A 25% Lower-Protein Diet Slowed Pancreatic Cancer in Mice

A pancreatic cancer diagnosis often comes with an instinct to change what you eat, and a new mouse study describes a genuinely surprising mechanism by which diet can influence this notoriously hard-to-treat cancer, but it is also a reminder that a finding being real does not mean it is ready to act on. Researchers at Fudan University in Shanghai found that cutting dietary protein by 25% slowed pancreatic tumor growth in mice by reshaping their gut bacteria and turning tumor-fighting immune cells back on, and the same protein-restricted diet made a checkpoint immunotherapy drug work better. The mechanism, published in Nature Cancer on 17 August 2026, is worth understanding, but this is a mouse finding: pancreatic cancer patients should not restrict protein on the strength of it, since unintended weight loss (cachexia) is already one of the disease's biggest dangers.

Key Takeaways

• A 25% protein-reduced diet slowed pancreatic tumor growth in mice by enriching a gut bacterium, Blautia coccoides, whose metabolite UDP-galactose reprograms tumor-fighting immune cells.
• The effect depended on gut bacteria, macrophages, and T cells together; removing any one erased the benefit, and the diet also boosted an anti-PD1 immunotherapy drug's effectiveness in mice.
• In two small human cohorts, lower levels of B. coccoides and UDP-galactose tracked with more advanced-stage pancreatic cancer, but that piece of the evidence is correlational, not proof the diet works in people.
• This is a mouse and cell-culture study; pancreatic cancer patients should not restrict protein without oncology guidance, since unplanned weight loss is already a serious risk in this disease.

Blautia Coccoides and UDP-Galactose: The Bacterial Link to Tumor Suppression

Blautia coccoides, a gut bacterium that became far more abundant in mice fed a diet with 25% less protein, suppressed pancreatic tumor growth by producing a metabolite called UDP-galactose (uridine diphosphate-galactose, a sugar-based molecule bacteria use in their own metabolism), according to the study from Zhongshan Hospital at Fudan University, published in Nature Cancer on 17 August 2026 (examined here via the study's Research Square preprint, posted January 2025, ahead of peer-reviewed revision). Researchers tested three levels of protein restriction, 12.5%, 25%, and 50%, and found the 25% reduction had the strongest anti-tumor effect; only the 50% reduction caused any weight loss. Mechanistically, UDP-galactose activates a receptor called P2Y14R on immune cells called macrophages, which switches on a signaling protein, STAT1, and shifts tumor-associated macrophages from a tumor-protecting state into a tumor-attacking one. B. coccoides is not a well-known probiotic species, but prior research has linked it to metabolic changes under low-protein conditions, consistent with what this study found.

Pancreatic Cancer Immunotherapy: Why the Finding Matters

Pancreatic ductal adenocarcinoma (PDAC, the most common and lethal form of pancreatic cancer) is notoriously resistant to checkpoint immunotherapy drugs like anti-PD1, largely because its tumor microenvironment is dominated by immune cells that protect the tumor rather than attack it, a pattern documented across the broader pancreatic cancer microbiome literature. In this study, combining the protein-reduced diet with anti-PD1 significantly reduced tumor burden and improved survival compared with anti-PD1 alone, and giving mice B. coccoides or UDP-galactose directly produced a similar boost. That echoes a growing body of gut microbiome and pancreatic cancer immunotherapy research suggesting specific bacteria or their metabolites, rather than diet composition alone, may determine whether checkpoint drugs work in this cancer, similar to how a separate microbial metabolite, TMAO, was shown to boost anti-PD1 response in pancreatic tumors in mice.

Microbiota Depletion and Fecal Transplant: How Researchers Proved Causation

The researchers used male mice with orthotopic and subcutaneous pancreatic tumors (grown from KPC cancer cells, a standard model of the disease's most common genetic drivers). Wiping out the mice's gut bacteria with antibiotics eliminated the protein-restricted diet's anti-tumor effect entirely, and transplanting fecal bacteria from protein-restricted donor mice into microbiota-depleted recipients reduced tumor burden even without any diet change, showing the effect travels with the bacteria. Depleting macrophages or CD4/CD8 T cells with targeted treatments also erased the diet's benefit, confirming the immune system carries out the effect rather than the diet acting directly on tumor cells (the researchers separately ruled out direct effects on tumor cell division, metabolism, and growth-signaling pathways). A more targeted experiment, knocking down the P2Y14R receptor specifically in macrophages, also abolished the diet's benefit, tying the entire mechanism to that one receptor. In humans, a cohort of 50 healthy people and 50 PDAC patients found B. coccoides was the single most differentially abundant bacterial species between the groups, markedly lower in patients, and a separate cohort of 60 PDAC patients across cancer stages found both B. coccoides and UDP-galactose were lower in advanced-stage disease.

What This Means for Pancreatic Cancer Patients and Their Families

This is a mouse and cell-culture study, and nobody should restrict dietary protein because of it, especially not a person already diagnosed with pancreatic cancer. Unintended weight loss and muscle wasting (cachexia) are already among the most dangerous complications of pancreatic cancer, and cutting protein intake without medical supervision could make that worse rather than better; any dietary change in a pancreatic cancer diagnosis should go through an oncologist or a clinical dietitian, not this study. If there is a hopeful practical thread here, it's that the researchers found giving mice the bacterial metabolite UDP-galactose directly, without changing their diet at all, reproduced the diet's tumor-suppressing effect, hinting that if this translates to people, the eventual application might be a drug or engineered probiotic delivering the metabolite itself, not a protein-restricted diet that patients would need to follow.

Limitations: Single Mouse Model and Cross-Sectional Human Data

The mechanistic work was done in a single mouse strain (C57BL/6J) and one tumor model derived from KPC cancer cells, so it is not yet clear whether the same pathway holds in other pancreatic cancer models or in people. Both human cohorts, the 50-versus-50 comparison and the 60-patient staging cohort, are cross-sectional, meaning they show a correlation between bacterial and metabolite levels and cancer stage at a single point in time, not a controlled test of the diet, the bacterium, or the metabolite in human patients. The preprint text used to report on this study predates the peer-reviewed Nature Cancer version, which lists additional co-authors and more pages, suggesting some details may have shifted during review even though the core findings and title stayed consistent. The authors also explicitly flag that there is no data yet on the long-term safety of extended low-protein diets in pancreatic cancer patients, which is precisely the open question that makes self-directed diet changes inadvisable right now.

What's Next: From Mouse Diet to Possible Drug Target

Because directly dosing mice with UDP-galactose reproduced the diet's benefit, the more plausible next step toward any human application is a targeted drug or engineered probiotic aimed at the P2Y14R pathway, rather than recommending pancreatic cancer patients change what they eat. Larger, prospective human studies would also be needed to confirm that low B. coccoides or UDP-galactose levels actually predict outcomes, rather than simply reflecting more advanced disease. No clinical trial or timeline is described in the available study text.

Common Questions

Should pancreatic cancer patients eat less protein because of this study? No. This finding comes from mice, not people, and pancreatic cancer patients already face a high risk of unintended weight loss; any dietary change should be discussed with an oncologist, not started based on an animal study.

What is Blautia coccoides? It's a gut bacterium that became markedly more abundant in mice fed a protein-reduced diet, and this study found it produces a metabolite, UDP-galactose, that helped slow pancreatic tumors by reprogramming immune cells.

Did the low-protein diet cure cancer in the mice? No. It slowed tumor growth and improved survival, especially when combined with anti-PD1 immunotherapy, but it did not eliminate the tumors in the models tested.

Could this lead to a new pancreatic cancer treatment? Possibly, though not the diet itself. Giving mice the bacterial metabolite UDP-galactose directly reproduced the diet's effect, which points toward a future drug or engineered probiotic rather than protein restriction as the more plausible path to any human treatment, pending clinical trials.

References

1. Yueying Chen, Fulin Nian, Shengdi Wu, et al. (2026). Low-protein diet enhances antitumor immunity in pancreatic cancer through microbiota-derived UDP-galactose. Nature Cancer. https://doi.org/10.1038/s43018-026-01222-2
2. Yuguo Niu, Xiaoming Hu, Yali Song, et al. (2024). Blautia Coccoides is a Newly Identified Bacterium Increased by Leucine Deprivation and has a Novel Function in Improving Metabolic Disorders. Advanced Science. https://doi.org/10.1002/advs.202309255
3. Holly Attebury, Donnele Daley (2023). The Gut Microbiome and Pancreatic Cancer Development and Treatment. The Cancer Journal. https://doi.org/10.1097/ppo.0000000000000647
4. Gauri Mirji, Alison Worth, Sajad Ahmad Bhat, et al. (2022). The microbiome-derived metabolite TMAO drives immune activation and boosts responses to immune checkpoint blockade in pancreatic cancer. Science Immunology. https://doi.org/10.1126/sciimmunol.abn0704