Researchers at Columbia University engineered the probiotic strain Escherichia coli Nissle 1917 to carry three treatment functions into solid tumours. The bacteria express cytosine deaminase, which converts the prodrug 5-fluorocytosine into the chemotherapy drug 5-fluorouracil, while also producing an IL-15 superagonist and a PD-L1-blocking nanobody inside the tumour.
The study was published on 16 September in the peer-reviewed journal Science Translational Medicine. Its central idea is spatial control: combine local chemotherapy activation with two immune interventions in one living delivery platform rather than exposing the whole body to the same mixture.
Mouse models showed effects beyond the treated tumour
The authors tested the platform in murine MC38 colorectal and B16-F10 melanoma models. The paper reports antitumour effects in both models and found increased activation of antigen-presenting cells, T cells and natural-killer cells when the immune payloads were added to the enzyme-prodrug therapy.
Abscopal and rechallenge experiments also indicated systemic antitumour immunity and durable immune memory in the treated mice. Those experiments are useful evidence that a locally delivered bacterial therapy can influence disease outside the injected tumour, but they do not establish that the same response will occur in people.
The evidence remains preclinical
The published work is peer reviewed, but its efficacy evidence comes from mouse tumour models. The public abstract does not establish human dosing, bacterial containment, long-term safety or clinical benefit, and no human trial result is part of this paper.
The result therefore supports a preclinical delivery strategy rather than a cancer treatment ready for patients. Human studies would need to show that the engineered bacteria can be controlled safely, localise treatment as intended and produce a meaningful benefit without unacceptable infection, immune or chemotherapy toxicity.