Bacteria as cancer therapy
Tumor-colonising bacteria as cancer therapy (Ewingella americana). Cancer. Reviewed Oct 2026.
Beat two standard drugs in mice in a single unreplicated study, and the strain it proposes to inject is a documented opportunistic pathogen in exactly the immunocompromised patients who would receive it.
What it is
Some bacteria grow only in the oxygen-starved core of a tumor, which healthy tissue does not have. Inject them and they multiply where the tumor is, kill cells directly, and the resulting inflammation drags the immune system into a fight it had been ignoring. A Japanese group (JAIST) screened nine strains from the guts of tree frogs, fire-bellied newts and a Japanese grass lizard, and found Ewingella americana was by far the most potent — shrinking mouse tumors more than anti-PD-L1 antibody or doxorubicin, with T-cell, neutrophil and B-cell activation on top of direct killing.
The evidence
Mouse tumor models only, one lab, not replicated. Outperformed anti-PD-L1 and doxorubicin in regression experiments; the authors report minimal pathogenicity and no significant adverse effects at effective doses. That safety claim is the weakest part and does not survive the human literature: E. americana is an established opportunistic human pathogen with 19 documented infections in a 2024 narrative review (Antibiotics, PMID 38927225) — bloodstream, respiratory and peritoneal-dialysis infections, one death — plus meningitis in a term newborn (PMID 32596194), fatal Waterhouse-Friderichsen syndrome (PMID 12604997), sepsis in an immunocompromised patient (Cureus 2025) and a multi-drug-resistant isolate (PMID 18626541). Mortality across reported cases is low and the organism is usually antibiotic-susceptible, but “minimal pathogenicity in mice” is not “safe in humans”, and cancer patients are precisely the immunocompromised population these case reports come from. No human trial of this strain exists, and the comparator effect sizes are mouse effect sizes, which historically translate badly.
What you can do today
Not available for this use, and years away if it arrives at all. Anything sold today with this claim is not backed by human evidence for it.
Sources
- Coverage, not a sourcetandfonline.com
No primary paper or registry record is linked for this entry yet. The evidence summary names the studies it relies on.
Change history
Tumor-colonising bacteria as cancer therapy (Ewingella americana) added to Cancer at grade D.
A strain from frog and newt guts beat anti-PD-L1 and doxorubicin in mouse tumors, which is a genuinely interesting result from a place nobody had searched. It enters at the bottom of the ladder anyway: one lab, mice, no replication. The paper's safety conclusion is the part that fails review — Ewingella americana has a documented human case record including neonatal meningitis, fatal sepsis and a multi-drug-resistant isolate, which the write-ups omit. The approach has one real human datapoint (Clostridium novyi-NT plus pembrolizumab, 25% response in 16 refractory patients), and that is what the grade for the field rests on, not this strain.
Related
More in Cancer
- Pelareorep for pancreatic cancerResponse rates in later-line settings beat the approved comparator, but that is not randomised survival data and the Phase 3 has not reported.
- mRNA-4157 melanoma vaccinePhase 3 beat Keytruda alone on recurrence, but only a press release exists, hazard ratios are unpublished and overall survival is immature.
- mRNA vaccine for pancreatic cancerNearly 90% of patients whose immune system responded were alive up to six years later in a small early trial, and randomised survival results have not been reported.
- CAR-T for solid tumoursEncouraging early responses in gastric and some childhood tumours, but it is not curative for most solid cancers.
- Creatine and cancer immunotherapyImproved anti-tumour immune signalling in mouse melanoma and has not been tested in people with cancer.
- CRISPR-Cas12a2 for p53 cancersSelectively killed p53-mutant cells in the lab, and nobody yet knows how to deliver it to every tumour cell in a living body.