Pamako Olive Oil Polyphenols vs Triple-Negative Breast Cancer — the Pasteur Institute Study and the Arithmetic of Dose
A phenolic extract isolated from Cretan Pamako olive oil, given to mice before cancer cells were implanted, reduced tumour mass by 46%. Given late — it did nothing at all. That is the single most important finding of the paper.
Scope of this article. We describe a pre-clinical study in an animal model that tested isolated phenolic compounds, a concentrated phenolic extract and — in one experiment — extra virgin olive oil itself as a food. We report only what the publication states and flag which results the authors themselves call preliminary. This text is not medical advice and does not constitute a health claim for any product.
At the end of August 2026 the journal Nutrients published a paper from three Athens research institutions that has interested us for years — because the raw material was an olive oil we have sold since the 2019/20 season. A team from the Hellenic Pasteur Institute, the Department of Pharmacy of the University of Athens and the National Hellenic Research Foundation examined what olive oil polyphenols do to triple-negative breast cancer in mice — and, more importantly, when they do it.
The result is less spectacular than the widely reported Louisiana study on colorectal cancer we covered earlier, but in one respect it is more interesting: for the first time, a single experimental framework compared pure oleocanthal, a mixed phenolic extract and an ordinary bottle of olive oil. And it measured how much depends on the moment you start.
Who did the research
| Institutions | Hellenic Pasteur Institute (Athens); Laboratory of Pharmacognosy and Natural Products Chemistry, Department of Pharmacy, National and Kapodistrian University of Athens; Institute of Chemical Biology, National Hellenic Research Foundation (NHRF) |
| Authors | Nikoleta Anna Madelou, Marianna Kapetanou, Katerina Papakonstantinou, Olga Koutsoni, Zacharias Kakazanis, Eleni Melliou, Prokopios Magiatis, Vasilis Zoumbourlis, Efstathios S. Gonos, Haralabia Boleti (corresponding author) |
| Journal | Nutrients 2026, 18(17), 2756 — open access |
| Published | 23 August 2026 |
| DOI | 10.3390/nu18172756 |
| Funding | Greek Ministry of Development (European Social Fund), Peloponnese regional recovery fund, World Olive Center for Health and the Leventis Foundation (scholarship), Hellenic Pasteur Institute internal funding. The authors declare no conflicts of interest. |
Where the material came from. Everything tested — pure oleocanthal, oleuropein aglycone, the concentrated phenolic extract and the oil fed to the animals — came from Pamako extra virgin olive oil from the Chania region of Crete, provided by the producer. The dephenolised oil used as the control was prepared from the same batch. The purity of the isolated compounds (≥97%) was confirmed by quantitative NMR in the laboratory of Prof. Prokopios Magiatis.
Why triple-negative breast cancer
Breast cancer accounts for roughly 2.3 million new diagnoses a year worldwide. Clinically it is classified by the presence of three receptors: oestrogen, progesterone and HER2. The triple-negative subtype (TNBC) has none of them — it makes up 15–20% of cases, behaves aggressively, recurs often, and because the absence of receptors means the absence of targets for hormonal and anti-HER2 therapies, chemotherapy remains the backbone of treatment.
The study used the MDA-MB-231 cell line, the standard TNBC model: highly invasive, metastasis-prone human breast adenocarcinoma cells. They were implanted into severely immunodeficient CB17-SCID mice, which do not reject human cells — and implanted orthotopically, into the mammary fat pads, where the tumour grows naturally, rather than under the skin.
What exactly was administered
This deserves a pause, because it is the first study to place three different “forms” of olive oil polyphenols side by side in one experimental design:
- Pure compounds — oleocanthal (5 mg/kg) and oleuropein aglycone (12.5 mg/kg), alone and combined (each at half dose). Given intraperitoneally, three times a week, from day 5 after cell implantation.
- Concentrated phenolic extract from the same oil — at a dose of 5 mg total phenolics per kg body weight, by oral gavage, three times a week.
- Extra virgin olive oil itself — 100 µl per mouse, a volume delivering the same 5 mg/kg of phenolics, also by gavage. The control was the same oil with its phenolics removed: identical fat, zero secoiridoids.
The composition of the extract matters, because it determines how the results translate to real bottles. The 5 mg/kg dose of total phenolics consisted of 2.7 mg/kg oleocanthal, 1.3 mg/kg oleacein, 0.64 mg/kg ligstroside aglycone, 0.27 mg/kg oleuropein aglycone and 0.05 mg/kg each of hydroxytyrosol and tyrosol. More than half was oleocanthal — exactly the profile typical of Pamako, in which oleocanthal dominates oleacein by roughly 5:1 (2025/26 season: 1,318 vs 264 mg/kg according to the qNMR certificate).
Result one: oleocanthal alone beats the combination
In the first experiment (intraperitoneal, 30 days) each of the three treated groups had smaller tumours than the control. Only pure oleocanthal, however, reached statistical significance.
| Group (intraperitoneal) | Dose | Tumour mass reduction | Significance |
|---|---|---|---|
| Oleocanthal | 5 mg/kg | −33.1% | p = 0.020 — significant |
| Oleuropein aglycone | 12.5 mg/kg | −22.0% | p = 0.13 — not significant |
| Oleocanthal + oleuropein aglycone | 2.5 + 6.25 mg/kg | −23.7% | p = 0.10 — not significant |
Note the third row: the combination at half doses produced almost the same effect as the full dose of oleuropein aglycone. The synergy this team had seen earlier in cell cultures did not translate into an advantage in the living animal, but activity was retained despite half the amount of each compound. Body weight was stable in all groups — none of the preparations caused signs of toxicity.
Result two: the extract works by the oral route — and timing matters
Intraperitoneal injections are a laboratory tool. The real question is whether polyphenols taken by the oral route work the same way. In the second experiment the phenolic extract was given by gavage from day 5 after implantation. Tumour mass fell by 44.8% — numerically better than injections of pure oleocanthal, though with four mice per group it did not cross the significance threshold (p = 0.078). The authors describe the effect size (Hedges' g = 1.30) as potentially large but themselves urge caution.
The third experiment is the most interesting. Same extract, same dose, three different starting points:
| Start of extract administration | Tumour mass reduction | Significance | Effect size |
|---|---|---|---|
| 7 days BEFORE implantation (prophylactic) | −46.0% | p = 0.0008 | large (g = 1.9) |
| 5 days after implantation (early tumour) | −35.0% | p = 0.044 | moderate-to-large (g = 1.46) |
| 20 days after implantation (established tumour) | no detectable effect | p = 0.26 | — |
The gradient is clear: the earlier the better, and after three weeks — nothing. One mouse in the prophylactic group never developed a tumour at all. The authors repeated the prophylactic group in an independent experiment and pooled both data sets, so this particular result rests on 7–8 animals rather than four.
What follows from this. In this model olive oil polyphenols act as a factor that hinders the establishment and early growth of a tumour, not as a drug for a tumour that has already developed. The authors put it carefully: the compounds “may primarily limit tumour initiation and early progression rather than reverse established disease, at least at the studied doses.” An important caveat for anyone tempted to read such papers as a promise of therapy.
Result three: the oil itself, no extraction
This is the experiment closest to the plate. Mice received 100 µl of Pamako olive oil per dose — the amount corresponding to 5 mg/kg of phenolics — starting 7 days before cell implantation. The control received the same oil with the phenolics removed.
Tumour mass fell from 0.500 g to 0.355 g, a reduction of 29.0%. With 3–4 animals per group the difference is not statistically significant (p = 0.18) and the effect is described as moderate (g = 0.77). The authors call this result “exploratory” in so many words. The direction, however, agrees with the extract, and the weaker effect has two natural explanations: polyphenols are absorbed differently from a fat matrix than from an aqueous solution, and the control itself — oil without phenolics — was not inert, as we will see in a moment.
Mechanisms: four levers instead of one
The paper's title speaks of a “multimechanistic” effect, and that is not decoration. In the tumours and serum of the prophylactic group, four independent parameters were measured.
| Parameter | What it means | Effect of the phenolic extract |
|---|---|---|
| Proteasome activity (chymotrypsin-like) in the tumour | Cancer cells are “addicted” to a high rate of protein degradation; the proteasome is their recycling machine | −30% vs control (p = 0.03) |
| γH2A.X in the tumour | Marker of DNA double-strand breaks — genomic instability | −47% vs control |
| Oxidised (carbonylated) proteins in serum | Systemic marker of oxidative stress | −33% vs control (p < 0.05) |
| Cell migration of MDA-MB-231 (in vitro) | The cells' ability to move — a prerequisite for metastasis | at 24 h, “wound” closure of 8–48% depending on concentration vs 88% in control |
The migration assay deserves its own sentence. In cell culture a scratch is made in the cell layer and the researchers watch how quickly it closes. At the lowest extract concentration — four times below the threshold at which it begins to inhibit cell division at all — the scratch was 48% closed after 24 hours versus 88% in the control. The cells were alive and dividing normally, but they stopped moving. The authors note that the effect exceeded what the literature reports for individual compounds, suggesting the mixture behaves differently from its parts.
A surprise: oil without phenolics does something too
An honestly designed study has a habit of producing controls that do not behave as they should. That happened here. The dephenolised oil:
- lowered γH2A.X in tumours to a degree comparable to the full oil;
- lowered oxidised proteins in serum — though the full oil lowered them by a further 21%;
- but raised proteasome activity in tumours — clearly above both the aqueous control and the extract group.
The authors draw two conclusions. First, olive oil is not only polyphenols: squalene, triterpenes and other non-phenolic constituents have biological activity of their own, and some effects cannot be attributed to the phenolic fraction alone. Second, not everything the dephenolised oil does is beneficial — higher proteasome activity in a tumour is an unwelcome signal in this context, and the phenolics correct it (in the full oil it was significantly lower than in the dephenolised oil, p = 0.007). In other words: the fat of olive oil and the polyphenols of olive oil are two different factors that work best together.
How much is that in olive oil? The arithmetic of dose
Unlike the Louisiana team, the authors of this paper do not convert the dose to a human equivalent — rightly, since they have no pharmacokinetic data. The calculation below is therefore ours, done with the standard interspecies conversion factor (Km mouse→human = 0.081), purely to show orders of magnitude.
- Mouse: 5 mg total phenolics per kg, three times a week
- Human equivalent: ~0.41 mg/kg
- 70 kg adult: ~28 mg of phenolics per dose, three times a week — about 12 mg a day on average
A surprisingly modest dose — smaller than in the colorectal study (57 mg of oleocanthal alone per day). But modest only if the oil is genuinely rich in polyphenols. From the methods it can be calculated that the oil fed to the mice contained roughly 800–1,100 mg/kg total phenolics (100 µl delivered 5 mg/kg to a 15–20 g mouse). Below: how much oil of a given phenolic content corresponds to 28 mg. Density 0.916 g/ml, 884 kcal per 100 g.
| Total phenolics in the oil | Oil per dose | Volume | Calories |
|---|---|---|---|
| 2,081 mg/kg — Pamako Organic, 2025/26 season (qNMR) | 13.6 g | ~15 ml (one tablespoon) | 120 kcal |
| 1,000 mg/kg — order of magnitude of the oil used in the study | 28.4 g | ~31 ml | 251 kcal |
| 500 mg/kg | 56.8 g | ~62 ml | 502 kcal |
| 250 mg/kg — threshold of the EU health claim | 113.6 g | ~124 ml | 1,004 kcal |
| 200 mg/kg — average supermarket oil | 142 g | ~155 ml | 1,255 kcal |
| 100 mg/kg | 284 g | ~310 ml | 2,511 kcal |
A point of reference. With an oil like the one used in the study, the dose from the paper is about two tablespoons. With the oil on this year's Pamako certificate — one tablespoon. With a supermarket oil — two-thirds of a cup and 1,255 kcal, three times a week. The same amount of polyphenols, a tenfold difference in calories. This is not a dietary recommendation, merely an illustration of why, without an analytical result on the label, any conversation about “healthy” doses of olive oil is meaningless.
Methodological note: the conversion rests on interspecies scaling, not on human studies. The authors stress that after oral intake polyphenols are extensively metabolised in the gut and liver, and that what circulates in blood is mainly their metabolites — free hydroxytyrosol after an ordinary serving of olive oil is very low (≤0.05 µM); high-phenolic oils raise metabolite levels to 1–4 µM. What fraction of a dose reaches the tissues, and in what form, remains the subject of ongoing research.
A co-author's voice. Prof. Prokopios Magiatis of the Department of Pharmacy, University of Athens — the same person whose signature appears on the qNMR certificates of the olive oils in our shop — commented on the results in the press release that followed publication:
“The 46% reduction in tumor burden is a particularly interesting finding, but it must be interpreted within the appropriate scientific context. This is a preclinical study and not a clinical trial in patients. For us, it is equally important that the study demonstrates that we can obtain a well-characterized and standardized mixture of natural active compounds from an excellent Greek olive oil and study it using the tools of modern biomedical research.”
And on the outlook: Greek olive oil could acquire a new dimension — beyond being an excellent food, it could serve as a raw material for high-value-added products and, provided the necessary clinical evidence is established in the coming years, potentially for future pharmaceutical applications.
What this study does not say
The authors gave the limitations a chapter of their own, and it is worth summarising faithfully — a rare and commendable practice.
- Small groups. Following the 3Rs principle of reducing animal use, 4–5 mice per group were used. Several large effects failed to reach significance for exactly this reason; the authors call them “hypothesis-generating”, not “proof of efficacy”.
- Mice without an immune system. The SCID model allows human cells to be implanted but removes immunity from the picture — and immunity plays an enormous role in real cancer.
- No pharmacokinetics. Concentrations of oleocanthal or its metabolites in blood and tumours were not measured, so injections and oral dosing cannot be compared directly, nor real exposure determined.
- Investigators were not blinded to group allocation or tumour measurement (only the biochemical analyses were blinded). Randomisation was manual.
- Outlier exclusion. In the timing experiment, four unusually large tumours (one each in four mice) were removed by the ROUT method. The authors provide a sensitivity analysis: without exclusions the pattern of results is unchanged, the significance levels are not.
- In vitro / in vivo discrepancy. In tumours the extract lowered γH2A.X; in cell culture it did not — and pure oleuropein aglycone actually raised it. The authors attribute this to the different contexts (chronic exposure to metabolites in the body versus 24-hour exposure to parent compounds in a dish) but honestly list it as a limitation.
- Prevention, not therapy. With an established tumour the extract did not work. Nothing in this paper applies to people who are already ill.
The authors' conclusion
Olive oil polyphenols show antitumour activity in a triple-negative breast cancer model, and the timing of administration is a critical determinant of efficacy — the greatest effect comes from prophylactic dosing begun before the tumour appears. Complex phenolic mixtures appear to act more strongly than single compounds, probably through complementary or synergistic interactions between constituents and their metabolites. The mechanism is multi-pronged: proteasome modulation, reduced oxidative stress, less DNA damage, inhibited cell migration. The authors propose olive oil polyphenols as candidates for further investigation as prophylactic agents and nutraceuticals, with the caveat that larger pre-clinical studies and clarification of bioavailability are needed.
What this means for the bottle on the shelf
Three things from this paper can be taken into the kitchen without over-interpretation.
First, the profile counts, not just the total. The extract in the study was dominated by oleocanthal, because that is what Pamako is. Two oils with the same “polyphenol content” on the label can have completely different compositions — and, as this study shows, different activity. A qNMR certificate with separate values for oleocanthal, oleacein and the D1 index says more than one aggregate number.
Second, oil without phenolics is not “the same oil, just milder”. The non-phenolic fraction has its own activity, partly unfavourable in this context. A late-harvest, long-stored or heated oil is, biochemically, a different product from the same oil fresh.
Third, regularity before intensity. The dose in the study was small but constant, starting a week before the “event”. If this model has any bearing on humans — which we do not know today — it argues for a daily spoonful of good oil over years, not a glassful in a crisis.
Pamako olive oil from Crete — the producer whose oil was the source of every compound used in the study described above:
Every olive oil in our shop comes with its qNMR result showing separate values for oleocanthal, oleacein and the D1 index.
Sources
Source publication (full text, open access): Madelou N.A., Kapetanou M., Papakonstantinou K., Koutsoni O., Kakazanis Z., Melliou E., Magiatis P., Zoumbourlis V., Gonos E.S., Boleti H. Time-Dependent Multimechanistic Antitumor Effects of Olive Oil Phenolics in a Triple-Negative Breast Cancer Mouse Model. Nutrients 2026, 18(17), 2756. Source ↗ · PubMed
Earlier work by the same team: Papakonstantinou A. et al., Int. J. Mol. Sci. 2022 (large-scale isolation of secoiridoids) and Nutrients 2023 (synergy of secoiridoid combinations in cell cultures).
Values for Pamako Organic Monovarietal (2025/26 season) are taken from the World Olive Center for Health qNMR certificate (University of Athens analysis, 21 Nov 2025): total phenolics 2,081 mg/kg, oleocanthal 1,318 mg/kg, oleacein 264 mg/kg, D1 index 1,582 mg/kg. These data are not part of the cited publication; phenolic content varies from season to season.
This article reports the results of a pre-clinical study in an animal model. It is not medical advice and does not constitute a health claim for any product. Under Regulation (EC) No 1924/2006, foods may not be attributed properties of preventing, treating or curing disease. Olive oil is a food, not a medicinal product.
Written by: Tadeusz Gruszczyński,
certified olive oil sommelier.
Copying or using this text without the author's permission is prohibited and may result in legal action.
