Tanshinone I

Tanshinone I at a Glance

Tanshinone I is one of the four orange-red tanshinones that give the root of dan shen (Salvia miltiorrhiza) its colour. Chemically it is an abietane diterpene quinone — C₁₈H₁₂O₃, molecular weight 276.3, CAS 568-73-0, formally 1,6-dimethylphenanthro[1,2-b]furan-10,11-dione — a flat, fully aromatic three-ring system carrying a furan on one end and an ortho-quinone on the other ^{1,2Reference 1Ke L et al. · 2023Tanshinone I: pharmacological activities, molecular mechanisms against diseases and future perspectivesView study →Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}. It is the smallest and the most aromatic of the four major tanshinones, and it sits firmly in the lipophilic half of danshen’s chemistry (XLogP ≈ 3.7), which is the single most important practical fact about it.

Two things matter most about this molecule.

First, it is genuinely the least-studied of the four. As of August 2026, the published record holds 438 papers on tanshinone I, against 1,984 for tanshinone IIA and 885 for cryptotanshinone. More than half of tanshinone I’s own literature (234 of the 438) also names tanshinone IIA in the same title or abstract — these molecules are studied side by side, and the risk of reading one’s evidence onto the other is not hypothetical. Where a finding here belongs to a sibling or to a whole tanshinone fraction, it is labelled as such.

Second, it does have an independent literature, and it is a specific one. There is a dedicated 2023 Phytomedicine review of tanshinone I alone ^{1Reference 1Ke L et al. · 2023Tanshinone I: pharmacological activities, molecular mechanisms against diseases and future perspectivesView study →} and a 2022 review of its pharmacology and total synthesis ^{2Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}. Its own best-verified threads are an anti-inflammatory pharmacology that includes a direct sPLA₂ inhibition result and a specific NLRP3 inflammasome result, an anti-metastatic thread in which it acts on cell adhesion and invasion rather than by killing cells, and an Aurora A kinase story in which it is repeatedly the most potent of the tanshinones tested. That last point is the useful correction to the assumption that tanshinone I is simply a weaker tanshinone IIA.

Tanshinone I Is Easily Confused With Its Siblings

The name is a prefix of another compound’s. In plain text matching, tanshinone I sits inside tanshinone IIA, so a careless search folds the two together — and tanshinone IIA has four times the literature. Much of the field compounds the problem by writing tanshinones collectively and never separating them, and all four molecules are routinely isolated from the same fraction and assayed in the same experiment, so co-mention is the norm.

Kept apart properly, tanshinone I has roughly 438 papers to tanshinone IIA’s 1,984 — and about 204 of those never mention IIA at all.

Occurrence in danshen is not in doubt. It is a well-attested constituent of Salvia miltiorrhiza, named as such in the Chinese Pharmacopoeia assay ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →}, with several hundred papers reporting the pairing.

It is a diterpene, not a flavonoid. Tanshinone I is an abietane-derived diterpene quinone ^{2,21Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →Reference 21Zhang Y et al. · 2012Tanshinones: sources, pharmacokinetics and anti-cancer activitiesView study →}, and a discrete molecule rather than a group name. At least one otherwise-solid paper calls it “one of major bioactive flavonoids” of danshen ^{8Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →}; that is an error in the source, and the compound has no flavonoid character whatsoever.

What counts as evidence for it. A result belongs to tanshinone I only where the isolated molecule was the test article. Findings reported for “tanshinones”, for a total-tanshinone fraction or for an ethanolic danshen extract are fraction-level, and a result on a named sibling belongs to that sibling.

Tanshinone I vs. Its Three Siblings

The four major tanshinones differ by very little — a hydrogenated ring here, a gem-dimethyl there — but the differences are not cosmetic, and in at least two assays tanshinone I is the outlier rather than the also-ran.

	Tanshinone I	Tanshinone IIA	Cryptotanshinone	Dihydrotanshinone I
Formula / MW	C₁₈H₁₂O₃ / 276.3	C₁₉H₁₈O₃ / 294.3	C₁₉H₂₀O₃ / 296.4	C₁₈H₁₄O₃ / 278.3
Core	Fully aromatic; aromatic furan	Aromatic quinone; saturated A-ring	Dihydrofuran; saturated A-ring	Dihydrofuran; aromatic A-ring
Literature (tiab)	438	1,984	885	273
Pharmacopoeial marker	Yes (in the 0.25% combined assay) ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →}	Yes (the headline marker) ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →}	Yes ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →}	No
Distinctive claim	Aurora A downregulation; anti-invasive without direct cytotoxicity ^{11,12,15Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}	Nrf2 induction; the cardioprotection literature	Antibacterial / STAT3	AChE inhibition; direct cytotoxicity

Three relationships matter for reading the evidence.

Tanshinone I is often the most potent of the four in growth-inhibition work. In a breast-cancer panel, tanshinones inhibited growth dose-dependently “with tanshinone I the most potent agent”, and tanshinone I was the only one of the group with potent activity against the triple-negative line MDA-MB-231 ^{11Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →}. In lung-cancer lines tested head to head — cryptotanshinone, tanshinone I, tanshinone IIA — tanshinone I was again the most potent ^{12Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}. This is the opposite of what the citation counts would lead you to expect.

Tanshinone I is not simply a weaker cytotoxin — it has a different mechanism. In the highly invasive lung line CL1-5, the authors state plainly that “unlike tanshinone IIA, which induces cell apoptosis, tanshinone I did not have direct cytotoxicity”; it suppressed migration, invasion and gelatinase activity instead ^{15Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}. An anti-invasive agent and a cytotoxic agent are different drugs with different development paths, and conflating them is the commonest error in this literature.

Tanshinone I can be made from dihydrotanshinone I, which contaminates the evidence base. Tanshinone I is obtainable by simple oxidation of dihydrotanshinone ^{2Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}, and a 2023 stability study found that in DMSO — the standard stock solvent for cell assays — dihydrotanshinone concentrations fell rapidly and were converted to tanshinone I (cryptotanshinone likewise converted to tanshinone IIA) ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}. Some published “dihydrotanshinone I” activity may therefore have been partly tanshinone I activity, and the reverse artefact is possible for aged aqueous stocks: all four tanshinones lost concentration in aqueous solution within 24 hours ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}.

Which Fraction? Why A Danshen Decoction Delivers Essentially None

Danshen is two chemically separate herbs sharing one root. The hydrophilic fraction — salvianolic acid B, salvianolic acid A, danshensu and their precursor rosmarinic acid — is what water pulls out. The lipophilic fraction — the tanshinones — is what ethanol, ethyl acetate or supercritical CO₂ pull out. Tanshinone I is squarely in the second group ^{2,4,21Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →Reference 21Zhang Y et al. · 2012Tanshinones: sources, pharmacokinetics and anti-cancer activitiesView study →}.

This is not a theoretical distinction; it has been demonstrated repeatedly as a negative result on the water extract. In rat liver microsome work, an ethyl-acetate danshen extract containing the major tanshinones inhibited warfarin hydroxylation while “the aqueous extract of Danshen had no effect” ^{25Reference 25Wu WW · 2010In vitroInhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivoView study →}. In primary rat hepatocytes, the ethanolic extract altered CYP1A2 markedly while “the water extract containing mainly salvianolic acid B and danshensu had no effect”, and salvianolic acid B and danshensu tested individually did nothing ^{22Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →}. In a preparative study, only the 90% alcohol eluent carried the tanshinones, at over 97% total-tanshinone content; the 0% and 45% eluents did not ^{10Reference 10Gao H et al. · 2018Simultaneous purification of dihydrotanshinone, tanshinone I, cryptotanshinone, and tanshinone IIA from Salvia miltiorrhiza and their anti-inflammatory activities investigationView study →}.

The practical consequences:

A traditional decoction of danshen root — simmered in water at 9–15 g/day — is a salvianolic-acid preparation. Whatever it does, it does not do it through tanshinone I in any meaningful amount.
The compound danshen dripping pill and most danshen injections are also water-type preparations. The large clinical literature on danshen for angina and coronary heart disease therefore says nothing about this molecule.
Tanshinone-containing products are a separate category — ethanolic extracts, total-tanshinone capsules, CO₂ extracts — and they are the only oral route that delivers tanshinone I at all.
“Tanshinone IIA sodium sulfonate” injection, used clinically in China, is a semi-synthetic water-soluble derivative of a different molecule. Its trial data belongs neither to tanshinone IIA nor, obviously, to tanshinone I.

A related caution about dose-scaling: even a tanshinone-rich preparation is delivering a 0.02–0.07% root constituent ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →}, and its oral bioavailability is poor (see Pharmacokinetics). The gap between the plasma exposure a capsule produces and the 10–50 µM used in a cell-culture dish is very large.

What It's Used For

The anti-inflammatory pharmacology is the best-characterised thread, and some of it is unusually specific.Tanshinone I inhibits human recombinant type IIA sPLA₂ at IC₅₀ 11 µM and cPLA₂ at 82 µM, cuts PGE₂ from LPS-stimulated macrophages at IC₅₀ 38 µM, and — the informative detail — does this without inhibiting COX-2 activity or expression, i.e. it works upstream at arachidonate release ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}. It was active in carrageenan paw oedema and adjuvant arthritis in vivo ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}.

It is a specific NLRP3 inflammasome inhibitor.It blocks NLRP3 but not AIM2 or NLRC4 activation, and does so by disrupting the NLRP3–ASC association; protective in mouse LPS septic shock and NASH models ^{7Reference 7Zhao J et al. · 2023Tanshinone I specifically suppresses NLRP3 inflammasome activation by disrupting the association of NLRP3 and ASCView study →}.

Aurora A kinase is a genuine functional target, not a bioinformatic guess.siRNA knockdown of Aurora A abolished tanshinone I’s activity in both breast and lung cancer cells — the causal test, run twice by the same group in two tissues ^{11,12Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}.

Its anticancer signature is anti-invasive and anti-angiogenic rather than cytotoxic.It suppresses ICAM-1/VCAM-1, VEGF, IL-8 transcription, gelatinase activity and endothelial tube formation, and reduces metastasis in mouse models — while explicitly lacking direct cytotoxicity in at least one line ^{14,15Reference 14Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}.

Nothing here is clinical.There is not one human trial of tanshinone I. Every number below is in vitro or in an animal.

What is not established:any direct antiplatelet activity of the isolated molecule (see Safety), any Nrf2 result belonging to tanshinone I specifically rather than to tanshinone IIA, and any pharmacokinetic profile in humans.

Evidence by IndicationStrength of support · Beta

Support is an experimental score — weighted by study type (human > animal > in vitro) and volume. A beta, not a validated metric. With zero human data on the isolated molecule, every score here is capped low by design. Each application links to its write-up.

This is the oldest and the most mechanistically satisfying thread, and it starts with an activity-guided isolation rather than a hunch. In a 2002 screen of 300 plant extracts for inhibitors of PGD₂ production, the methanol extract of S. miltiorrhiza scored, its ethyl-acetate subfraction scored hardest, and fractionation of that subfraction delivered tanshinone I as the active principle. The isolated molecule inhibited PGE₂ formation from LPS-stimulated RAW 264.7 macrophages at IC₅₀ 38 µM but did not affect COX-2 activity or COX-2 expression — instead it inhibited type IIA human recombinant sPLA₂ at IC₅₀ 11 µM and rabbit recombinant cPLA₂ at IC₅₀ 82 µM, i.e. it throttles arachidonate release rather than its conversion. It was also active in rat carrageenan paw oedema and adjuvant-induced arthritis ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}.

Two decades later a 2023 study identified a second, sharper mechanism. Tanshinone I inhibited NLRP3 inflammasome activation in macrophages while leaving AIM2 and NLRC4 inflammasomes untouched — the selectivity control that separates a real inhibitor from a general anti-inflammatory. Immunoprecipitation localised the effect to disruption of the NLRP3–ASC interaction, blocking assembly. In vivo it lowered IL-1β in peritoneal lavage and serum in LPS-induced septic shock and reduced liver inflammation and fibrosis in a NASH model ^{7Reference 7Zhao J et al. · 2023Tanshinone I specifically suppresses NLRP3 inflammasome activation by disrupting the association of NLRP3 and ASCView study →}.

A third strand is microglial. In LPS-stimulated BV-2 microglia, tanshinone I suppressed NO, TNF-α, IL-1β and IL-6 and inhibited NF-κB nuclear translocation and DNA binding — but did not suppress the anti-inflammatory M2 markers IL-10 and IL-1Ra, which is a more interesting result than blanket suppression ^{8Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →}. A 2025 study extended the anti-inflammatory picture to a psoriasis-like dermatitis model, with effects on keratinocyte differentiation as well as inflammation ^{30Reference 30Wang Q et al. · 2025Tanshinone I ameliorates psoriasis-like dermatitis by suppressing inflammation and regulating keratinocyte differentiationView study →}.

Isolate vs. plant: all of the above used the purified molecule. The countervailing datum is fraction-level and points the other way: in a systematic purification study, the total tanshinone fraction was more anti-inflammatory in LPS-stimulated RAW264.7 cells than dihydrotanshinone, tanshinone IIA, cryptotanshinone or tanshinone I individually ^{10Reference 10Gao H et al. · 2018Simultaneous purification of dihydrotanshinone, tanshinone I, cryptotanshinone, and tanshinone IIA from Salvia miltiorrhiza and their anti-inflammatory activities investigationView study →}. Whatever tanshinone I contributes, it is not the sole author of danshen’s lipophilic anti-inflammatory effect.

Gap: the enzyme IC₅₀s (11–82 µM) are far above any plausible plasma concentration from an oral preparation, and no study has measured whether the in-vivo anti-inflammatory effects occur at exposures a person could reach. No human data of any kind.

Two 2008 papers, published independently within months of each other, established the shape of this thread — and both carried it through to an in-vivo endpoint, which is rare for this class.

The first worked from the endothelial side. Tanshinone I dose-dependently inhibited ICAM-1 and VCAM-1 expression in TNF-α-stimulated human umbilical vein endothelial cells, and consequently reduced adhesion of both monocytes and MDA-MB-231 breast cancer cells to the endothelium — an effect reproduced by siRNA knockdown of ICAM-1 and VCAM-1, confirming the route. It also inhibited TNF-α-induced VEGF production and VEGF-mediated tube formation in HUVECs, inhibited VEGF production in the tumour cells themselves, and reduced migration through extracellular matrix. In vivo, tumour mass volume and metastasis incidence both fell ^{14Reference 14Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →}.

The second worked from the tumour side, in the highly invasive lung adenocarcinoma line CL1-5. Tanshinone I inhibited migration, invasion and gelatinase activity in macrophage-conditioned-medium-stimulated cells and reduced both tumorigenesis and metastasis in SCID mice. Mechanistically it reduced IL-8 transcriptional activity by attenuating AP-1 and NF-κB DNA binding, with microarray pointing to Ras-MAPK and Rac1 signalling. The authors state explicitly that, unlike tanshinone IIA, tanshinone I had no direct cytotoxicity in this system — and they add a caution of their own, that “its potential anticoagulant effect should be noted and evaluated” ^{15Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}.

Later work is consistent. In osteosarcoma, tanshinone I at 10 and 20 mg/kg suppressed growth in subcutaneous and orthotopic xenografts and inhibited metastasis in an intracardiac inoculation model, abolishing IL-6-driven JAK/STAT3 activation with Bax up and Bcl-2 down ^{16Reference 16Wang W et al. · 2019Tanshinone I inhibits the growth and metastasis of osteosarcoma via suppressing JAK/STAT3 signalling pathwayView study →}.

In cervical cancer cells it reduced adhesion, invasion and migration by inducing BNIP3/NIX-mediated mitophagy and reprogramming mitochondrial metabolism, with siRNA against BNIP3 and NIX preventing the effect ^{17Reference 17Cui S et al. · 2022Tanshinone I inhibits metastasis of cervical cancer cells by inducing BNIP3/NIX-mediated mitophagy and reprogramming mitochondrial metabolismView study →}.

Isolate vs. plant: all isolate. None of it has been shown for a danshen extract, and none of it transfers to the herb.

Gap: every model is a xenograft or a cell line; there is no orthotopic-to-clinical bridge, no combination work with standard chemotherapy beyond one hepatocellular study, and no human exposure data to tell you whether the effective mouse doses (10–200 mg/kg) correspond to anything achievable orally in people.

Aurora A (AURKA) is a mitotic kinase overexpressed in many solid tumours, and it is the one molecular target for which tanshinone I has been subjected to the proper causal test rather than a correlation.

In breast cancer cells, a panel of tanshinones inhibited growth dose-dependently with tanshinone I the most potent, and tanshinone I was the only member with potent activity against the triple-negative line MDA-MB-231. Among the biomarkers tracked, Aurora A tracked the anti-growth activity most consistently, and siRNA knockdown of Aurora A dramatically reduced tanshinone I’s growth-inhibition and apoptosis-induction — the knockdown test that distinguishes a functional target from a bystander. The regulation turned out to be epigenetic: Aurora A overexpression in these cells was not driven by promoter DNA methylation but by histone acetylation, and tanshinone I significantly reduced acetylated histone H3 at the Aurora A gene. Normal mammary epithelial cells were much less affected ^{11Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →}.

The same group replicated the logic in lung cancer. Of cryptotanshinone, tanshinone I and tanshinone IIA, tanshinone I was again the most potent; Aurora A siRNA again dramatically eliminated its activity. In an H1299 xenograft, tanshinone I at 200 mg/kg reduced final tumour weight by 34%, proliferation by 54%, angiogenesis by 72% and Aurora A expression by 67%, without altering food intake or body weight ^{12Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}.

A third study, in colorectal cancer, found a p53-dependent split: tanshinone I induced apoptosis in CRC cells with functional p53 via an Aurora A → p53 axis, had much less effect on normal colon epithelium or on p53-mutant CRC cells, and in the p53-mutant setting acted instead through survivin downregulation ^{13Reference 13Lu M et al. · 2016Tanshinone I induces human colorectal cancer cell apoptosis: the potential roles of Aurora A-p53 and survivin-mediated signaling pathwaysView study →}.

Isolate vs. plant: isolate throughout, with the caveat that the two 2012–2013 papers are from one laboratory and the breast paper’s title says “tanshinones” even though tanshinone I carries the finding ^{11,12Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}. A separate report of AURKA suppression via miR-32 in non-small-cell lung cancer is fraction-level (“tanshinones”) and is not credited here.

Gap: no direct kinase-inhibition assay establishing whether tanshinone I binds Aurora A or merely lowers its expression; no independent laboratory has replicated the siRNA result; and no selectivity panel against the wider kinome exists, so “Aurora A inhibitor” remains a functional description rather than a pharmacological classification.

Both neuroprotection studies use the isolated molecule and both take the anti-inflammatory mechanism as their route, which makes this thread an extension of application 1 rather than an independent claim.

In an MPTP mouse model of Parkinson’s disease, tanshinone I improved motor function on the rotarod, normalised striatal neurotransmitters on HPLC-ECD and protected nigrostriatal dopaminergic neurons, alongside modulation of microglial activation and attenuation of the TNF-α rise. Critically, tanshinone I did not alter MPP⁺ concentration — so it was not simply blocking the toxin’s activation, which is the standard confound in this model. Oral tanshinone I at 10 mg/kg daily for two weeks produced no liver toxicity on biochemistry or histology ^{8Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →}.

In a gerbil transient global cerebral ischemia model (5 minutes of occlusion), pre-treatment with tanshinone I protected hippocampal CA1 pyramidal neurons from delayed death. The mechanism proposed was maintenance of endogenous cytokine balance — the anti-inflammatory cytokines IL-4 and IL-13 rose with treatment in sham animals without a corresponding rise in IL-2 or TNF-α, and after ischemia all cytokine levels were maintained rather than collapsing. The authors also showed that exogenous IL-4 alone did not protect, which usefully rules out the simplest reading of their own result ^{9Reference 9Park JH et al. · 2014Anti-inflammatory effect of tanshinone I in neuroprotection against cerebral ischemia-reperfusion injury in the gerbil hippocampusView study →}.

Isolate vs. plant: isolate in both. Danshen’s own clinical stroke literature tests injectable, often compound, water-type preparations and cannot be used to support this.

Gap: each model has been run once, by one group, with pre-treatment rather than post-insult dosing in the ischemia study — the design that flatters neuroprotectants most. There is no blood–brain-barrier penetration measurement for tanshinone I, so the exposure at the target site is unknown.

Danshen’s reputation is cardiovascular, and this is precisely where tanshinone I’s own file is thinnest. The bulk of the mechanistic cardiovascular work on tanshinones belongs to tanshinone IIA (Nrf2 induction, ischemia-reperfusion infarct-size reduction, antifibrotic effects) or to the water-soluble salvianolic acid B (antiplatelet, free-radical scavenging, I/R protection). A 2025 review of vascular smooth muscle cell modulation treats tanshinone I, tanshinone IIA and cryptotanshinone together, describing class effects on calcium signalling, programmed cell death and pro-inflammatory cascades in the contractile-to-synthetic phenotype switch — it is a tanshinone-class account, not a tanshinone I account ^{29Reference 29Li J et al. · 2025Tanshinone I, tanshinone IIA, and cryptotanshinone: key bioactive components modulating vascular smooth muscle cell functionView study →}.

The honest position: tanshinone I is a co-marker in the Chinese Pharmacopoeia’s combined-tanshinone assay ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →} and a genuine member of the fraction credited with lipophilic cardioprotection, but the specific cardiovascular experiments almost all name a sibling. Its documented anti-inflammatory and anti-adhesion-molecule activity ^{6,7,14Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →Reference 7Zhao J et al. · 2023Tanshinone I specifically suppresses NLRP3 inflammasome activation by disrupting the association of NLRP3 and ASCView study →Reference 14Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →} is plausibly relevant to vascular inflammation, and that is as far as the evidence goes.

Gap: no isolated-molecule study of tanshinone I in myocardial ischemia-reperfusion, no blood-pressure study, no lipid study, and — see Safety — no platelet-aggregation study of the isolated molecule at all. Any cardiovascular claim made for tanshinone I on the strength of danshen’s clinical record is an unattributed inheritance.

Mechanisms

Target / pathway	Effect	Relevant to	Evidence level
Type IIA secretory PLA₂	inhibition, IC₅₀ 11 µM (human recombinant)	anti-inflammatory	isolate, in vitro ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}
Cytosolic PLA₂	inhibition, IC₅₀ 82 µM (rabbit recombinant)	anti-inflammatory	isolate, in vitro ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}
PGE₂ production (LPS macrophages)	inhibition, IC₅₀ 38 µM — not via COX-2	anti-inflammatory	isolate, in vitro ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}
NLRP3 inflammasome	specific inhibition; disrupts NLRP3–ASC binding; AIM2/NLRC4 unaffected	septic shockNASH	isolate, in vitro + mouse ^{7Reference 7Zhao J et al. · 2023Tanshinone I specifically suppresses NLRP3 inflammasome activation by disrupting the association of NLRP3 and ASCView study →}
NF-κB	inhibits p65 nuclear translocation and DNA binding	neuroinflammationIL-8	isolate, in vitro ^{8,15Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}
Aurora A (AURKA)	expression downregulated via reduced histone H3 acetylation; siRNA-validated as functional target	antiproliferative	isolate, in vitro + xenograft ^{11,12Reference 11Gong Y et al. · 2012Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and functionView study →Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}
ICAM-1 / VCAM-1	dose-dependent suppression in TNF-α-stimulated endothelium	anti-metastatic	isolate, in vitro + mouse ^{14Reference 14Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →}
VEGF / tube formation	suppressed in endothelium and tumour cells	anti-angiogenic	isolate, in vitro + mouse ^{12,14Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →Reference 14Nizamutdinova IT et al. · 2008Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion moleculesView study →}
IL-8 transcription	reduced via AP-1 and NF-κB DNA binding	anti-invasive	isolate, in vitro ^{15Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}
JAK/STAT3	abolishes IL-6-driven activation	osteosarcoma metastasis	isolate, in vitro + mouse ^{16Reference 16Wang W et al. · 2019Tanshinone I inhibits the growth and metastasis of osteosarcoma via suppressing JAK/STAT3 signalling pathwayView study →}
BNIP3 / NIX mitophagy	induced; siRNA-reversible	anti-metastatic	isolate, in vitro ^{17Reference 17Cui S et al. · 2022Tanshinone I inhibits metastasis of cervical cancer cells by inducing BNIP3/NIX-mediated mitophagy and reprogramming mitochondrial metabolismView study →}
CYP1A1 / CYP1A2	induction on repeat exposure via AhR upregulation (5.2× activity, 6.5× protein, rat hepatocytes)	drug interactions	isolate, ex vivo + mouse ^{22,24Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →Reference 24Feng C et al. · 2013AnimalTanshinone I protects mice from aristolochic acid I-induced kidney injury by induction of CYP1AView study →}
CYP1A1, CYP2C6, CYP2C11	acute inhibition of warfarin 4′-, 6- and 7-hydroxylation	warfarin interaction	isolate, rat microsomes ^{25Reference 25Wu WW · 2010In vitroInhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivoView study →}
ERα / ERβ	dose-dependent transactivation	endocrine caution	isolate, in vitro ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}
Androgen receptor (yeast assay)	antiandrogenic	endocrine caution	isolate, in vitro ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}

Pharmacokinetics

There is no human pharmacokinetic study of tanshinone I. What exists is metabolism work in human tissue fractions and animal PK, and both say the same discouraging thing.

Absorption and formulation. Tanshinone I has extremely low water solubility, poor pharmacokinetic properties, a short half-life and low oral bioavailability — a limitation that is the explicit motivation for the substantial derivative-chemistry literature built on it ^{2Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}. Solid-dispersion formulation with povidone K-30 or poloxamer 407 measurably improves solubility and dissolution of tanshinone I from a S. miltiorrhiza extract relative to commercial tablets ^{19Reference 19Yu H et al. · 2012Enhancement of solubility and dissolution rate of cryptotanshinone, tanshinone I and tanshinone IIA extracted from Salvia miltiorrhizaView study →}. Particle size matters too: rats given an ultrafine granular powder (D90 < 45 µm) of danshen showed significantly higher AUC and Cmax for tanshinone I than rats given traditional decoction slices ^{20Reference 20Zhi H et al. · 2019AnimalComparative in vivo constituents and pharmacokinetic study in rats after oral administration of ultrafine granular powder and traditional decoction slices of Chinese SalviaView study →}. A broader review of tanshinone pharmacokinetics notes that in-vivo potency assessments “vary greatly due to lack of uniformity of solvent vehicles and routes of administration” — which is a polite way of saying that much of the animal literature is not comparable across studies ^{21Reference 21Zhang Y et al. · 2012Tanshinones: sources, pharmacokinetics and anti-cancer activitiesView study →}.

Metabolism. The one study in human material is unusually clear. In human liver microsomes and S9 fractions, tanshinone I generated six metabolites by three routes — hydroxylation, reduction and glucuronidation. CYP2A6 was primarily responsible for phase I oxidation. Hydroxylated tanshinone I was then glucuronidated by UGT1A1, UGT1A3, UGT1A7 and UGT1A9, plus the extrahepatic isoforms UGT1A8 and UGT1A10. Separately, NQO1 reduced the quinone to a relatively unstable hydroquinone that was immediately conjugated by UGT1A9, UGT1A1 and UGT1A8 ^{18Reference 18Li Y et al. · 2019Metabolic characteristics of tanshinone I in human liver microsomes and S9 subcellular fractionsView study →}. That last route is worth flagging: a two-electron quinone reduction handled by NQO1 with rapid trapping is the detoxifying branch of quinone metabolism, and it is consistent with the absence of a redox-cycling toxicity signal for this molecule.

Stability. Tanshinone I concentrations decline in aqueous solution over 24 hours, as do all four tanshinones — relevant to any in-vitro result and to product shelf-life ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}.

The bottom line for a reader: an oral tanshinone-containing danshen product delivers a small fraction of a constituent that is present at 0.02–0.07% of the root ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →}, through a poorly-absorbing lipophilic molecule, into first-pass CYP2A6 and UGT metabolism. The distance between that and a 38 µM cell-culture IC₅₀ is the honest reason nothing here has become a medicine.

Clinical Trials

None. There are no clinical trials of tanshinone I as an isolated molecule, and no registered study of it under any name. Human exposure occurs only incidentally, via ethanolic or CO₂ danshen extracts and total-tanshinone products in which tanshinone I is a minor co-constituent.

Completed	Ongoing	Terminated	Preclinical
—(none)	—	—	~438indexed papers, all in vitro or animal

Last checked: August 2026.

Isolate vs. Plant

The transfer problem here runs in both directions and is worth stating carefully.

Plant → molecule is invalid. Danshen has one of the largest clinical evidence bases of any herb — over 230 randomised trials in angina alone. Essentially none of it is relevant to tanshinone I, because the trialled products are compound formulas, dripping pills and injections built on the water-soluble fraction. Reading danshen’s cardiovascular record onto this molecule is the single most likely error a reader could make here.

Molecule → plant is also invalid, for a different reason. Every anticancer, anti-inflammatory and neuroprotective finding above used the purified compound at micromolar concentrations. A danshen preparation delivering 0.02–0.07% tanshinone I ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →} does not reproduce those exposures, and in the one head-to-head test the whole tanshinone fraction outperformed tanshinone I alone on an anti-inflammatory endpoint ^{10Reference 10Gao H et al. · 2018Simultaneous purification of dihydrotanshinone, tanshinone I, cryptotanshinone, and tanshinone IIA from Salvia miltiorrhiza and their anti-inflammatory activities investigationView study →}. If the fraction is what works, the molecule is a component of a mixture, not the mixture’s explanation.

Sibling → molecule is the third trap. Where a claim mentions Nrf2 induction, myocardial infarct-size reduction, pulmonary antifibrosis or cerebral-ischemia meta-analyses, the molecule in question is almost always tanshinone IIA. Danshen’s own compound-extraction notes here carried exactly that error — attributing an Nrf2 finding to tanshinone I on the strength of a review whose title is Pharmacological activity and mechanism of tanshinone IIA in related diseases. That review is a tanshinone IIA paper; the attribution has been dropped here.

The CYP1A2 Janus Problem

This is where tanshinone I most clearly separates from its siblings, and it is the most decision-relevant pharmacology it has.

In primary rat hepatocytes, the individual tanshinones cryptotanshinone, dihydrotanshinone and tanshinone IIA inhibited CYP1A2-catalysed phenacetin O-deethylation, with IC₅₀ values of 12.9, 17.4 and 31.9 µM respectively. But after four days of exposure, the ethanolic danshen extract and tanshinone I increased CYP1A2 activity 6.8- and 5.2-fold and CYP1A2 protein 13.5- and 6.5-fold, with concomitant upregulation of aryl hydrocarbon receptor mRNA. The authors called it the “Janus action” of the tanshinones on CYP1A2 — inhibition on acute exposure, induction on repeat exposure — and tanshinone I was the induction arm ^{22Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →}. The water extract, and salvianolic acid B and danshensu individually, did nothing ^{22Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →}.

The in-vivo corroboration is independent and unambiguous. In mice, tanshinone I — described in that paper as “a strong inducer of Cyp1a” — raised hepatic CYP1A1 and CYP1A2 mRNA and protein, increased transcriptional activity of both at luciferase reporters, and thereby accelerated the clearance of aristolochic acid I, reducing its AUC and its concentration in liver and kidney and protecting the mice against aristolochic-acid nephrotoxicity ^{24Reference 24Feng C et al. · 2013AnimalTanshinone I protects mice from aristolochic acid I-induced kidney injury by induction of CYP1AView study →}. The pharmacology is real enough to change a second drug’s kinetics in a whole animal.

For comparison, an earlier study established tanshinone IIA as a potent and selective CYP1A2 inhibitor (Kᵢ 7.2 nM against mouse MROD, with a 48-fold CYP1A1/1A2 selectivity ratio); tanshinone I and cryptotanshinone inhibited too but with a selectivity ratio of only 4 ^{23Reference 23Ueng YF et al. · 2003AnimalDiterpene quinone tanshinone IIA selectively inhibits mouse and human cytochrome P4501A2View study →}.

Why this matters. CYP1A2 clears caffeine, theophylline, clozapine, olanzapine, tizanidine, melatonin and a good part of duloxetine. A constituent that induces it on repeat dosing would tend to reduce exposure to those drugs, while one that inhibits it acutely would raise it. Tanshinone I appears to do the first and its siblings the second — in a preparation that contains all four. This has never been tested in humans, and the direction of the net effect of a tanshinone-rich product on CYP1A2 in a person is genuinely unknown. It is a reason to treat concentrated tanshinone extracts as an interaction risk of unresolved sign, not a reason to make a confident prediction.

Danshen, Warfarin and Bleeding — What Belongs to This Molecule

Danshen’s interaction with warfarin is one of the best-documented herb–drug interactions in the materia medica, and it must not be soft-pedalled: there are human case reports of raised INR, over-anticoagulation and bleeding on the combination ^{26Reference 26Chan TY · 2001Interaction between warfarin and danshen (Salvia miltiorrhiza)View study →}, and rat pharmacokinetic/pharmacodynamic work showing danshen increases warfarin absorption and prolongs prothrombin time ^{27Reference 27Lo AC et al. · 1992AnimalThe effects of danshen (Salvia miltiorrhiza) on pharmacokinetics and pharmacodynamics of warfarin in ratsView study →}. Danshen also has antiplatelet activity in its own right, which compounds the concern independently of warfarin.

The attribution question is whether any of that belongs to tanshinone I. The answer is: part of the pharmacokinetic arm does, and the antiplatelet arm does not.

What is attributed to tanshinone I specifically. In rat liver microsomes, tanshinone I as an individual test compound decreased the formation of 4′-, 6- and 7-hydroxywarfarin — reactions mediated by CYP1A1, CYP2C6 and CYP2C11 respectively — as did tanshinone IIA, cryptotanshinone and the ethyl-acetate danshen extract containing them. The aqueous danshen extract had no effect at all. In the same study, whole danshen given to rats decreased warfarin’s Cmax, prolonged its Tmax, reduced hydroxywarfarin formation in vivo and raised steady-state plasma warfarin by 23% ^{25Reference 25Wu WW · 2010In vitroInhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivoView study →}. So the tanshinone fraction, including this molecule, is a plausible carrier of the metabolic half of the interaction.

What is not attributed to tanshinone I. The human case reports and the classic rat PK/PD work were done with whole danshen ^{26,27Reference 26Chan TY · 2001Interaction between warfarin and danshen (Salvia miltiorrhiza)View study →Reference 27Lo AC et al. · 1992AnimalThe effects of danshen (Salvia miltiorrhiza) on pharmacokinetics and pharmacodynamics of warfarin in ratsView study →} and cannot be assigned to any single constituent. The antiplatelet and fibrinolytic activity that underpins danshen’s bleeding risk is predominantly a property of the water-soluble salvianolic acids. And — importantly — no platelet-aggregation study of isolated tanshinone I was found in this review of the literature. The only direct hint is a note by the authors of the CL1-5 lung-cancer paper that tanshinone I’s “potential anticoagulant effect should be noted and evaluated” ^{15Reference 15Lee CY et al. · 2008Anticancer effects of tanshinone I in human non-small cell lung cancerView study →}; that is a flag for future work, not a finding.

The practical position. Anyone taking warfarin, a DOAC, an antiplatelet drug, or facing surgery should avoid danshen preparations — that advice stands on the herb-level evidence and is unaffected by which fraction they buy. For this molecule specifically, the defensible statement is narrower: tanshinone I inhibits several of the CYP isoforms that clear warfarin in rat liver, so a concentrated tanshinone product is a pharmacokinetic interaction candidate; the antiplatelet limb of danshen’s bleeding risk sits with a different fraction; and no one has measured what an isolated tanshinone I dose does to human coagulation.

Major Natural Sources

Tanshinone I is, for practical purposes, a Salvia miltiorrhiza compound. Reported concentrations are low and vary with cultivar, growing conditions and — measurably — harvest timing.

Source	Part	Reported level	Basis
Dan shen (Salvia miltiorrhiza)	Root, rhizome	0.02–0.07% of dried root	HPLC across two cultivation years, consistent with prior reports ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →}
Danshen (Pharmacopoeial spec)	Dried root	≥0.25% combined tanshinone IIA + cryptotanshinone + tanshinone I	Chinese Pharmacopoeia marker assay ^{4Reference 4Su CY et al. · 2015ReviewSalvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — reviewView study →}
S. miltiorrhiza f. alba	Root	Present; not quantified here	White-flowered form, chemically studied separately
Total-tanshinone extract	90% alcohol eluent	>97% total tanshinones (all four)	Preparative resin + semi-preparative HPLC ^{10Reference 10Gao H et al. · 2018Simultaneous purification of dihydrotanshinone, tanshinone I, cryptotanshinone, and tanshinone IIA from Salvia miltiorrhiza and their anti-inflammatory activities investigationView study →}

Two agronomic notes with real consequences for product variability: harvesting at 120 or 180 days after transplanting raised tanshinone I content by 30–37% over a 60-day harvest, while tanshinone IIA rose only 9% — the tanshinones do not accumulate in lockstep ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →}. And extraction solvent decides everything: water gets you the salvianolic acids, alcohol gets you these.

Discovery & Synthesis

The tanshinones were named for the drug, not the plant. In 1934, Manzo Nakao and Tadakatsu Fukushima of the Shanghai Science Institute published On the Chemical Composition of Salvia Miltiorrhiza (Chinese drug Tan-Shen) in Yakugaku Zasshi, the work in which the red pigments of the root were first characterised and christened — tanshinone I among them, also recorded in the older literature as tanshinone A ^{3Reference 3Nakao M · 1934On the chemical composition of Salvia miltiorrhiza (Chinese drug Tan-Shen)View study →}. More than 40 lipophilic tanshinones and structurally related compounds have been isolated from danshen since ^{21Reference 21Zhang Y et al. · 2012Tanshinones: sources, pharmacokinetics and anti-cancer activitiesView study →}.

Structurally, tanshinone I is an abietane-derived diterpene quinone in which the parent skeleton has been almost completely aromatised: a planar phenanthrene core, a furan ring fused at one end, and an ortho-quinone at C-10/C-11, giving 1,6-dimethylphenanthro[1,2-b]furan-10,11-dione (PubChem CID 114917). It is the most oxidised and most rigid of the four majors, and it can be prepared by simple oxidation of dihydrotanshinone I ^{2Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}.

It has been made from scratch: a facile three-step total synthesis of tanshinone I was reported in 2016 ^{31Reference 31Yang W et al. · 2016A facile three-step total synthesis of tanshinone IView study →}, and it has become a popular scaffold for medicinal chemistry precisely because the natural product’s own drug-likeness is poor. Derivatisation is concentrated at three sites — the ortho-quinone, C-17, and C-15 of the furan ring — and the returns are substantial: one 7-aza B-ring derivative reached IC₅₀ 0.12 and 0.33 µM against KB and KB/VCR cells (roughly 49- and 13-fold better than tanshinone I), with water solubility of 15.7 mg/mL, improved microsomal stability, a 2.58 h half-life and 21% oral bioavailability ^{2Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →}. That such gains were available says something plain about the parent molecule.

Toxicity & Safety

The honest verdict is not characterised, and the unknown grading is deliberate: there is no acute or repeat-dose toxicology study of tanshinone I, no in-vivo genotoxicity work, no reproductive or developmental data, and no human exposure data. What exists is fragmentary and mostly incidental — a negative in-vitro cytotoxicity/genotoxicity screen ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}, a two-week mouse study at 10 mg/kg orally that reported no liver toxicity on biochemistry and histology ^{8Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →}, and a xenograft study at 200 mg/kg that noted no change in food intake or body weight ^{12Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}. Those are reassuring as far as they go, but they are not a safety dataset, and unlike a common dietary polyphenol this molecule has no meaningful history of human dietary exposure to fall back on — the traditional water decoction of danshen does not deliver it.

Safety at a Glance

Toxicity unknownDrug interactionsBleeding risk

Bleeding and anticoagulants. Danshen strongly potentiates warfarin in humans ^{26,27Reference 26Chan TY · 2001Interaction between warfarin and danshen (Salvia miltiorrhiza)View study →Reference 27Lo AC et al. · 1992AnimalThe effects of danshen (Salvia miltiorrhiza) on pharmacokinetics and pharmacodynamics of warfarin in ratsView study →}, and tanshinone I inhibits the rat CYP isoforms that clear warfarin ^{25Reference 25Wu WW · 2010In vitroInhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivoView study →}. Avoid danshen and tanshinone products with warfarin, DOACs and antiplatelet drugs, and stop well before surgery.
CYP1A2 of unresolved direction. Tanshinone I induces CYP1A2 on repeat exposure via AhR while its siblings inhibit it ^{22,23,24Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →Reference 23Ueng YF et al. · 2003AnimalDiterpene quinone tanshinone IIA selectively inhibits mouse and human cytochrome P4501A2View study →Reference 24Feng C et al. · 2013AnimalTanshinone I protects mice from aristolochic acid I-induced kidney injury by induction of CYP1AView study →}. Products contain all four, and the net human effect is unknown — treat concentrated tanshinone extracts as an interaction risk with caffeine, theophylline, clozapine, olanzapine and tizanidine.
Estrogen-receptor activity. Dose-dependent ERα- and ERβ-mediated activity in vitro, plus antiandrogenic activity in a yeast transactivation assay ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →} — relevant to hormone-sensitive conditions and unstudied in vivo.
No reproductive or developmental data. Nothing is known about use in pregnancy or lactation; danshen itself is traditionally contraindicated in both.
Preclinical only. Every efficacy claim comes from a cell line or an animal. No human has knowingly taken this molecule as an isolate.
Poorly absorbed and unstable in water. Low oral bioavailability, short half-life, and measurable concentration loss in aqueous solution within 24 hours ^{2,28Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →} — an argument against assuming a product’s label content is what reaches you.
No cytotoxicity or genotoxicity detected in vitro. A 2023 screen found neither for tanshinone I or tanshinone IIA ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}. Absence of a signal in one screen is not a clean record.
Quinone metabolism runs down the detoxifying route. NQO1 performs a two-electron reduction with immediate glucuronidation, rather than the one-electron redox cycling that makes some quinones toxic ^{18Reference 18Li Y et al. · 2019Metabolic characteristics of tanshinone I in human liver microsomes and S9 subcellular fractionsView study →}.

Full Safety and Interactions Detail

No adverse effect specific to isolated tanshinone I has been reported in humans, because no human has been given it. The unknown toxicity grading reflects an absent literature rather than a clean one, and it is the appropriate reading for a molecule whose realistic human exposure route — concentrated ethanolic or CO₂ danshen extracts — is recent, unstandardised and untested.

Interaction risk is the dominant concern, and it has two limbs. The first is bleeding. The danshen–warfarin interaction is established at herb level by multiple human case reports of raised INR and bleeding ^{26Reference 26Chan TY · 2001Interaction between warfarin and danshen (Salvia miltiorrhiza)View study →} and by rat PK/PD work ^{27Reference 27Lo AC et al. · 1992AnimalThe effects of danshen (Salvia miltiorrhiza) on pharmacokinetics and pharmacodynamics of warfarin in ratsView study →}. Molecular attribution is partial: in rat liver microsomes tanshinone I, tanshinone IIA and cryptotanshinone individually inhibited CYP1A1-, CYP2C6- and CYP2C11-mediated warfarin hydroxylation, while the aqueous danshen extract did not ^{25Reference 25Wu WW · 2010In vitroInhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivoView study →} — so the tanshinone fraction plausibly carries the metabolic half of the interaction. The antiplatelet half belongs mainly to the salvianolic acids, and no platelet-aggregation study of isolated tanshinone I was identified. The practical advice does not depend on resolving this: avoid danshen and tanshinone products with any anticoagulant or antiplatelet drug, and discontinue before surgery.

The second limb is CYP1A2, and it is genuinely unusual. Repeat exposure to tanshinone I induces CYP1A2 through AhR upregulation — 5.2-fold activity and 6.5-fold protein in rat hepatocytes after four days ^{22Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →} — and in mice it induces hepatic CYP1A1/1A2 strongly enough to accelerate the clearance of a co-administered compound and change its toxicity ^{24Reference 24Feng C et al. · 2013AnimalTanshinone I protects mice from aristolochic acid I-induced kidney injury by induction of CYP1AView study →}.

Its siblings do the opposite acutely, inhibiting CYP1A2 with IC₅₀s of 12.9–31.9 µM ^{22Reference 22Lee WY et al. · 2012AnimalTanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytesView study →} and, for tanshinone IIA, with nanomolar potency ^{23Reference 23Ueng YF et al. · 2003AnimalDiterpene quinone tanshinone IIA selectively inhibits mouse and human cytochrome P4501A2View study →}.

Since real products contain all four in varying ratios and the timescales differ (inhibition acute, induction on repeat dosing), the net direction in a person cannot be predicted from this data. Substrates worth flagging include caffeine, theophylline, clozapine, olanzapine, tizanidine, melatonin and duloxetine. Tanshinone I is itself cleared by CYP2A6 and a panel of UGTs ^{18Reference 18Li Y et al. · 2019Metabolic characteristics of tanshinone I in human liver microsomes and S9 subcellular fractionsView study →}, so inhibitors of those pathways would raise its own exposure.

Endocrine activity is a real, under-discussed finding. In a comparative in-vitro study, tanshinone I showed dose-dependent bioactivity mediated by both ERα and ERβ, antiandrogenic activity in a yeast transactivation assay, and significant anabolic activity in C2C12 muscle cells ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}. Danshen preparations are used for menopausal complaints in Chinese medicine, so this is not an idle laboratory curiosity — but it has never been followed into an animal or a person, and anyone with a hormone-sensitive condition should treat concentrated tanshinone products as unstudied rather than assumed safe.

Pregnancy and lactation: no reproductive or developmental toxicology exists for tanshinone I, and danshen is traditionally contraindicated in both — a precautionary position, not a studied finding, but the appropriate one in the absence of data.

One methodological caution that affects how to read any tanshinone safety claim: all four tanshinones are chemically unstable in solution. In DMSO, dihydrotanshinone converts to tanshinone I and cryptotanshinone to tanshinone IIA; in aqueous media all four decline within 24 hours ^{28Reference 28Roth A et al. · 2023In vitroChemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systemsView study →}. Results reported for one tanshinone may reflect a mixture, and product content may drift with age.

Dosage

There is no dose. Tanshinone I has never been administered to a person as an isolated molecule, no clinical study exists, and nothing here should be read as a recommendation.

For orientation only, the animal doses at which effects were reported are 10–20 mg/kg orally in mouse Parkinson’s and osteosarcoma models ^{8,16Reference 8Wang S et al. · 2015AnimalTanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s diseaseView study →Reference 16Wang W et al. · 2019Tanshinone I inhibits the growth and metastasis of osteosarcoma via suppressing JAK/STAT3 signalling pathwayView study →} and 200 mg/kg in a lung xenograft ^{12Reference 12Li Y et al. · 2013Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A functionView study →}; in vitro, active concentrations run from about 11 µM (sPLA₂ inhibition) to tens of micromolar ^{6Reference 6Kim SY et al. · 2002AnimalEffects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responsesView study →}. These are not convertible to a human intake, and the compound’s poor oral bioavailability makes naive body-surface-area scaling especially misleading ^{2,21Reference 2Huang X et al. · 2022Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modificationsView study →Reference 21Zhang Y et al. · 2012Tanshinones: sources, pharmacokinetics and anti-cancer activitiesView study →}.

Real-world exposure is only ever indirect, through alcohol-extracted or CO₂-extracted danshen products — total-tanshinone capsules, ethanolic tinctures, standardised lipophilic extracts — in which tanshinone I is one minor constituent among four, at a root level of 0.02–0.07% ^{5Reference 5Xing Z et al. · 2024Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhizaView study →}. A water decoction of danshen root delivers essentially none of it. Because of the danshen–anticoagulant interaction, none of these preparations should be combined with warfarin, DOACs or antiplatelet drugs, and all should be stopped before surgery.

References

Ke L, Zhong C, Chen Z, Zheng Z, Li S, Chen B, Wu Q, Yao H (2023). Tanshinone I: pharmacological activities, molecular mechanisms against diseases and future perspectives. Phytomedicine. https://pubmed.ncbi.nlm.nih.gov/36608501/
Huang X, Jin L, Deng H, Wu D, Shen QK, Quan ZS, Zhang CH, Guo HY (2022). Research and development of natural product tanshinone I: pharmacology, total synthesis, and structure modifications. Frontiers in Pharmacology. https://pubmed.ncbi.nlm.nih.gov/35903340/
Nakao M, Fukushima T (1934). On the chemical composition of Salvia miltiorrhiza (Chinese drug Tan-Shen). Yakugaku Zasshi 54(9):844. https://doi.org/10.1248/yakushi1881.54.9_844
Su CY, Ming QL, Rahman K, Han T, Qin LP (2015). Salvia miltiorrhiza: traditional medicinal uses, chemistry, and pharmacology — review. Chinese Journal of Natural Medicines. https://pubmed.ncbi.nlm.nih.gov/25835361/
Xing Z, Bi G, Li T, Zhang Q, Knight PR (2024). Effect of harvest time on growth and bioactive compounds in Salvia miltiorrhiza. Plants (Basel). https://pubmed.ncbi.nlm.nih.gov/38999628/
Kim SY, Moon TC, Chang HW, Son KH, Kang SS, Kim HP (2002). Effects of tanshinone I isolated from Salvia miltiorrhiza Bunge on arachidonic acid metabolism and in vivo inflammatory responses. Phytotherapy Research. https://pubmed.ncbi.nlm.nih.gov/12410540/
Zhao J, Liu H, Hong Z, Luo W, Mu W, Hou X, et al. (2023). Tanshinone I specifically suppresses NLRP3 inflammasome activation by disrupting the association of NLRP3 and ASC. Molecular Medicine. https://pubmed.ncbi.nlm.nih.gov/37400760/
Wang S, Jing H, Yang H, Liu Z, Guo H, Chai L, Hu L (2015). Tanshinone I selectively suppresses pro-inflammatory genes expression in activated microglia and prevents nigrostriatal dopaminergic neurodegeneration in a mouse model of Parkinson’s disease. Journal of Ethnopharmacology. https://pubmed.ncbi.nlm.nih.gov/25666429/
Park JH, Park OK, Cho JH, Chen BH, Kim IH, Ahn JH, et al. (2014). Anti-inflammatory effect of tanshinone I in neuroprotection against cerebral ischemia-reperfusion injury in the gerbil hippocampus. Neurochemical Research. https://pubmed.ncbi.nlm.nih.gov/24760430/
Gao H, Huang L, Ding F, Yang K, Feng Y, Tang H, Xu QM, Feng J, Yang S (2018). Simultaneous purification of dihydrotanshinone, tanshinone I, cryptotanshinone, and tanshinone IIA from Salvia miltiorrhiza and their anti-inflammatory activities investigation. Scientific Reports. https://pubmed.ncbi.nlm.nih.gov/29855534/
Gong Y, Li Y, Abdolmaleky HM, Li L, Zhou JR (2012). Tanshinones inhibit the growth of breast cancer cells through epigenetic modification of Aurora A expression and function. PLoS One. https://pubmed.ncbi.nlm.nih.gov/22485147/
Li Y, Gong Y, Li L, Abdolmaleky HM, Zhou JR (2013). Bioactive tanshinone I inhibits the growth of lung cancer in part via downregulation of Aurora A function. Molecular Carcinogenesis. https://pubmed.ncbi.nlm.nih.gov/22389266/
Lu M, Wang C, Wang J (2016). Tanshinone I induces human colorectal cancer cell apoptosis: the potential roles of Aurora A-p53 and survivin-mediated signaling pathways. International Journal of Oncology. https://pubmed.ncbi.nlm.nih.gov/27279458/
Nizamutdinova IT, Lee GW, Lee JS, Cho MK, Son KH, Jeon SJ, et al. (2008). Tanshinone I suppresses growth and invasion of human breast cancer cells, MDA-MB-231, through regulation of adhesion molecules. Carcinogenesis. https://pubmed.ncbi.nlm.nih.gov/18586687/
Lee CY, Sher HF, Chen HW, Liu CC, Chen CH, Lin CS, Yang PC, Tsay HS, Chen JJ (2008). Anticancer effects of tanshinone I in human non-small cell lung cancer. Molecular Cancer Therapeutics. https://pubmed.ncbi.nlm.nih.gov/19001436/
Wang W, Li J, Ding Z, Li Y, Wang J, Chen S, Miao J (2019). Tanshinone I inhibits the growth and metastasis of osteosarcoma via suppressing JAK/STAT3 signalling pathway. Journal of Cellular and Molecular Medicine. https://pubmed.ncbi.nlm.nih.gov/31293090/
Cui S, Chen T, Wang M, Chen Y, Zheng Q, Feng X, Li S, Wang J (2022). Tanshinone I inhibits metastasis of cervical cancer cells by inducing BNIP3/NIX-mediated mitophagy and reprogramming mitochondrial metabolism. Phytomedicine. https://pubmed.ncbi.nlm.nih.gov/35124382/
Li Y, Fan Y, Su H, Wang Q, Li GF, Hu Y, Jiang J, Tan B, Qiu F (2019). Metabolic characteristics of tanshinone I in human liver microsomes and S9 subcellular fractions. Xenobiotica. https://pubmed.ncbi.nlm.nih.gov/29357726/
Yu H, Subedi RK, Nepal PR, Kim YG, Choi HK (2012). Enhancement of solubility and dissolution rate of cryptotanshinone, tanshinone I and tanshinone IIA extracted from Salvia miltiorrhiza. Archives of Pharmacal Research. https://pubmed.ncbi.nlm.nih.gov/22941489/
Zhi H, Li Z, Deng Y, Qi P, Hou J, Deng W, Cheng J, Wu W, Guo D (2019). Comparative in vivo constituents and pharmacokinetic study in rats after oral administration of ultrafine granular powder and traditional decoction slices of Chinese Salvia. Biomedical Chromatography. https://pubmed.ncbi.nlm.nih.gov/30242797/
Zhang Y, Jiang P, Ye M, Kim SH, Jiang C, Lü J (2012). Tanshinones: sources, pharmacokinetics and anti-cancer activities. International Journal of Molecular Sciences. https://pubmed.ncbi.nlm.nih.gov/23202971/
Lee WY, Zhou X, Or PM, Kwan YW, Yeung JH (2012). Tanshinone I increases CYP1A2 protein expression and enzyme activity in primary rat hepatocytes. Phytomedicine. https://pubmed.ncbi.nlm.nih.gov/21899995/
Ueng YF, Kuo YH, Peng HC, Chen TL, Jan WC, Guengerich FP, Lin YL (2003). Diterpene quinone tanshinone IIA selectively inhibits mouse and human cytochrome P4501A2. Xenobiotica. https://pubmed.ncbi.nlm.nih.gov/12851037/
Feng C, Xie X, Wu M, Li C, Gao M, Liu M, Qi X, Ren J (2013). Tanshinone I protects mice from aristolochic acid I-induced kidney injury by induction of CYP1A. Environmental Toxicology and Pharmacology. https://pubmed.ncbi.nlm.nih.gov/23981375/
Wu WW, Yeung JH (2010). Inhibition of warfarin hydroxylation by major tanshinones of danshen (Salvia miltiorrhiza) in the rat in vitro and in vivo. Phytomedicine. https://pubmed.ncbi.nlm.nih.gov/19577456/
Chan TY (2001). Interaction between warfarin and danshen (Salvia miltiorrhiza). Annals of Pharmacotherapy. https://pubmed.ncbi.nlm.nih.gov/11302416/
Lo AC, Chan K, Yeung JH, Woo KS (1992). The effects of danshen (Salvia miltiorrhiza) on pharmacokinetics and pharmacodynamics of warfarin in rats. European Journal of Drug Metabolism and Pharmacokinetics. https://pubmed.ncbi.nlm.nih.gov/1301354/
Roth A, Zhao P, Soukup ST, Guigas C, Stärke J, Kulling SE, Diel P (2023). Chemical stability and bioactivity of tanshinone I, tanshinone IIA, cryptotanshinone and dihydrotanshinone in in vitro test systems. Toxicology Letters. https://pubmed.ncbi.nlm.nih.gov/36563867/
Li J, Wang W, Huang H, Zhou R, Yang Z, Cui Z, Niu Z, Shi S, Wang P (2025). Tanshinone I, tanshinone IIA, and cryptotanshinone: key bioactive components modulating vascular smooth muscle cell function. Frontiers in Pharmacology. https://pubmed.ncbi.nlm.nih.gov/41142250/
Wang Q, et al. (2025). Tanshinone I ameliorates psoriasis-like dermatitis by suppressing inflammation and regulating keratinocyte differentiation. Drug Design, Development and Therapy. https://pubmed.ncbi.nlm.nih.gov/39876988/
Yang W, Wang JJ, Shao PP, Yuan SY, Li XQ (2016). A facile three-step total synthesis of tanshinone I. Journal of Asian Natural Products Research 18:677–683. https://doi.org/10.1080/10286020.2015.1136906