Antioxidant

Carnosic acid is the defining lipophilic antioxidant of the Lamiaceae, and it works by two distinct routes at once.

Directly, it is a stoichiometric radical scavenger — peroxyl 2.7 × 10⁷ M⁻¹s⁻¹ (roughly 10–20× faster than carnosol), hydroxyl 5.9 × 10¹⁰ M⁻¹s⁻¹ — and inhibits lipid peroxidation in microsomal and liposomal systems more strongly than propyl gallate; with carnosol it accounts for over 90% of rosemary extract’s antioxidant capacity ^{6Reference 6Aruoma OI et al. · 1992In vitroAntioxidant and pro-oxidant properties of active rosemary constituents: carnosol and carnosic acid — in vitroView study →}.

Being lipid-soluble puts it inside membranes, which is why it outperforms the water-soluble rosmarinic acid in fat-phase applications and underperforms it in aqueous ones.

The direct action is sacrificial: EPR spin-probe work confirms carnosic acid, not carnosol, is the ROS quencher, and it is consumed into carnosol and other oxidized derivatives as it quenches — in vitro and in the living leaf ^{7Reference 7Loussouarn M et al. · 2017In vitroCarnosic acid and carnosol, two major antioxidants of rosemary, act through different mechanisms — in vitro and in plantaView study →}.

Indirectly, oxidation converts its catechol into an electrophilic ortho-quinone that S-alkylates specific cysteines on Keap1, releasing Nrf2 to switch on the antioxidant-response-element battery (glutathione synthesis, HO-1, NQO1); this raises brain reduced glutathione in vivo ^{8Reference 8Satoh T et al. · 2008In vitroCarnosic acid, a catechol-type electrophilic compound, protects neurons both in vitro and in vivo through activation of the Keap1/Nrf2 pathway via S-alkylation of targeted cysteines on Keap1 — miceView study →}.

The same redox activity cuts both ways: carnosic acid stimulated DNA damage in the iron-dependent bleomycin assay while scavenging hydroxyl radicals in the deoxyribose assay ^{6Reference 6Aruoma OI et al. · 1992In vitroAntioxidant and pro-oxidant properties of active rosemary constituents: carnosol and carnosic acid — in vitroView study →} — the classic dual antioxidant/pro-oxidant signature of a redox-active polyphenol.

Rosemary extract is an authorized EU food antioxidant (E392) whose specification and use limits are written as the sum of carnosic acid and carnosol ^{9,10Reference 9EFSA Panel on Food Additives et al. · 2018ReviewRefined exposure assessment of extracts of rosemary (E 392) from its use as food additive — review, including the Regulation (EU) No 231/2012 and JECFA specificationsView study →Reference 10European Commission · 2012ReviewCommission Regulation (EU) No 231/2012 laying down specifications for food additives listed in Annexes II and III to Regulation (EC) No 1333/2008 — monographView study →}.

Carnosol is not a radical scavenger, and carnosic acid is.

A stability-controlled plant-physiology study tracked both compounds through defined oxidative conditions and found that carnosic acid quenches ROS and is consumed doing so (confirmed by spin-probe EPR), while carnosol resists — but under oxidative conditions that generate no ROS, carnosol inhibited lipid peroxidation and carnosic acid did not ^{7Reference 7Loussouarn M et al. · 2017In vitroCarnosic acid and carnosol, two major antioxidants of rosemary, act through different mechanisms — in vitro and in plantaView study →}.

Two antioxidants, two mechanisms; this is the cleanest dissociation in the literature and the best evidence the pair are pharmacologically distinct.

Carnosol’s cellular antioxidant action is indirect and electrophilic: it interrupts the Nrf2–Keap1 protein–protein interaction, drives ARE-luciferase activity, and induces HO-1, eNOS and VE-cadherin in human microvascular endothelium under tert-butyl hydroperoxide stress ^{11Reference 11Li X et al. · 2019Carnosol as a Nrf2 activator improves endothelial barrier function through antioxidative mechanismsView study →}.

The same Keap1-cysteine S-alkylation mechanism underlies its phase-2 enzyme induction ^{12Reference 12Takahashi T et al. · 2009AnimalCarnosic acid and carnosol inhibit adipocyte differentiation in mouse 3T3-L1 cells through induction of phase 2 enzymes and activation of glutathione metabolismView study →}.

Route decides the result: intraperitoneal carnosol at 100–400 mg/kg raised rat liver GST 1.6–1.9× and quinone reductase 3.1–4.8×, but dietary carnosol at 0.01–1.0% raised GST not at all ^{13Reference 13Singletary KW · 1996AnimalRosemary extract and carnosol stimulate rat liver glutathione-S-transferase and quinone reductase activitiesView study →}, a result most citing reviews omit.

Punicalagin is the most potent in vitro antioxidant polyphenol of pomegranate juice, and commercial juices out-scored red wine and green tea across ABTS, DPPH, DMPD and FRAP in direct proportion to their punicalagin content ^{68Reference 68Gil MI et al. · 2000Antioxidant activity of pomegranate juice and its relationship with phenolic composition and processingView study →}.

Purified from husk it scavenges DPPH and ABTS at IC50 ≈ 110 and 150 µg/mL ^{69Reference 69Aguilar-Zárate P et al. · 2017Characterisation of pomegranate-husk polyphenols and semi-preparative fractionation of punicalaginView study →}, and its metal chelation suppresses Cu/Zn-induced amyloid-β(25–35) and PrP(106–126) aggregation in vitro ^{70Reference 70Navale GR et al. · 2023Effect of cycloastragenol and punicalagin on PrP(106–126) and Aβ(25–35) oligomerization and fibrillizationView study →}.

In human placental trophoblasts punicalagin — but not ellagic acid — reduced oxidative stress and stimulus-induced apoptosis, reproducing whole pomegranate juice’s effect ^{71Reference 71Chen B et al. · 2012Pomegranate juice and punicalagin attenuate oxidative stress and apoptosis in human placenta and in human placental trophoblastsView study →}.

The human study found the in-vivo generated metabolites — the only species that reach the blood — showed no significant antioxidant activity compared with punicalagin, and no relevant change appeared in any of 34 hematological and serobiochemical parameters over five days at ~4.4 g/day ^{72Reference 72Cerdá B et al. · 2004In vitroThe potent in vitro antioxidant ellagitannins from pomegranate juice are metabolised into bioavailable but poor antioxidant hydroxy-6H-dibenzopyran-6-one derivatives by the colonic microflora of healthy humansView study →}.

The antioxidant chemistry is real, and its venue is the food and the gut lumen rather than human plasma.

Salvianolic acid A carries six phenolic hydroxyls arranged as three catechols, and the field’s standard review names it the most potent antioxidant of the salvianolic acids — the counterpart to salvianolic acid B, which is the most abundant ^{81Reference 81Ho JH · 2011Salvianolic acids — small compounds with multiple mechanisms for cardiovascular protectionView study →}.

The antioxidant capacity is not merely stoichiometric: in a galactose-cataract rat model, 0.05% Sal A eye drops lowered lens malondialdehyde and hydrogen peroxide, raised protein and non-protein thiols, and — separately — inhibited aldose reductase in vitro, so the redox effect and a defined enzyme target run together in the same tissue ^{82Reference 82Du GH et al. · 1995AnimalPrevention of galactose-induced cataractogenesis in rats by salvianolic acid A [in Chinese]View study →}.

The whole-root data point in the same direction from the opposite end: danshen’s DPPH, hydroxyl-radical, superoxide and lipid-peroxidation scavenging rise in lockstep with total salvianolic acid content during drying ^{83Reference 83Zhou GJ et al. · 2014Post-harvest induced production of salvianolic acids and significant promotion of antioxidant properties in roots of Salvia miltiorrhiza (Danshen)View study →}.

No head-to-head IC50 against salvianolic acid B has been published, so “Sal A is more potent” is consensus rather than a measured ratio.

At sub-1% oral bioavailability ^{84Reference 84Sun J et al. · 2018AnimalSome pharmacokinetic parameters of salvianolic acid A following single-dose oral administration to ratsView study →} direct systemic radical scavenging from a decoction is not a plausible human mechanism, and the cataract result works because the drops were applied to the eye.

Structurally Sal B is a caffeic-acid tetramer — three danshensu units plus one caffeic-acid unit, 718.6 Da, carrying eight phenolic hydroxyls against rosmarinic acid’s four.

That doubling is why it out-scavenges its own biosynthetic precursor.

The cleanest demonstration is agronomic rather than pharmacological: across a thermal dehydration of danshen root, Sal B rose from 0.01% to 5.51% and antioxidant activity improved in four separate assays (lipid peroxidation, DPPH, hydroxyl, superoxide) in step with it — IC50s falling from 4.85→2.69, 7.75→0.43, 2.57→1.13 and 17.25→1.10 mg/mL respectively ^{83Reference 83Zhou GJ et al. · 2014Post-harvest induced production of salvianolic acids and significant promotion of antioxidant properties in roots of Salvia miltiorrhiza (Danshen)View study →}.

Nrf2/HO-1 induction is the recurring cell-level antioxidant mechanism in the cardiovascular reviews ^{85,86Reference 85He G et al. · 2023ReviewSalvianolic acid B — a review of pharmacological effects, safety, combination therapy, new dosage forms and novel drug delivery routesView study →Reference 86Shao Y et al. · 2025Salvianolic acid B — a promising cardioprotective agentView study →}.

At 1–2% oral bioavailability ^{87,88Reference 87Wu YT et al. · 2006AnimalBioavailability of salvianolic acid B in conscious and freely moving ratsView study →Reference 88Gao DY et al. · 2009Bioavailability of salvianolic acid B and effect on blood viscosities after oral administration of salvianolic acids in beagle dogsView study →} direct systemic radical scavenging from an oral dose is not a plausible in-vivo mechanism.

The antioxidant activity is the foundation of the organ-protective effects in parenteral and animal settings rather than a standalone oral claim.

Salvianolic acids A and B are potent free-radical scavengers by virtue of multiple catechol groups, and the family’s vascular protection has been characterized in vascular endothelial cells, aortic smooth muscle and cardiomyocytes: reduced ROS, reduced leukocyte–endothelial adherence, suppressed matrix-metalloproteinase expression, and — a distinct mode for a polyphenol — competitive binding to target proteins that interrupts protein–protein interactions ^{81,89Reference 81Ho JH · 2011Salvianolic acids — small compounds with multiple mechanisms for cardiovascular protectionView study →Reference 89Wu Y et al. · 2020The effect of salvianolic acid on vascular protection and possible mechanismsView study →}.

Potency in this family tends to track oligomer size, since each added phenylpropanoid unit brings more free catechol hydroxyls.

The evidence is in vitro and animal, from two narrative reviews ^{81,89Reference 81Ho JH · 2011Salvianolic acids — small compounds with multiple mechanisms for cardiovascular protectionView study →Reference 89Wu Y et al. · 2020The effect of salvianolic acid on vascular protection and possible mechanismsView study →}, with no standalone human antioxidant outcome.

Polyphenol radical-scavenging rank orders largely track the count of free phenolic hydroxyls, so much of this is generic to large catechol oligomers rather than distinguishing pharmacology.

Tanshinone IIA induces the Nrf2/ARE antioxidant response, raising superoxide dismutase and lowering malondialdehyde ^{101Reference 101Guo R et al. · 2020ReviewPharmacological activity and mechanism of tanshinone IIA in related diseases — reviewView study →}.

This is the mechanism that anchors both major organ-protection literatures rather than a standalone claim, and it is unusually well replicated in pooled animal data: - Myocardial ischemia–reperfusion: meta-analysis of 15 rat studies — significantly increased SOD, decreased MDA, reduced myocardial infarct area, with effect size varying by ischemia duration ^{102Reference 102Zhang X et al. · 2024Meta-analysisEfficacy of tanshinone IIA in rat models with myocardial ischemia-reperfusion injury — systematic mini-review and meta-analysis of 15 studiesView study →}. - Cerebral ischemia: meta-analysis of 13 rodent studies — raised brain SOD, lowered GFAP and HMGB1, inhibited brain NF-κB activation ^{103Reference 103Xie X et al. · 2023Meta-analysisThe protective effect of tanshinone IIA on cerebral ischemic injury in animal models — systematic review and meta-analysis of 13 studiesView study →}.

The redox story has a second face: as a quinone, tanshinone IIA is reduced by intestinal NQO1 to an unstable catechol that autoxidises back to the parent, generating hydrogen peroxide ^{104Reference 104Hao H et al. · 2007In vitroIdentification of a novel intestinal first pass metabolic pathway: NQO1 mediated quinone reduction and subsequent glucuronidation — in vivo and in vitro metabolism studyView study →}.

The same electrophilicity that induces Nrf2 also drives redox cycling.

There is no human antioxidant-biomarker study of the molecule.

Theaflavins are stoichiometric radical scavengers, and the benzotropolone core plus the pyrogallol and catechol hydroxyls make them better than EGCG in head-to-head chemiluminescence assays.

The activity is congener-specific rather than uniform: theaflavin-3,3’-digallate is the strongest scavenger of hydrogen peroxide and the hydroxyl radical, while the ungalloylated theaflavin (TF1) is the best superoxide scavenger, and theaflavin-3’-gallate performs broadly, including protection against hydroxyl-radical-induced DNA damage ^{109Reference 109Wu YY et al. · 2011Evaluation of the antioxidant effects of four main theaflavin derivatives through chemiluminescence and DNA damage analysesView study →}.

In cells, theaflavins also protect hepatocyte and hepatoma lines from oxidative and benzo[a]pyrene-induced DNA damage, partly by suppressing CYP1A1 ^{110Reference 110Feng Q et al. · 2002Black tea polyphenols, theaflavins, prevent cellular DNA damage by inhibiting oxidative stress and suppressing cytochrome P450 1A1 in cell culturesView study →}.

At ~2 nM plasma ^{111Reference 111Mulder TP et al. · 2001Analysis of theaflavins in biological fluids using liquid chromatography-electrospray mass spectrometryView study →} direct tissue scavenging is not a plausible in-vivo mechanism — the only human redox datapoint is a raised GSH:GSSG ratio after nine days of a mixed theaflavin-enriched extract in an 18-person exercise trial ^{112Reference 112Arent SM et al. · 2010RCTThe effects of theaflavin-enriched black tea extract on muscle soreness, oxidative stress, inflammation, and endocrine responses to acute anaerobic interval training: a randomized, double-blind, crossover studyView study →}.

In cell-culture medium at pH 7.4 the same molecules act as prooxidants, generating hydrogen peroxide and depleting glutathione; catalase and pyruvate abolish their cytotoxicity ^{113Reference 113Babich H et al. · 2008Theaflavin-3-gallate and theaflavin-3’-gallate, polyphenols in black tea with prooxidant propertiesView study →}, so a good share of the in-vitro antioxidant and anticancer literature is reporting an assay artefact.

Across three strawberry cultivars’ agro-industrial by-products agrimoniin was the principal polyphenol (0.38–1.56 g/kg) and showed the highest correlation with antioxidant capacity of any compound measured (R² 0.87, p < 0.001) ^{1Reference 1Villamil-Galindo E et al. · 2021Strawberry agro-industrial by-products as a source of bioactive compounds — effect of cultivar on the phenolic profile and the antioxidant capacityView study →}; across 90 Fragaria × ananassa cultivars, ellagitannins with agrimoniin in particular had the greatest single influence on ABTS activity ^{2Reference 2Nowicka A et al. · 2019Comparison of polyphenol content and antioxidant capacity of strawberry fruit from 90 cultivars of Fragaria × ananassa DuchView study →}.

The isolate scavenges DPPH (IC₅₀ 13.0 µM) ^{3Reference 3Kim SB et al. · 2017In vitroPhytochemical analysis of Agrimonia pilosa Ledeb, its antioxidant activity and aldose reductase inhibitory potential — in vitro and ex vivoView study →}, H₂O₂ (SC₅₀ 0.20 µM), superoxide and hypochlorous acid concentration-dependently ^{4Reference 4Bazylko A et al. · 2013In vitroIn vitro antioxidant and anti-inflammatory activities of extracts from Potentilla recta and its main ellagitannin, agrimoniinView study →}, and chelates iron ^{5Reference 5Fedotcheva TA et al. · 2021In vitroNew properties and mitochondrial targets of polyphenol agrimoniin as a natural anticancer and preventive agent — in vitroView study →}.

The chemistry is generic polyphenol redox activity shared with every large hydrolysable tannin, and the apportionment is a correlation within a mixture.

Castalagin is routinely the largest identified phenolic in the extracts whose antioxidant capacity is being measured — camu-camu seed ^{14Reference 14Fidelis M et al. · 2020Camu-camu seed (Myrciaria dubia) — from side stream to an antioxidant, antihyperglycemic, antiproliferative, antimicrobial, antihemolytic, anti-inflammatory, and antihypertensive ingredientView study →}, jabuticaba seed ^{15Reference 15Fidelis M et al. · 2020Response surface optimization of phenolic compounds from jabuticaba (Myrciaria cauliflora [Mart.] O.Berg) seeds: antioxidant, antimicrobial, antihyperglycemic, antihypertensive and cytotoxic assessmentsView study →}, Kakadu plum fruit powder during colonic fermentation ^{16Reference 16Adiamo OQ et al. · 2024In-vitro colonic fermentation of Kakadu plum (Terminalia ferdinandiana) fruit powder: microbial biotransformation of phenolic compounds and cytotoxicityView study →} — and ellagitannins are documented metal chelators (Al(III) binding sites mapped ^{17Reference 17Frešer F et al. · 2026Ellagitannins as Al(III) chelators: determined binding sites with a predictive modelView study →}); iron chelation is one of the two proposed explanations for how castalagin protects R. bromii at the bacterial surface ^{18Reference 18Messaoudene M et al. · 2022A natural polyphenol exerts antitumor activity and circumvents anti-PD-1 resistance through effects on the gut microbiotaView study →}.

This is an extract-level property credited to a molecule that is almost never assayed as the isolate.

It is real in the gut lumen, where the concentration exists, and it is not a systemic antioxidant claim.

Casuarinin protected cultured MDCK kidney cells from H₂O₂-induced oxidative stress and DNA oxidative damage and prevented intracellular glutathione depletion, more effectively than trolox ^{24Reference 24Chen CH et al. · 2004Casuarinin protects cultured MDCK cells from hydrogen peroxide-induced oxidative stress and DNA oxidative damageView study →}.

It showed the strongest DPPH scavenging of the phenolics isolated from Quercus acutissima bark ^{25Reference 25Kim EB et al. · 2022Antiacne and anti-inflammatory effects of phenolic compounds from Quercus acutissimaView study →} and strong dose-dependent antioxidant activity in Osbeckia octandra ^{26Reference 26Perera HDSM et al. · 2025In vitroInvestigating the antioxidant, pro-inflammatory enzyme inhibitory, neuroprotective, and antineuroinflammatory bioactives of Osbeckia octandra using in vitro and in silico approachesView study →}.

The in-vivo gastric study contributes matching tissue markers — glutathione +194%, catalase +586%, malondialdehyde −56% in rat gastric mucosa ^{27Reference 27Al-Sayed E et al. · 2020AnimalProtective role of casuarinin from Melaleuca leucadendra against ethanol-induced gastric ulcer in ratsView study →}.

Cell-free and cell-culture antioxidant activity in a non-absorbed polyphenol does not establish a systemic antioxidant effect.

Cryptotanshinone’s antioxidant action is induction rather than scavenging, which fits a molecule that is itself a redox-cycling quinone.

It works through the Keap1/Nrf2 axis in cigarette-smoke COPD ^{28Reference 28Song H et al. · 2023AnimalCryptotanshinone alleviates lipopolysaccharide and cigarette smoke-induced chronic obstructive pulmonary disease in mice via the Keap1/Nrf2 axis — mouse in vivo studyView study →}, through AMPK/SIRT1 and Nrf2 in ethanol-induced liver injury ^{29Reference 29Nagappan A et al. · 2019AnimalCryptotanshinone from the Salvia miltiorrhiza Bunge attenuates ethanol-induced liver injury by activation of AMPK/SIRT1 and Nrf2 signaling pathways — mouse in vivo studyView study →}, and through PI3K/AKT/Nrf2 with SLC7A11/GPX4 ferroptosis signaling in cerebral ischemia–reperfusion ^{30Reference 30Gao N et al. · 2025AnimalCryptotanshinone alleviates cerebral ischemia reperfusion injury by regulating ferroptosis through the PI3K/AKT/Nrf2 and SLC7A11/GPX4 signaling pathway — rodent in vivo studyView study →}; it reduces MPP⁺-induced oxidative stress and apoptosis in SH-SY5Y cells, with STAT3 implicated ^{31Reference 31Wang Q et al. · 2022In vitroCryptotanshinone ameliorates MPP⁺-induced oxidative stress and apoptosis of SH-SY5Y neuroblastoma cells: the role of STAT3 in Parkinson’s disease — in vitro studyView study →}.

This is a mechanism operating inside other disease models rather than a standalone action, and its antibacterial activity is the mirror image, where the same molecule is a prooxidant.

Curcumin’s antioxidant action is mostly indirect and inductive rather than simple radical scavenging: it activates the Nrf2/ARE program, inducing heme oxygenase-1 and the cell’s own antioxidant enzymes.

That is a more durable mechanism than stoichiometric scavenging, and it overlaps heavily with the anti-inflammatory story (Nrf2 and NF-κB are reciprocally regulated).

The same bioavailability limit applies: the effect is well established in cells and animals, and thinly evidenced as a clinical antioxidant outcome in people.

Danshensu has two antioxidant mechanisms.

Direct scavenging is real but modest: in a head-to-head, danshensu and salvianolic acid B both out-scavenged vitamin C on hydroxyl, superoxide, DPPH and ABTS radicals, while both were weaker than vitamin C on iron chelation and hydrogen peroxide; the authors’ explicit conclusion was that “salvianolic acid B is superior to danshensu”, and that the condensation of danshensu with caffeic acid is what drives the activity ^{32Reference 32Zhao GR et al. · 2008Characterization of the radical scavenging and antioxidant activities of danshensu and salvianolic acid BView study →}.

One catechol against Sal B’s four.

The Nrf2 mechanism is the stronger claim.

In mouse transient MCAO, sodium danshensu binds Nrf2 directly — confirmed by docking, molecular dynamics and surface plasmon resonance — raising Nrf2 nuclear expression and downstream SLC7A11, GPX4, FTH1 and HO-1 while lowering ACSL4 and TFRC, cutting Fe²⁺, ROS and MDA and raising glutathione; in Nrf2-deficient mice the protection was partially lost ^{33Reference 33Guo C et al. · 2025Therapeutic time window of sodium danshensu on cerebral ischemia and its mechanism of inhibiting oxidative stress and ferroptosis through the Nrf2 pathwayView study →}.

The same Nrf2 route carries the cisplatin-nephrotoxicity result, with NF-κB suppression alongside ^{34Reference 34Yu C et al. · 2021Danshensu attenuates cisplatin-induced nephrotoxicity through activation of the Nrf2 pathway and inhibition of NF-κBView study →}, and Keap1-Nrf2/NQO1 the doxorubicin-cardiotoxicity result ^{35Reference 35Zhang J et al. · 2019Therapeutic potentials and mechanisms of the Chinese traditional medicine DanshensuView study →}.

At herb level, across six danshen water extracts made under different heating regimes, multiple linear regression identified danshensu as the most significant marker for both the antioxidant and the vasodilation effects of the extract, with antioxidant and relaxation effects correlating at r = 0.895-0.977 ^{36Reference 36Zhou X et al. · 2012Danshensu is the major marker for the antioxidant and vasorelaxation effects of danshen (Salvia miltiorrhiza) water-extracts produced by different heat water-extractionsView study →}.

That is correlational, strong evidence of contribution and weak evidence of sufficiency, but it is the best available link between this molecule and a traditional preparation.

Epirosmanol is a competent lipophilic chain-breaking antioxidant, and it has independent replication on purified compound across four groups and eighteen years.

Haraguchi et al. reached it by bioassay-directed fractionation of rosemary leaf — the activity led them to the molecule — and showed it inhibited superoxide production in the xanthine/xanthine oxidase system and completely blocked NADH- and NADPH-driven mitochondrial and microsomal lipid peroxidation at 3–30 µM ^{42Reference 42Haraguchi H et al. · 1995Inhibition of lipid peroxidation and superoxide generation by diterpenoids from Rosmarinus officinalisView study →}.

Zeng et al. measured inhibition of Cu²⁺-mediated human LDL oxidation and apo-B modification at IC₅₀ 7–10 µmol/L, with ESR tracing the mechanism to lipid-free-radical scavenging ^{43Reference 43Zeng HH et al. · 2001Antioxidant properties of phenolic diterpenes from Rosmarinus officinalisView study →}.

Miura et al. ran the oil stability index across a large sage/thyme panel and placed epirosmanol — with carnosol, rosmanol, isorosmanol, galdosol and carnosic acid — at activity “comparable to that of α-tocopherol”, explicitly attributing it to the ortho-dihydroxy group the whole series shares ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →}.

The original 1984 isolation paper already had it at roughly four times BHA and BHT in a lard substrate ^{45Reference 45Nakatani N · 1984Two antioxidative diterpenes from rosemary (Rosmarinus officinalis L.) and a revised structure for rosmanolView study →}.

The mechanistic reading is a class property rather than a molecular signature.

These papers report epirosmanol as one name in a list every time, and the SAR conclusion in ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →} is that the catechol does the work — which every member of the carnosic acid cascade carries.

Where the abietanes were discriminated, at the Nrf2/antioxidant-response element, epirosmanol was weaker than its own C-7 epimer rosmanol, and the authors concluded that “the stereochemistry of the hydroxyl group at C-7 is important” ^{46Reference 46Fischedick JT et al. · 2013Structure activity relationship of phenolic diterpenes from Salvia officinalis as activators of the nuclear factor E2-related factor 2 pathwayView study →}.

It was not carried into that paper’s confirmatory experiments (Nrf2-dependent gene expression, H₂O₂ neuroprotection) — those were run on carnosol and carnosaldehyde only.

So the indirect/Nrf2 arm of antioxidant action stops at a reporter readout for this molecule.

Every measurement is cell-free or ex-vivo (lard, isolated mitochondria and microsomes, isolated LDL) — there is no cellular antioxidant assay and no animal oxidative-stress model.

There is no pharmacokinetic study at all, so no exposure achievable in a person is known.

And the active range overlaps the cytotoxic one (7–10 µM active ^{43Reference 43Zeng HH et al. · 2001Antioxidant properties of phenolic diterpenes from Rosmarinus officinalisView study →} vs ~14 µM cytotoxic ^{47Reference 47Llurba Montesino N et al. · 2021Salvia officinalis L.: antitrypanosomal activity and active constituents against Trypanosoma brucei rhodesienseView study →}), so “antioxidant at 10 µM” is a chemical observation rather than a therapeutic one.

Galdosol is a competent lipophilic chain-breaking antioxidant, and this is the only action attributed to it.

Miura, Kikuzaki and Nakatani isolated eleven abietane diterpenoids, three apianane terpenoids, an anthraquinone and eight flavonoids from sage leaf and evaluated them by the oil stability index in a methyl-linoleate/silicone-oil substrate at 90 °C. Carnosol, rosmanol, epirosmanol, isorosmanol, galdosol and carnosic acid all showed “remarkably strong activity, which was comparable to that of α-tocopherol”, an order of magnitude above the flavonoids in the same panel, which matched BHT. DPPH radical scavenging reproduced the rank order, and the authors attributed the activity explicitly to the ortho-dihydroxy groups these compounds share ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →}.

The mechanistic reading is a class property rather than a molecular signature.

Every member of the carnosic acid cascade carries the catechol, and galdosol appears as one name in a list of six — never singled out, never ranked first, never carried into a confirmatory experiment.

It retains the catechol despite being the most oxidized member of the series, which is why the activity survives to this end of the cascade, and that is the one mechanistic point specific to it.

The evidence is cell-free only.

There is no cellular antioxidant assay, no Nrf2/ARE reporter data (galdosol was not in the sage diterpene SAR panel that produced those for its siblings), no animal oxidative-stress model, no human biomarker data, and no plant concentration figure — so there is no basis for saying galdosol contributes meaningfully to any real preparation’s antioxidant capacity.

As a chemical antioxidant geraniin is potent — DPPH IC₅₀ 0.92–1.27 µM, hydroxyl-radical IC₅₀ 0.11–1.44 µM, superoxide IC₅₀ 2.65 µM and xanthine-oxidase inhibition at IC₅₀ 30.49 µM for the isolate ^{48Reference 48Lin SY et al. · 2008Antioxidant, anti-semicarbazide-sensitive amine oxidase, and anti-hypertensive activities of geraniin isolated from Phyllanthus urinariaView study →} — and it was the strongest nitric-oxide scavenger among the compounds isolated from amla fruit rind ^{49Reference 49Kumaran A · 2006Nitric oxide radical scavenging active components from Phyllanthus emblica LView study →}.

In vivo it restored protein carbonyls, SOD, glutathione peroxidase/reductase and the GSH/GSSG ratio in high-fat-diet rats at 10–50 mg/kg ^{50Reference 50Chung APYS et al. · 2018AnimalGeraniin protects high-fat diet-induced oxidative stress in Sprague Dawley ratsView study →}.

The active species in vivo is not the parent molecule: after oral geraniin, seven dibenzopyran (urolithin-type) microbial metabolites appear, four of which out-perform intact geraniin and corilagin in ORAC, and plasma antioxidant capacity tracks metabolite concentration rather than parent ^{51Reference 51Ito H · 2011Metabolites of the ellagitannin geraniin and their antioxidant activitiesView study →}.

Geraniin is an antioxidant precursor.

A two-arm antioxidant: it scavenges superoxide with an IC50 of ~1.31 µM — essentially equivalent to gallic acid in the tube — and separately induces SOD, catalase and glutathione peroxidase ^{52,53Reference 52Janarthanam VA et al. · 2025In vitroIn vitro and in silico evaluation of the antioxidant and anticancer potential of hamamelitannin from Hamamelis virginiana — in vitro/in silicoView study →Reference 53Masaki H et al. · 1995In vitroProtective activity of hamamelitannin on cell damage induced by superoxide anion radicals in murine dermal fibroblasts — in vitroView study →}.

The interesting part is cellular rather than chemical: at matched scavenging potency it protects dermal fibroblasts at half gallic acid’s concentration, and under UVB gives 72.6% survival vs 35.5%, which the authors attribute to the hamamelose sugar anchoring the molecule at the cell surface — a delivery advantage, not a redox one ^{53,54Reference 53Masaki H et al. · 1995In vitroProtective activity of hamamelitannin on cell damage induced by superoxide anion radicals in murine dermal fibroblasts — in vitroView study →Reference 54Masaki H et al. · 1995In vitroProtective activity of hamamelitannin on cell damage of murine skin fibroblasts induced by UVB irradiation — in vitroView study →}.

All of it is in vitro, and the same galloylated tannin class turns mildly pro-oxidant and cytotoxic to skin cells above the effective range ^{55Reference 55Touriño S et al. · 2008In vitroHighly galloylated tannin fractions from witch hazel (Hamamelis virginiana) bark: electron transfer capacity, in vitro antioxidant activity, and effects on skin-related cells — in vitroView study →}.

The role is contributor rather than primary because witch hazel’s antioxidant capacity is shared with, and largely carried by, its polymeric proanthocyanidins.

Dopamine’s o-dihydroxy (catechol) ring makes it a direct radical scavenger, and this is the one action where plant dopamine is doing real chemistry at real concentrations.

In Cavendish banana it was identified as the principal water-soluble antioxidant, out-performing glutathione, BHA, BHT, luteolin, quercetin and catechin at suppressing linoleic-acid oxygen uptake and scavenging DPPH, and matching gallocatechin gallate and ascorbate — at 80–560 mg/100 g in peel and 2.5–10 mg/100 g in pulp ^{56Reference 56Kanazawa K · 2000High content of dopamine, a strong antioxidant, in Cavendish bananaView study →}.

The same chemistry underlies its documented role in plant tolerance to drought, salt and nutrient stress ^{57,58Reference 57Liu Q et al. · 2020ReviewFunctions of dopamine in plants: a reviewView study →Reference 58Kulma A · 2007Catecholamines are active compounds in plantsView study →}.

The outcomes are in-vitro assay and plant physiology only, and because swallowed dopamine is largely conjugated at the gut wall ^{59Reference 59Eisenhofer G et al. · 1999Dopamine sulphate: an enigma resolvedView study →}, any dietary antioxidant contribution is plausible in the gut lumen and unproven systemically.

No human antioxidant endpoint rests on it.

Isorosmanol has two mechanisms, and only the second is its own.

The class mechanism: isorosmanol carries the ortho-dihydroxy (catechol) ring shared by the whole carnosic acid cascade, and it performs like the rest of the series in conventional lipid substrates.

In the oil-stability-index panel, carnosol, rosmanol, epirosmanol, isorosmanol, galdosol and carnosic acid all showed activity “comparable to that of α-tocopherol”, with DPPH scavenging giving the same ranking, and the authors attributed the activity explicitly to the shared catechol ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →}.

The 1984 isolation paper had already placed isorosmanol and epirosmanol at roughly four times BHA and BHT in a lard substrate ^{45Reference 45Nakatani N · 1984Two antioxidative diterpenes from rosemary (Rosmarinus officinalis L.) and a revised structure for rosmanolView study →}.

That part is a scaffold property and does not distinguish the molecule.

The mechanism that does: Higashino, Karatsu and Masuda asked what happens to these diterpenes in the presence of a cysteine thiol.

Of the polyphenols tested, only carnosol and isorosmanol enhanced antioxidant capacity against radical-induced lipid oxidation, and both did so catalytically — using the thiol as a reduction source so the polyphenol is regenerated rather than consumed ^{60Reference 60Higashino H et al. · 2024Catalytic antioxidant activity of two diterpenoid polyphenols of rosemary, carnosol, and isorosmanol, against lipid oxidation in the presence of cysteine thiolView study →}.

Density-functional calculations attributed the difference to the ortho-quinones of carnosol and isorosmanol being less stable than those of the other diterpenes, and to the regioselectivity of their reaction with thiol ^{60Reference 60Higashino H et al. · 2024Catalytic antioxidant activity of two diterpenoid polyphenols of rosemary, carnosol, and isorosmanol, against lipid oxidation in the presence of cysteine thiolView study →}.

A 2025 follow-up from the same laboratory reproduced the effect in an aqueous micellar system ^{61Reference 61Hamasaki H · 2025Enhancement of antioxidant activity of isorosmanol in aqueous micellar solution by a cysteine thiolView study →}.

This matters mechanistically: carnosic acid is a sacrificial antioxidant, consumed in the act.

A catalytic antioxidant is a different kind of object, and isorosmanol is one of only two compounds in this series shown to behave that way — and the two are exactly the two sharing the C-20→C-7 lactone.

Every one of these measurements is cell-free — lard, model oil, micellar solution — with an added thiol.

There is no cellular antioxidant assay outside the muscle study ^{62Reference 62Berger A et al. · 2025Advances in identifying functional groups in carnosol analogues to address their efficacy in skeletal muscle functionView study →}, no animal oxidative-stress model, no human data, and no demonstration that physiological glutathione supports the same turnover.

Direct chemical antioxidant activity, demonstrated on the isolated compound.

In HPLC-coupled assays of Rubus matsumuranus leaf metabolites, lambertianin C — with lambertianin A, sanguiin H-6 and sanguiin H-11 — was an effective DPPH radical scavenger and Fe²⁺ chelator ^{63Reference 63Kashchenko NI et al. · 2021Metabolites of Siberian raspberries — LC-MS profile, seasonal variation, antioxidant activity and thermal stability of Rubus matsumuranus phenolomeView study →}; isolated raspberry/cloudberry ellagitannin fractions dominated by sanguiin H-6 and lambertianin C were highly effective radical scavengers and protected human LDL and methyl-linoleate emulsions from oxidation, though only weakly in bulk oil ^{20Reference 20Kähkönen M et al. · 2012In vitroAntioxidant activity of isolated ellagitannins from red raspberries and cloudberries — in vitroView study →}.

The molecule never reaches plasma, so this cannot be a systemic antioxidant action in a person — any in-vivo antioxidant role would run through ellagic acid and the urolithins, and even that is gated by urolithin metabotype ^{64,65Reference 64Tomás-Barberán FA et al. · 2014Ellagic acid metabolism by human gut microbiota — consistent observation of three urolithin phenotypes in intervention trials, independent of food source, age and health statusView study →Reference 65Cortés-Martín A et al. · 2018The gut microbiota urolithin metabotypes revisited — the human metabolism of ellagic acid is mainly determined by agingView study →}.

In whole red raspberry the bulk of antioxidant capacity was attributed to sanguiin H-6, vitamin C and the anthocyanins rather than to the trimer ^{66Reference 66Mullen W et al. · 2002Ellagitannins, flavonoids, and other phenolics in red raspberries and their contribution to antioxidant capacity and vasorelaxation propertiesView study →}.

Rosmanol is a competent lipophilic chain-breaking antioxidant, replicated on purified compound by four independent groups across four decades, and it is the strongest member of its own series in the two studies that ranked the series rather than lumping it.

Haraguchi et al. reached it by bioassay-directed fractionation of rosemary leaf — the activity led to the molecule — and showed it inhibited superoxide production in the xanthine/xanthine oxidase system and completely blocked NADH- and NADPH-driven mitochondrial and microsomal lipid peroxidation at 3–30 µM, alongside carnosic acid, carnosol and epirosmanol ^{42Reference 42Haraguchi H et al. · 1995Inhibition of lipid peroxidation and superoxide generation by diterpenoids from Rosmarinus officinalisView study →}.

Zeng et al. measured inhibition of Cu²⁺-mediated human LDL oxidation and apo-B modification at IC₅₀ 7–10 µmol/L — reported for the carnosol / rosmanol / epirosmanol set collectively, not per compound — with ESR tracing the mechanism to lipid-free-radical scavenging ^{43Reference 43Zeng HH et al. · 2001Antioxidant properties of phenolic diterpenes from Rosmarinus officinalisView study →}.

Miura et al. ran the oil stability index across a large sage/thyme panel and placed rosmanol, with carnosol, epirosmanol, isorosmanol, galdosol and carnosic acid, at activity “comparable to that of α-tocopherol”, attributing it explicitly to the ortho-dihydroxy group the whole series shares ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →}.

Four abietanes from the Chilean Sphacele salviae — including rosmanol — all outperformed BHT on DPPH ^{76Reference 76Escuder B et al. · 2002In vitroAntioxidant capacity of abietanes from Sphacele salviae — in vitroView study →}.

Two panels do discriminate, and rosmanol wins both.

In Salvia chamelaeagnea, rosmanol showed the lowest anodic peak potential of six isolated compounds (Epa 0.11 V), which the authors read as the highest antioxidant power, in agreement with its ORAC and lipid-peroxidation results ^{77Reference 77Etsassala NGER et al. · 2019Investigation of in-vitro antioxidant and electrochemical activities of isolated compounds from Salvia chamelaeagnea extractView study →}.

In Salvia aurita, rosmanol gave the highest ORAC of five isolated compounds, 25.79 mM Trolox equivalent/g, marginally ahead of carnosol (23.96) ^{78Reference 78Etsassala NGER et al. · 2020In vitroAlpha-glucosidase and alpha-amylase inhibitory activities, molecular docking, and antioxidant capacities of Salvia aurita constituents — in vitroView study →}.

The mechanistic reading is still a class property rather than a molecular signature.

The SAR conclusion in ^{44Reference 44Miura K et al. · 2002Antioxidant activity of chemical components from sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index methodView study →} is that the catechol does the work, and every member of the carnosic acid cascade carries one.

Where the cascade was discriminated at a cellular target — Nrf2/ARE reporter activation in mouse primary cortical cultures — rosmanol was stronger than its C-7 epimer epirosmanol (the authors concluded C-7 stereochemistry matters) but weaker than carnosol and carnosaldehyde, which reached ~50-fold activation at 25 µM and were the only two carried into the confirmatory Nrf2-dependent gene-expression and H₂O₂-neuroprotection experiments ^{46Reference 46Fischedick JT et al. · 2013Structure activity relationship of phenolic diterpenes from Salvia officinalis as activators of the nuclear factor E2-related factor 2 pathwayView study →}.

Rosmanol was not.

Every antioxidant measurement is cell-free or ex vivo — a model oil, isolated mitochondria, isolated LDL, an electrochemical cell.

No cellular antioxidant assay, no reliable animal oxidative-stress model on the isolate, no human data.

And the active range (7–30 µM) overlaps the cytotoxic range (≈42 µM, ^{79Reference 79Cheng 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 →}) closely enough that “antioxidant at 10 µM” is a chemical observation rather than a therapeutic one.

The 1981 isolation was itself antioxidant-driven ^{80Reference 80Nakatani N · 1981Structure of rosmanol, a new antioxidant from rosemary (Rosmarinus officinalis L.)View study →}.

Preparative HPLC fractionation of raspberry extract apportioned the fruit’s antioxidant capacity and found sanguiin H-6 to be a major contributor alongside vitamin C and the anthocyanins, with vasodilation activity confined to the sanguiin-H-6/lambertianin-C fractions ^{66Reference 66Mullen W et al. · 2002Ellagitannins, flavonoids, and other phenolics in red raspberries and their contribution to antioxidant capacity and vasorelaxation propertiesView study →}.

Ellagitannin isolates from red raspberry and cloudberry (60% and 80% ellagitannin by weight, dimer and trimer dominant) were highly effective DPPH radical scavengers and significantly protected human LDL and methyl-linoleate emulsions, though only weakly in bulk oil ^{20Reference 20Kähkönen M et al. · 2012In vitroAntioxidant activity of isolated ellagitannins from red raspberries and cloudberries — in vitroView study →}.

The apportionment is shared with vitamin C, anthocyanins and the trimer, and the chemistry is generic polyphenol redox activity.

These are cell-free and emulsion systems, and the molecule never enters the circulation — it is degraded in the colon and is undetectable in plasma ^{90,91Reference 90González-Barrio R et al. · 2010Clinical trialBioavailability of anthocyanins and ellagitannins following consumption of raspberries by healthy human volunteers and subjects with an ileostomyView study →Reference 91Dobani S et al. · 2021Ex vivo fecal fermentation of human ileal fluid collected after raspberry consumption modifies (poly)phenolics and modulates genoprotective effects in colonic epithelial cellsView study →}, so no systemic antioxidant effect can be claimed for it.

This is an enzyme-level antioxidant action rather than stoichiometric scavenging.

In high-glucose retinal pigment epithelial cells it reduced intracellular ROS and mitochondrial superoxide while restoring glutathione content and reactivating MnSOD and glutathione peroxidase ^{92Reference 92Choi YH · 2023Reduction of high glucose-induced oxidative injury in human retinal pigment epithelial cells by sarsasapogenin through inhibition of ROS generation and inactivation of NF-κB/NLRP3 inflammasome pathwayView study →}, and its direct binding partner GPX4 is itself the lipid-peroxide-detoxifying enzyme that gates ferroptosis ^{93Reference 93Wang F et al. · 2025Sarsasapogenin stimulates angiogenesis and osteogenesis coupling to treat estrogen deficiency-induced osteoporosis by activating the GPX4/SLIT3/ROBO1 axisView study →}.

Consistent across the diabetic-complication and colitis models ^{94Reference 94Mandlik DS et al. · 2021AnimalProtective effect of sarsasapogenin in TNBS induced ulcerative colitis in rats associated with downregulation of pro-inflammatory mediators and oxidative stressView study →}.

The evidence is cell and rodent only, with no human antioxidant endpoint.

Shatavarin IV is an inducer of endogenous defences rather than a stoichiometric scavenger.

In C. elegans, shatavarin IV lowered intracellular ROS and protein carbonylation while raising transcription of sod-1, sod-2, sod-3, gst-4, gst-7 and ctl-2 ^{95Reference 95Smita SS et al. · 2017AnimalShatavarin IV elicits lifespan extension and alleviates Parkinsonism in Caenorhabditis elegansView study →}; the same laboratory showed the longevity arm of this effect requires SKN-1/NRF-2 and partly DAF-16/FOXO ^{96Reference 96Smita SS et al. · 2020AnimalA bioactive compound Shatavarin IV-mediated longevity as revealed by dietary restriction-induced autophagy in Caenorhabditis elegansView study →}, which is the canonical stress-response transcription-factor route.

In LPS-stimulated human SH-SY5Y cells the isolate reduced ROS and nitric oxide at 10 ng/mL (≈11 nM) ^{97Reference 97Das D et al. · 2026In vitroShatavarin IV, a bioactive constituent of Asparagus racemosus, exerts antioxidant and anti-inflammatory effects in LPS-treated cultured SH-SY5Y cells via TrkB-BDNF axisView study →} — a striking potency for a saponin that has not been independently replicated and is provisional until it is.

Tanshinone I has the weakest antioxidant evidence of the danshen diterpenes, and much of it is inherited from tanshinone IIA. Both dedicated tanshinone I reviews list “anti-oxidative stress” / “antioxidant” among its activities ^{98,99Reference 98Ke L et al. · 2023Tanshinone I: pharmacological activities, molecular mechanisms against diseases and future perspectivesView study →Reference 99Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}, and the 2023 review’s target list includes Akt/Nrf2 among the pathways associated with tanshinone I across the literature ^{98Reference 98Ke L et al. · 2023Tanshinone I: pharmacological activities, molecular mechanisms against diseases and future perspectivesView study →}.

No isolated-molecule antioxidant assay of tanshinone I has been published, and the specific Nrf2 induction claim usually attached to it traces to a tanshinone IIA review ^{101Reference 101Guo R et al. · 2020ReviewPharmacological activity and mechanism of tanshinone IIA in related diseases — reviewView study →}.

The action is review-asserted rather than demonstrated.

What is mechanistically supported and adjacent: tanshinone I’s quinone is handled by NQO1 via a two-electron reduction with immediate glucuronide trapping ^{100Reference 100Li Y et al. · 2019Metabolic characteristics of tanshinone I in human liver microsomes and S9 subcellular fractionsView study →} — the detoxifying branch of quinone metabolism, consistent with the absence of a redox-cycling toxicity signal, but that is a metabolic fact rather than an antioxidant action.

That makes it a supporting observation until an isolate assay is published.

The best-characterised antioxidant action is cellular rather than chemical, not DPPH radical scavenging but suppression of stimulated oxidant production.

Urolithin A was the most potent of the urolithins at inhibiting reactive-oxygen-species release from human neutrophils, active at 1 µM, and inhibited myeloperoxidase release at 20 µM ^{116Reference 116Piwowarski JP et al. · 2014Influence of gut microbiota-derived ellagitannins’ metabolites urolithins on pro-inflammatory activities of human neutrophils — urolithin A cut stimulated ROS by 42.6–53.7% at 1 µMView study →}; it also acts through Nrf2 in the gut-barrier work ^{117Reference 117Singh R et al. · 2019Enhancement of the gut barrier integrity by a microbial metabolite through the Nrf2 pathway — urolithin A and UAS03 via AhR–Nrf2 and tight-junction proteinsView study →}.

Human data are thin and mixed: an 8-week athlete trial found superoxide dismutase fell versus placebo ^{118Reference 118Zhao H et al. · 2024RCTAssessment of urolithin A effects on muscle endurance, strength, inflammation, oxidative stress and protein metabolism in male athletes with resistance training — an 8-week randomized, double-blind, placebo-controlled study (n = 20)View study →}, and a soccer-player pilot found antioxidant status declined over preseason in placebo but was preserved on urolithin A ^{119Reference 119Monsalve Acevedo A et al. · 2025RCTEffects of urolithin A supplementation on performance and antioxidant status in academy soccer players during preseason — a pilot randomised controlled trialView study →}.

Direct radical-scavenging claims transferred from the parent ellagitannins do not belong to this molecule, and the glucuronide conjugate is far less active than the aglycone tested in these assays ^{120Reference 120Ávila-Gálvez MÁ et al. · 2019Tissue deconjugation of urolithin A glucuronide to free urolithin A in systemic inflammationView study →}.

A consistent Nrf2-axis story rather than direct radical scavenging, which matters because urolithin B has only one phenolic hydroxyl and is a poor stoichiometric antioxidant by construction.

In microglia it reduced intracellular ROS and NADPH-oxidase subunit expression and upregulated heme oxygenase-1 through Nrf2/ARE signaling ^{121Reference 121Lee G et al. · 2019Anti-inflammatory and antioxidant mechanisms of urolithin B in activated microgliaView study →}.

In myocardial ischemia/reperfusion it drove p62 accumulation and p62–Keap1 interaction, promoting Nrf2 nuclear translocation — and its protection against superoxide production and apoptosis was lost when Nrf2 was silenced, which is a genuine mechanistic dependency rather than a correlation ^{122Reference 122Zheng D et al. · 2020Urolithin B, a gut microbiota metabolite, protects against myocardial ischemia/reperfusion injury via p62/Keap1/Nrf2 signaling pathwayView study →}.

In RANKL-stimulated osteoclast precursors it lowered ROS by upregulating Nrf2 and downstream scavenging enzymes ^{123Reference 123Qu Z et al. · 2022Urolithin B suppresses osteoclastogenesis via inhibiting RANKL-induced signalling pathways and attenuating ROS activitiesView study →}, and in D-galactose-aged mice it raised Cu/Zn-SOD and catalase activity and lowered advanced glycation end-products ^{124Reference 124Chen P et al. · 2021AnimalThe gut microbiota metabolite urolithin B improves cognitive deficits by inhibiting Cyt C-mediated apoptosis and promoting the survival of neurons through the PI3K pathway in aging miceView study →}.

It is an indirect, inducible antioxidant action rather than a chemical one.

Vescalagin and castalagin were isolated from camu-camu seed and peel specifically as the C-glycosidic ellagitannins carrying that material’s antioxidant activity ^{125Reference 125Kaneshima T et al. · 2016In vitroAntioxidant activity of C-glycosidic ellagitannins from the seeds and peel of camu-camu (Myrciaria dubia) — in vitroView study →}, and vescalagin is repeatedly one of the two dominant phenolics in antioxidant camu-camu seed and seed-coat extracts ^{14,126,127Reference 14Fidelis M et al. · 2020Camu-camu seed (Myrciaria dubia) — from side stream to an antioxidant, antihyperglycemic, antiproliferative, antimicrobial, antihemolytic, anti-inflammatory, and antihypertensive ingredientView study →Reference 126Willemann JR et al. · 2020In vitroResponse surface optimization of phenolic compounds extraction from camu-camu (Myrciaria dubia) seed coat based on chemical properties and bioactivity — in vitroView study →Reference 127Fidelis M et al. · 2020In vitroFrom byproduct to a functional ingredient: camu-camu (Myrciaria dubia) seed extract as an antioxidant agent in a yogurt model — in vitroView study →}.

The mechanism is generic polyphenol chemistry — radical scavenging plus Cu²⁺ chelation in the assays used — and the assays are chemical (DPPH, FRAP, Folin) run on extracts, with vescalagin’s contribution inferred from co-occurrence rather than isolated and dose-ranged.

It is a class effect that vescalagin participates in rather than a molecule-level finding.

The mechanistic thread under nearly every protective result vincamine produces, and the frame a 2023 review uses for the whole molecule ^{128Reference 128Ren et al. · 2023ReviewVincamine, from an antioxidant and a cerebral vasodilator to its anticancer potential — reviewView study →}.

Across hepatic, renal, cardiac, gastric and neural rodent models the readout is remarkably uniform: superoxide dismutase and glutathione rise, malondialdehyde falls, and the route is Nrf2/HO-1 activation rather than stoichiometric radical scavenging ^{128,129,130Reference 128Ren et al. · 2023ReviewVincamine, from an antioxidant and a cerebral vasodilator to its anticancer potential — reviewView study →Reference 129Abouelhamd et al. · 2025AnimalVincamine induces cytoprotective autophagy via regulation of AMPK/mTOR signalling in gentamicin-induced hepatotoxicity and nephrotoxicity in rats — animal modelView study →Reference 130Abouelhamd et al. · 2026AnimalMechanistic insights into vincamine’s cardio-protection against gentamicin toxicity: crosstalk between Nrf2/HO-1 and related pathways — animal modelView study →}.

Inductive antioxidant action is a more durable mechanism than direct scavenging, and it is reciprocally coupled to the NF-κB suppression below — one upstream switch, two downstream phenotypes.

The work is entirely cell and animal on the isolate, there is no human antioxidant-biomarker study of vincamine, and Nrf2/NF-κB modulation is the single most frequently reported activity in natural-product pharmacology, so on its own it discriminates poorly between molecules.

Casuarictin was the most potent DPPH scavenger of ten polyphenols isolated from Syzygium guineense leaf, IC₅₀ 3.9 ± 0.1 µM, narrowly ahead of casuarinin (4.5 µM) and pentagalloylglucose (5 µM) — though pentagalloylglucose was much the better xanthine oxidase and 15-lipoxygenase inhibitor in the same panel ^{19Reference 19Nguyen TL et al. · 2016In vitroFlavonoids, gallotannins and ellagitannins in Syzygium guineense and the traditional use among Malian healers — in vitroView study →}.

It is among the minor ellagitannins of red raspberry identified in the fractionation work showing berry ellagitannins to be effective radical scavengers and protective of human LDL and lipid emulsions ^{20Reference 20Kähkönen M et al. · 2012In vitroAntioxidant activity of isolated ellagitannins from red raspberries and cloudberries — in vitroView study →}, and it is one of the abundant leaf polyphenols of Combretum racemosum, whose extracts were active in FRAP and DPPH ^{21Reference 21Dah-Nouvlessounon D et al. · 2023Pharmacological property and cytotoxic effect showing antiproliferative potency in human melanoma cell lines (A375) of Combretum racemosum leaf and root extracts used in BeninView study →}.

Stoichiometric scavenging rank order in cell-free assays tracks free-hydroxyl count more than it tracks any distinctive pharmacology, and a molecule with twelve phenolic OH groups will always score well.

No cellular antioxidant assay, no in-vivo endpoint, and no plasma exposure to make a systemic role plausible ^{22,23Reference 22Piwowarski JP et al. · 2016Differences in metabolism of ellagitannins by human gut microbiota ex vivo culturesView study →Reference 23Daniel EM et al. · 1991In vitroThe effects of pH and rat intestinal contents on the liberation of ellagic acid from purified and crude ellagitannins — in vitroView study →}.

Dioscin raises superoxide dismutase, glutathione and glutathione peroxidase and lowers malondialdehyde across the ethanol, CCl4 and renal injury models ^{37,38,39Reference 37Lu B et al. · 2012AnimalMechanism investigation of dioscin against CCl4-induced acute liver damage in miceView study →Reference 38Xu T et al. · 2014Protective effects of dioscin against alcohol-induced liver injuryView study →Reference 39Qiao Y et al. · 2018Protective effects of dioscin against fructose-induced renal damage via adjusting Sirt3-mediated oxidative stress, fibrosis, lipid metabolism and inflammationView study →}, and activates Nrf2/HO-1 directly in cisplatin kidney injury ^{40Reference 40Wang S et al. · 2022Dioscin protects against cisplatin-induced acute kidney injury by reducing ferroptosis and apoptosis through activating Nrf2/HO-1 signalingView study →}.

In every one of these studies the antioxidant readout is a downstream marker within a disease model, not an isolated demonstration — there is no standalone antioxidant assay of dioscin, and a saponin is not a redox-active scaffold, so the effect is most likely secondary to CYP2E1 suppression and reduced injury rather than direct scavenging.

Inductive rather than scavenging: diosgenin activates the Nrf2/HO-1 antioxidant-response program, raising SOD, catalase and glutathione and lowering lipid-peroxidation markers ^{41Reference 41Leng J et al. · 2020AnimalNeuroprotective effect of diosgenin in a mouse model of diabetic peripheral neuropathy involves the Nrf2/HO-1 pathway — mouse in vivo studyView study →}.

The finding is almost never isolated — it appears as a mechanism inside neuroprotection and metabolic disease models rather than as a standalone antioxidant result, and there is no human or whole-herb antioxidant study of the molecule.

The evidence is isolate-only.

In the adjuvant-arthritis model, isolated protodioscin restored glutathione, catalase and superoxide dismutase and lowered lipid peroxidation and nitrite ^{67Reference 67Liu JY et al. · 2017AnimalProtodioscin ameliorates oxidative stress, inflammation and histology outcome in complete Freund’s adjuvant induced arthritis rats — [animal model]View study →} — an indirect, enzyme-restoring antioxidant signature rather than direct radical scavenging, which is what a large polar glycoside with no phenolic groups should be expected to do.

There is no reported ORAC/DPPH potency for the isolate, and the antioxidant reputation of Tribulus extracts belongs mainly to their flavonoids (rutin and relatives), not their saponins.

Ranked among the more active ellagitannins — with amariin and phyllanthusiin D, and comparable to rutin and quercetin 3-O-glucoside — across DPPH, ABTS/ferrylmyoglobin, FRAP and pulse-radiolysis panels of Phyllanthus amarus polyphenols ^{73Reference 73Londhe JS et al. · 2008Antioxidant activity of some polyphenol constituents of the medicinal plant Phyllanthus amarus LinnView study →}, and protected rat liver mitochondria against lipid and protein oxidation and plasmid DNA against radiation strand breaks, acting mainly through superoxide scavenging while the flavonoids in the same panel scavenged hydroxyl radicals ^{74Reference 74Londhe JS et al. · 2009Radioprotective properties of polyphenols from Phyllanthus amarus LinnView study →}.

Stoichiometric scavenging rank order in cell-free assays tracks the number of free phenolic hydroxyls more than it tracks any distinctive pharmacology, and none of it is reachable in plasma ^{75Reference 75Mihaylova R et al. · 2025ReviewUnlocking the therapeutic potential of ellagitannins: a comprehensive review of key representativesView study →}.

Both papers are from the same laboratory.

Provisional in both directions.

On the protective side the isolate scavenges DPPH radicals ^{105Reference 105Yin J et al. · 2019Anti-acne vulgaris effect including skin barrier improvement and 5α-reductase inhibition by tellimagrandin I from Carpinus tschonoskiiView study →}, is non-mutagenic and non-genotoxic while protecting DNA against 4-nitroquinoline-1-oxide, sodium azide and doxorubicin ^{106Reference 106Fernandes AS et al. · 2024Tellimagrandin-I and camptothin A exhibit chemopreventive effects in Salmonella enterica serovar Typhimurium strains and human lymphocytesView study →}, and blocks hydroquinone/Cu(II)- and H₂O₂/Cu(II)-mediated strand breaks ^{107Reference 107Yi ZC et al. · 2009In vitroChebulinic acid and tellimagrandin I inhibit DNA strand breaks by hydroquinone/Cu(II) and H2O2/Cu(II), but potentiate DNA strand breaks by H2O2/Fe(II)View study →}.

On the other side, the same molecule reduces Cu(II) to Cu(I) and causes copper-dependent strand breaks in plasmid and human fibroblast DNA on its own ^{108Reference 108Yi ZC et al. · 2009In vitroProoxidant action of chebulinic acid and tellimagrandin I — causing copper-dependent DNA strand breaksView study →}, and against H₂O₂/Fe(II) it protects at 1–10 µM but potentiates damage above 50 µM ^{107Reference 107Yi ZC et al. · 2009In vitroChebulinic acid and tellimagrandin I inhibit DNA strand breaks by hydroquinone/Cu(II) and H2O2/Cu(II), but potentiate DNA strand breaks by H2O2/Fe(II)View study →}.

The accurate mechanistic label is a redox switch whose direction depends on the transition metal present and on concentration, not an antioxidant.

No in-vivo antioxidant endpoint has ever been measured for this molecule.

The evidence is not a direct radical-scavenging measurement.

It is secondary biochemical endpoints inside the cardioprotection experiment — malondialdehyde down and superoxide dismutase activity restored in tribulosin-pretreated hearts ^{114Reference 114Zhang S et al. · 2010AnimalTribulosin protects rat hearts from ischemia/reperfusion injury — isolated Langendorff-perfused rat heart, animal modelView study →} — which as plausibly measure less tissue injury as any intrinsic antioxidant chemistry.

No cell-free assay (DPPH, ORAC, FRAP) has ever been run on this molecule, and a fully glycosylated spirostanol has no phenolic chemistry to scavenge with, which makes the indirect explanation the more likely one.

A whole-extract correlation between antioxidant capacity and tribulosin content across fourteen Indian populations ^{115Reference 115Limboo A et al. · 2026Pharmacopoeial standardisation and bioactive metabolite-based identification of elite sources of Tribulus terrestris L. from Rajasthan and adjoining regions, IndiaView study →} is a sample-level association across extracts differing in dozens of constituents, not a molecular finding.