Quick answer
A molecule made by gut bacteria called cis-Bac429 — an N-acyl amide produced by Bacteroides — significantly boosted anti-PD-1 immunotherapy against lung cancer in mouse models. University of Florida researchers isolated it by working backward from 183 bacterial strains to the 6 that stimulated anti-tumor T-cell activity, then to the single metabolite responsible. It is preclinical only: the molecule has been tested directly in mouse tumors, never in humans. The broader field it belongs to — using the gut microbiome to improve immunotherapy — does have human trial results now, and they are genuinely encouraging but more complicated than any single-molecule story.
Key takeaways:
- From 183 Bacteroides isolates taken from the feces of humanized anti-PD-1 responder mice, exactly 6 strains stimulated IFNγ production from primary CD8+ T cells; together they enhanced anti-PD-1 efficacy in syngeneic and orthotopic lung cancer models (Cell Reports Medicine 2026).
- The active compound is cis-Bac429, an N-acyl amide from Bacteroides; its chemically saturated synthetic version (sat-Bac429) is inactive, meaning the effect is structurally specific — not just "any bacterial fat molecule."
- Intratumoral cis-Bac429 plus anti-PD-1 significantly shrank subcutaneous lung and colon tumors in mice and drove IFNγ+ CD8+ T-cell infiltration into tumors.
- In the phase 2 FMT-LUMINate trial (2026), a single fecal-microbiota-transplant capsule before treatment produced an 80% objective response rate in first-line NSCLC (16/20) and 75% in melanoma — but the benefit came from loss of specific harmful taxa, not from donor strain engraftment.
- Strain-level gut microbiome signatures predict immunotherapy response better than species-level or clinical data — but only when the signature matches the treatment regimen, not the cancer type (Nature Medicine 2024).
- No microbiome-derived molecule, including cis-Bac429, is approved or in human trials; patients should not change treatment based on this research.
The discovery pipeline: from feces to a single molecule
The study behind this article, published in Cell Reports Medicine (volume 7, issue 1, published online December 19, 2025), started with a prior observation: enrichment of the Bacteroides genus was associated with better anti-PD-1-mediated tumor therapy. Rather than stopping at that correlation, the team reverse-engineered it.
Step 1: Build a strain library. The researchers isolated 183 Bacteroides strains from the feces of humanized mice that responded to anti-PD-1 therapy.
Step 2: Test each one. Supernatants — the liquid the bacteria grew in, containing their metabolic output — from just 6 of the 183 isolates stimulated IFNγ production from primary CD8+ T cells, the interferon that supercharges killer T cells against tumors. This six-strain consortium ("6-consort") enhanced anti-PD-1's anti-tumor efficacy in both syngeneic and orthotopic lung cancer models compared to mice colonized with feces from non-responders, and the effect depended on IFNγ production.
Step 3: Follow the chemistry. Bioassay-guided fractionation and comparative metabolomics traced the activity to a single N-acyl amide the researchers named cis-Bac429, produced by Bacteroides.
Step 4: Prove specificity. The critical control: a synthetic saturated version (sat-Bac429) did nothing. Only the cis configuration stimulated IFNγ from CD8+ T cells. When cis-Bac429 was administered directly into lung and colon tumors in mice alongside anti-PD-1, tumor growth dropped significantly and IFNγ+ CD8+ T cells infiltrated the tumors.
That last step is why the result matters beyond one molecule: it demonstrates that a defined bacterial metabolite can act as an adjuvant to immunotherapy, and that the immune-stimulating chemistry is precise enough that a small structural change erases it. First author Rachel Newsome and senior author Christian Jobin (UF Health Cancer Institute) have co-founded a spinout, Bebi Therapeutics, to develop Bacteroides-type N-acyl amides as adjuvant treatments for anti-PD-1-refractory NSCLC — but that is a company formation, not a clinical program with human data.
Why lung cancer is the right target
Lung cancer remains the leading cause of cancer-related death, and advanced NSCLC is treated with anti-PD-1/PD-L1 immunotherapy — yet many patients derive no direct benefit, with over half showing primary resistance to checkpoint inhibitors. Anything that converts non-responders into responders without adding broad toxicity would change outcomes at scale.
A 2025 Nature Communications study showed the concept isn't limited to small molecules: extracellular vesicles shed by commensal Bifidobacterium were taken up by lung cancer cells (via dynamin-dependent endocytosis), upregulated PD-L1 expression through the TLR4-NF-κB pathway, accumulated in tumors after oral gavage in mice, and synergized with anti-PD-1 to enhance anti-tumor responses in a syngeneic mouse model. Between bacterial metabolites (cis-Bac429) and bacterial vesicles (Bif.BEVs), the pipeline of microbiome-derived immunotherapy adjuvants in lung cancer is filling in — entirely in animal models so far.
The human evidence: what FMT trials actually showed
The molecular work stands on a decade of human evidence that the gut microbiome genuinely influences checkpoint-inhibitor outcomes. The honest version of that story has three acts.
First, observation. The foundational 2018 Science study (Gustave Roussy) found that the gut microbiome influenced the efficacy of PD-1-based immunotherapy in epithelial tumors, and that recent antibiotic use was associated with worse outcomes — the first large-scale signal that depleting the ecosystem could cost patients benefit.
Second, first-in-human FMT. In 2021, two pioneering trials — one at the University of Pittsburgh Medical Center, one at Sheba Medical Center in Israel — gave FMT from responders to anti-PD-1-refractory melanoma patients, then re-dosed anti-PD-1. The Pittsburgh trial provided clinical benefit in 6 of 15 patients, with responders showing increased CD8+ T-cell activation and objective responses including one complete and two partial responses; the Sheba phase 1 trial (n=10) likewise saw responses in 3 of 10 patients (two partial, one complete), alongside favorable shifts in immune-cell infiltrates and gene expression in both the gut and the tumor microenvironment. Modest numbers, and a landmark proof of concept.
Third, the surprise. The phase 2 FMT-LUMINate trial (Nature Medicine, 2026) moved earlier: a single FMT capsule from healthy donors before first-line immunotherapy in NSCLC (n=20, anti-PD-1) and melanoma (n=20, anti-PD-1 plus anti-CTLA-4). The response rates were striking — 80% (16/20) in NSCLC, meeting the trial's primary endpoint, and 75% (15/20) in melanoma. But the mechanism was not what the field expected. Shotgun metagenomics showed responders did not achieve benefit by acquiring donor strains; instead, they lost specific baseline bacterial species (frequently Enterocloster citroniae, E. lavalensis, and Clostridium innocuum), and that loss pattern reproduced across three published FMT oncology trials. In a mouse re-colonization experiment, reintroducing the species that FMT had eliminated abrogated the anti-tumor effect of immunotherapy entirely. The benefit came from removing immunotherapy-hostile taxa, not adding helpful ones.
That finding reframes the Bac429 story too. If FMT works by clearing bad actors as much as by supplying good ones, the dream of one booster molecule may be incomplete — a drug like cis-Bac429 might need to be paired with strategies that deplete the species that suppress response.
One more caution from the biomarker side. A 2024 Nature Medicine analysis of a 106-patient trial cohort found that strain-resolved microbiome signatures predicted response to combination checkpoint blockade better than species-level data or clinical factors — but the signatures only transferred between cohorts when the immunotherapy regimen matched (anti-PD-1 monotherapy versus anti-PD-1 plus anti-CTLA-4). A microbiome biomarker for immunotherapy, it turns out, must be matched to the treatment regimen rather than the cancer type. Any single-number "microbiome score" sold to patients without that context is not based on this evidence.
What this means going forward
For cancer treatment: microbiome-derived adjuvants and FMT-based interventions are moving from correlation to randomized human trials, with striking response rates in the first-line FMT setting — results that still need confirmation in larger, blinded studies before they change standard care.
For microbiome science: the 2025–2026 papers share a theme — the field's early "which bacteria predict response" lists were unstable because strain-level detail and regimen context matter. Defined metabolites (cis-Bac429), defined vesicles (Bif.BEVs), and defined depletion targets (Enterocloster species) are the new currency.
For patients: nothing here is actionable today. cis-Bac429 has never been given to a human. FMT in oncology is experimental and carries real risks — in the FMT-LUMINate melanoma cohort, 13 of 20 patients experienced grade 3 or higher adverse events, which is why independent safety monitoring exists. If you are on or considering immunotherapy, the microbiome-relevant decisions with human evidence behind them are ordinary ones: avoiding unnecessary antibiotic courses during treatment and discussing any probiotic with your oncology team.
Frequently asked questions
Can I take a probiotic or eat fermented foods to make immunotherapy work better? There is no human evidence that any commercially available probiotic or dietary change improves checkpoint-inhibitor outcomes, and some oncologists advise against live-probiotic supplements during immunotherapy because their safety in immunocompromised patients is unproven. The human evidence that exists comes from screened, medical-grade fecal microbiota transplantation in controlled trials. Diet does shape the microbiome's metabolic output — that is biologically real — but no diet has been tested against immunotherapy response in a randomized trial. Keep your oncology team informed about everything you take.
Related articles
- Gut microbiome and kidney cancer immunotherapy
- The gut-immune axis: mechanisms
- Fermented foods: science, benefits, risks
- Long-term antibiotics and the gut microbiome
Medical disclaimer: This article is for informational purposes only and does not constitute medical advice. Cancer treatment requires professional medical guidance; never modify your treatment without consulting your oncologist.