An Oral DHB Prodrug Clears Mouse Tumors With PD-1

University of Michigan and Sungkyunkwan researchers turned a poorly absorbed gut chemical into an oral nano-prodrug that, with PD-1 blockade, cleared tumors in mice. The chemical is 3,4-dihydroxybenzoic acid, or DHB, made when gut bacteria process dietary fiber and also known as protocatechuic acid.

James J. Moon, J. G. Searle Professor of Pharmaceutical Sciences at Michigan, published the work on 10 August 2026 in Nature Nanotechnology with first author Kai Han and co-corresponding author Young Seok Cho of Sungkyunkwan University. The headline around the paper has been the microbiome. The object that moved tumors was a lipid-coated precursor, Prodrug 201, built because free DHB does not stay in blood long enough to act like a drug.

Free DHB Lasts Minutes in Blood

The team did not start with a new molecule. It started with a delivery failure. Many gut metabolites are small and water-loving, and the paper says they can vanish from circulation within minutes, which is why eating fiber is a weak way to dose a T cell in a tumor.

A 2024 review of protocatechuic acid cites a mouse study in which an oral 50 mg/kg dose of the free acid hit a plasma peak of 73.6 μM in 5 minutes and showed a half-life of 2.9 minutes. Sungkyunkwan University later said natural DHB was cleared within minutes after it was swallowed. That is a food chemical, not a pill.

The fix was chemistry plus a lipid shell. The group converted DHB into an inactive precursor, then packed it in an oral nano-emulsion. Sungkyunkwan University described that product as an oleic acid-based nano-emulsion named Prodrug 201. In rats given a single 9.7 mg/kg DHB-equivalent oral dose, the oral nano-emulsion of DHB in mice work’s companion rat study put plasma half-life at 4.1 hours, peak concentration at 182 ng/ml, area under the curve 14.3-fold higher than free DHB, and oral bioavailability 9.72-fold higher.

PRODRUG 201 IN RAT BLOOD

• Half-life: Oral Prodrug 201 lasted 4.1 hours in plasma versus free DHB, which the paper treats as a minutes-scale problem.
• Peak level: Cmax reached 182 ng/ml after the 9.7 mg/kg DHB-equivalent oral dose.
• Total exposure: Area under the curve rose 14.3-fold compared with free DHB.
• Absorbed share: Oral bioavailability (%F) rose 9.72-fold compared with free DHB.

Conference abstracts had compressed that blood-level gain into a single “14-fold” line. The journal paper splits it, and the split is the whole point: more of the dose has to reach T cells, or the metabolite never gets a vote in a tumor.

Why Most Patients Still Do Not Respond to PD-1

Checkpoint drugs take the brakes off T cells so they can see cancer. They changed melanoma and several other cancers, and they still leave most patients without a lasting response. The team’s own patent application on metabolite prodrugs states that immune checkpoint therapies work in only a subset of patients, typically 10-30%.

Moon joined Michigan in 2012, a year after the U.S. Food and Drug Administration cleared the first checkpoint inhibitor for advanced skin cancer in March 2011. His lab’s brief, as he later described it, was to raise those response rates with new drugs and with dietary fibers that change immune tone.

One reason PD-1 antibodies stall is T-cell exhaustion. Killer CD8 cells that see antigen for too long stop dividing, make fewer killing proteins, and pile on extra brakes. Stem-like CD8 cells, marked by Tcf1 and CD62L, can still renew and feed a response. The paper treats those stem-like cells as the population PD-1 blockade needs, and DHB as a way to push T cells toward that state instead of a short, spent effector burst.

That is a companion-drug problem, not a salad problem. If DHB only appears in the gut after fiber is fermented, the dose at a tumor is whoever’s microbes happen to make, and then rapidly clear, a phenolic acid. An oral prodrug with a known composition is an attempt to put that dose on a schedule.

A Quiet Metabolic State in Killer T Cells

DHB was not pulled from a catalog at random. Moon’s group had already reported that an oral inulin gel, a fiber from plant roots, synergized with PD-1 blockade through short-chain fatty acids. The same gel also shifted a set of non-SCFA metabolites in feces. The new paper screened those, including glycerol, GABA, pantothenate, ferulate, and DHB, on OT-I CD8 cells activated with the SIINFEKL peptide.

DHB stood out. It raised CD44+ CD62L+ central memory CD8 cells and Tcf1+ cells, kept more cells alive after IL-2 was pulled for 36 hours, and preserved a strong interferon-gamma recall when antigen came back. It did not push cells into a terminal effector look. Several DHB isomers, including 2,4-DHB and 2,3-DHB, also raised central memory cells, but only 3,4-DHB raised IL-2 secretion in that assay.

The metabolic readouts are quieter still. Targeted metabolomics showed drops in glycolytic intermediates such as glucose-6-phosphate, fructose-6-phosphate, phosphoenolpyruvate, and pyruvate. In CT26-bearing mice given oral DHB at 2.4 mg per dose every other day, CD8 cells from lymph nodes ran a lower extracellular acidification rate, a proxy for glycolysis. A glycolysis blocker, 2-deoxyglucose, copied the DHB memory phenotype. Inhibitors of glutamine use, pyruvate import, and fatty-acid oxidation did not.

HOW DHB CHANGES KILLER T CELLS

• Glutathione trap: DHB forms an adduct with glutathione and lowers intracellular GSH, a handle the paper ties to the later signaling shift.
• Glycolysis down: Seahorse assays and metabolomics both show a less glycolytic, more resting CD8 cell, the metabolic look of stem-like memory cells.
• Akt-mTORC1-Myc: Western blots after DHB showed changes in phosphorylated Akt, mTOR, and Myc, the growth circuit that keeps effectors burning sugar.
• Stem markers up: Tcf1, CD62L, Sca-1, and nuclear Foxo1 rose, while Foxo1 phosphorylation fell.
• PD-1 gets fuel: Those stem-like, antigen-specific CD8 cells are the cells the paper says anti-PD-1 needs in order to work in cold or resistant tumors.

Single-cell RNA and ATAC sequencing backed the same direction: more progenitor-like programs, more open chromatin at Tcf7, Sell, and Ccr7, and a drop in glycolysis and hypoxia gene sets. DHB is doing pharmacology on T-cell fuel use. It is not a general antioxidant story in this paper, even though protocatechuic acid has been sold that way in food research for years.

Mouse Tumors Vanished Only With the Combo

Free DHB given by mouth helped some models. It was not the winning product. In CT26 colon tumors, several oral DHB prodrugs were tested at a molar equivalent of 0.11 mg DHB per dose, with anti-PD-1 at 100 µg injected in the belly on a fixed calendar. Prodrug 201 plus the antibody gave the strongest tumor control and raised CD8 cells specific for AH1, a CT26 epitope, in blood.

The paper then ran that combo through tumors that often shrug at PD-1 alone, including B16F10 melanoma and NOOC1 squamous tumors the authors call anti-PD-1-resistant. In both, oral Prodrug 201 plus the antibody beat free DHB plus the antibody. In MMTV-PyMT breast tumors, the combo was compared with antibody alone and with saline, with mammary tumors scored on day 88. A genetic colon model, CDX2-Cre APC, was used to count tumors in the bowel after DHB plus PD-1 blockade.

MOUSE TUMOR MODELS IN THE STUDY

CT26 survivors from the Prodrug 201 plus anti-PD-1 group were rechallenged on day 90 with 1.5 × 10^5 CT26 cells on the opposite flank and watched for another 60 days, a test of immune memory. Depleting CD8 cells wiped out the benefit. The paper also says DHB improved adoptive T cell therapy, the lab version of giving a mouse tumor-specific T cells. Moon, speaking about the work, said DHB improved CAR T-cell therapies as well, the clinical method that re-engineers a patient’s own T cells.

Group sizes in these figures were small, often five mice, which is normal for a mechanism paper and far from a clinical result. No registry entry turned up for a Prodrug 201 or DHB checkpoint trial in people. Moon said he hopes the mouse findings will hold in human trials. That sentence is still a hope.

Michigan Filed the Prodrug Patent in 2022

The legal file is older than the journal article. A U.S. provisional application, 63/213,990, was filed on 23 June 2021. The full Michigan invention listing for the metabolites sits in the university’s tech-transfer catalog as technology 2021-483, naming Deepak Nagrath, Moon, Han, Kim Hutchings, and Martin Clasby, and describing an oral adjunct for patients on checkpoint blockade.

The utility application was filed on 22 June 2022 and published on 12 September 2024 as US20240299299A1, assigned to the University of Michigan System. It covers metabolites, derivatives, and prodrugs, including 3,4-dihydroxybenzoate, meant to raise the yield of immunotherapies and vaccines, and it lists DHB and 3-methylhistidine as hits from the screen. The Nature paper repeats that patent applications on a microbial-metabolite prodrug oral form to improve checkpoint blockade have been filed with Moon, Han, Nagrath, Hutchings, and Clasby as inventors.

FROM PATENT FILE TO JOURNAL PAPER

1. 23 June 2021: U.S. provisional 63/213,990 is filed on metabolite compositions for immunotherapy.
2. 22 June 2022: The utility application is filed and later published as US20240299299A1.
3. 12 September 2024: The application publishes, still pending, with the University of Michigan as assignee.
4. 4 November 2025: A Society for Immunotherapy of Cancer abstract restates the Prodrug 201 nano-emulsion and the 14-fold blood-level claim later refined in the journal.
5. 10 August 2026: Nature Nanotechnology publishes the full mouse, sequencing, and pharmacokinetic package.

Moon discloses board roles, paid consulting, research funding, and equity in EVOQ Therapeutics, founded in 2016 for autoimmune work, and Saros Therapeutics, launched in 2021 for cancer. Those companies are not named as the vehicle for Prodrug 201 in the paper. The filing still means the oral metabolite idea was treated as property years before the mouse survival curves ran in a journal.

Olives, Onions and a Compound You Already Eat

DHB is not an exotic toxin from a rare microbe. It is protocatechuic acid, a phenolic acid found in plants and made in the bowel when bacteria break down anthocyanins and related flavonoids. A 2024 review lists protocatechuic acid in olives and grapes, and also in hibiscus, onions, and potatoes, among more than thirty plant foods. The same review puts its molecular weight at 154.12 g/mol and calls it only slightly soluble in water, about 1 part in 50.

Food chemists have chased this molecule as an antioxidant and as a possible cancer-preventive in cell dishes and rodent diets. Gut bacteria such as those that process cyanidin-3-glucoside can generate it in the colon, so blood levels after a polyphenol-rich meal can exceed what the food itself contained as free acid. That is interesting nutrition. It is not a dosing plan for PD-1-resistant melanoma.

Cooking wipes a lot of it out; one cited figure is as much as 90% lost after boiling at 100 °C. Even when it is swallowed as the free acid, the mouse oral half-life of 2.9 minutes explains why Moon’s group refused to stop at “a metabolite from fiber.” If the clinical goal is stem-like CD8 cells next to a tumor, the relevant product is the prodrug that held 4.1 hours in rat plasma, not a second helping of chicory.

Public posts that picked up the paper mostly repeated the fiber-metabolite hook and moved on. The broader research feed is already crowded with other microbial chemicals, from older T-cell-stemness metabolite papers to newer work on riboflavin-related compounds, that also claim to help checkpoint drugs in mice. None of those headlines is an approved oral medicine either. The difference this group is selling is a defined nano-emulsion a factory could, in principle, make twice the same way.

The Same Lab’s Fiber Gel Is Already in Patients

Moon has said the gut makes many useful compounds that can be turned into drugs, and that this is the first time natural microbial metabolites have been developed as a new oral form for immunotherapy. That claim is about formulation. The biology of metabolites shaping T-cell stemness was already in the literature. His sentence is the manufacturing claim.

This is the first time anyone has shown that natural microbial metabolites can be developed as a new oral formulation for immunotherapy.

James J. Moon, J. G. Searle Professor of Pharmaceutical Sciences, University of Michigan

The same laboratory’s fiber work is further along than Prodrug 201. Moon’s group built an edible inulin gel, described as similar to Greek yogurt, that feeds gut bacteria and, in mice, improved checkpoint therapy. A Michigan research profile says that gel is inulin gel already in a kidney cancer trial led by Ulka Vaishampayan, and that Moon is in the early stages of a third company to sell an inulin-based supplement. That trial is not a Prodrug 201 trial. The gel changes a community of microbes. The nano-emulsion is a single chemical with a structure the paper says is set up for scale, quality control, and a pill a patient might actually take.

The authors also write that the same nano-medicine methods might later help in autoimmune disease, the other half of Moon’s shop. For cancer, the near work they describe is more metabolite screening, not a posted first-in-human date. Until a trial enrolls, the second effect of this paper is narrower than the fiber headlines: it shows that a cheap phenolic acid can be rebuilt so it behaves like an oral partner to PD-1 blockade in mice, including in tumors that ignored the antibody alone.

Whether human T cells, human guts, and human tumors will copy those mouse curves is the open measurement. The patent is filed. The gel is in patients. The DHB nano-emulsion is still a mouse drug with a blood-level problem that, for once, the chemists actually solved.

Disclaimer: This article is news reporting on a laboratory study in mice and related patents, and it is for information only. It is not medical advice, a treatment plan, or a claim that DHB, Prodrug 201, inulin gel, or any fiber supplement treats cancer in people. Readers should consult a licensed oncologist or other qualified cancer specialist before changing immunotherapy, diet, or supplements. Figures, trial statuses, and disclosures reflect the Nature Nanotechnology paper, the University of Michigan filings, and the university pages cited here, and they may change if human studies of the prodrug begin.