaceae species from the Canary Islands — in vitroView study →}.

The abstract does not name the cell line, there is no normal-cell comparator and therefore no selectivity index, and nothing has followed in 36 years: no replication, no mechanism, no in-vivo tumor model.

Galdosol is a redox-active catechol, a chemotype that commonly reduces viability through generalised oxidative stress rather than any tumour-selective mechanism, and its close sibling epirosmanol was later measured as cytotoxic to normal rat L6 myoblasts at about 14 µM, a selectivity index of 1.6.

An unexplained low-micromolar loss of cell viability is a safety observation rather than an efficacy claim.

The bioenergetic mechanism is the best supported.

Gracillin came out of a large natural-product screen as a mitochondria-targeting hit and was shown to disrupt complex II by abrogating succinate dehydrogenase activity while leaving the succinate:ubiquinone reductase step intact, a specific lesion rather than generic mitochondrial poisoning, confirmed because cell death was potentiated by two structurally unrelated complex II inhibitors binding different sites (3-nitropropionic acid at SDHA, thenoyltrifluoroacetone at the ubiquinone site).

ATP fell, ROS rose, and the effect carried through to mutant-Kras-driven lung tumorigenesis and patient-derived xenografts ^{72Reference 72Min HY et al. · 2019The natural compound gracillin exerts potent antitumor activity by targeting mitochondrial complex IIView study →}.

A companion study covered the other half of the energy budget: glycolysis-associated metabolites were suppressed, and docking plus enzyme assay nominated phosphoglycerate kinase 1 (PGK1), again with patient-derived xenograft efficacy ^{73Reference 73Min HY et al. · 2020Potent anticancer effect of the natural steroidal saponin gracillin is produced by inhibiting glycolysis and oxidative phosphorylation-mediated bioenergeticsView study →}.

Attacking OXPHOS and glycolysis at once explains the broad-spectrum viability data better than any single signalling target.

In-vivo replication is unusually good, with xenograft or allograft suppression in lung, colorectal, breast, liver, glioblastoma, melanoma and pancreatic models from at least eight independent laboratories, typically at 1–10 mg/kg/day ^{72,73,74,75,76,77,78,79Reference 72Min HY et al. · 2019The natural compound gracillin exerts potent antitumor activity by targeting mitochondrial complex IIView study →Reference 73Min HY et al. · 2020Potent anticancer effect of the natural steroidal saponin gracillin is produced by inhibiting glycolysis and oxidative phosphorylation-mediated bioenergeticsView study →Reference 74Su JX et al. · 2025Gracillin suppresses cancer progression through inducing Merlin/LATS protein–protein interaction and activating Hippo signaling pathwayView study →Reference 75Liao J et al. · 2026Gracillin overcomes osimertinib resistance in NSCLC via dual inhibition of EGFR and Mcl-1View study →Reference 76Lee W et al. · 2025Gracillin induces mitochondria-mediated apoptosis on pancreatic ductal adenocarcinoma through disruption of redox homeostasis via inhibiting NRF2/HO-1 antioxidant axisView study →Reference 77Li JK et al. · 2023In vitroGracillin exerts anti-melanoma effects in vitro and in vivo: role of DNA damage, apoptosis and autophagyView study →Reference 78Yang L et al. · 2021Gracillin shows potent efficacy against colorectal cancer through inhibiting the STAT3 pathwayView study →Reference 79Li Y et al. · 2022Gracillin shows potential efficacy against non-small cell lung cancer through inhibiting the mTOR pathwayView study →}, and one properly controlled protein–protein-interaction mechanism (Merlin/LATS driving MST-independent Hippo activation, abolished by a competitive peptide from the binding interface) ^{74Reference 74Su JX et al. · 2025Gracillin suppresses cancer progression through inducing Merlin/LATS protein–protein interaction and activating Hippo signaling pathwayView study →}.

Two limits remain.

Ten different primary targets have been nominated by ten different groups, complex II ^{72Reference 72Min HY et al. · 2019The natural compound gracillin exerts potent antitumor activity by targeting mitochondrial complex IIView study →}, PGK1 ^{73Reference 73Min HY et al. · 2020Potent anticancer effect of the natural steroidal saponin gracillin is produced by inhibiting glycolysis and oxidative phosphorylation-mediated bioenergeticsView study →}, Merlin/LATS ^{74Reference 74Su JX et al. · 2025Gracillin suppresses cancer progression through inducing Merlin/LATS protein–protein interaction and activating Hippo signaling pathwayView study →}, EGFR/Mcl-1 ^{75Reference 75Liao J et al. · 2026Gracillin overcomes osimertinib resistance in NSCLC via dual inhibition of EGFR and Mcl-1View study →}, CCL24 ^{80Reference 80Huang S et al. · 2026CCL24 recruits CCR3+ TAMs to promote immunosuppression via YAP1 activation and serves as a therapeutic target for gracillin in colorectal cancerView study →}, NRF2 ^{76Reference 76Lee W et al. · 2025Gracillin induces mitochondria-mediated apoptosis on pancreatic ductal adenocarcinoma through disruption of redox homeostasis via inhibiting NRF2/HO-1 antioxidant axisView study →}, STAT3 ^{77,78Reference 77Li JK et al. · 2023In vitroGracillin exerts anti-melanoma effects in vitro and in vivo: role of DNA damage, apoptosis and autophagyView study →Reference 78Yang L et al. · 2021Gracillin shows potent efficacy against colorectal cancer through inhibiting the STAT3 pathwayView study →}, mTOR ^{79Reference 79Li Y et al. · 2022Gracillin shows potential efficacy against non-small cell lung cancer through inhibiting the mTOR pathwayView study →}, STC1 ^{81Reference 81Cui A et al. · 2026Study on the anti-lung cancer effects of gracillin from Reineckia carnea by targeting STC1View study →} and TIPE2 ^{82Reference 82Liu W et al. · 2021Beneficial effects of gracillin from Rhizoma Paridis against gastric carcinoma via the potential TIPE2-mediated induction of endogenous apoptosis and inhibition of migration in BGC823 cellsView study →}, with no cross-validation between them, and a membrane-active saponin producing a new target in each laboratory is what non-specific activity looks like.

The NCI screen is negative on selectivity: gracillin (NSC-698787) was cytotoxic at micromolar levels across most of the 60-line panel with no selectivity pattern, no activity against EKVX, HT29, OVCAR-5 or SN12C, and no COMPARE match, and the investigators advanced its furostanol relative instead ^{83Reference 83Hu K · 2003In vitroThe cytotoxicity of methyl protoneogracillin (NSC-698793) and gracillin (NSC-698787), two steroidal saponins from the rhizomes of Dioscorea collettii var. hypoglauca, against human cancer cells in vitroView study →}.

In Paris vietnamensis material gracillin gave IC₅₀ values of 1.17, 1.20 and 1.18 µM against a lung, a liver and a prostate line ^{84Reference 84Hoang LTA et al. · 2025Diosgenin glycosides from Paris vietnamensis (Takht) H.Li with cytotoxic and inhibitory NO production activitiesView study →}, indistinguishable across three tissues, which is the signature of a membrane effect.

Wild yam is the herb that carries it.

Repression of NF-κB p65 by helenalin induces autophagic cell death in cancer cells ^{85Reference 85Lim CB et al. · 2012In vitroNF-κB p65 repression by the sesquiterpene lactone, helenalin, contributes to the induction of autophagy cell death — in vitroView study →}, and isolated helenalin at 10–50 µM triggers apoptosis in Jurkat leukemia T cells that is independent of the CD95 death receptor and not blocked by Bcl-2 or Bcl-xL overexpression, via caspase-3/-8 ^{86Reference 86Dirsch VM et al. · 2001In vitroHelenalin triggers a CD95 death receptor-independent apoptosis that is not affected by overexpression of Bcl-x(L) or Bcl-2 — in vitroView study →}.

Potency is high, with an IC50 of about 0.44 µM against a human small-cell lung carcinoma line ^{87Reference 87Woerdenbag HJ et al. · 1994In vitroCytotoxicity of flavonoids and sesquiterpene lactones from Arnica species against the GLC4 and the COLO 320 cell lines — in vitroView study →}, and extends across a wide panel of murine and human tumor lines for helenalin and its mono- and difunctional esters ^{88Reference 88Grippo AA et al. · 1992In vitroThe cytotoxicity of helenalin, its mono and difunctional esters, and related sesquiterpene lactones in murine and human tumor cells — in vitroView study →}.

None of this is selective.

The same apoptotic behaviour is shown by ambrosin, alantolactone and hymenin, so it is thiol-alkylation chemistry shared across cytotoxic sesquiterpene lactones ^{89Reference 89Dirsch VM et al. · 2001In vitroCytotoxic sesquiterpene lactones mediate their death-inducing effect in leukemia T cells by triggering apoptosis — in vitroView study →}; no selectivity index against normal cells has been established, and the whole-animal expression of the same potency is an oral LD50 of 85–105 mg/kg ^{90Reference 90Witzel DA et al. · 1976AnimalMammalian toxicity of helenalin, the toxic principle of Helenium microcephalum DC (smallhead sneezeweed) — animal model, five speciesView study →}.

There is no in-vivo tumor model and no clinical work, and the toxicity precludes a systemic route.

Arnica is the herb that carries it.

Isolated protodioscin induces both apoptosis and autophagy in glioblastoma cells through LC3B/MCL-1 modulation, with LC3 knockdown abolishing both, and suppressed GBM8401 xenograft growth and Ki67 in mice without weight loss or organ toxicity ^{96Reference 96Tai CT et al. · 2025In vitroTargeting LC3B/MCL-1 expression by protodioscin induces autophagy and apoptosis in human glioblastoma cells in vitro and in vivo — [animal model]View study →}.

Against ER-positive and triple-negative breast lines it is cytotoxic at IC50 1.53–6 µM while normal PBMCs are unaffected below 50 µM ^{97Reference 97Bouchmaa N et al. · 2023In vitroIn vitro evaluation of dioscin and protodioscin against ER-positive and triple-negative breast cancer — [in vitro]View study →}; protodioscin purified from fenugreek killed HL-60 leukemia cells (100% hypodiploid nuclei at 10 µM) but was only weakly active against KATO III gastric cells ^{98Reference 98Hibasami H et al. · 2003In vitroProtodioscin isolated from fenugreek (Trigonella foenumgraecum L.) induces cell death and morphological change indicative of apoptosis in leukemic cell line H-60, but not in gastric cancer cell line KATO III — [in vitro]View study →}.

That selectivity index (above 8) and the cell-type discrimination matter for this class, because most saponin cytotoxicity results are membrane-disruption artefacts with no normal-cell comparator, and protodioscin is one of the few where the comparator was run.

The micromolar potencies are not reachable in plasma after an oral dose, and there is no human oncology data.

Tribulus, fenugreek and wild yam are the herbs that carry it.

Punicalagin inhibited proliferation of PC-3 and LNCaP prostate cancer cells and induced caspase-3/-8-mediated apoptosis, with an anti-angiogenic signal in the chick chorioallantoic membrane assay, and viability of normal BPH-1 prostate cells was not significantly affected at the same 10–100 µM concentrations ^{107Reference 107Adaramoye O et al. · 2017Punicalagin, a polyphenol from pomegranate fruit, induces growth inhibition and apoptosis in human PC-3 and LNCaP cellsView study →}.

Among ellagitannins from Terminalia calamansanai, punicalagin induced HL-60 apoptosis at IC50 74.8 µM, weaker than sanguiin H-4 (38.0 µM) and 2-O-galloylpunicalin (42.2 µM) from the same plant ^{108Reference 108Chen LG et al. · 2009In vitroEllagitannins from Terminalia calamansanai induced apoptosis in HL-60 cellsView study →}.

Punicalagins inhibit CYP1A1 with 5–10 times selectivity over CYP1B1, though urolithin A was the more potent CYP1B1 inhibitor in that study ^{109Reference 109Kasimsetty SG et al. · 2009Effects of pomegranate chemical constituents/intestinal microbial metabolites on CYP1B1 in 22Rv1 prostate cancer cellsView study →}.

There is no animal tumor study and no human trial, and the active concentrations of 10–100 µM belong to a molecule that is undetectable in human plasma.

Terminalia chebula is the herb that carries it.

Rosmanol has a moderate cell-line literature that converges on ROS-driven mitochondrial apoptosis, and it is consistently the most cytotoxic of the three main rosemary diterpenes, which is a statement about potency rather than selectivity.

Cheng et al. found 50 µM rosmanol produced a 51% apoptotic ratio in human colon adenocarcinoma COLO 205 cells at 24 h (IC₅₀ about 42 µM), engaging both apoptotic arms: the death-receptor arm (Fas and FasL upregulation, pro-caspase-8 and Bid cleavage) and the mitochondrial arm (Bax translocation to mitochondria, loss of membrane potential, cytochrome c and AIF release, pro-caspase-9 and -3 processing, PARP and DFF-45 cleavage) ^{110Reference 110Cheng AC et al. · 2011In vitroRosmanol potently induces apoptosis through both the mitochondrial apoptotic pathway and death receptor pathway in human colon adenocarcinoma COLO 205 cells — in vitroView study →}.

In breast cancer, rosmanol inhibited MCF-7 and MDA-MB-231 proliferation, arrested cells in S phase, and drove ROS-dependent apoptosis abolished by N-acetylcysteine, via loss of PI3K/AKT, ERK and JAK2/STAT3 signalling with a rise in p-p38 and PIAS3, and did not significantly affect normal human breast MCF-10A cells ^{111Reference 111Jiang D et al. · 2021In vitroRosmanol induces breast cancer cells apoptosis by regulating PI3K/AKT and STAT3/JAK2 signaling pathways — in vitroView study →}.

Acunha et al. tested carnosic acid, carnosol and rosmanol side by side on two colon cancer lines (HT-29, HCT116) and on undifferentiated and differentiated HepaRG hepatocyte-like cells at 10–100 µM: rosmanol exerted the strongest effect in both cancer lines and produced the greatest metabolome alteration in hepatocytes, followed by carnosol then carnosic acid, with the ranking of metabolic disturbance tracking the ranking of cytotoxicity, and differentiated hepatocytes were more resistant ^{112Reference 112Acunha T et al. · 2018In vitroMetabolomics study of early metabolic changes in hepatic HepaRG cells in response to rosemary diterpenes exposure — in vitroView study →}.

The active range of 20–50 µM sits against a 1.8 µM in-vivo ceiling ^{113Reference 113Wang L et al. · 2017AnimalDetermination and pharmacokinetic study of three diterpenes in rat plasma by UHPLC-ESI-MS/MS after oral administration of Rosmarinus officinalis L. extract — ratsView study →}, a 20- to 30-fold shortfall.

The one normal-cell comparator ^{111Reference 111Jiang D et al. · 2021In vitroRosmanol induces breast cancer cells apoptosis by regulating PI3K/AKT and STAT3/JAK2 signaling pathways — in vitroView study →} is qualitative, with no IC50 for the normal line, so there is no selectivity index.

The three in-vivo tumor models that exist come from a single 2024–2026 cluster of uncertain reliability ^{114,115,116Reference 114Chang J et al. · 2026AnimalProtective effects of rosmanol on DEN-induced lung cancer via targeting Nrf2/NF-κB/PI3K/Akt and COX-2/CYP2E1/VEGF pathways — ratsView study →Reference 115Tao W et al. · 2025In vitroRosmanol suppresses nasopharyngeal carcinoma cell proliferation and enhances apoptosis, the regulation of MAPK/NF-κB signaling pathway — in vitroView study →Reference 116Niu YB et al. · 2025AnimalRosmanol triggers apoptosis against MNNG-induced gastric carcinogenesis in rats through attenuation of the PI3K/AKT/HMGB1 signaling pathwayView study →}, and there is no clinical work.

Artecanin is a guaianolide sesquiterpene lactone with a modest in-vitro cytotoxicity signal from a single research group and no identified molecular target.

Tested as the isolate on HL-60 promyelocytic leukemia cells it sat within a four-lactone set spanning IC₅₀ 3.6–13.5 µM and was among the two less potent; it accumulated cells in G1 and G2/M at the expense of S phase, produced concentration-dependent nuclear condensation, activated caspase-9 and caspase-3, raised the Bax/Bcl-2 ratio and lowered CDK1 and cyclin B2 at 10 µM, an intrinsic mitochondrial apoptosis signature, and was substantially less active against normal human fibroblasts than against HL-60 ^{6Reference 6Molnár J et al. · 2016In vitroInvestigation of the antiproliferative properties of natural sesquiterpenes from Artemisia asiatica and Onopordum acanthium on HL-60 cells — in vitroView study →}.

An earlier paper from the same group rated it moderately active against A2780, A431, HeLa and MCF7 ^{7Reference 7Hajdú Z et al. · 2014In vitroAntiproliferative activity of Artemisia asiatica extract and its constituents on human tumor cell lines — in vitroView study →}.

The enzyme paper that mapped feverfew’s guaianolide pathway cites artecanin’s antiproliferative property on leukemia cells as part of the case for producing it biotechnologically ^{8Reference 8Liu Q et al. · 2018In vitroKauniolide synthase is a P450 with unusual hydroxylation and cyclization-elimination activity — in vitro enzyme characterisationView study →}.

Two papers, one research group and one plant source (Artemisia asiatica), with no independent replication.

All readouts are downstream, no target has been identified, and low-micromolar cytotoxicity is close to the default for any α-methylene-γ-lactone.

No animal study and no pharmacokinetics exist, so there is no basis for supposing the dish concentrations are reachable in vivo.

Parthenolide’s leukemia-stem-cell selectivity and its IKKβ Cys179 mechanism ^{9Reference 9Kwok BH et al. · 2001In vitroThe anti-inflammatory natural product parthenolide from the medicinal herb feverfew directly binds to and inhibits IκB kinase — in vitroView study →} belong to a different compound and have never been tested for artecanin.

Feverfew is the herb that carries it, as a minor guaianolide behind parthenolide.

Two different mechanisms sit under this molecule.

Directly, purified castalagin is a potent PARP1 inhibitor (IC50 0.86 µM, Ki 1.64 µM, mixed-type) and also inhibits DNA topoisomerase II, making it a dual DNA-repair/replication enzyme inhibitor in cells ^{20Reference 20Kamada Y et al. · 2018Castalagin and vescalagin purified from leaves of Syzygium samarangense (Blume) Merrill & L.MView study →}; the classic 57-tannin screen found hydrolysable tannins selectively cytotoxic to melanoma while inactive above 10 µg/mL against four other tumor lines ^{21Reference 21Kashiwada Y et al. · 1992Antitumor agents, 129View study →}, and the galloylated derivative 1-O-galloylcastalagin induces HL-60 apoptosis at IC50 12.5 µM ^{22Reference 22Yang LL et al. · 2000Induction of apoptosis by hydrolyzable tannins from Eugenia jambos L. on human leukemia cellsView study →}.

Indirectly, it acts through the gut microbiota: the one time castalagin was given to tumor-bearing animals its effect was entirely microbiota-dependent and absent in germ-free mice ^{23Reference 23Messaoudene M et al. · 2022A natural polyphenol exerts antitumor activity and circumvents anti-PD-1 resistance through effects on the gut microbiotaView study →}, so the enzyme inhibition, however clean, is not how the molecule works in vivo.

Its anticancer activity in a living animal is the immunomodulator one.

Camu camu is the herb that carries it.

Two 2024–25 studies give coclaurine a named molecular target.

In colorectal HCT116 cells it drives VDR nuclear localisation and expression, induces late apoptosis with PARP and caspase-3 cleavage, raises Bax/TP53, lowers BCL-2 and arrests growth in S-phase, and loses every one of those effects in CRISPR VDR-knockout cells, an unusually clean receptor-dependency demonstration for a plant alkaloid ^{24Reference 24Alghamdi HA et al. · 2025In vitroReticuline and coclaurine exhibit vitamin D receptor-dependent anticancer and pro-apoptotic activities in the colorectal cancer cell line HCT116 — in vitroView study →}.

In non-small-cell lung cancer it was identified by fractionation as the Stephania tetrandra constituent that suppresses EFHD2, a cisplatin-resistance protein, by disrupting FOXG1 binding at the EFHD2 promoter; it also downregulated NOX4–ABCC1 signalling and reduced stemness and metastatic behaviour, sensitising the cells to cisplatin ^{25Reference 25Hu SY et al. · 2024In vitroStephania tetrandra and its active compound coclaurine sensitize NSCLC cells to cisplatin through EFHD2 inhibition — in vitroView study →}.

Both are cell culture, neither has an animal model, neither published a potency figure comparable to attainable plasma levels, and no non-malignant comparator was tested.

No pharmacokinetic study of coclaurine exists, so it is unknown whether any of these concentrations is reachable in vivo.

Graviola, magnolia and cassia cinnamon are the herbs that carry it.

Costunolide induces apoptosis in leukemia lines through a ROS burst, mitochondrial permeability transition and cytochrome c release, a cascade blocked by NAC and cyclosporin A ^{26Reference 26Lee MG et al. · 2001In vitroCostunolide induces apoptosis by ROS-mediated mitochondrial permeability transition and cytochrome C release — in vitroView study →}, with ROS-dependent JNK activation upstream of Bcl-2 ^{27Reference 27Choi JH · 2009In vitroCostunolide-induced apoptosis in human leukemia cells — involvement of c-jun N-terminal kinase activation — in vitroView study →} and a caspase-8 receptor arm in NALM-6 cells ^{28Reference 28Kanno S et al. · 2008In vitroCostunolide-induced apoptosis is caused by receptor-mediated pathway and inhibition of telomerase activity in NALM-6 cells — in vitroView study →}.

It lowers hTERT transcription (hTR unchanged) in breast carcinoma lines and is now used as a pharmacological TERT tool compound ^{29,30Reference 29Choi SH et al. · 2005In vitroInhibitory effects of costunolide on the telomerase activity in human breast carcinoma cells — in vitroView study →Reference 30Ahmad F et al. · 2016In vitroNrf2-driven TERT regulates pentose phosphate pathway in glioblastoma — in vitro and mouse xenograftView study →}.

It inhibits VEGFR-2 (KDR/Flk-1) autophosphorylation while sparing VEGFR-1 ^{31Reference 31Jeong SJ et al. · 2002In vitroCostunolide, a sesquiterpene lactone from Saussurea lappa, inhibits the VEGFR KDR/Flk-1 signaling pathway — in vitro and mouse corneal micropocket assayView study →}, and it and parthenolide selectively reduce detyrosinated tubulin independent of NF-κB inhibition ^{32Reference 32Whipple RA et al. · 2013In vitroParthenolide and costunolide reduce microtentacles and tumor cell attachment by selectively targeting detyrosinated tubulin independent from NF-κB inhibition — in vitroView study →}, the cleanest mechanistic result in the set.

In-vivo support is thin: orthotopic MDA-MB-231 growth and metastasis reduction in mice ^{33Reference 33Choi YK et al. · 2013In vitroSaussurea lappa Clarke-derived costunolide prevents TNFα-induced breast cancer cell migration and invasion by inhibiting NF-κB activity — in vitro and orthotopic mouse modelView study →} and reduced tumor burden in DMBA hamster cheek-pouch carcinogenesis ^{34Reference 34Ohnishi M et al. · 1997AnimalInhibitory effects of dietary protocatechuic acid and costunolide on 7,12-dimethylbenz[a]anthracene-induced hamster cheek pouch carcinogenesis — animal modelView study →}.

An indiscriminate cysteine electrophile kills cells in culture by construction, so selectivity rather than potency is the unanswered question, and the class reviews say so ^{35Reference 35Merfort I · 2011ReviewPerspectives on sesquiterpene lactones in inflammation and cancer — reviewView study →}.

There is no human trial, no human pharmacokinetics, and no evidence that assay concentrations are reachable in tissue ^{36Reference 36Zhang J et al. · 2014AnimalPharmacokinetic study on costunolide and dehydrocostuslactone after oral administration of traditional medicine Aucklandia lappa Decne. by LC/MS/MS — animal model in ratsView study →}.

Feverfew and magnolia are the herbs that carry it.

Dioscin isolated from Dioscorea villosa reduced viability of MCF-7 and MDA-MB-231 cells, upregulated GATA3, shifted MDA-MB-231 toward epithelial morphology and cut invasion by 65% at 5.76 µM ^{66Reference 66Aumsuwan P et al. · 2016In vitroThe anticancer potential of steroidal saponin, dioscin, isolated from wild yam (Dioscorea villosa) root extract in invasive human breast cancer cell line MDA-MB-231 in vitroView study →}; other work reports caspase-independent apoptosis via AIF and maps inhibitory targets in HepG2 ^{67Reference 67Zhang YS et al. · 2018Molecular mechanism and inhibitory targets of dioscin in HepG2 cellsView study →}.

None of it distinguishes targeted cytotoxicity from detergent lysis.

No isolate study reports a selectivity index against normal cells; the D. villosa study compared two tumor lines with no normal comparator ^{66Reference 66Aumsuwan P et al. · 2016In vitroThe anticancer potential of steroidal saponin, dioscin, isolated from wild yam (Dioscorea villosa) root extract in invasive human breast cancer cell line MDA-MB-231 in vitroView study →}, cytotoxicity tracks the same sugar chain that causes haemolysis ^{68Reference 68Li W et al. · 2007Synthesis, cytotoxicity, and hemolytic activity of 6′-O-substituted dioscin derivativesView study →}, the authors of the whole-extract transcriptomic study named the detergent saponins an “interfering variable” and worked sub-lethally to avoid them ^{69Reference 69Mazzio E et al. · 2021Effects of wild yam root (Dioscorea villosa) extract on the gene expression profile of triple-negative breast cancer cellsView study →}, and the one paper claiming a receptor mechanism (estrogen receptor-β in prostate cancer) was retracted in 2021 ^{70Reference 70Tao X et al. · 2017Dioscin induces prostate cancer cell apoptosis through activation of estrogen receptor-βView study →}.

Wild yam and tribulus are the herbs that carry it.

Two isolate measurements exist: IC₅₀ 10.61 µM against osteosarcoma cells (against 6.43 µM for protodioscin and 6.90 µM for dioscin in the same run, with the study attributing the herb’s effect to those two rather than to protogracillin) ^{99Reference 99Hwang JT et al. · 2019Development of an oriental medicine discrimination method through analysis of steroidal saponins in Dioscorea nipponica Makino and their anti-osteosarcoma effectsView study →}, and group-level cytotoxicity against K562 leukemia cells shared with seven co-isolated furostanols ^{100Reference 100Hu K et al. · 1997Antineoplastic agentsView study →}.

Neither reported a selectivity index against normal cells, and steroidal saponins are surfactants whose cholesterol-dependent membrane permeabilisation is a standing source of false-positive cytotoxicity in this class ^{101Reference 101Podolak I et al. · 2010ReviewSaponins as cytotoxic agents: a reviewView study →}.

A structure–activity comparison in the same series found the F-ring, the feature that distinguishes protogracillin from gracillin, critical to what selectivity there is ^{83Reference 83Hu K · 2003In vitroThe cytotoxicity of methyl protoneogracillin (NSC-698793) and gracillin (NSC-698787), two steroidal saponins from the rhizomes of Dioscorea collettii var. hypoglauca, against human cancer cells in vitroView study →}.

At about 1,065 Da with rapid plasma clearance ^{102,103Reference 102Liao M et al. · 2016AnimalSimultaneous determination of four furostanol glycosides in rat plasma by UPLC-MS/MS and its application to PK study after oral administration of Dioscorea nipponica extractsView study →Reference 103Tang YN et al. · 2015AnimalUPLC-QTOF-MS identification of metabolites in rat biosamples after oral administration of Dioscorea saponins: a comparative studyView study →}, a sustained 10 µM tissue concentration is not a realistic oral exposure.

Tribulus is the herb that carries it.

Screening the NCI natural-products repository for G2 DNA-damage-checkpoint inhibitors flagged a ragweed extract, and assay-guided fractionation identified psilostachyins A and C as the actives; they also blocked cells in mitosis with aberrant spindles without inhibiting tubulin polymerisation in vitro, and removing the α,β-unsaturated carbonyl abolished activity, direct evidence that the mechanism is covalent Michael addition ^{104Reference 104Sturgeon CM et al. · 2005In vitroModulation of the G2 cell cycle checkpoint by sesquiterpene lactones psilostachyins A and C isolated from the common ragweed Ambrosia artemisiifolia — in vitroView study →}.

In human colon adenocarcinoma, psilostachyin inhibited the multidrug-resistant Colo320 line at IC50 5.29 µM but with only about 5-fold selectivity over normal MRC-5 fibroblasts, and did not inhibit ABCB1/P-gp ^{105Reference 105Kovács B et al. · 2022In vitroAntiproliferative and cytotoxic effects of sesquiterpene lactones isolated from Ambrosia artemisiifolia on human adenocarcinoma and normal cell lines — in vitroView study →}.

In murine BW5147 lymphoma it was cytotoxic through apoptosis but was outperformed on both potency and tolerability by psilostachyin C ^{106Reference 106Martino R et al. · 2015In vitroSesquiterpene lactones from Ambrosia spp. are active against a murine lymphoma cell line by inducing apoptosis and cell cycle arrest — in vitroView study →}.

All of it is in vitro, there is no animal tumor model, and the selectivity is poor enough that cytotoxic is the accurate word.

Cell-free, sanguiin H-6 inhibits the catalytic strand-passing activity of topoisomerase I (IC₅₀ 1 µM) and topoisomerase II (IC₅₀ 0.01 µM) irreversibly, by preventing covalent enzyme–DNA complexes rather than stabilising them, so it is not a cleavable-complex poison of the camptothecin/etoposide type; on natural chromatin potency falls to 5 µM and above 10 µM, and HeLa growth inhibition needs ED₅₀ 12 µM “without any marked selectivity” ^{117Reference 117Bastow KF et al. · 1993In vitroInhibition of DNA topoisomerases by sanguiin H-6, a cytotoxic dimeric ellagitannin from Sanguisorba officinalis — in vitroView study →}.

It blocks VEGF₁₆₅ binding to KDR/Flk-1 at about 0.3 µg/mL and VEGF-driven HUVEC proliferation at about 7.4 µg/mL ^{118Reference 118Lee SJ · 2005In vitroSanguiin H-6 blocks endothelial cell growth through inhibition of VEGF binding to VEGF receptor — in vitroView study →}, blocks tube formation at 6.25 µM ^{119Reference 119Park EH et al. · 2016In vitroAssessment of the anti-metastatic properties of sanguiin H-6 in HUVECs and MDA-MB-231 human breast cancer cells — in vitroView study →}, drives caspase-8/-3/PARP apoptosis with a raised Bax:Bcl-2 ratio in MCF-7 and MDA-MB-231 independently of the estrogen receptor ^{120Reference 120Park EJ et al. · 2017In vitroCytotoxic effect of sanguiin H-6 on MCF-7 and MDA-MB-231 human breast carcinoma cells — in vitroView study →}, caspase-8/BID and p38 MAPK apoptosis in A2780 ovarian cells at 20–40 µM ^{121Reference 121Lee D et al. · 2016In vitroInhibition of A2780 human ovarian carcinoma cell proliferation by a Rubus component, sanguiin H-6 — in vitroView study →}, and suppresses TGF-β1/Smad2/3-driven EMT in A549 ^{122Reference 122Ko H et al. · 2015In vitroSanguiin H6 suppresses TGF-β induction of the epithelial-mesenchymal transition and inhibits migration and invasion in A549 lung cancer — in vitroView study →}.

Every one of these is cell culture at 6–200 µM, and the molecule is undetectable in plasma ^{123Reference 123González-Barrio R et al. · 2010Clinical trialBioavailability of anthocyanins and ellagitannins following consumption of raspberries by healthy human volunteers and subjects with an ileostomyView study →}; Park’s own data show no effect on MDA-MB-231 viability up to 25 µM, with 200 µM needed to reduce it ^{119Reference 119Park EH et al. · 2016In vitroAssessment of the anti-metastatic properties of sanguiin H-6 in HUVECs and MDA-MB-231 human breast cancer cells — in vitroView study →}.

There is no xenograft, no in-vivo tumor model of the isolate and no clinical work.

Raspberry leaf is the herb that carries it.

Shatavarin IV inhibited AGS gastric-adenocarcinoma proliferation at IC50 2.46 µM under hyperglycaemic conditions, arresting cells in G0/G1, driving apoptosis by 36 h, suppressing migration and invasion, and shifting the MMP-9/TIMP-1 balance ^{124Reference 124Chatterjee A et al. · 2024In vitroShatavarin-IV, a steroidal saponin from Asparagus racemosus, inhibits cell cycle progression and epithelial-to-mesenchymal transition in AGS cells under hyperglycemic conditionsView study →}; earlier work found the 84.7%-pure isolate cytotoxic to MCF-7, HT-29 and A-498 ^{125Reference 125Mitra SK et al. · 2012In vitroShatavarins (containing Shatavarin IV) with anticancer activity from the roots of Asparagus racemosusView study →}.

No study in this literature has included a non-malignant comparator cell line, and no selectivity index has been reported.

For a spirostanol saponin, a surfactant, membrane permeabilisation is the default explanation for cytotoxicity, and it has not been excluded.

The one in-vivo tumor result (reduced tumor volume in Ehrlich-ascites-carcinoma mice at 250–500 mg/kg) used AR-2B, a fraction containing 5.05% shatavarin IV, rather than the isolate ^{125Reference 125Mitra SK et al. · 2012In vitroShatavarins (containing Shatavarin IV) with anticancer activity from the roots of Asparagus racemosusView study →}.

Shatavari is the herb that carries it.

In a screen of 45 ellagitannins against sarcoma-180 in mice, tellimagrandin II was the most active of the monomeric ellagitannins, and the structure–activity conclusion was that galloyl groups at glucose O-2 and O-3 are required; ellagitannins built on casuarictin or potentillin units were largely inactive ^{133Reference 133Miyamoto K et al. · 1993AnimalAntitumor activities of ellagitannins against sarcoma-180 in miceView study →}.

Recent cell work adds apoptosis and reduced invasion in ectopic endometrial cells via 3-aminopropanal accumulation and inflammasome activation ^{134Reference 134Fan W et al. · 2025Tellimagrandin II stimulates inflammasomes by causing an accumulation of 3-aminopropanal, which promotes apoptosis of endometriotic cells while inhibiting invasionView study →}.

The mouse work was a single intraperitoneal injection, which bypasses the absorption problem, it is thirty years old with no isolate follow-up, and no molecular target has been identified.

Meadowsweet, clove and peony are the herbs that carry it.

Terrestrosin D depolarises the mitochondrial membrane and induces caspase-independent apoptosis in PC-3 castration-resistant prostate cancer cells while lowering VEGF secretion, with xenograft support ^{135Reference 135Wei S et al. · 2014In vitroTerrestrosin D, a steroidal saponin from Tribulus terrestris L., inhibits growth and angiogenesis of human prostate cancer in vitro and in vivo — [animal model]View study →}, and separately inhibits PSMD1, a 19S proteasome regulatory subunit, stabilising P53 and driving autophagy plus apoptosis in triple-negative breast cancer ^{136Reference 136Jia LL et al. · 2025AnimalTerrestrosin D promotes autophagy and apoptosis of breast cancer cells through PSMD1-dependent activation of the P53 pathway — [animal model]View study →}.

The PSMD1–P53 axis is a novel named target.

Selectivity is open: in the prostate study the compound suppressed growth and induced apoptosis in normal human umbilical vein endothelial cells at the same concentrations, framed as anti-angiogenesis but indistinguishable from non-selective membrane and mitochondrial toxicity, and no selectivity index against normal cells was reported in either paper ^{135,136Reference 135Wei S et al. · 2014In vitroTerrestrosin D, a steroidal saponin from Tribulus terrestris L., inhibits growth and angiogenesis of human prostate cancer in vitro and in vivo — [animal model]View study →Reference 136Jia LL et al. · 2025AnimalTerrestrosin D promotes autophagy and apoptosis of breast cancer cells through PSMD1-dependent activation of the P53 pathway — [animal model]View study →}.

Spirostanol saponins are surfactants, and this is the standard confound of the assay class.

Tribulus is the herb that carries it.

Vescalagin preferentially inhibits the decatenation activity of topoisomerase IIα by a redox-independent, catalytic (non-poisoning) mechanism: siRNA knockdown of Top2α but not Top2β conferred resistance in CEM cells, and vescalagin produced no DNA double-strand breaks, instead preventing etoposide-driven Top2α–DNA covalent complexes, reported as the first catalytic inhibitor whose cytotoxicity is attributable at least partly to preferential Top2α inhibition ^{137Reference 137Auzanneau C et al. · 2012In vitroThe polyphenolic ellagitannin vescalagin acts as a preferential catalytic inhibitor of the α isoform of human DNA topoisomerase II — in vitroView study →}.

It is separately a PARP1 inhibitor (IC50 2.67 µM; its epimer castalagin is more potent at 0.86 µM), with cellular poly(ADP-ribosyl)ation attenuated below 5 µM ^{20Reference 20Kamada Y et al. · 2018Castalagin and vescalagin purified from leaves of Syzygium samarangense (Blume) Merrill & L.MView study →}, and is cytotoxic to oral squamous-carcinoma lines with relative sparing of normal oral cells ^{138Reference 138Orabi MAA et al. · 2021In vitroTwo new C-glycosidic ellagitannins and accompanying tannins from Lawsonia inermis leaves and their cytotoxic effects — in vitroView study →}.

Every number is low-micromolar in vitro, and intact vescalagin does not reach those concentrations in plasma ^{139Reference 139Piwowarski JP et al. · 2014In vitroRole of human gut microbiota metabolism in the anti-inflammatory effect of traditionally used ellagitannin-rich plant materials — ex vivo human faecal incubation and in vitro macrophage assayView study →}.

Camu camu is the herb that carries it.