Ellagic acid — a naturally occurring hydrolysable polyphenol and dilactone derived from hexahydroxydiphenic acid. It occurs as free ellagic acid and, more commonly, as a structural component or hydrolysis product of ellagitannins in pomegranate, raspberries, blackberries, strawberries, walnuts, muscadine grapes, and related foods. It is formally classified as a dietary polyphenolic phytochemical rather than an approved anticancer drug. Oral EA has low aqueous solubility, limited absorption, rapid conjugation, and extensive microbiome-dependent conversion into urolithins; therefore, free EA and its microbial metabolites should be treated as related but pharmacokinetically distinct agents.
Primary mechanisms (ranked):
Bioavailability / PK relevance: Free EA has very poor oral bioavailability because of low water solubility, limited intestinal permeability, rapid methylation and glucuronidation, and short systemic persistence. Human plasma exposure to unchanged EA after ellagitannin-rich foods is generally transient and substantially below the micromolar concentrations commonly used in cancer-cell experiments. Gut bacteria convert unabsorbed EA and ellagitannins into urolithins, which circulate mainly as glucuronide or sulfate conjugates and may account for much of the systemic biological activity. Urolithin production varies markedly among individuals according to microbiome metabotype. Liposomal, nanoparticle, phospholipid, and other delivery systems improve exposure experimentally, but none has established clinical anticancer efficacy.
In-vitro vs systemic exposure relevance: Many anticancer studies use approximately 5–100 µM EA for hours to days. These concentrations commonly exceed achievable circulating free-EA exposure after food or conventional oral supplementation. Direct relevance is greater for gastrointestinal luminal exposure, local tissue exposure, specialized formulations, or experiments involving metabolites. Results obtained at high micromolar concentrations should not be interpreted as evidence that an ordinary dietary dose will produce equivalent systemic tumor effects.
Clinical evidence status: Predominantly preclinical. EA has extensive cell-culture evidence and several animal tumor studies, but human oncology evidence remains limited to an older small adjunctive prostate-cancer trial, observational or combination-product studies, tissue-disposition studies, and trials of ellagitannin-rich pomegranate, berry, or muscadine extracts. These mixed extracts cannot establish EA-specific efficacy. Human randomized trials in metabolic, inflammatory, gastrointestinal, and neurologic conditions provide preliminary safety and biomarker data but do not validate EA as a cancer treatment. EA is not an FDA-approved anticancer agent and should be classified as investigational or dietary adjunct evidence, not established oncology therapy.
Safety / interaction constraints: Food-derived exposure is generally well tolerated, and small supplementation trials have not identified a consistent major toxicity signal. Long-term safety of purified high-dose EA in oncology populations is insufficiently characterized. Potential interaction concerns include inhibition or modulation of drug-metabolizing enzymes and transporters, antiplatelet activity, and uncertain effects when combined with cytotoxic chemotherapy or radiotherapy. The antioxidant-versus-pro-oxidant direction is dose-, tissue-, and treatment-dependent; concurrent use during cancer therapy therefore requires clinical review rather than assuming either protection or sensitization.
| Rank | Pathway / Axis | Cancer Cells | Normal Cells | TSF | Primary Effect | Notes / Interpretation |
|---|---|---|---|---|---|---|
| 1 | Mitochondrial apoptosis | Mitochondrial membrane potential ↓; Bax and cytochrome-c ↑; Bcl-2 and Bcl-xL ↓; caspase-3 and caspase-9 ↑ | Apoptotic injury ↓ or ↔ (context-dependent) | R, G | Programmed tumor-cell death | A recurrent downstream mechanism, usually following survival-pathway inhibition, mitochondrial stress, or ROS accumulation. |
| 2 | PI3K AKT mTOR survival signaling | PI3K ↓; phosphorylated AKT ↓; mTOR ↓; PTEN ↑ (model-dependent) | AKT survival signaling ↔ or ↑ during tissue injury (context-dependent) | R, G | Growth and survival suppression | The direction can differ in nonmalignant injury models, where EA may preserve physiological AKT signaling. |
| 3 | Cell-cycle and p53 p21 control | p53 ↑; p21 ↑; cyclin D1 and cyclin E ↓; CDK2, CDK4, and CDK6 ↓; arrest ↑ | Aberrant proliferation ↓; normal cycling cells usually less affected | G | Cytostasis and apoptosis priming | G0/G1 arrest is frequent, although G2/M or S-phase effects occur depending on tumor type and exposure. |
| 4 | NF-κB inflammatory survival signaling | NF-κB activation ↓; COX-2 ↓; iNOS ↓; IL-6 and IL-8 ↓; anti-apoptotic transcription ↓ | Inflammatory signaling ↓ | R, G | Anti-inflammatory and anti-survival activity | NF-κB suppression can contribute to reduced proliferation, invasion, treatment resistance, and inflammatory cytokine production. |
| 5 | Mitochondrial ROS increase | ROS ↑; oxidative DNA damage ↑; mitochondrial stress ↑ (dose-dependent) | ROS injury ↓ or ↔ through antioxidant activity (context-dependent) | P, R | Selective oxidative stress | Pro-oxidant effects are most evident in redox-unstable cancer cells, high-concentration experiments, and combination treatments. EA is not uniformly pro-oxidant. |
| 6 | NRF2 antioxidant response | NRF2 and HO-1 ↔, ↑, or ↓ (model-dependent) | NRF2 ↑; HO-1 and antioxidant enzymes ↑; lipid peroxidation ↓ | R, G | Secondary redox adaptation | Normal-tissue protection is relatively consistent, whereas cancer-cell NRF2 modulation is heterogeneous and may depend on baseline NRF2 activity. |
| 7 | Glycolysis and intracellular pH regulation | Glucose consumption ↓; lactate production ↓; LDH and PKM2 ↓; NHE1 ↓; intracellular pH ↓ (model-dependent) | ↔ | R, G | Metabolic restriction and acidification | Demonstrated in selected cancer models and should not be generalized to every tumor type. |
| 8 | STAT3 and MAPK growth signaling | STAT3 ↓; phosphorylated STAT3 ↓; ERK signaling ↓; JNK or p38 modulation (context-dependent) | Stress-responsive MAPK signaling ↔ or normalized | R, G | Proliferation and survival inhibition | MAPK direction varies with tumor type, concentration, and whether stress-mediated apoptosis is being induced. |
| 9 | EMT invasion and metastasis | TGF-β and SMAD signaling ↓; Snail and Twist ↓; vimentin ↓; E-cadherin ↑; MMP2 and MMP9 ↓ | Pathological matrix remodeling ↓ | G | Reduced migration and invasion | Includes reversal of EMT and, in some models, partial resensitization of drug-resistant cancer cells. |
| 10 | HIF-1α VEGF angiogenesis | HIF-1α ↓; VEGF ↓; VEGFR2 signaling ↓; angiogenesis ↓ | Pathological angiogenesis ↓; physiological effect uncertain | G | Anti-angiogenic activity | Supported mainly by cell, endothelial-assay, and animal evidence rather than validated clinical anti-angiogenic activity. |
| 11 | DNA damage and repair balance | Oxidative DNA damage ↑; PARP cleavage ↑; p53-dependent damage response ↑ (dose-dependent) | Genotoxic injury ↓ through antioxidant and antimutagenic activity (context-dependent) | R, G | Tumor-cell damage with normal-cell protection potential | The bidirectional profile reflects differing redox conditions and exposure levels in malignant versus normal systems. |
| 12 | Chemosensitization | Drug sensitivity ↑; EMT and survival signaling ↓; apoptosis ↑ (combination-dependent) | Chemotherapy-associated injury may ↓ in some models | R, G | Adjunctive treatment modulation | Reported with agents including gemcitabine and several experimental combinations. Clinical confirmation is inadequate. |
| 13 | Radiosensitization | Radiation response ↑; ROS and DNA damage ↑; clonogenic survival ↓ (model-dependent) | Radiation injury may ↓ through antioxidant and anti-inflammatory effects (model-dependent) | P, R, G | Differential radiation response | Recent preclinical work supports possible simultaneous tumor sensitization and normal-tissue protection, but this has not been established clinically. |
| 14 | Microbiome conversion to urolithins | Direct EA exposure limited; systemic effects may be mediated partly by urolithin metabolites | Urolithin production and conjugate exposure vary markedly among individuals | G | Metabolite-dependent systemic activity | Metabotype 0, A, or B and microbial composition can materially alter metabolite identity, quantity, and biological response. |
| 15 | Clinical Translation Constraint | Free EA exposure commonly below active in-vitro concentrations; tumor delivery unproven | Food exposure generally tolerated; high-dose long-term safety and drug interactions incompletely defined | G | Bioavailability and evidence limitation | Poor solubility, rapid metabolism, microbiome heterogeneity, mixed-extract trials, limited oncology RCT evidence, and lack of an approved pharmaceutical formulation restrict translation. |
TSF: P: 0–30 min R: 30 min–3 hr G: >3 hr